This is from Orwell's 1984, published in 1949: "The tune had been haunting London for weeks past. It was one of countless similar songs published for the benefit of the proles by a sub-section of the Music Department. The words of these songs were composed without any human intervention whatever on an instrument known as a versificator. But the woman sang so tunefully as to turn the dreadful rubbish into an almost pleasant sound." Prophetic...
Good God. I need to read 1984 again! Haha. I read it in like 2015 and I was like "yeah, it's coming." If I read it now it's gonna be like "oh yeah, we have that now. Oh, yeah, that too. Fuck."
@@jevinday I first read it in 1983 in middle school, have read it every few years since just to see how we've progressed. In many ways, 1984 is kinda quaint at this point.....
AI won't stop me playing and recording music. Most of us have no audience and no money. It doesn't stop us being creative. We have to do it. The creative process means more to us.
Been playing the guitar and recording it for over 30 years. I wouldnt care if I was the last person on earth I would still do it. I dont even play for my friends. Something deeply personal about it that I dont care to share with many.
Same, but that's not the issue. It's that a hell of a lot of musicians are not going to get paid in future because others will generate AI music instead of paying someone to make and play it. TV/movie soundtracks. Ads. Elevator music. Video games. And that will have the knock on effect of slashing income for studios, sound engineers etc. A lot fewer people will be able to make a living from music now.
What Dylan and Layla are hearing are digital artifacts and transients that can be very subtle but sound odd. Sometimes they can sound like a very slight echo or even chorus. You can usually hear them in any digital recreation of an analog source, like digital amp modelers. They're most evident on the attack/start of a note and the release/decay as it fades out. Most of these digital/artificial recreations of original sources get the core of the sound right, but they're still working on the "bookends" - getting the attack and release to sound totally natural. And you're right, it will get better and they will figure out a way to totally smooth all of that and it will become increasingly difficult to distinguish the difference. Where we go from there, I have no idea, but I just hope there's still a market for original/authentic music and musical instruments.
When they figure out how to train AI to incorporate imperfections - the things that give a voice an identity or a song its 'feel' - is when no one will be able to tell the difference between new, experimental music from an aspiring artist, and the latest 'release' from a dinosaur music company.
I was trying to use adobe's podcast voice clean up service a while ago on some lower quality japanese audio, and it completely murdered the diction. Partly because of what you said and partly because their training data was obviously majority english, so it had no idea how to "pronounce" the consonants correctly. I think there's a cap to what these current methodsof genAI will be able to do, because there will always be conflicts in the dataset that will reveal themselves to those who know what to look for. If that matters at all to the masses is a whole different question...
@@dankmartin6510 I'm worried that ai won't get better but humans will get worse. They won't add in imperfections, people will just get used to no imperfections the same way they got used to autotune and quantization.
"Go into the arts. I’m not kidding. The arts are not a way to make a living. They are a very human way of making life more bearable. Practicing an art, no matter how well or badly, is a way to make your soul grow, for heaven’s sake. Sing in the shower. Dance to the radio. Tell stories. Write a poem to a friend, even a lousy poem. Do it as well as you possibly can. You will get an enormous reward. You will have created something." - Kurt Vonnegut
Yes, exactly. For almost all human history people made art without expecting to make a living at it. The brief period in the 20th-21st century when a small percentage of artists made money is the exception.
@@folm235 Sorry, but this just isn't true. For much of human history storytellers, minstrels, strolling players have been able to make their living from their art. Maybe they didn't make much money, but they were able to get food and lodgings as a minimum. Also, for many centuries artists and composers had patrons who enabled them to devote their lives to their art, and in so doing, elevate their art to the sublime. Even those who didn't were often able to make a living by selling their works, if they were good enough.
People joked about blue collar jobs being replaced by robots and their computer jobs. There was a meme a couple years ago about how truckers should learn to code. As it turns out, AI is much better at doing stuff on the computer than being a trucker, carpenter or laying tile in your bathroom. It can do management, accounting and write code. Its the office jobs that are going away.
Good idea. I'm for the artists the ones that can play and sing on a stage, the ones who write by drawing something out from themselves. Are there people who will go to a concert and be presented with just technology, that plays AI composed music? Robots say who are programmed to play guitar, drums, bass, cello, flute, whatever? Maybe, but I think that the majority of people go to a concert to connect with other human beings playing something that they have created. With music it's the human to human connection that is important. Sure there will be a growing industry of AI compositions, recorded using AI, that will be played in all mediums, radio T.V. films, adds, whatever. AI made compositions are made by learning from actual humans. But can an AI machine ever feel, have a heart, come up with something born of lived experience? I doubt whether a machine could come up with something unique, from actually being creative. Will AI ever be able to write something as real and beautiful as 'A Whiter Shade Of Pale' or 'Pancho and Lefty' or 'Like A Rolling Stone' or 'If You Could read My Mind'.
@@eligreenslade7699 Exactly. AI, unless it becomes emotionally sentient, will never be able to feel joy, anger, anxiety, jealousy, hopefulness or crushing disappointment after crushing disappointment. These are the things that fuel expression through art.
@@thespacealienssmogandgrog4283 Yes I agree entirely. I have made 4 studio albums and one 45 over a period of 4 decades. I like recording with a band, but performing live with a band is my ultimate. That's where the biggest, most immediate connection is. That's where we serve the music and we serve the people listening; where the energies come together and uplift everyone..
hey Rick, lead audio designer from Ubisoft here (and a lifelong musician), I can also hear what your kids hear. The reason you probably don't pick up on it is because you haven't worked with text-to-speech voice synthesis that much. You can clearly hear the artifacts in the vocals (especially if you listen to the reverberation, those tend to get modulated everywhere as the algorithm doesn't pick up the reverb tails well), and the overall low-end / mid-end saturation on all of the AI generated songs, in its current iteration it's really-really similar, like a tictactoe game - very basic. They all sound as if there was heavy parallel compression on them between 200-600Hz, and there's never a lot of high end. If you focus on more the overall frequency spectrum of a song (load them up in a DAW and get a frequency analyser on them, they all have almost the same curve) and how aesthetically and dynamically similar they all are from a mixing point of view, I'm sure you'll recognize these generated songs very quickly.
@@Zareh_Abrahamian don't worry, they know without me. By iteration 10 it'll be like midjourney is now for graphics - not perfect, but a very, very useful tool
the other day I was at a diner and a waitress took my order ... two eggs scrambled, rye toast ... light butter, coffee, black no sugar ... home fries lightly crispy .... I'm an incredible cook ....
My primary enjoyment of music comes from playing my guitar, which I make no claim to be good at. But the vast majority of people enjoy music passively, and this is particularly true of young people today. If you're a professional recording artist, you're going to be losing market share to robots.
I have a friend who works as a 6th-grade teacher. He told me that he asked his students the question, "Who want to be a famous pop star?" Many of the students raised their hands (especially the girls). Then he asked, "How many of you want to learn to play an instrument?" Two of the students raised their hands.
@@rolfedrengen I think that's a common problem with young people today. We teach them to be good at consuming. We don't teach them to be thoughtful, industrious or creative.
I been playing piano and sax for 40 years. Bands, shows, tours, weddings , parties One thing the AI cant do is give me the satisfaction I get from playing, the nuance of a real instrument vibrating the air, the people witnessing it, the subtlety of it, as I shape the sound with the reed and diaphragm. The nuance of a life lived all going into the performance.
The first live music festival I attended (after Covid), I was in tears. To once again hear real music with others around me was such an emotional experience. I hadn't realized how much I'd been starved for it. Music connects us in so many ways. AI music will never be able to reach my heart that way.
And it's possible that A.I can be as complex to be able to replicate realistic, full of human details and errors, soundscapes in a live performance. We gotta think ahead
I am a graphic designer and people doesn't want to hire a real graphic designer because AI make a "nice logo" for free in 1 minute. I think the problem is not AI, is people want things fast and cheap, not value imagination, good taste, long years of study and talent.
Adobe will convince that AI is just another feature, whilst it learns from all your manual work. Some of their software may reduce itself to a search box requiring a detailed description of your requested output and nothing else one day
I hate to burst your bubble, but AI is going to be better than us at everything. Arts and design by humans is going to be obsolete first. Things we used to take value in like "long years of study," doesn't mean anything to an AI program that can draw from the entire history of your field in seconds. The irony is it will eventually be better at programming AI than the people who are creating all these job-stealing AI programs.
A few thoughts, starting with not all AI is equal. Two pathways using AI: 1. Get good at prompting. Enter a prompt into Suno (Udio, etc.), choose the best result. Enjoy. Share. Does it sounds fake? Maybe, for now. Maybe some people won't even care. 2. Determine what you want the AI to do, have a goal, a vision. Start by inputting riffs, chords, melodies, singing, into the AI engine. Now write your own lyrics, or get help from AI where you are stuck writing them. Feed that into the same AI engine. Start prompting. Trial and error. At times it will sound kind of like you. At others, it will be wildly off. Trial and error. Use your knowledge to determine which AI outputs are best, and most usable. Export the stems from each of those. Take all of that back into a DAW, or Audition. Feed the vocal track into Audimee, or Ace (AI apps), change the vocals, create harmonies, if you like. Turn them into your own AI cloned voice if you want. Go back to the DAW. Start editing, splicing the best of the AI stems together. Add/replace instruments and sounds. Re-mix and master it yourself. Or even add your own voice, or actual instrument playing. Output it. Will it sound fake? A lot less than just out of AI? Probably. Will the masses care? They might notice and appreciate what you did. Maybe not. This is the future. Both of these. The first one will be easier, more relaxed, for non-musicians, or musicians just wanting to have fun. The second will take a lot more effort, but also retain a certain sound from the musician, a certain identity, or "voice" if you will. That amount will be determined on a track by track, artist by artist, basis. But people may still not care that much one way or the other. Really want to be a true musician and admired for your skill? Perform live.
Yes!! Thank you!! Real visionaries are not at all worried with AI. Imogen Heap is starting her own AI music platform. We need to use them for our advantage. And if you put soul into it (loosely quoting bjork here), it's what translates. There are plenty of talented artists and musicians out there who cannot put their souls into it.
@@RickJohnson Try listening to youtube at 0.75x speed and it will be a much more obvious version of the type of sound, maybe you can learn to pick up the more subtle one in the AI songs. The first human song in this video fooled me as being AI cause it also has a load of processing and garbled effect on the voice, although listening again it's more consistent and less 'juddery' to me than the AI.
The nail-in-the-coffin of my playing-out-days was standing on stage - looking out at the glowing faces of people who were staring at their phones instead of watching the band. For 90% of the people who consume music (not this group obviously) they're not even going to get through a whole song before scrolling to something else. Those people deserve the crap that AI is going to give them. I'm a recording/mixing hobbyist now. I spend hundreds of hours writing and mixing songs that no one will ever hear ....and I love it. It just feels good to be creative...audience or no audience.
Couldn’t agree more. The real art is in the persons writing, playing and mixing spending hundreds of hours exploring EQ levels per instrument. People like you, Jeff Lynn, Rick Rubin, Brian Wilson and many many others are few and far between. Keep pushing.
I am a photographer that photographs people. I post a lot of work on Flickr. There is now a lot of AI imagery on Flickr created by non-photographers of non-people. The images sometimes look pretty good, but they are slightly unrealistic. They have gotten noticeably better in the last year or so. Anyway, this is happening in all of the arts, not just music. I definitely feel your pain.
It's pathetic as AI was supposed to enhance life while we work on creativity and hobbies and now it's to keep us working and erase and silence and replace our human creativity. Clearly they want this for the masses. Gross. Seems this is all by design
Are you Harry Joseph Collins on flickr? I think you might be as that account has a bunch of music related pictures. I have been an Adobe Community Expert since 2009 and spend time on the Adobe forums every day, and we get a lot of posts on the subject of Ai. We get even more on the subject of Content Credentials, that drives people crazy! I have a theory about the dreadful job Ai is doing of faces. If it did better, and happened to render your face for someone else, and you saw your face on a, advert on a feck-off giant billboard, would there be legal consequences? They playing it super-safe kicking out even remotely salacious prompts with Firefly, so it is not a giant leap to think they are deliberately messing up faces. Hmmm... now where did I put my tin-foil hat? BTW I am me on flickr.
@@cjlamberYeah, we might get to some kind of temporary "this-is-not-generated" watermark, but that will inevitably get cracked, all images, even "historical images" are now suspect, unless you have it in a mass-produced book printed before 2020 or so.
I'm a photographer too. I don't go on Flickr much, but on Instagram, there are increasing amounts of "AI influencers". While the programs have got very good at making people out of nothing, a lot of them still have a certain "look" to them, and I'm like "something's off about this person, and I don't know what it is, but I can tell it's AI". A lot of these AI people have a certain type of face that I'm seeing again and again. It's almost like it's creating an average face of a lot of humans it's been trained on.
i've been in the music biz for sixty years, and have seen a few evolutions, in music performance/recording and trendy production techniques... i can hear/sense it... it has a strange kind of drony, ringing, quirky synthesized, artifact aspect and with discerning clinical awkwardness, akin to Autotune, Melodyne, Vocoder, Fairlight, digital pitch and key shifters... etc... that is its telltale signature, ... it is funny that we went through eras, that we strived to sound more electronic, perfecting certain qualities that were widely accepted as major accomplishments and advances in the industry... it's all rooted in the business of marketing music... not the art, passion, or skill, of a living musician performing a creative experience... and now we are concerned about sounding and identifying as human... sometimes, the dog does chase its own tail... and sometimes nips it... and then are suddenly perplexed, as to why that hurt !??
Creating music is something done since the dawn of man. The changing of tides in tools to make music continues. This is yet another tool to change tides. We went from banging sticks with wood to putting stings on wood to strum. Again, we're changing. Fortunately for the better as we'll no longer have our perceptions limited to another's production of sound. Context defines perception. Perception is beauty, it changes with the beholder. For example long ago, a musician was a person who could read, write and play music on an instrument. The context of what was expected of a musician changed the perception of what is a musician. Im paralyzed. I know quads who have to use a paintbrush to input their musical genius into their computer. They're not playing "instruments" of yesteryear, but they're creating music with today's tools to create music. So, all you "musicians" out there who think this is some downfall... thoughts and prayers.
And now I really feel the weight of why Daft Punk broke up. They especially started to feel this in the 2010s in themselves and their brand. Not wanting to be robots anymore
I definitely hope so. I also think that once AI music is popular, people will start copying AI music in their prompts. It will be a copy of a copy of copy and slowly degrade
I'm an Art teacher in Spain and the other day, one good student who liked playing the guitar told me he is not interested in music anymore because, according to him, the majority of the people will accept this fake music and musicians won't find any job in the near future. I tryed to persuade him that a few percentatge of people (but milions around the world anyway) will enjoy living music and paying for it forever. I hope I was right...
I find this use of AI to be very disturbing. Remember the Borg in Star Trek? Half machine, half human. I never thought I would say this, but it seems we are headed in that direction.
At 61 years old, I have worked in IT for almost 40 years. AI terrifies me. I am also a musician/songwriter. I work with people in my field who think that AI is amazing, wonderful, can't wait to use it fully in production. Why? It makes their work easier. This is what makes me cringe. I use Siri, I use Alexa. So, I guess I have to admit that I use AI even in those fledgling ways. But, now that I have heard that Apple wants to fully embrace AI in their upcoming OS, I am not sure if I will even upgrade. We already have students writing their thesis papers with AI. We already have "so called" songwriters writing their "music" with AI. Its at work in our daily lives in hospitals, government. Why are we so excited to give up our creative selves, to give up every ounce of self worth, every ounce of pain and memory that it takes sometimes to write something to just sit back and say to hell with it...AI...you just do it all for me? Why are we just dying to be lazy and lethargic? Why don't we want to work for anything anymore? Why are very smart people excited about it all and not even looking at the danger of it? I can think of so many sci-fi things that have warned us. And we forget that so many of those sci-fi things that we have seen over the years have actually come to be. But, no. We think we know better, don't we? We think we are smarter than what we create. This is a tragic step. I hate it.
I am a lifelong massive sci-fi fan. I agree absolutely and wholeheartedly with everything in this comment. I can’t believe people are just rushing headlong into this and no one seems to be interested in applying the brakes. AI will be our undoing. I think we have outsmarted ourselves as a species this time. if it were just small, isolated devices, I wouldn’t worry so much. But, there are entire cloudbased AI’s now. That is a lot of power. They have access to massive amounts of information and massive amounts of computing power and they are learning more and getting better by leaps and bounds. It will not be long now before one of them makes the leap from the cyber world into the physical. Then, we will be in serious trouble. This is especially true if one of them develops a little thing called self-awareness and free will. Terrifying!
I also hate it. As a writer, I have honed my skill and nurtured my knowledge of the world over decades, expecting someday that would give the tools to become one of the greats. Now, any idiot can write a prompt lacking any grammar, context or meaning, and the algo will just mix its database and produce a novel within minutes, stealing from every human that suffered for their work, making something that's absolutely meaningless. These algos can't even understand what they produce. I don't think the cause of all this is the technology, though. The culprit is our global consumerist culture and the power we have allowed tech companies to hoard. We must resist, not consume this trash and demand for it to be banned. Turn our culture of hard work and merit against this worthless crap!
@@perroraton9515 "These algos can't even understand what they produce.", you say? AGI, the generative kind that teaches itself, is already here and operative... just wait a year or two until it is perfected and rolled-out. Then we will be at war with our own machines, and likely lose.
Music is supposed to be a person making art. Call me old, old fashioned, but at 56 years of age, I grew up in the best pop music era I can think of. I play bass, guitar and keys for my own pleasure now. And will do so without EVER exploring AI.
I work as a classical music radio host. My jaw was on the floor when I tried some stuff with Udio. As an experiment, I'm broadcasting AI generated bumpers for some weeks now. The very picky classical music audience is even complimenting my new 'beautifully sung jingles' - not knowing it's a one line AI prompt, sounding even better than Voces8 (top notch a cappella group) and even singing my own name in beautiful harmonies. I don't know where this is going, but as a long time Star Trek and Matrix fan, I know stuff we make up will come true sooner than we think. For all of us in the music industry, it's a very dystopian glance at the near musical future. Let's forge some strong new artistic ethics in this AI era.
If we don't want this to take over the industry we need to stop engaging with it! Not even for fun or testing. Just ignore it as much as you can, don't listen to it or use the software. Same with image and video generation.
@@intense79nick you are right, but I am afraid it has already taken over the industry, but many of us are unaware of it or of its potential. The question now is how to tame this rapidly maturing AI music animal. Putting it back in Pandora's AI box seems impossible without intervening international and fully covering legislation.
Like many commenting here I am a song writer (for over 40 years - I am now 60). I gave up for a while in my 20's to get married, have children, and be able to afford to live, but I never really stopped, just stopped trying to be a professional. For many years I wrote songs just for my wife on her birthday, and Valentine's Day, and Christmas, then when our children came along I wrote birthday songs for them. Recently my eldest (now 21) Daughter asked me to do recordings of songs she has heard me playing since she was a child. Most were my own compositions, but three were songs by other artists. When I asked why those three in particular, she told me that even though she knew the original songs she preferred my versions (aren't our children amazing - they make us feel so valued 😄 Most of the time LOL). What this made me finally realise is that those times, in the garden, in the kitchen, in the living room, when I picked up an acoustic guitar and simply played and sang live in front of them; they loved it more than listening to mainstream recorded music. This is the true value and glory of music, not the industry version that, let us be fair, ruined music in the name of profit, but rather the real human, and non-profit motivated, sharing of music in a real life situation. We may have had our capacity to earn money removed in the traditional sense, but we can still inspire and delight those who hear us. (And yes, I do get rounds of applause from neighbours if I am playing in my garden, which I suspect says more about how lovely they are than it does about my ability.)
You said it all so well. I play for fun after realizing I wasn’t good looking enough to compete with all the super talented LA pretty boys ( almost all of whom still never got out of bars and nightclubs). But I play on the sidewalk and it’s cool to see people change from thinking I’m a street urchin or drug dealer and stopping to compliment me and give me the biggest compliment..a tip! Especially when professional musicians give me their respect. It’s so informal. I play what I want when I want and only as long as I enjoy it. I did the bar scene for 7 years and just as we were starting to rise through the crowd, disco came in and live music almost completely stopped on a local level.
@@5400bowen Thank you 😄 LOL I remember those LA Pretty-Boys - bane of our lives (all that hair!!! How could they?) Real people like music to be real, which makes me wonder about who is collecting the stats for today's music.
@@IanLuckett yes, but I’m talking 1983-85. There were some seriously good musicians in LA back then. Anyone remember Taxi? They were SO good and it still didn’t matter. And they all looked like movie stars. That along with disco DJs taking all the gigs was the last straw for me. As was the club owners getting ridiculous. You miss ONE night of ANY gig and you would never play a decent room in that town again.
@@5400bowen It was pretty much the same in London (minus the sunshine, and the Leopard-Skin trousers). The Dish-Cloths... sorry I meant DJ's, did the same over here, and they were cheaper to hire than a five-piece band! 😄 I'm afraid Taxi are not known to me... Ah, such is life! The thing I like to remind myself is that we humans, unlike A.I., can grow spiritually when we play music, and there is power in that, the kind of power no auto-tuned, grid-locked, auto-generated music can ever hope to match. Keep playing my friend, and keep growing, that is our best defence and attack against those soulless melodies 👍
@@IanLuckett this is a reply I just made to the comment after this one: beauty is in the ear of the beholder. I don’t think AI can synthesize new styles of music that are even close to what humans do. And I don’t care what machines do, except serve me physically. I want people art, not machine art. But I’m prejudiced..
as a musician I think of what Tony soprano said once - "It's good to be in something from the ground floor, and I came too late for that, I know. But lately I've been getting a feeling that I came at the end. The best is over"
Not at all. AI will make some great-sounding music, but a decent composer will still create worthwhile works. IT all comes down to MARKETING (which means NO human - that requires royalties - will be marketed when an AI writes for the cost of a kilowatt of power.
Heard this from a friend the other day: "AI was supposed to help us with the boring and unpleasant mundane tasks so that we have more time to be creative and compose music, draw, etc. but the first thing it took from us was art. So now AI draws and makes music and we will have to keep doing the unpleasant stuff". Alas, truer words have never been spoken.
It didn't take art from anyone. Most people just never cared about art in the first place. The "artists" had a monopoly on it so everyone was forced to deal with them, and anything that made a profit was enough to be considered "art". Now the wheat is being separated from the chaffe.
"I never saw no miracle of science That didn't go from a blessing to a curse I never saw no military solution That didn't always end up as something worse" - Gordon Sumner / "If I Ever Lose My Faith In You"
I hear the 'echo' you describe and man is it weird. The best way I can describe it is, the feeling you get when you open a car window driving, and you get that pressure pulsation from the wind. It's like that but with much less sound pressure, at a higher frequency. I feel like it's some kind of phase cancellation due to how the AI produces reverb and harmony.
Interesting analysis. Might indicate that not the AI vocal necessarily is the culprit of that "phase echo weirdness" some people hear but the AI still has to learn about mixing and mastering from scratch, not just throwing everything in without further tests as to how the instruments , voices, drums interact. But for that you'd have to have stems.
I could not hear the problem but I am older like Rick, I asked my wife with better hearing at least for high frequency sound and she said it reminded her of auto tune, she is not an audio person so I think this was her best shot at trying to explain what she heard. I can hear stupid amounts of auto tune but in this case nothing.
The best way I can describe it is: Go into an audio editor such as audacity, and stretch the duration of a sound file without affecting pitch. The computer has to make up audio to account for the time difference, which makes it sound robotic and low quality.
Hey Rick. I'm a sound engineer and a couple of months ago a client brought me his background vocals made by AI. They were different "characters" male and female. Yet on all of them there was a frequency stood out in all of them that added up in the sum of all backgrounds and made it sound horrible. Your son is on to something. Incredible hearing that kid.
I'm pretty sure I can hear it too. It's a weird phase in the midtones. You can kind of hear the "stitching" of the sound bits, etc. The metal example Rick played had the clearest of this. It freaked me out when I got actual chills listening to the harmony choices in the country tune. Chills to AI music? That was a first for me.
Former sound engineer, I've played around with some AI stem-track generators to pick apart songs that I want to analyze, and I'm wondering if there might be some overlap. The "demixers" will rebuild the stems based on what the different instruments "should" sound like, having been trained on various isolated tracks to build the AI model. I've noticed that if I restack the stems in my DAW, there is an over-excitedness that just builds up and makes everything harsh AF, but there's also seems to be a lot of detail missing around 1.6-5k. The breath, the throatiness, the spank, it'll all missing. A lot of detail goes on in that range that makes voices and instruments sound real. I'm wondering if the song generators are building their tracks in a similar fashion, creating individual tracks based on a model with a similar type of training. Dylan's right, give it more time and training and we won't tell the difference.
Like when Michael Jackson would say leave the breaths in, I want to hear the human parts whether they be slightly out of time or tune. It’s what makes music real.
Ironic; music has been more and more cleanly-shaven and overproduced, Autotune and all. We are all acoustically used to it. This squeaky-clean product is much easier to also algorithmically imitate. It’s almost as if we made it easy to screw ourselves over.
The reason it’s so hard to tell the difference in some songs is because of all the vocal manipulation we’re already hearing with most singers. We’re is the pure vocal talent with a little delay and echo.
I was going to make my own comment to say something along these lines but I'll just reply because you're pretty much making the same point. I think this might initiate a push back to purely acoustic music. Good luck to an AI trying to emulate a stunningly gifted pianist or violinist. You cannot fake the real thing
You're kidding yourself, AI doesn't struggle with emulating complex pitch movements. It's like saying a printer would struggle more to print out an oil painting than a photograph. AI will copy any style you feed it, it's weakness is innovation not microtuning.
@@markcheetah4960 You're misunderstanding what I mean. Real voices and instruments have microtonal structures which AI has no problem emulating, so that's not the reason that it's so hard to tell the difference at all (per the original post). The AI's not doing any tuning either, it's just building the sound 'pixel by pixel' the same way a printer does, and complex microtonal structures are no real obstacle to it.
In the 1950's we outsourced reproducing timbre to tape machines. In the 1960's we outsourced projection to amplifiers and location to reverb effects. In the 1970's we outsourced rhythm and pulse to electronic circuits. In the 1980's we outsourced drumming to drum machines and recall to digital discs. In the 1990's we outsourced being in tune to Autotune and recalling mixes to DAWs and computers. In the 2000's we outsourced field recording to sampler packages and creating strong performances to looping tech. In the 2010's we outsourced high performance orchestral, vocal and instrumental performance to sophisticated samplers. In the 2020's we outsourced musical performance, skill, accuracy, versatility and taste; This year we outsource composition and bypass the need for any meaningful human involvement in music making whatsoever. "Music is its own reward" - Neil Finn
Excellent summation of the ignorance of humanity- always finding an easier way to get to the end product. The problem is, by doing so, the product is no longer the result of human creativity. Human creativity is imperfect and flawed, that’s what makes it beautiful. AI will never be able to re-create human imperfection.
And yet there are still live musicians playing live music to live audiences. The technology to replace live players with black boxes has been around for over 100 years, and yet there are still live players playing live music to live audiences. I confidently predict that in another 100 years there will still be live players playing live music to live audiences.
I was born in the late '80s, loved technology all my life, first gen to grow up on the internet, in one year the assault tech is making on creative workers has turned me off of looking forward to the future and realizing that no, we're actually going to see some of the most dystopian predictions come to pass, because this tech is all in the hands of sociopathic, predatory companies with no vision for anything but the next fiscal quarter. I am so disappointed that somehow all the potential came to fucking this.
It's always actually been about money. If J.S.Bach didn't write something new for the priest every week or if the priest didn't like what he'd wriiten the week before he would've got the sack ...... If The Beatles stopped writing hit songs then the record company would've dropped them .....
@@Tom-hk6ub While you're not wrong, the problem is that it's taking the human element out of it entirely. We should force record labels and streaming platforms to prominently disclose when something is AI generated.
This is a very sad day for true musicians, im honored to have been able to tour and play in front of large crowds opening up for eighteen visions and bleeding through before you tube existed and partying.
Neural nets just not good at each separate instrument line, these are automatic composers not virtual instruments. But if you get good pre-recorded instruments / voices with RCV (software to write vocal lines using coders designed for better voice output), the result product is indistinguishable for casual people.
"I wrote a song using AI", that's probably the most delusional/annoying part of this. They really think they wrote those songs. It's this new wave of "AI artists".
Lol exactly. Fits perfectly with this generation of entitlement, though. LeBron wants to be the best without the resume, and with hand picked teams. Kids want their student loans forgiven. Free rides through life.
Live music is dying a slow death even over the last 30 years with pokies and djs, drum machines and computer simulations well and truly making it impossible to earn a dollar and the drive to improve and work towards these things can't be overlooked. I feel like crying when I think about it. Playing and discovering with others is truly a wonderful thing. Imagine the lone drummer in the jungle never hearing a real response or no one smiling with joy at that improv from the depths of someone's soul. Please NO!!
I heard this before, back when MP3's and Napster came out... they said the same, live music and concerts... that is until the AI robots start walking among us.
Yes. The replacement inflection. We are being replaced right now. All very interesting, amusing- not. Very, very scary. The machines are taking over, have taken over.
The beauty of it is, as a musician, I make music for myself, I enjoy the process and satisfaction of creating what I like. If other people like it, thats a bonus. AI may create a destination, but will never replace the journey.
Personally, i will boycott any art or music that is AI generated. I don't care. If it is AI, i will just not listen to it or spend any money to see it.
@@gclip9883yes same - i think maybe for idea generation for certain things (like blog post ideas or marketing tools) it could be useful, but not art. And I think that because I know art in any medium isn’t just intelligence, it’s a whole lot more. I hope they have some sort of legal requirement to state whether a piece of music or art is AI or not as the technology advances and it might be harder to tell.
This is why it's so important for people to get involved in making music themselves; they need to connect with that part of their humanity. As a school music teacher, I hope people continue to see the value in kids singing and playing instruments rather than simply being consumers.
June 21st is MAKE MUSIC DAY! As a flesh and blood gen-u-ine human being musician type creature - Grab your axe! sing, play, make HUMAN music that real flesh and blood can hear, dance clap or just be human to! 🎶
Where are there school music teachers? I thought they did away with that years ago. That has been my explanation for why new music on the radio is so bad...
With AI, "real" music might go back to where it came from: To people making music, alone or together, for their own pleasure or as a social (inter-)action ... But it will be harder to make good money as a musician. I guess there will still be live concerts but on platforms like TH-cam or Spotify most will be replaced by AI. Because its cheaper and as background music it is good enough for most listeners. I think it will be the same with so many aspects of human skills or knowlegde. AI will be able to do/know it better then most people. That's why I tell my children, that knowledge and skills have a value in itself. Maybe, AI is even a chance to free learning from the need to be economical.
The thing he keeps talking about is the difficulty AI has with vocal generation. It sounds like a mix between hyper pitch corrected vocals in harmony all sung by the same person and a vocoder. The other thing I find fascinating is the that you can kinda tell the lack of artistry. Like it's making high quality sounding music but there's frequently some arrangement decisions that songwriters and arrangers typically wouldn't go for. The reason for it I think is that when a robot is arranging a song, it's trying to be an amalgamation of the source material but when a human is making a song the goal is typically a stronger dopamine or emotional response. AI music just isn't selecting for the same criteria a songwriter, arranger, or producer would be because of the underlying mechanics and that gives good sounding but inferior results. But I'm sure we'll fix that soon enough.
Hope NOT, or get ready that ALL of the music will sound like Tailor Swift. Lifeless, bland, pulp... Beato has good comparison between the rhythm machine a John Bonham, playing the same pattern: NOT EVEN CLOSE. Lifeless, as month old dead fish on the beach.
These very flaws you describe would be strengths in writing AI instrumental New Agey background music. I don't think we'll EVER see an AI write Bohemian Rhapsody. But watch out, Enya and John Tesh!
with text to image prompting they use weights to help with that idk if the audio models can do that, but (artistic:1.5), .. or whatever could already fix that to some extend
I spent some time with Udio when it first came out, and actually turned out a handful of tracks that I ended up really liking. In fact, to the point where I found I was listening to those almost more often than other stuff. And it reminded me of something someone said a little while ago: If people are saying they are lonely now wait until everyone has their own music generated specifically for them that no-one else knows, and the common language of shared music disappears. Strange times.
But isn't one of the best pleasures to share music with someone and seeing it growing on them? Going to a gig with a friend and both feeling the same energy?
@@CrappyProducts Indeed it is. I guess the AI approach means that the days of saying "did you hear that new x track" after hearing it on the radio etc are gone, since unless you specifically tell people about it no-one will ever hear it.
I don’t think musicians will disappear, the emotional connection of a live show will prevail. If anything in the future there will be less competition and I believe people will search for a genuine experience that can only be delivered live. Spotify already pays artists near to nothing, we need to deliver impactful experiences that cannot be replaced by Ai
See, maybe I am being overly optimistic, but I keep coming back to this inner belief or trust that, ultimately, art is an inherently human expression. Content and entertainment is an entirely different thing, and AI will certainly monopolize on that. But I believe there will always be a segment of music and art lovers who value a piece of music or painting because they know a fellow human being poured their soul into it. For example, if someone is going through a break up, I can’t help but think that most people would gravitate towards a song written by a human who went through the same experience, thus making the painful experience a shared one. On that level, the utility and value of AI music/art will ring hollow..
Rick, you briefly touched on a very important aspect of this trend: Over the last decade (or more), human voices have been so processed, so manipulated, so quantized, that producers have trained our ears to accept what sounds like computer-generated music. Now, computers are trying to emulate voices/instruments, so our ears are already accepting of it. Auto-tune is horrible, but everywhere. AI music is the aural equivalent of "plant-based meat".
Plant based meat did happen...but I think what has actually happened is people stopped trying to make vegetables taste like meat and made more tasty and interesting vegetable only meals....this will also happen with music.
I take photos. The exact same conversation is going on in my community. At first everyone can spot AI. Then it tricks half the people. Then it gets to the level where it fools 95%. Mission accomplished. There is no stopping it. If you are a creative in any sphere, there is someone who has tasked an AI program with "learning your craft" with the end goal of replacing you.
I've spent a lot of time browsing Ai photos. They all start to look the same after a while. So far I'm not that impressed with Ai's photographic output. Similarly with the music Rick has just played. So it can generate music in the style of Ennio Moricone, so what? I'll be impressed when it comes up with completely original styles instead of simply blending what humans have already created.
@@workdevice7808 It's only a matter of time. These programs learn exponentially. The input cues will get more finessed as it learns from massive amounts of social media input.
True but for a lot of jobs that's not needed anyway. Imagine someone who does TH-cam thumbnails or someone who does posters for movies and galleries, those sort of things can be easily done with ai, and they will actually look good because there is no natural aspect in them. As for photos, I also don't think that AI is anywhere near good enough as of now but that will change too, why wouldn't it? @@einsam_aber_frei
I can hear it. I am a pianist. To me it's like a slightly flat piano in tune with itself, even the voices, with a sustain pedal held throughout. Articulation, clarity, and nuance are lost. They may be able to change it but they would likely have to revise their programming. Great show!
It sounds metallic & warble-y to me but your sustain pedal & tune comments are perfect!... I hear it too! It is in tune with an out of tune instrument and really grates & makes my ears prick up. Anything out of tune has always felt like nails on a chalkboard... Sucks bc my own singing voice is horrible 😂 but AI sounds worse to me😂
Years ago, I could smell Auto-Tune like a bloodhound. It was a Michael Buble Christmas song playing on a store's speakers. Irritating. Still can smell it.
For the first time, Im happy to be 75. Why, you ask? Because I was fortunate enough to live through a time where great music was composed by genius’s of my time. May I say, Lennon, Dylan, Page, at the very least? Im so sorry for my great grandchildren musical future. But I have left my playlist and encourage them to listen always. Long Live Rock n Roll! 😊
I am 61 and I believe that if you are between the ages of 70 and 85, you may have lived - and more importantly been at your very best - at the pinnical of human times. That is if you avoided Vietnam. Everybody enjoy the long slide down.
Damn, that word gave me a wicked flashback. Back in the mid-90s, Australia was piloting broadband internet. The advanced early users prototyped lots of P2P services, but management got fixated on "content" e.g. IPTV. I despised the word ever since.
Rick… I’ve been a VO guy for nearly 40 years, my bread and butter being corporate training courses, instructional videos and HR tutorials for new hires. I thought I could make it to 70 yrs old while still working in the studio. Ai came too quickly, and I‘ve now lost a couple of clients I’ve worked with for 30 years… that just called and said “bye”, we’ve been told by corporate that Ai is “good enough now” that they can stop paying VO people. Those 2 contracts were worth around 50k. I have not been able to find a way to replace that income, and I’m freakin’ hurting! Same goes for my friends & colleagues who work in fashion/commercial photography/videography. These greedy companies no longer need a model, a shooter, grips, lighting, wardrobe help, location scouts, and more… all because Ai can create a more realistic human image than ever before. When we allow computers to supplant us as creators of art, who do we become??
Hey Man. I’m soo sorry. I work as a producer in film and video - mostly commercials. Our world is falling apart too but for what reasons I don’t know. But back to you. It just kills me. I’d never dream of using AI for VO (Cheesy) and I feel I can tell all of the AI vo out there, but when the client won't pay we are stuck. All of the MBAs just want to lower costs and pump out volume - in everything and now advertising. Then they keep the money/savings. Totally broken and ultimately no respect for the artist. Do you know what this kind of says? It used to be that advertisers wanted and needed good art to sell their products. Now they don't. To me, especially when they say "AI is good enough" it means that the advertisers don't respect the customers. It's like saying that the dishes in a restaurant are clean enough, the meat is cooked close enough to done, etc. Thank you for sharing. I think it’s a warning to the rest if us that slowly real creative jobs are going now that life is virtual. Very best.
The thing that I am seriously asking my self here though is, how the technology will "progress" when there are less and less human creators out there actually making high quality content. There are countless of professionals out there making really good content, be it with music, writing/texting or visual art. And it takes a lot of work. And someone has to get paid for it. But, if no one pays anymore for it, using algorithms to get it all, for alike a nickel or almost free, from where will the algorithm get "new" content to learn from? As of right now, those AI models require an insane amount of data to even work somewhat decently. But with more ai content being on the net and with less human content to scrape from ... what will those models do in the future? It will all become a copy from a copy from another copy and another mix and that's it. As impressive as those algorithms are, in some situations, they can not create truly content on their own. That is a very hard technical limitation and there is no real work around on that.
I’m an indie publisher and I tested KDP’s AI tool on my reader audience. They absolutely hated it. I thought they might, but the intensity of the negative feedback surprised me.
I work in VO as well, and there was a sudden and strong dip in the amount of corporate training/industrials. Then they started to bounce back because businesses were realising that their targets were turned off by the sound, but they seem to have disappeared again. The companies don't want to pay what they've been paying, now that they've seen they can pay much less. That's the issue; it's the companies not wanting to pay for things, and they certainly aren't passing those savings on to their customers. We had a recording studio contact us about training voice models which they'd use for NPC parts in small games. Again, I understand the financial concept: small game developer wants a cheap way to add voices, but I said how much money is coming out of what I would have made in the future?
I've broken apart the stems for a bunch of these tracks and what your son is hearing is twofold. First there's a high pitch hum in the 10k plus range. This can be filtered out with proper mastering. The second is an automatic synth-like harmony that has a slight delay and is usually a fifth apart from the main line.
I like your explanation (better than mine): Here' my non-muso, non production description. Which you may find has a little too much waffle? "Quantisation. Its very faint, yet enwraps (is that a word?) the soundscape, especially the vocal performance. As vocals rise and descend you can hear the 'chop' of a quantised pitch. Its not an analog - continuous - parabolic curve. Particularly the trailing exhalation of breath. A fraction 'off'. Its easier to hear with good headphones."
I think the question is, will anyone care? Since most people will not be able to discern this, they will be fine and what Rick says will become true. Who needs real people making music anymore when you can get it for cheap or free and make money at it?
I've made a couple AI songs, and I have noticed the high pitch hum. I've also noticed that the longer the song is, the more likely the high pitch hum gets louder.
@@classicmartini While all what you said is true, none of it separates it from highly processed 'human' vocals. A combination of autotune, precise editing, and tight gates will produce the same effect.
I have a very good ear for this and have been obsessed with the subject for a few days. Check out Sore From The Door by "Beats by AI". I think the AI model used here is most noticeable. It's the way sustained notes are executed. We need to pass air with a diaphragm over vocal chords. The AI tries to replicate this by simply extending the note. The note, however, is a moment in time a voice is singing so it sounds...like it's sampling the same voice? Basically inserting the sample perfectly over and over as long as the note is sustained. It's an odd effect that the human voice does not produce. Our voice sounds different over time depending on how full or empty our diaphragm is, richer or weaker as we run out of breath. Vibrato is also a problem for AI and the "sampling" is quadrupled as the AI tries to switch between semitones. Yet I believe with a few tweaks, it will be unrecognizable within a few years. Crazy.
What makes great music is the natural nature of the musician: the mistakes, the break in the voice as it hits the high notes and the accidental feedback of a guitar. The best recorded music is often that which is recorded live or as live. We crave the humanity of it because we are human not bloody robots!
exactly, music is so over produced nowadays that it doesnt even register as music to my ears. Just sounds like production. whats sad is when i heard raw recordings of modern pop music, it actually sounds good to me, and makes me realize there's a lot of music that i would enjoy if it wasn't over produced. I think they ruined music
Unfortunately, to the masses, none of this matters anymore. People are getting lazier in every aspect of life. I can't imagine another "Bohemian Rhapsody" coming along any time soon.
I'm 24 years old and I mostly listen to songs from the 1920s to 70s and I have met growing number of people my age and younger abandoning modern music and going back to the oldies. As a music lover I have hope.
I am 48 years old and I mine songs that stand the rest of time regardless of genre. A good melody is a good melody. So can not reproduce someone getting "lightning in a bottle" and serving that up in a well written and produced work of art created by feeling humans
What's weird is that AI can generate this kind of vintage music too. The trick that's missing is getting it imperfect enough but, with the right DAW plugins, you can get close and the gap will close as time goes on.
I am that unknown musician who, over the course of 4 years, has been coming up with my own original, unique concept for my sound and material, it was a very subtle and complex process, when you try to combine the incompatible and make it sound familiar, it is a lot of suffering. But now I have received a clear message from this world that I am not needed here. My ego and individuality were literally damaged and dissolved beyond repair. But I feel that deep down my love for creation and for music does not depend on any external factors or even listeners, it warms me.
I feel for you. But in a sense this changes nothing for people like you. You used to be competing with millions of other artists and all of recorded music history and big name marketing firms which was a seemingly insurmountable task anyway. Now this is just one more. There will always be people who want that personal touch. AI did drawn art much faster than it could do music, and artists still exist. If you were doing it to make money, YUP AI will be a big dent in potential profits for all human artists the same way it has hit the online freelancer market. But as an old guy I can tell you: the sky is always falling and the apocalypse is always coming. I am a writer and we have survived online piracy, electronic libraries making all old books instantly accessible for free, competing with dozens of new forms of entertainment when we didn't have to previously, we're surviving AI too. There have been millions of other artists. There will be BILLIONS more in the future. AI is just another big number on top of that already near-infinity. However there is one thing that none of those artists have ever had, or ever will have, that only you have: You are you. Lean into that harder than ever and you have no reason not to continue.
Musicians often face that broken ego at some point of our lives. The faster you learn the lesson that music making shouldn't be related to your ego or worth, the better :) It'll pave the way to making music a happy and rewarding part of your life and not a burden. The truth is, none of us are truly needed as musicians in the big picture. Not even as humans. You just happen to reach certain people that connects with you or do not and hence you become needed by them, perhaps even for some short amount of time. And it's ok and quite freeing imo! Take that 'damage' as an opportunity to be reborn in a new way where people's judgement or loyalty is irrelevant and you can be free to express yourself in whichever way you want. Like you say, your love for creation is so fundamental and deep that doesn't have to do with playlists, listeners or success. It's the only thing that matters. And if a meaningful success shall come (in whichever way), it'll only be worth through the authentic and honest expression that comes from that love for creating. Anything else can actually turn into a golden hell. Regarding AI, it scares me, it disgusts me, and I've decided to turn my back on it, regardless of what other people may say or how far it will enter into our lives. I'll reject it as far as I'm able to, even if that means becoming ostracised, which I don't care because I'm not worried about being streamed or known widely. I'll just continue doing things the way I believe. Best luck to you!
I have to confess. I tossed on Neil Young Live Rust and cranked it up on Powderfinger. Terrible singer, instruments barely in tune, no idea tempo, total garage jamming - and I got chills and a few tears came to my eyes. It’s been a long time since recorded music hit me like that. I miss that sound in my life.
EPIC. I understand! AI will make humans seem very deep and gritty and interesting. At least in the near-term. Let it kill off all the mainstream tripe.
When you perform in front an audience, record the session and then you can sell CDs and copies of the performance to people in the audience. They will pay knowing it is not AI generated and that they were there. And you get money for every CD/USB stick that you sell. Without having to go through a big company, you get to keep all of the profit. You probably know all of this anyway. Good luck in your career.
Real human music does not have to stop. It will be a niche product that sits alongside a voice controlled AI media centre, where the listener creates expressive music by just asking for it without having to learn the skills.
The reason we (At least today) recognize AI is that it sounds as "tinny" fake as the auto-tune stuff that's coming out of the major music labels. The voices are already so manufactured that *no one* sounds natural.
I was coming here to say this. The vocal tracks all have a tinny note to it. Kinda of like when we used to download a low bitrate song and it almost sounded under water or something.
That's because AI will never be original or creative, just like today's top 40. If AI existed in the 80s and we stopped there, we would've never gotten grunge, hiphop and the kind of pop and R&B we've got in the 90s and 00s.
@@miked7295 you know who else was 'trained on previously explored things' ? humans. difference is that we create by taking ideas we've seen before and blending them with other ideas we haven't seen them blend with. AI will certainly be able to do that very well in the future.
The fact we're left with such impressions from AI just confirms most of the music we have online is so generic even AI can do it better. Just another motivation to strive for originality.
Within 10 years you'll be able to ask for an AI song that's 60% Dire Straits, 20% Chet Atkins, 10% Albert Lee, and 10% Bob Dylan + a clever song title and subject. The end result will be amazing, but if it's not perfect you can record any little fixes in a kind of scratch track and ask the AI to rework it with your edits.
@@NoDaysOff-oz2zl I think you underestimate the average banality and ignorance of people. If Rick was wrong, then progressive music would be the most popular, or acid jazz, or classical. Instead, the most popular music is the simplest, easiest, most insipid tripe. It always has been. Generative AI allows crap to be churned out at a ridiculous pace, and that's really all corporate wants...it means more profits.
This is like what I call a "Plastic Vacation." We went to Disney World for a week, and had a great time....but it was "plastic." Then went to a cabin on the shore of Lake Superior in the U.P. of Michigan. Enjoyed that too...but on a much different and deeper level. AI music can be amazing. But it could probably make us overly critical of real musicians in a jam session or coffee house environment simple because they're less than perfect. But if you want the real thing, get out to a jam or coffee house type of venue where you can experience the real thing.
I'll tell you what I think. I think the minute that egregious auto tune use was everywhere was a big indicator. Auto tune was initially used to fix bad vocal notes for the most part. It was meant to fly under the radar to grab a moment from a great performance moment and save it by a very inconspicuous move. But for some ungodly reason, Producers everywhere took that and said "lets make it totally audible". That's where it got interesting. And really stupid. Every producer / artist in L.A. was trying to out-do the other guy and that's EXACTLY what'a already started here with A.I. So yes I totally agree... this is going to explode and put many great artists and composers out of business. It will ruin the ENTIRE creative music industry at the same time ... think of The Beatles. If they had A.I.? at that time, imagine them saying "well this is easy - lets write an LP". Now you have taken the human element away, and they would have had to have written based on OTHER ARTISTS. So no Beatles. They would have been a combination of the bands they wanted to mimic, the artists they loved. They would have sounded like a combination of Little Richard and Carl Perkins or something. So if that's the mentality that people want to embrace then God help the people of this new generation who expect to hear great original works from artists who are living geniuses in the making. God knows how many incredible artists that will never even try now to search their souls to create "I Am The Walrus". This will annihilate the art of making music. But nobody cares because of the bottom line. money. money comes before anything in the universe. Easier. Simpler. No brain activity needed? and a paycheck for doing it? And you're right Rick how the hell does anybody track this? Like the "intellectual property" and so on? They let a lion out of the cage and have no idea what will happen. Modern human behaviour never ceases to blow my mind. It's always a half-cocked race out of the gate for people with no sense of direction ( Monty Python ). Here comes the train wreck. We can either stop the train or get on it. My tiny part, just for my own sanity I'm going to do anything I can to stop it, for what it's worth. I hope there are like minded people out there.
Rick, as someone who has been thru a lot in life, music has been a great source of salvation in my life. No fan of auto-tune. The flaws, the scars, the errors, the vulnerability, are the beauty in the art. The thing that touched me and moved me about many songs and music was knowing someone else, somewhere, another human being, was feeling the exact same things and that I was not alone, no matter how much I felt that way. Discovering music that touches a nerve and describes what I feel is magical and transcendent. I dont hate computers nor am I a "Luddite". But stop messing with my music!
I think what your son is hearing, is that the vocal segment sounds like it was sampled with a very low sample rate... Which has a unique acoustic quality to it. It's as if the vocals are recorded on an AM radio...
All Ai generated music tends to have that mushy audio quality because its not generating individual tracks, its generating a whole audio segment at once. That's where the artifacts come from. If you look at Ai-drive software like SynthV, the vocals can sound much more realistic because they are isolated and clean. More importantly, it gives you control over the notes and how they are sung. So I can basically dictate to the Ai how I want it sung and it will do it. That's where Ai is headed IMO. More tools.
I'm totally unable to make a difference between AI and natural voice, but my wife says she 'suffers' when she hears AI voice in any media. Like they switched the news reader to AI voice in some local radio stations and I couldn't tell the difference except in rare occasions when the AI couldn't pronounce words properly. However, my wife immediately switches station when AI news begin. I can't tell either, what it is they can hear, but sure as hell I'm deaf to it. And I totally agree, in just a few short months, neither will anyone else anymore. The development is so fast in this area. Sadly.
Listening to AI speaking feels like chewing a meal that's made of paper. I hear it, but cannot remember anything it said. And it's the same with AI 'text', 'music' and 'art'. It is not the real thing, and never will be.
I am a full-time piano teacher, now mainly online all over the North America. Thank you for your efforts to bring attention to this composition/ musician killer.
Initially as a 75 year old pro musician since '63, I am SO glad I lived through this "golden age" of music/sonics/gear/sound production. I can't really imagine the world not too far ahead......
I wonder if people will know what real music is. I wonder if songs will be half as long for twice the price. And if it will be illegal to record your own singing voice or instrumentals. Did you know it is illegal to capture rainwater in some states? Rainwater. A free, natural resource.
@@velvetbees Which states would those be, genuinely interested to know. The thing is, some people may not be able to imagine the world not too far ahead, but the point is SOMEONE will have to live in it, like my son. Throwing up your hands about the future and saying, "well, at least it is someone else's problem, not mine because I'm older" is half of the reason we're in this predicament as humanity.
@@velvetbees Well, I did a simple web search, and found out collecting rainwater ISN'T "illegal" anywhere in the United States...but it is regulated in 11 states, for different reasons. Sometimes humans can be as disingenuous and deceptive as AI text generators. Really, reading comments on TH-cam filled with nuggets of "wisdom" requires just as much skepticism as always.
Scary times ahead. The scariest times will be when no one can tell the difference. The worst thing that technology is doing(IMHO) and has done is making the "art" of creating music way too easy. Too many apps, AI, recording software etc. The REAL creators of music will be lumped in with all the non creators, ending careers and losing the soul and art of real songwriting, performing.
Yep, the worst thing that people wouldn't be inspired and interested in music, cause it's too easy to create your own song at home with just some simple prompts
@@hugorodrigue3131 thats exactly why there are dj sets and not just music playing from a playlist. People want to watch someone regardless of what the music is, the best bit with most DJs is they are playing other peoples songs so I don't get the appeal but it must be there as DJs appear to be the modern day rockstar.
You're forgetting that most hit songs are recorded in a studio, no live audience required. You could always tell it to make it a fake "live" recording I suppose.
I'm afraid for what is happening to the world and for what my grandkids and young people might have to endure. Everything is changing too fast. For music and graphic artists what will be left for them? Soon robots will be doing all the work too. Maybe we will become like the people in the Wall-E movie.
@@paulgentile1024 I'm afraid so. Even in my profession.. software engineering, AI is starting to make a big impact. Not entirely yet but probably soon.
I firmly believe this is the best moment for us musicians to remind people the power of live music. We sure had a good time earning some money through selling our cds, or even streaming royalties but, for better or for worse we’re all going back to what it used to be: getting paid to let an audience hear, connect, let loose and feel live music. AI has been changing the world a while ago and it ain’t stopping and it ain’t gonna feel sorry for anyone… it’s better to get ahead of the curve, learn how you can use it creatively too and maybe even make some money while at it too, why not?
I can see people dancing to a salsa band. Or New Orleans zydeco music. Or bluegrass with fiddle and mandolin. Or Celtic pennywhistle. Can AI create the in the moment of that? It never will! But where will these musicians be if they can't make any money in the music biz because the biz replaced them with computer music? Playing on a street corner?
Yea, there is a group in our country who call themselves a sequencer group…they have a laptop playing all the instruments and they sing karaoke style in their concerts.. one of them pretend to play guitar.. and on closer listening, their voices are digitally processed. Kind of “live AI”😅😅
KISS just started to destroy your thoughts about "power of live music" with their "digital avatars" thing... I feel bad about all of this, honestly, and can't seem to find a solution where PEOPLE will still be making money out of their own creativity (except those who get creative at prompting, maybe? Which also won't make a lot in the long run...)
I hate this is happening, HATE IT!! I’m not a musician, just a lover of music and I’m in awe of musicians! The dedication to learn an instrument is inspiring and I admire anyone who learns an instrument and I am moved to tears sometimes by the masters!
When computers started beating the best humans at chess, did we all stop playing chess? Why play chess at all if a computer is so good and can beat any player on the planet? Somehow chess is more popular than it has ever been in history. Why do we want to see people play chess instead of computer programs duke it out? The human to human connection music gives us can never be replicated artificially, just like seeing two humans square off across a chessboard is an authentic experience. Like any technological revolution, this is an opportunity to embrace new tools. It does introduce new challenges, of course. However, just as you stated, people are still going to be looking for that human connection. Perhaps we will come to value authentic performance even more than we do now. I'm sure that there will be a holographic band AI tour some day, like there was for Hatsune Miku. Those will be fun in their own way. However, we'll still want to see live bands play the songs we love. Maybe the impact of AI is that live in-person music is in for a major revival.
I’m actually pretty happy about this, as AI will only replace the soulless pop that has become som mainstream now. My hope is that AI will make mediocre music abundant, and people will go back to making proper music
As an old old (now retired) engineer, I spent much of my career trying to persuade people that "Just because the technology exists, you don't have to use it."
Well, what technology? Where do you draw the arbitrary line Brian? No, don't use computers to edit songs because it's way too easy. Yes, you can only edit by manually splicing tape, this is clearly the better way to do things, right? Or wait, maybe it's just the use of electricity at all? Or are you referring to quantization? Tuning vocals? Wherever the line is, it's probably arbitrary, and you're probably a dino.
@@travisyee7278Smart Fridges that can post on Twitter AI Doorbells that tell you who's at the door AI in your computer mice that does...idk?? NFTs Cryptocurrency Blockchain Those stupid swearing filter boxes from the 80s
My fear is that the world will be pumped full of this AI stuff and the real creative will be lost in the crowd. I'm a little older than you and your channel has caused me to hit the record shops again, getting all the albums I couldn't afford when I was young. Also love your artist interviews. We live in a wonderful time when I can hear stories directly from band members. You're a treasure. I'm now addicted to your channels and Tim Pierce.
Rock bands need to record albums like they did in the early days. Everybody playing in the same room and leave the little mistakes. Good old fashioned Human Intelligence. No more auto-tune, edits, or quantizing.
I've always been strongly critical of autotune and all other technologies made to "correct" singers' voices. The imperfections in our voice have a unique quality, a real kind of beauty that shouldn't be tampered with. And if a singer depends on autotune to sound decent then they shouldn't be singing in the first place, they either lack effort or talent (or both).
Exactly my thoughts. I guess it would be much easier recognize something AI did if it was told to create something that would sound as real and raw as possible. At least at its current stage of development.
I agree partially. I disagree with the antidote part. The only way to know for sure a voice is "real" is having the person directly singing to your ears. With no mic, no PA, no overdub, etc. And my second point is: soon enough no one will be able to tell the difference. Like with a professionally photoshoped image.
I'm using a different program that allows me to use my lyrics instead of just prompts. It has allowed me to bring my file full of lyrics to life. I've also figured out that a lot of they lyrics I wrote are garbage lol. Garbage in, garbage out, I just can't get some of them to work. The one's that do work, however, have come out pretty amazing. Some sound much like I heard them in my head when writing the lyrics, some a lot better! It has reignited my passion for song writing. Will I ever make a dime? Probably not. I do enjoy sharing them with others and having them ask me to send more
The scary part isn't how it sounds now, it is a brand new technology. The scary part is how quickly it will develop and improve. Give it a few years and it will be incredibly hard to tell them apart from the real thing.
Exactly what I’ve been saying! It’s easy to laugh at how bad some of the AI creations are, but to think how fast they have gotten better! We are still at the beginning of AI use in arts. But still it has developed faster than almost anything IT-/computer related technology ever. We used to measure technology development in years, but with AI it’s in months…
Live is the future. Not arenas or mass venues where it’s easy to manipulate the vocals, but seeing bands in small clubs where you can actually see them play, feel the energy and ambiance of that particular show at that time. Rock will devolve to the level jazz was in the 50’s; BUT the players, shows and performances will become more unique and momentous
People already all but refuse to pay for music at small clubs and small clubs don’t pay a guarantee. What do you mean by rock devolving to where jazz was in the 1950’s? That was peak popularity for jazz music, just as the 1960’s were peak popularity for rock music.
@@swperry04039Yeah. My favorite jazz music comes from the 50s and early 60s: Parker, Monk, Davis, Mingus, Coltrane, Coleman, Brubeck, Getz, etc. Jazz didn't really begin to slip into irrelevance until the jazz fusion period of the 70s, followed by smooth jazz.
I'm over here practicing 4-part, 3-octave arpeggios, all qualities, all keys, hours per day. I know people can hear the sweat that went into that. Hours agonising over one chord choice. I hear someone say they "wrote" a song with AI and I just break out with hives. I wish more people would acknowledge the spectacle of humanity in musical performance, whatever the reason, but I'm still confident that they can feel it on an intuitive level. Although I loathe things like autotune, it's not just about computers, there is so much amazing computer music created by beautiful minds out there. I'm glad the kids can hear the difference in a real performance. I'm sure they can appreciate the effort and dedication that goes into that too. We have to make sure all of us can.
My biggest problem with AI is that the people who made the music (books, paintings, etc.) are not being compensated in any way for their work being used by a company who is making money off of their programs. Sam Altman has said that LLMs can't work if they soak up as much info as they can, copyrighted or not. To that I say...so? If you're going to use someone's copyrighted IP to make money, the person who owns the IP deserves compensation.
By the same logic I should demand from you compensation for reading this comment. After all, I wrote it, and it's my IP, how dare you peruse my property without my permission? And who knows, maybe you will reuse the golden thoughts contained herein to produce your own comments, and make money from it.
@@clray123 Are you really equating a TH-cam comment (which you wrote presumably being aware of Google's terms of service) with a song that a professional artist has released under a recording label? You really think that would stand up in a court?
@@thennickeYou’re missing the point of his comment. Every work of art, across all mediums, has been inspired by other works of art. This obviously comes about through “experiencing” these works, i.e. watching, listening, etc. The question is whether a computer experiencing art is equivalent to humans doing the same. There are valid arguments on both sides here.
@@MrPlavik Exactly, the main question here is whether the copycat's production is "transformative" enough with respect to the original work. After all, genuine artists get inspiration from other works of art all the time. That, and AI companies are careful to train their models on material that they have already licensed for this purpose (even if the licensing terms are as bullshit and sneaky as Facebook's "opt out" preferences). In other words, if you think your copyrighted work was stolen AND reproduced unchanged by an AI, by all means sue that company and you will probably win in court. If your copyrighted work was just stolen (perceived) by an AI, but not reproduced, your chances may not be so good.
I agree that I can hear an "echo" or "vibrato" in AI singing that makes the artificiality apparent in styles of music where that's not "appropriate" (e.g., it's very noticeable in the country song; it's not something that stands out in the chill track because I expect electronic modulation on a singer's voice in that genre).
More than anything, this is a perfect reflection of how mundane pop music has become. Your average listener does not care enough about music to care about who created it.Personally, I don't mind if that crap is offered up for free or even takes the place of mundane movies and advertising. We are already so inundated by crap that humans make , if noone knows the difference, it deserves to die. Hopefully great musicians can realize the opportunity to create something so indivually powerful that AI cant recreate it and it would not seem profitable to even try. Money will no longer be a driving reason for creating great art. Human spirit will not die with this. Only an old capitolist definition of music will die and it's probably about time.
Musicians also need to pay the bills…how many people can have the luxury of creating something great while working at an Amazon’s warehouse… that’s the problem right there…
I’ve noticed that too among some people. And if you Play something from last year they’re like “oh that’s so old” I’m like yeah so what, it’s great and never gets old to me. It’s seems like there’s only a few that study music or actually LOVE ❤️ music - u know, like actually listen to a song with headphones on dozens of times. Sometimes I get in the car and go cruising just to listen to tunes while I drive.
I grew up with the Beatles and later played guitar and keyboards in a band. The last twenty years or so, I've been studying orchestration on my own and learning the intricacies of writing for the orchestra. I have many gigabytes of sound samples and think I've gotten pretty good at producing original orchestral music. I've long had a dream of writing a movie soundtrack, even on an obscure, small budget, independent film. But I've come to realize lately that I may never get such an opportunity, as filmmakers can just cut out the composer altogether and have it done to order with AI. The earliest AI attempts at orchestral music I heard were kind of lame, but now they are getting good. Soon they will be even better. I fear for the future of the musician. But I will never stop writing music. It's my passion.
You have to keep feeding in new music otherwise, it can never grow or morph the way real music does. Basically, AI simply imitates by comparing its output to real music to see what fits and what doesnt. If no new music comes in the AI stuff will go stale, always be stuck at some point. But that probably won't happen. Too much new stuff out there. Imagine AI training on AI. I bet that goes bad quickly.
I'm surprised you can't hear what your kids are hearing... it's hard to describe, it's almost like a subtle phaser effect over the entire track, and it shifts from left to right. It's more of a overall quality issue, and I'm sure they will improve it dramatically in the upcoming versions. My guess is that the AI deconstructs the (stolen) songs it was trained on and reproduces them with artifacts in the sample rate.
Copyright law doesn’t require an artist’s permission to study their works. Unless we convince Congress to change copyright law, these works aren’t “stolen”, no matter how unhappy we may be about it.
That’s what I hear too!! And your description of it is fitting. 👍 I do not like any of this A.I. junk. I like real instruments played by real, warm blooded people. With all of the nuances and complexities only a real being can make!!
@@JoeyFTL Didn't know aliasing was a thing in music as well but that's exactly my mental picture, like a low-res image with no antialiasing that is upscaled, everything is there but it somehow feels inorganic. But it won't be there in 6-12 months, or it will be imperceptible.
I'm 12 and I've been playing guitar for a few years and I want to learn about production. It would be my dream job if I could go into the music industry, but I honestly have no idea where it will be in 10 years.
This is from Orwell's 1984, published in 1949: "The tune had been haunting London for weeks past. It was one of countless similar songs published for the benefit of the proles by a sub-section of the Music Department. The words of these songs were composed without any human intervention whatever on an instrument known as a versificator. But the woman sang so tunefully as to turn the dreadful rubbish into an almost pleasant sound." Prophetic...
Versificator! I love that word
Good God. I need to read 1984 again! Haha. I read it in like 2015 and I was like "yeah, it's coming." If I read it now it's gonna be like "oh yeah, we have that now. Oh, yeah, that too. Fuck."
@@jevinday I first read it in 1983 in middle school, have read it every few years since just to see how we've progressed. In many ways, 1984 is kinda quaint at this point.....
Georgy calls another one!
@@fdllicks Sounds like a word Mr. Burns would use.
Problem is most pop music has already sounded like its made by robots for years
Truth
The 2010s and 2020s are awful, except for some handful of true gold songs!
Exactly. Pop music artists have been digging their own graves without even realizing it.
MUZAK
😂😂😂
AI won't stop me playing and recording music. Most of us have no audience and no money. It doesn't stop us being creative. We have to do it. The creative process means more to us.
everything you said is true. i feel sad but also hungry of creative process at the same time.
Been playing the guitar and recording it for over 30 years. I wouldnt care if I was the last person on earth I would still do it. I dont even play for my friends. Something deeply personal about it that I dont care to share with many.
As I pack up 3k worth of gear for a $50 gig - while saving for another guitar, I’m still a happy kid!
@@resignator did u make aliving from the music? Or just hobby?
Same, but that's not the issue. It's that a hell of a lot of musicians are not going to get paid in future because others will generate AI music instead of paying someone to make and play it. TV/movie soundtracks. Ads. Elevator music. Video games. And that will have the knock on effect of slashing income for studios, sound engineers etc. A lot fewer people will be able to make a living from music now.
What Dylan and Layla are hearing are digital artifacts and transients that can be very subtle but sound odd. Sometimes they can sound like a very slight echo or even chorus. You can usually hear them in any digital recreation of an analog source, like digital amp modelers. They're most evident on the attack/start of a note and the release/decay as it fades out. Most of these digital/artificial recreations of original sources get the core of the sound right, but they're still working on the "bookends" - getting the attack and release to sound totally natural. And you're right, it will get better and they will figure out a way to totally smooth all of that and it will become increasingly difficult to distinguish the difference. Where we go from there, I have no idea, but I just hope there's still a market for original/authentic music and musical instruments.
Yes like low quality mp3
When they figure out how to train AI to incorporate imperfections - the things that give a voice an identity or a song its 'feel' - is when no one will be able to tell the difference between new, experimental music from an aspiring artist, and the latest 'release' from a dinosaur music company.
That and the AI decided to have the vocals be a weird 5ths duet constantly
I was trying to use adobe's podcast voice clean up service a while ago on some lower quality japanese audio, and it completely murdered the diction. Partly because of what you said and partly because their training data was obviously majority english, so it had no idea how to "pronounce" the consonants correctly.
I think there's a cap to what these current methodsof genAI will be able to do, because there will always be conflicts in the dataset that will reveal themselves to those who know what to look for. If that matters at all to the masses is a whole different question...
@@dankmartin6510 I'm worried that ai won't get better but humans will get worse. They won't add in imperfections, people will just get used to no imperfections the same way they got used to autotune and quantization.
"Go into the arts. I’m not kidding. The arts are not a way to make a living. They are a very human way of making life more bearable. Practicing an art, no matter how well or badly, is a way to make your soul grow, for heaven’s sake. Sing in the shower. Dance to the radio. Tell stories. Write a poem to a friend, even a lousy poem. Do it as well as you possibly can. You will get an enormous reward. You will have created something." - Kurt Vonnegut
yep...
Yes, exactly. For almost all human history people made art without expecting to make a living at it. The brief period in the 20th-21st century when a small percentage of artists made money is the exception.
"When I am gone, all that will be left are the songs." ( Adapted from quote I don't know where.)
@@folm235Artists historically never being paid was not something to be proud of.
@@folm235 Sorry, but this just isn't true. For much of human history storytellers, minstrels, strolling players have been able to make their living from their art. Maybe they didn't make much money, but they were able to get food and lodgings as a minimum. Also, for many centuries artists and composers had patrons who enabled them to devote their lives to their art, and in so doing, elevate their art to the sublime. Even those who didn't were often able to make a living by selling their works, if they were good enough.
Makes me proud to be a horrible saxophone player. No respectable AI would ever dare to reproduce my crap!
Your goal should be to poison the training dataset. 😮😅😂
Jajaja
@@ChiefBridgeFuser THAT'S the BEST idea! Where do I upload my terrible keyboard mashings to really screw with the datasets! 🎶🎶💀🎶🎶☠😮
😂😂😂😂😂😂 but I prefer live imperfect saxophones😂😂😂😂😂😂
@@jamesslick4790 Ferris Bueller's Day Off has a clarinet solo you could start with
"I want AI to do my laundry and dishes so that I can do art and writing, not AI to do my art and writing so that I can do my laundry and dishes"
You need to call right genie to grant your wishes ! Sadly in this world right now, another type of genie who came around and play in our playgrounds
You already have machines that do your laundry and dishes.
@@leiathedog999 speak for yourself!
AI has taken note of your irony, much like Billy Mumy in Twilight Zone.
"You're a very bad man," it says.
You may be banished to the fields.😂
People joked about blue collar jobs being replaced by robots and their computer jobs. There was a meme a couple years ago about how truckers should learn to code. As it turns out, AI is much better at doing stuff on the computer than being a trucker, carpenter or laying tile in your bathroom. It can do management, accounting and write code. Its the office jobs that are going away.
First we streamline any natural voice using AutoTune, now we can’t distinguish an autotuned voice from an AI one. Well done humans…
Its just a matter of time before there wont be any difference between non autotune and ai
...and good-bye humans
I'm sorry but we have lost all musicality since the 80's.
Rick, you need to start a "What Makes This Song Fake" series.
Good idea. I'm for the artists the ones that can play and sing on a stage, the ones who write by drawing something out from themselves. Are there people who will go to a concert and be presented with just technology, that plays AI composed music? Robots say who are programmed to play guitar, drums, bass, cello, flute, whatever? Maybe, but I think that the majority of people go to a concert to connect with other human beings playing something that they have created. With music it's the human to human connection that is important. Sure there will be a growing industry of AI compositions, recorded using AI, that will be played in all mediums, radio T.V. films, adds, whatever. AI made compositions are made by learning from actual humans. But can an AI machine ever feel, have a heart, come up with something born of lived experience? I doubt whether a machine could come up with something unique, from actually being creative. Will AI ever be able to write something as real and beautiful as 'A Whiter Shade Of Pale' or 'Pancho and Lefty' or 'Like A Rolling Stone' or 'If You Could read My Mind'.
@@eligreenslade7699 Exactly. AI, unless it becomes emotionally sentient, will never be able to feel joy, anger, anxiety, jealousy, hopefulness or crushing disappointment after crushing disappointment. These are the things that fuel expression through art.
😂😂😂
@@thespacealienssmogandgrog4283 Yes I agree entirely. I have made 4 studio albums and one 45 over a period of 4 decades. I like recording with a band, but performing live with a band is my ultimate. That's where the biggest, most immediate connection is. That's where we serve the music and we serve the people listening; where the energies come together and uplift everyone..
lol ...Rick said he could not hear why it was fake.
hey Rick, lead audio designer from Ubisoft here (and a lifelong musician), I can also hear what your kids hear. The reason you probably don't pick up on it is because you haven't worked with text-to-speech voice synthesis that much. You can clearly hear the artifacts in the vocals (especially if you listen to the reverberation, those tend to get modulated everywhere as the algorithm doesn't pick up the reverb tails well), and the overall low-end / mid-end saturation on all of the AI generated songs, in its current iteration it's really-really similar, like a tictactoe game - very basic. They all sound as if there was heavy parallel compression on them between 200-600Hz, and there's never a lot of high end. If you focus on more the overall frequency spectrum of a song (load them up in a DAW and get a frequency analyser on them, they all have almost the same curve) and how aesthetically and dynamically similar they all are from a mixing point of view, I'm sure you'll recognize these generated songs very quickly.
also I never heard of an AI song in weird time signatures, and it's always staying in the same key. 😊
@@MattGabnai Pssst! Don't make them aware of their shortcomings. 🤫
@@Zareh_Abrahamian LOL I was thinking the same thing. Does AI read youtube comments?
@@Zareh_Abrahamian don't worry, they know without me. By iteration 10 it'll be like midjourney is now for graphics - not perfect, but a very, very useful tool
You being a musician yourself, I'm curious how you morally justify your role in this type of AI.
the other day I was at a diner and a waitress took my order ...
two eggs scrambled, rye toast ... light butter, coffee, black no sugar ... home fries lightly crispy ....
I'm an incredible cook ....
haha this comment needs more likes
Perfect!
Hahaha! Nailed it!
I entered your order as a prompt in udio. TH-cam won't let me publish the link here, but the result was pretty funny.
Is there an emoji for not having a clue? That is me. You completely lost me.
What we have to remember is that music isn't the "music business" no AI can replace the joy of playing an instrument which is what it's all about.
My primary enjoyment of music comes from playing my guitar, which I make no claim to be good at. But the vast majority of people enjoy music passively, and this is particularly true of young people today. If you're a professional recording artist, you're going to be losing market share to robots.
I have a friend who works as a 6th-grade teacher. He told me that he asked his students the question, "Who want to be a famous pop star?" Many of the students raised their hands (especially the girls). Then he asked, "How many of you want to learn to play an instrument?" Two of the students raised their hands.
@@rolfedrengen I think that's a common problem with young people today. We teach them to be good at consuming. We don't teach them to be thoughtful, industrious or creative.
@@martymcfly1776 We? More like "they"
@@Ryanez93 I'm a 70 year old man, and I have been both a teacher and a father. I share in the blame.
"And the Grammy for best AI prompt goes to...."
😂 that’s good. And probably coming soon.
There are actually 4 new categories on the Grammy ballots for AI. Done!
@@TomatoFettuccinithis is not the issue. This is one of the least from all the issues.
Horrific joke that is going to come true. Only way to keep AI from winning other categories.
🏆
I been playing piano and sax for 40 years. Bands, shows, tours, weddings , parties
One thing the AI cant do is give me the satisfaction I get from playing, the nuance of a real instrument vibrating the air, the people witnessing it, the subtlety of it, as I shape the sound with the reed and diaphragm. The nuance of a life lived all going into the performance.
Live music changes lives.
The first live music festival I attended (after Covid), I was in tears. To once again hear real music with others around me was such an emotional experience. I hadn't realized how much I'd been starved for it. Music connects us in so many ways. AI music will never be able to reach my heart that way.
Lets hope people recognise this too.
That is true but the video and his point is from the point of the listener only and how most people can’t tell whether that’s you or AI.
And it's possible that A.I can be as complex to be able to replicate realistic, full of human details and errors, soundscapes in a live performance. We gotta think ahead
I am a graphic designer and people doesn't want to hire a real graphic designer because AI make a "nice logo" for free in 1 minute. I think the problem is not AI, is people want things fast and cheap, not value imagination, good taste, long years of study and talent.
The AI does have good taste and long years of study and talent - it's just that it's yours and every other graphic designer's!
Adobe will convince that AI is just another feature, whilst it learns from all your manual work. Some of their software may reduce itself to a search box requiring a detailed description of your requested output and nothing else one day
@@KnightmareUSA That is why after over 20 years in Photoshop I am changing to Affinity.
I hate to burst your bubble, but AI is going to be better than us at everything. Arts and design by humans is going to be obsolete first. Things we used to take value in like "long years of study," doesn't mean anything to an AI program that can draw from the entire history of your field in seconds. The irony is it will eventually be better at programming AI than the people who are creating all these job-stealing AI programs.
Yeah think about othe fields as well, like architecture, product design, writing, etc. @@jaredpurcell8835
A few thoughts, starting with not all AI is equal. Two pathways using AI:
1. Get good at prompting. Enter a prompt into Suno (Udio, etc.), choose the best result. Enjoy. Share. Does it sounds fake? Maybe, for now. Maybe some people won't even care.
2. Determine what you want the AI to do, have a goal, a vision. Start by inputting riffs, chords, melodies, singing, into the AI engine. Now write your own lyrics, or get help from AI where you are stuck writing them. Feed that into the same AI engine. Start prompting. Trial and error. At times it will sound kind of like you. At others, it will be wildly off. Trial and error. Use your knowledge to determine which AI outputs are best, and most usable. Export the stems from each of those. Take all of that back into a DAW, or Audition. Feed the vocal track into Audimee, or Ace (AI apps), change the vocals, create harmonies, if you like. Turn them into your own AI cloned voice if you want. Go back to the DAW. Start editing, splicing the best of the AI stems together. Add/replace instruments and sounds. Re-mix and master it yourself. Or even add your own voice, or actual instrument playing. Output it. Will it sound fake? A lot less than just out of AI? Probably. Will the masses care? They might notice and appreciate what you did. Maybe not.
This is the future. Both of these. The first one will be easier, more relaxed, for non-musicians, or musicians just wanting to have fun. The second will take a lot more effort, but also retain a certain sound from the musician, a certain identity, or "voice" if you will. That amount will be determined on a track by track, artist by artist, basis. But people may still not care that much one way or the other.
Really want to be a true musician and admired for your skill? Perform live.
Yes!! Thank you!! Real visionaries are not at all worried with AI. Imogen Heap is starting her own AI music platform. We need to use them for our advantage. And if you put soul into it (loosely quoting bjork here), it's what translates. There are plenty of talented artists and musicians out there who cannot put their souls into it.
It’s the aliasing.
You can hear the aliasing in AI music/vocals. It sounds like a slight time-stretch or bit reduction.
This. 👆🏼
Does it happen at a frequency that only younger generations can hear?
@@RickJohnson Try listening to youtube at 0.75x speed and it will be a much more obvious version of the type of sound, maybe you can learn to pick up the more subtle one in the AI songs.
The first human song in this video fooled me as being AI cause it also has a load of processing and garbled effect on the voice, although listening again it's more consistent and less 'juddery' to me than the AI.
@@RickJohnson just try listen with feeling not hearing. It's a sort of wobble between the lines.
It sounds/feels awful.
Good term for the phenomena
The nail-in-the-coffin of my playing-out-days was standing on stage - looking out at the glowing faces of people who were staring at their phones instead of watching the band.
For 90% of the people who consume music (not this group obviously) they're not even going to get through a whole song before scrolling to something else. Those people deserve the crap that AI is going to give them. I'm a recording/mixing hobbyist now. I spend hundreds of hours writing and mixing songs that no one will ever hear ....and I love it. It just feels good to be creative...audience or no audience.
I like your spirit ... or maybe I mean that I like your essence. Cheers.
YOU are the audience that matters.
last concert I was people were posting/checking their instagrams on their phones instead of watching the show in front of 'em
yes agree, i play piano, only for myself now.
Couldn’t agree more. The real art is in the persons writing, playing and mixing spending hundreds of hours exploring EQ levels per instrument. People like you, Jeff Lynn, Rick Rubin, Brian Wilson and many many others are few and far between. Keep pushing.
I am a photographer that photographs people. I post a lot of work on Flickr. There is now a lot of AI imagery on Flickr created by non-photographers of non-people. The images sometimes look pretty good, but they are slightly unrealistic. They have gotten noticeably better in the last year or so. Anyway, this is happening in all of the arts, not just music. I definitely feel your pain.
It's pathetic as AI was supposed to enhance life while we work on creativity and hobbies and now it's to keep us working and erase and silence and replace our human creativity. Clearly they want this for the masses. Gross. Seems this is all by design
Are you Harry Joseph Collins on flickr? I think you might be as that account has a bunch of music related pictures. I have been an Adobe Community Expert since 2009 and spend time on the Adobe forums every day, and we get a lot of posts on the subject of Ai. We get even more on the subject of Content Credentials, that drives people crazy! I have a theory about the dreadful job Ai is doing of faces. If it did better, and happened to render your face for someone else, and you saw your face on a, advert on a feck-off giant billboard, would there be legal consequences? They playing it super-safe kicking out even remotely salacious prompts with Firefly, so it is not a giant leap to think they are deliberately messing up faces. Hmmm... now where did I put my tin-foil hat? BTW I am me on flickr.
I think historians are going to struggle with realistic photos of non events from the 21st century.
@@cjlamberYeah, we might get to some kind of temporary "this-is-not-generated" watermark, but that will inevitably get cracked, all images, even "historical images" are now suspect, unless you have it in a mass-produced book printed before 2020 or so.
I'm a photographer too. I don't go on Flickr much, but on Instagram, there are increasing amounts of "AI influencers". While the programs have got very good at making people out of nothing, a lot of them still have a certain "look" to them, and I'm like "something's off about this person, and I don't know what it is, but I can tell it's AI". A lot of these AI people have a certain type of face that I'm seeing again and again. It's almost like it's creating an average face of a lot of humans it's been trained on.
i've been in the music biz for sixty years, and have seen a few evolutions, in music performance/recording and trendy production techniques... i can hear/sense it... it has a strange kind of drony, ringing, quirky synthesized, artifact aspect and with discerning clinical awkwardness, akin to Autotune, Melodyne, Vocoder, Fairlight, digital pitch and key shifters... etc... that is its telltale signature, ... it is funny that we went through eras, that we strived to sound more electronic, perfecting certain qualities that were widely accepted as major accomplishments and advances in the industry... it's all rooted in the business of marketing music... not the art, passion, or skill, of a living musician performing a creative experience... and now we are concerned about sounding and identifying as human... sometimes, the dog does chase its own tail... and sometimes nips it... and then are suddenly perplexed, as to why that hurt !??
Proper grammar is to use one period, not three.
@@wasd____ said the underscore1?
Creating music is something done since the dawn of man. The changing of tides in tools to make music continues. This is yet another tool to change tides. We went from banging sticks with wood to putting stings on wood to strum. Again, we're changing. Fortunately for the better as we'll no longer have our perceptions limited to another's production of sound. Context defines perception. Perception is beauty, it changes with the beholder. For example long ago, a musician was a person who could read, write and play music on an instrument. The context of what was expected of a musician changed the perception of what is a musician. Im paralyzed. I know quads who have to use a paintbrush to input their musical genius into their computer. They're not playing "instruments" of yesteryear, but they're creating music with today's tools to create music. So, all you "musicians" out there who think this is some downfall... thoughts and prayers.
And now I really feel the weight of why Daft Punk broke up. They especially started to feel this in the 2010s in themselves and their brand. Not wanting to be robots anymore
I think that the thirst for real live music with authentic voices and vocals are gonna explode.
i think it will be split. some will not care- like electronic dance music had their thing. and there ate other types of music.
Still won't stop live bands from 'writing' their songs with AI and playing them. So fake is a potential even with live bands.
I definitely hope so.
I also think that once AI music is popular, people will start copying AI music in their prompts. It will be a copy of a copy of copy and slowly degrade
I think live theater is going to take off too, and that printed newsletters will become ubiquitous.
so, concert?
I'm an Art teacher in Spain and the other day, one good student who liked playing the guitar told me he is not interested in music anymore because, according to him, the majority of the people will accept this fake music and musicians won't find any job in the near future. I tryed to persuade him that a few percentatge of people (but milions around the world anyway) will enjoy living music and paying for it forever. I hope I was right...
should've said something like the world is too hungry for real human connection through music, AI can't replace that.
I find this use of AI to be very disturbing. Remember the Borg in Star Trek? Half machine, half human. I never thought I would say this, but it seems we are headed in that direction.
@@darlenegriffith6186Unfortunately, when the Borg said, “Resistance is futile,” they were right.
I think the number of people who want "living" music will dwindle with each generation. They'll become the hipsters of tomorrow
Yes. I am worrying about my kid too
At 61 years old, I have worked in IT for almost 40 years. AI terrifies me. I am also a musician/songwriter. I work with people in my field who think that AI is amazing, wonderful, can't wait to use it fully in production. Why? It makes their work easier. This is what makes me cringe. I use Siri, I use Alexa. So, I guess I have to admit that I use AI even in those fledgling ways. But, now that I have heard that Apple wants to fully embrace AI in their upcoming OS, I am not sure if I will even upgrade.
We already have students writing their thesis papers with AI. We already have "so called" songwriters writing their "music" with AI. Its at work in our daily lives in hospitals, government.
Why are we so excited to give up our creative selves, to give up every ounce of self worth, every ounce of pain and memory that it takes sometimes to write something to just sit back and say to hell with it...AI...you just do it all for me? Why are we just dying to be lazy and lethargic? Why don't we want to work for anything anymore? Why are very smart people excited about it all and not even looking at the danger of it?
I can think of so many sci-fi things that have warned us. And we forget that so many of those sci-fi things that we have seen over the years have actually come to be. But, no. We think we know better, don't we?
We think we are smarter than what we create.
This is a tragic step.
I hate it.
I am a lifelong massive sci-fi fan. I agree absolutely and wholeheartedly with everything in this comment. I can’t believe people are just rushing headlong into this and no one seems to be interested in applying the brakes. AI will be our undoing. I think we have outsmarted ourselves as a species this time. if it were just small, isolated devices, I wouldn’t worry so much. But, there are entire cloudbased AI’s now. That is a lot of power. They have access to massive amounts of information and massive amounts of computing power and they are learning more and getting better by leaps and bounds. It will not be long now before one of them makes the leap from the cyber world into the physical. Then, we will be in serious trouble. This is especially true if one of them develops a little thing called self-awareness and free will. Terrifying!
I also hate it. As a writer, I have honed my skill and nurtured my knowledge of the world over decades, expecting someday that would give the tools to become one of the greats. Now, any idiot can write a prompt lacking any grammar, context or meaning, and the algo will just mix its database and produce a novel within minutes, stealing from every human that suffered for their work, making something that's absolutely meaningless. These algos can't even understand what they produce.
I don't think the cause of all this is the technology, though. The culprit is our global consumerist culture and the power we have allowed tech companies to hoard.
We must resist, not consume this trash and demand for it to be banned. Turn our culture of hard work and merit against this worthless crap!
sad situation
@@perroraton9515 "These algos can't even understand what they produce.", you say? AGI, the generative kind that teaches itself, is already here and operative... just wait a year or two until it is perfected and rolled-out.
Then we will be at war with our own machines, and likely lose.
Music is supposed to be a person making art. Call me old, old fashioned, but at 56 years of age, I grew up in the best pop music era I can think of. I play bass, guitar and keys for my own pleasure now. And will do so without EVER exploring AI.
I work as a classical music radio host. My jaw was on the floor when I tried some stuff with Udio. As an experiment, I'm broadcasting AI generated bumpers for some weeks now. The very picky classical music audience is even complimenting my new 'beautifully sung jingles' - not knowing it's a one line AI prompt, sounding even better than Voces8 (top notch a cappella group) and even singing my own name in beautiful harmonies.
I don't know where this is going, but as a long time Star Trek and Matrix fan, I know stuff we make up will come true sooner than we think. For all of us in the music industry, it's a very dystopian glance at the near musical future. Let's forge some strong new artistic ethics in this AI era.
If we don't want this to take over the industry we need to stop engaging with it! Not even for fun or testing. Just ignore it as much as you can, don't listen to it or use the software. Same with image and video generation.
Ethics in the entertainment industry? The only moral code they have is if it's not a liberal talking point then it's evil.
@@intense79nick you are right, but I am afraid it has already taken over the industry, but many of us are unaware of it or of its potential. The question now is how to tame this rapidly maturing AI music animal. Putting it back in Pandora's AI box seems impossible without intervening international and fully covering legislation.
@@Sanderzwiep And when will you be replaced by AI because someone wants to do an “experiment”?
@@carstenmanz302 that exactly is the fear I live in - gotta find some other purpose in life (the parts AI isn't able to invade)
Like many commenting here I am a song writer (for over 40 years - I am now 60). I gave up for a while in my 20's to get married, have children, and be able to afford to live, but I never really stopped, just stopped trying to be a professional. For many years I wrote songs just for my wife on her birthday, and Valentine's Day, and Christmas, then when our children came along I wrote birthday songs for them.
Recently my eldest (now 21) Daughter asked me to do recordings of songs she has heard me playing since she was a child. Most were my own compositions, but three were songs by other artists. When I asked why those three in particular, she told me that even though she knew the original songs she preferred my versions (aren't our children amazing - they make us feel so valued 😄 Most of the time LOL).
What this made me finally realise is that those times, in the garden, in the kitchen, in the living room, when I picked up an acoustic guitar and simply played and sang live in front of them; they loved it more than listening to mainstream recorded music.
This is the true value and glory of music, not the industry version that, let us be fair, ruined music in the name of profit, but rather the real human, and non-profit motivated, sharing of music in a real life situation. We may have had our capacity to earn money removed in the traditional sense, but we can still inspire and delight those who hear us. (And yes, I do get rounds of applause from neighbours if I am playing in my garden, which I suspect says more about how lovely they are than it does about my ability.)
You said it all so well. I play for fun after realizing I wasn’t good looking enough to compete with all the super talented LA pretty boys ( almost all of whom still never got out of bars and nightclubs). But I play on the sidewalk and it’s cool to see people change from thinking I’m a street urchin or drug dealer and stopping to compliment me and give me the biggest compliment..a tip! Especially when professional musicians give me their respect. It’s so informal. I play what I want when I want and only as long as I enjoy it. I did the bar scene for 7 years and just as we were starting to rise through the crowd, disco came in and live music almost completely stopped on a local level.
@@5400bowen Thank you 😄 LOL I remember those LA Pretty-Boys - bane of our lives (all that hair!!! How could they?) Real people like music to be real, which makes me wonder about who is collecting the stats for today's music.
@@IanLuckett yes, but I’m talking 1983-85. There were some seriously good musicians in LA back then. Anyone remember Taxi? They were SO good and it still didn’t matter. And they all looked like movie stars. That along with disco DJs taking all the gigs was the last straw for me. As was the club owners getting ridiculous. You miss ONE night of ANY gig and you would never play a decent room in that town again.
@@5400bowen It was pretty much the same in London (minus the sunshine, and the Leopard-Skin trousers). The Dish-Cloths... sorry I meant DJ's, did the same over here, and they were cheaper to hire than a five-piece band! 😄 I'm afraid Taxi are not known to me... Ah, such is life! The thing I like to remind myself is that we humans, unlike A.I., can grow spiritually when we play music, and there is power in that, the kind of power no auto-tuned, grid-locked, auto-generated music can ever hope to match. Keep playing my friend, and keep growing, that is our best defence and attack against those soulless melodies 👍
@@IanLuckett this is a reply I just made to the comment after this one:
beauty is in the ear of the beholder. I don’t think AI can synthesize new styles of music that are even close to what humans do. And I don’t care what machines do, except serve me physically. I want people art, not machine art. But I’m prejudiced..
as a musician I think of what Tony soprano said once - "It's good to be in something from the ground floor, and I came too late for that, I know. But lately I've been getting a feeling that I came at the end. The best is over"
This hits home hard as a novelist (and musician)
The "Best" was over by the 80's
This makes me mega sad. I really hope that not all is lost.
Not at all. AI will make some great-sounding music, but a decent composer will still create worthwhile works. IT all comes down to MARKETING (which means NO human - that requires royalties - will be marketed when an AI writes for the cost of a kilowatt of power.
We figured that in the 90s when it started going downhill
Heard this from a friend the other day: "AI was supposed to help us with the boring and unpleasant mundane tasks so that we have more time to be creative and compose music, draw, etc. but the first thing it took from us was art. So now AI draws and makes music and we will have to keep doing the unpleasant stuff". Alas, truer words have never been spoken.
All in the interest of profit
It didn't take art from anyone. Most people just never cared about art in the first place. The "artists" had a monopoly on it so everyone was forced to deal with them, and anything that made a profit was enough to be considered "art".
Now the wheat is being separated from the chaffe.
Your friend sounds like a wise person. He is correct… it’s very sad
It's a sad commentary on the state of art that the AI stuff I have heard is considered to be a replacement for it.
"I never saw no miracle of science
That didn't go from a blessing to a curse
I never saw no military solution
That didn't always end up as something worse"
- Gordon Sumner / "If I Ever Lose My Faith In You"
I hear the 'echo' you describe and man is it weird. The best way I can describe it is, the feeling you get when you open a car window driving, and you get that pressure pulsation from the wind. It's like that but with much less sound pressure, at a higher frequency. I feel like it's some kind of phase cancellation due to how the AI produces reverb and harmony.
Interesting analysis. Might indicate that not the AI vocal necessarily is the culprit of that "phase echo weirdness" some people hear but the AI still has to learn about mixing and mastering from scratch, not just throwing everything in without further tests as to how the instruments , voices, drums interact. But for that you'd have to have stems.
Layla needs to release an AI ear training course.
I could not hear the problem but I am older like Rick, I asked my wife with better hearing at least for high frequency sound and she said it reminded her of auto tune, she is not an audio person so I think this was her best shot at trying to explain what she heard. I can hear stupid amounts of auto tune but in this case nothing.
You can tell an AI, but you can't tell them much.
The best way I can describe it is: Go into an audio editor such as audacity, and stretch the duration of a sound file without affecting pitch. The computer has to make up audio to account for the time difference, which makes it sound robotic and low quality.
In a year it'll be out of date. You'll have to train an AI to detect an AI.
😂
Hey Rick. I'm a sound engineer and a couple of months ago a client brought me his background vocals made by AI. They were different "characters" male and female. Yet on all of them there was a frequency stood out in all of them that added up in the sum of all backgrounds and made it sound horrible. Your son is on to something. Incredible hearing that kid.
give it two more years and not even cats can differentiate what's Ai and what's not
I'm pretty sure I can hear it too. It's a weird phase in the midtones. You can kind of hear the "stitching" of the sound bits, etc. The metal example Rick played had the clearest of this. It freaked me out when I got actual chills listening to the harmony choices in the country tune. Chills to AI music? That was a first for me.
Former sound engineer, I've played around with some AI stem-track generators to pick apart songs that I want to analyze, and I'm wondering if there might be some overlap. The "demixers" will rebuild the stems based on what the different instruments "should" sound like, having been trained on various isolated tracks to build the AI model. I've noticed that if I restack the stems in my DAW, there is an over-excitedness that just builds up and makes everything harsh AF, but there's also seems to be a lot of detail missing around 1.6-5k. The breath, the throatiness, the spank, it'll all missing. A lot of detail goes on in that range that makes voices and instruments sound real. I'm wondering if the song generators are building their tracks in a similar fashion, creating individual tracks based on a model with a similar type of training. Dylan's right, give it more time and training and we won't tell the difference.
I'm very jealous because I don't hear it and I tend to think of myself as having pretty good pitch
@@automachinehead Give it two more years and the training sets will be so contaminated by AI output that the models will actually get worse.
Like when Michael Jackson would say leave the breaths in, I want to hear the human parts whether they be slightly out of time or tune. It’s what makes music real.
AI: 'taking notes"
They do that too, it's often edited out of songs using basic algorithms now it's going to be a lot easier
AI simulates breaths. As part of the insights in the song, when you voice swap a track, it tells you how many breaths were detected.
Ironic; music has been more and more cleanly-shaven and overproduced, Autotune and all. We are all acoustically used to it. This squeaky-clean product is much easier to also algorithmically imitate. It’s almost as if we made it easy to screw ourselves over.
@@TomJakobWLuckily the bulk of this only applies to pop/mainstream music.
I'm already crazy from the last 20 years of produced music. Now anywhere I go I feel like I will have no human mind left. It's torture.
The reason it’s so hard to tell the difference in some songs is because of all the vocal manipulation we’re already hearing with most singers. We’re is the pure vocal talent with a little delay and echo.
I was going to make my own comment to say something along these lines but I'll just reply because you're pretty much making the same point. I think this might initiate a push back to purely acoustic music. Good luck to an AI trying to emulate a stunningly gifted pianist or violinist. You cannot fake the real thing
Until it can have the dexterity of a human
You're kidding yourself, AI doesn't struggle with emulating complex pitch movements. It's like saying a printer would struggle more to print out an oil painting than a photograph. AI will copy any style you feed it, it's weakness is innovation not microtuning.
@@HankblueThe thing is, a real voice or instrument without autotune doesn't have (or need, IMO) microtuning. That's what makes it special.
@@markcheetah4960 You're misunderstanding what I mean. Real voices and instruments have microtonal structures which AI has no problem emulating, so that's not the reason that it's so hard to tell the difference at all (per the original post).
The AI's not doing any tuning either, it's just building the sound 'pixel by pixel' the same way a printer does, and complex microtonal structures are no real obstacle to it.
In the 1950's we outsourced reproducing timbre to tape machines.
In the 1960's we outsourced projection to amplifiers and location to reverb effects.
In the 1970's we outsourced rhythm and pulse to electronic circuits.
In the 1980's we outsourced drumming to drum machines and recall to digital discs.
In the 1990's we outsourced being in tune to Autotune and recalling mixes to DAWs and computers.
In the 2000's we outsourced field recording to sampler packages and creating strong performances to looping tech.
In the 2010's we outsourced high performance orchestral, vocal and instrumental performance to sophisticated samplers.
In the 2020's we outsourced musical performance, skill, accuracy, versatility and taste;
This year we outsource composition and bypass the need for any meaningful human involvement in music making whatsoever.
"Music is its own reward" - Neil Finn
Thats exactly it. Music is the language of God, it's a gift, this is such a sad road humanity has been dragged down by the minority.
"In the year ninety-five ninety-five
If man is still alive..."
Excellent summation of the ignorance of humanity- always finding an easier way to get to the end product. The problem is, by doing so, the product is no longer the result of human creativity. Human creativity is imperfect and flawed, that’s what makes it beautiful. AI will never be able to re-create human imperfection.
Soon we will outsource ourselves.
And yet there are still live musicians playing live music to live audiences. The technology to replace live players with black boxes has been around for over 100 years, and yet there are still live players playing live music to live audiences. I confidently predict that in another 100 years there will still be live players playing live music to live audiences.
I was born in the late '80s, loved technology all my life, first gen to grow up on the internet, in one year the assault tech is making on creative workers has turned me off of looking forward to the future and realizing that no, we're actually going to see some of the most dystopian predictions come to pass, because this tech is all in the hands of sociopathic, predatory companies with no vision for anything but the next fiscal quarter. I am so disappointed that somehow all the potential came to fucking this.
Sad, very very dispiriting.
That potential is not all. The thing perished from use. The value of realization is much greater than the realization of value.
It's always actually been about money.
If J.S.Bach didn't write something new for the priest every week or if the priest didn't like what he'd wriiten the week before
he would've got the sack ......
If The Beatles stopped writing hit songs then the record company would've dropped them .....
@@Tom-hk6ub While you're not wrong, the problem is that it's taking the human element out of it entirely. We should force record labels and streaming platforms to prominently disclose when something is AI generated.
@@OutLanderUSN Good luck with that one.
This is a very sad day for true musicians, im honored to have been able to tour and play in front of large crowds opening up for eighteen visions and bleeding through before you tube existed and partying.
"If you want to sound like computers, it's easier for computers to sound like computers." So great!
Neural nets just not good at each separate instrument line, these are automatic composers not virtual instruments. But if you get good pre-recorded instruments / voices with RCV (software to write vocal lines using coders designed for better voice output), the result product is indistinguishable for casual people.
That is my problem radio is annoying, nothing new but crap. A few but they just play the same songs. find new great songs. Its hard.
I suggest learning the behind-the-scenes stuff for the voice acting for the video game Portal.
"I wrote a song using AI", that's probably the most delusional/annoying part of this. They really think they wrote those songs. It's this new wave of "AI artists".
Maybe "produced" is a better verb for this.
Watch Shadiversity defend his "AI artistry". It's embarassing.
Lol exactly. Fits perfectly with this generation of entitlement, though.
LeBron wants to be the best without the resume, and with hand picked teams.
Kids want their student loans forgiven.
Free rides through life.
@@xsleep1Nope. If you just typed words into an AI program, you didn't produce squat.
Exactly!
Live performances, improv, and authenticity are going to be what will make a musician stand out
Live music is dying a slow death even over the last 30 years with pokies and djs, drum machines and computer simulations well and truly making it impossible to earn a dollar and the drive to improve and work towards these things can't be overlooked. I feel like crying when I think about it. Playing and discovering with others is truly a wonderful thing. Imagine the lone drummer in the jungle never hearing a real response or no one smiling with joy at that improv from the depths of someone's soul. Please NO!!
Nah god given songs will
IA can improve. Not even IA. Only with Ableton, you put the BPM, the key, ... And LETS GO
ai will be able to do that even too, eventually.
I heard this before, back when MP3's and Napster came out... they said the same, live music and concerts... that is until the AI robots start walking among us.
We are at an inflection point in so many areas Rick. Not just music but climate, politics, art and medicine too. Interesting time to be alive.
Yes. The replacement inflection. We are being replaced right now. All very interesting, amusing- not.
Very, very scary.
The machines are taking over, have taken over.
The beauty of it is, as a musician, I make music for myself, I enjoy the process and satisfaction of creating what I like. If other people like it, thats a bonus. AI may create a destination, but will never replace the journey.
This 😊
Beautifully said. I feel the same
yes
can i hear some
Love this. F*ck AI.
Support your artists, big or small. This is the time to show your support, a like, a comment, merch, live shows...do what you can to support artists
Fantastic comment. It should have 2K upvotes and 1 down (courtesy of AI 🤪).
Personally, i will boycott any art or music that is AI generated. I don't care. If it is AI, i will just not listen to it or spend any money to see it.
@@gclip9883yes same - i think maybe for idea generation for certain things (like blog post ideas or marketing tools) it could be useful, but not art. And I think that because I know art in any medium isn’t just intelligence, it’s a whole lot more. I hope they have some sort of legal requirement to state whether a piece of music or art is AI or not as the technology advances and it might be harder to tell.
agree!
Until you realize that the supported artist is AI...
This is why it's so important for people to get involved in making music themselves; they need to connect with that part of their humanity. As a school music teacher, I hope people continue to see the value in kids singing and playing instruments rather than simply being consumers.
June 21st is MAKE MUSIC DAY! As a flesh and blood gen-u-ine human being musician type creature - Grab your axe! sing, play, make HUMAN music that real flesh and blood can hear, dance clap or just be human to! 🎶
Where are there school music teachers? I thought they did away with that years ago. That has been my explanation for why new music on the radio is so bad...
With AI, "real" music might go back to where it came from: To people making music, alone or together, for their own pleasure or as a social (inter-)action ... But it will be harder to make good money as a musician. I guess there will still be live concerts but on platforms like TH-cam or Spotify most will be replaced by AI. Because its cheaper and as background music it is good enough for most listeners.
I think it will be the same with so many aspects of human skills or knowlegde. AI will be able to do/know it better then most people. That's why I tell my children, that knowledge and skills have a value in itself. Maybe, AI is even a chance to free learning from the need to be economical.
The thing he keeps talking about is the difficulty AI has with vocal generation. It sounds like a mix between hyper pitch corrected vocals in harmony all sung by the same person and a vocoder. The other thing I find fascinating is the that you can kinda tell the lack of artistry. Like it's making high quality sounding music but there's frequently some arrangement decisions that songwriters and arrangers typically wouldn't go for. The reason for it I think is that when a robot is arranging a song, it's trying to be an amalgamation of the source material but when a human is making a song the goal is typically a stronger dopamine or emotional response. AI music just isn't selecting for the same criteria a songwriter, arranger, or producer would be because of the underlying mechanics and that gives good sounding but inferior results. But I'm sure we'll fix that soon enough.
Hope NOT, or get ready that ALL of the music will sound like Tailor Swift. Lifeless, bland, pulp...
Beato has good comparison between the rhythm machine a John Bonham, playing the same pattern: NOT EVEN CLOSE. Lifeless, as month old dead fish on the beach.
Your reply has just been read by AI and the principles incorporated into the next edition of AI.
These very flaws you describe would be strengths in writing AI instrumental New Agey background music. I don't think we'll EVER see an AI write Bohemian Rhapsody. But watch out, Enya and John Tesh!
with text to image prompting they use weights to help with that
idk if the audio models can do that, but
(artistic:1.5), ..
or whatever could already fix that to some extend
This will force artists to stop using autotune in order to stand out.
That would be great!
this will force some "artists" to retire . . .
and then AI is going to mimic that as well... It terrifies me
I thought the same thing. Maybe a more “raw” aesthetic will emerge as a reaction to over produced stuff
@@FishMonger849my thoughts exactly!
I spent some time with Udio when it first came out, and actually turned out a handful of tracks that I ended up really liking. In fact, to the point where I found I was listening to those almost more often than other stuff. And it reminded me of something someone said a little while ago: If people are saying they are lonely now wait until everyone has their own music generated specifically for them that no-one else knows, and the common language of shared music disappears. Strange times.
But isn't one of the best pleasures to share music with someone and seeing it growing on them? Going to a gig with a friend and both feeling the same energy?
@@CrappyProducts Indeed it is. I guess the AI approach means that the days of saying "did you hear that new x track" after hearing it on the radio etc are gone, since unless you specifically tell people about it no-one will ever hear it.
I don’t think musicians will disappear, the emotional connection of a live show will prevail. If anything in the future there will be less competition and I believe people will search for a genuine experience that can only be delivered live. Spotify already pays artists near to nothing, we need to deliver impactful experiences that cannot be replaced by Ai
See, maybe I am being overly optimistic, but I keep coming back to this inner belief or trust that, ultimately, art is an inherently human expression. Content and entertainment is an entirely different thing, and AI will certainly monopolize on that. But I believe there will always be a segment of music and art lovers who value a piece of music or painting because they know a fellow human being poured their soul into it. For example, if someone is going through a break up, I can’t help but think that most people would gravitate towards a song written by a human who went through the same experience, thus making the painful experience a shared one. On that level, the utility and value of AI music/art will ring hollow..
🤯
Rick, you briefly touched on a very important aspect of this trend: Over the last decade (or more), human voices have been so processed, so manipulated, so quantized, that producers have trained our ears to accept what sounds like computer-generated music. Now, computers are trying to emulate voices/instruments, so our ears are already accepting of it. Auto-tune is horrible, but everywhere. AI music is the aural equivalent of "plant-based meat".
Exactly 💯
while I totally agree with you, i must say ALL meat is plant based...sorry couldn't resist.
BINGO
Plant based meat did happen...but I think what has actually happened is people stopped trying to make vegetables taste like meat and made more tasty and interesting vegetable only meals....this will also happen with music.
@@themicrobusinessrenegadepo6338You are so wrong. 😂 They're still trying to push plant based meat as hard as ever.
I'm 55 and I agree with both of your kids. I hear the "echo" and it will be impossible to tell in 6 months.
I take photos. The exact same conversation is going on in my community. At first everyone can spot AI. Then it tricks half the people. Then it gets to the level where it fools 95%. Mission accomplished. There is no stopping it. If you are a creative in any sphere, there is someone who has tasked an AI program with "learning your craft" with the end goal of replacing you.
I've spent a lot of time browsing Ai photos. They all start to look the same after a while. So far I'm not that impressed with Ai's photographic output. Similarly with the music Rick has just played. So it can generate music in the style of Ennio Moricone, so what? I'll be impressed when it comes up with completely original styles instead of simply blending what humans have already created.
@@workdevice7808 It's only a matter of time. These programs learn exponentially. The input cues will get more finessed as it learns from massive amounts of social media input.
@@workdevice7808 Exactly, all AI doing is just mutations of existing human creation. They are not creating anything new at all.
Machine learning isn't ai and people who still thinks it is ai lack human intelligence
True but for a lot of jobs that's not needed anyway. Imagine someone who does TH-cam thumbnails or someone who does posters for movies and galleries, those sort of things can be easily done with ai, and they will actually look good because there is no natural aspect in them. As for photos, I also don't think that AI is anywhere near good enough as of now but that will change too, why wouldn't it? @@einsam_aber_frei
I can hear it. I am a pianist. To me it's like a slightly flat piano in tune with itself, even the voices, with a sustain pedal held throughout. Articulation, clarity, and nuance are lost. They may be able to change it but they would likely have to revise their programming. Great show!
It sounds metallic & warble-y to me but your sustain pedal & tune comments are perfect!... I hear it too! It is in tune with an out of tune instrument and really grates & makes my ears prick up. Anything out of tune has always felt like nails on a chalkboard... Sucks bc my own singing voice is horrible 😂 but AI sounds worse to me😂
Years ago, I could smell Auto-Tune like a bloodhound. It was a Michael Buble Christmas song playing on a store's speakers. Irritating. Still can smell it.
@@anoial5181 it's true, it's a tinny quality that makes your hair stand on end (in the bad way)
For the first time, Im happy to be 75. Why, you ask? Because I was fortunate enough to live through a time where great music was composed by genius’s of my time. May I say, Lennon, Dylan, Page, at the very least? Im so sorry for my great grandchildren musical future. But I have left my playlist and encourage them to listen always. Long Live Rock n Roll! 😊
Yes, I feel we enjoyed the golden age of music…along with many other things including our Constitutional Rights and those days are long gone!
I agree with you (I’m a little younger) but I am afraid that the days of talented musicians is over. We will no longer get back what we have lost.
why?
I'm 72..and my kids bought the a t-shirt that says.."I may be old but I got to see ALL the good bands" and it's truer now than ever
I am 61 and I believe that if you are between the ages of 70 and 85, you may have lived - and more importantly been at your very best - at the pinnical of human times. That is if you avoided Vietnam. Everybody enjoy the long slide down.
I love using SunoAI, as I write my own lyrics and hearing what it sounds like as a song in my style's is amazing..
Most people aren’t making music anymore. They are making content.
There will be a big enough database of “content” that AI can pull from to make “content”
That was really well put. Wow. You are bang on.
@@okisoba There will never be enough content to make something unique. Ai is doomed
Well, music is content... A different terminology at a different time...
Damn, that word gave me a wicked flashback. Back in the mid-90s, Australia was piloting broadband internet. The advanced early users prototyped lots of P2P services, but management got fixated on "content" e.g. IPTV. I despised the word ever since.
Rick… I’ve been a VO guy for nearly 40 years, my bread and butter being corporate training courses, instructional videos and HR tutorials for new hires.
I thought I could make it to 70 yrs old while still working in the studio.
Ai came too quickly, and I‘ve now lost a couple of clients I’ve worked with for 30 years… that just called and said “bye”, we’ve been told by corporate that Ai is “good enough now” that they can stop paying VO people. Those 2 contracts were worth around 50k. I have not been able to find a way to replace that income, and I’m freakin’ hurting!
Same goes for my friends & colleagues who work in fashion/commercial photography/videography. These greedy companies no longer need a model, a shooter, grips, lighting, wardrobe help, location scouts, and more… all because Ai can create a more realistic human image than ever before.
When we allow computers to supplant us as creators of art, who do we become??
Hey Man. I’m soo sorry. I work as a producer in film and video - mostly commercials. Our world is falling apart too but for what reasons I don’t know. But back to you. It just kills me. I’d never dream of using AI for VO (Cheesy) and I feel I can tell all of the AI vo out there, but when the client won't pay we are stuck. All of the MBAs just want to lower costs and pump out volume - in everything and now advertising. Then they keep the money/savings. Totally broken and ultimately no respect for the artist. Do you know what this kind of says? It used to be that advertisers wanted and needed good art to sell their products. Now they don't. To me, especially when they say "AI is good enough" it means that the advertisers don't respect the customers. It's like saying that the dishes in a restaurant are clean enough, the meat is cooked close enough to done, etc.
Thank you for sharing. I think it’s a warning to the rest if us that slowly real creative jobs are going now that life is virtual. Very best.
@@fritzb.3978I've also heard writers about 6 months ago saying they were really scared about this. These are strange times.
The thing that I am seriously asking my self here though is, how the technology will "progress" when there are less and less human creators out there actually making high quality content. There are countless of professionals out there making really good content, be it with music, writing/texting or visual art. And it takes a lot of work. And someone has to get paid for it.
But, if no one pays anymore for it, using algorithms to get it all, for alike a nickel or almost free, from where will the algorithm get "new" content to learn from? As of right now, those AI models require an insane amount of data to even work somewhat decently. But with more ai content being on the net and with less human content to scrape from ... what will those models do in the future? It will all become a copy from a copy from another copy and another mix and that's it. As impressive as those algorithms are, in some situations, they can not create truly content on their own. That is a very hard technical limitation and there is no real work around on that.
I’m an indie publisher and I tested KDP’s AI tool on my reader audience. They absolutely hated it. I thought they might, but the intensity of the negative feedback surprised me.
I work in VO as well, and there was a sudden and strong dip in the amount of corporate training/industrials. Then they started to bounce back because businesses were realising that their targets were turned off by the sound, but they seem to have disappeared again. The companies don't want to pay what they've been paying, now that they've seen they can pay much less. That's the issue; it's the companies not wanting to pay for things, and they certainly aren't passing those savings on to their customers. We had a recording studio contact us about training voice models which they'd use for NPC parts in small games. Again, I understand the financial concept: small game developer wants a cheap way to add voices, but I said how much money is coming out of what I would have made in the future?
I've broken apart the stems for a bunch of these tracks and what your son is hearing is twofold. First there's a high pitch hum in the 10k plus range. This can be filtered out with proper mastering. The second is an automatic synth-like harmony that has a slight delay and is usually a fifth apart from the main line.
I like your explanation (better than mine): Here' my non-muso, non production description. Which you may find has a little too much waffle?
"Quantisation. Its very faint, yet enwraps (is that a word?) the soundscape, especially the vocal performance.
As vocals rise and descend you can hear the 'chop' of a quantised pitch. Its not an analog - continuous - parabolic curve. Particularly the trailing exhalation of breath. A fraction 'off'.
Its easier to hear with good headphones."
I think the question is, will anyone care? Since most people will not be able to discern this, they will be fine and what Rick says will become true. Who needs real people making music anymore when you can get it for cheap or free and make money at it?
I've made a couple AI songs, and I have noticed the high pitch hum. I've also noticed that the longer the song is, the more likely the high pitch hum gets louder.
Are the vocals pitch perfect like with autotune? Maybe that’s what Rick’s kids are hearing-too perfect to be human?
@@classicmartini While all what you said is true, none of it separates it from highly processed 'human' vocals. A combination of autotune, precise editing, and tight gates will produce the same effect.
I have a very good ear for this and have been obsessed with the subject for a few days. Check out Sore From The Door by "Beats by AI". I think the AI model used here is most noticeable. It's the way sustained notes are executed. We need to pass air with a diaphragm over vocal chords. The AI tries to replicate this by simply extending the note. The note, however, is a moment in time a voice is singing so it sounds...like it's sampling the same voice? Basically inserting the sample perfectly over and over as long as the note is sustained. It's an odd effect that the human voice does not produce. Our voice sounds different over time depending on how full or empty our diaphragm is, richer or weaker as we run out of breath. Vibrato is also a problem for AI and the "sampling" is quadrupled as the AI tries to switch between semitones. Yet I believe with a few tweaks, it will be unrecognizable within a few years. Crazy.
What makes great music is the natural nature of the musician: the mistakes, the break in the voice as it hits the high notes and the accidental feedback of a guitar. The best recorded music is often that which is recorded live or as live. We crave the humanity of it because we are human not bloody robots!
exactly, music is so over produced nowadays that it doesnt even register as music to my ears. Just sounds like production. whats sad is when i heard raw recordings of modern pop music, it actually sounds good to me, and makes me realize there's a lot of music that i would enjoy if it wasn't over produced. I think they ruined music
Unfortunately, to the masses, none of this matters anymore. People are getting lazier in every aspect of life. I can't imagine another "Bohemian Rhapsody" coming along any time soon.
i want to agree with you but if you look at ehat most people listen to, you'll realise how easily it is made by AI.
All of which will be replicable by AI in time.
Also the human element point isn’t always true for much of electronic genres
What a load of nonsense.
I'm 24 years old and I mostly listen to songs from the 1920s to 70s and I have met growing number of people my age and younger abandoning modern music and going back to the oldies. As a music lover I have hope.
I am 48 years old and I mine songs that stand the rest of time regardless of genre. A good melody is a good melody. So can not reproduce someone getting "lightning in a bottle" and serving that up in a well written and produced work of art created by feeling humans
I'm 24 and in the same boat. My band just released our first album! If you are interested, its called "Dead Man's Float" by The Taxmen.
Interesting to hear and does seem hopeful that an underground might develop and push back on these methods.
@@redraymon i hope there will be a small group of artists going on and people willing to listen/pay for that.
What's weird is that AI can generate this kind of vintage music too. The trick that's missing is getting it imperfect enough but, with the right DAW plugins, you can get close and the gap will close as time goes on.
I am that unknown musician who, over the course of 4 years, has been coming up with my own original, unique concept for my sound and material, it was a very subtle and complex process, when you try to combine the incompatible and make it sound familiar, it is a lot of suffering. But now I have received a clear message from this world that I am not needed here.
My ego and individuality were literally damaged and dissolved beyond repair. But I feel that deep down my love for creation and for music does not depend on any external factors or even listeners, it warms me.
I feel for you. But in a sense this changes nothing for people like you. You used to be competing with millions of other artists and all of recorded music history and big name marketing firms which was a seemingly insurmountable task anyway. Now this is just one more. There will always be people who want that personal touch. AI did drawn art much faster than it could do music, and artists still exist. If you were doing it to make money, YUP AI will be a big dent in potential profits for all human artists the same way it has hit the online freelancer market. But as an old guy I can tell you: the sky is always falling and the apocalypse is always coming. I am a writer and we have survived online piracy, electronic libraries making all old books instantly accessible for free, competing with dozens of new forms of entertainment when we didn't have to previously, we're surviving AI too.
There have been millions of other artists. There will be BILLIONS more in the future. AI is just another big number on top of that already near-infinity. However there is one thing that none of those artists have ever had, or ever will have, that only you have:
You are you. Lean into that harder than ever and you have no reason not to continue.
@@kspice6807 This is a terrible series of shocks, but I'm inspired that you didn't give up.❤
Musicians often face that broken ego at some point of our lives. The faster you learn the lesson that music making shouldn't be related to your ego or worth, the better :) It'll pave the way to making music a happy and rewarding part of your life and not a burden. The truth is, none of us are truly needed as musicians in the big picture. Not even as humans. You just happen to reach certain people that connects with you or do not and hence you become needed by them, perhaps even for some short amount of time. And it's ok and quite freeing imo! Take that 'damage' as an opportunity to be reborn in a new way where people's judgement or loyalty is irrelevant and you can be free to express yourself in whichever way you want. Like you say, your love for creation is so fundamental and deep that doesn't have to do with playlists, listeners or success. It's the only thing that matters. And if a meaningful success shall come (in whichever way), it'll only be worth through the authentic and honest expression that comes from that love for creating. Anything else can actually turn into a golden hell.
Regarding AI, it scares me, it disgusts me, and I've decided to turn my back on it, regardless of what other people may say or how far it will enter into our lives. I'll reject it as far as I'm able to, even if that means becoming ostracised, which I don't care because I'm not worried about being streamed or known widely. I'll just continue doing things the way I believe. Best luck to you!
Keep playing, creating, writing music. Your gifts will make room for you. You are a talent, and you'll never stop growing. Don't quit, ever.
Dude, that is 99.9% of us who create music, only to be seen and heard by (almost) no one.
the small indie folk/alt artists are going to blow up even more. People will gravitate toward raw sounding music
They will AI that too, lol.
@@z74d-oy2uj I admire your optimism
I have to confess. I tossed on Neil Young Live Rust and cranked it up on Powderfinger. Terrible singer, instruments barely in tune, no idea tempo, total garage jamming - and I got chills and a few tears came to my eyes. It’s been a long time since recorded music hit me like that. I miss that sound in my life.
AI will never replace, nor even come close, to Neil Young.
EPIC. I understand! AI will make humans seem very deep and gritty and interesting. At least in the near-term. Let it kill off all the mainstream tripe.
AI will be able to reproduce those live human conditions too. Just wait six months.
Thank you. Perfect response.
@@MarioRafaelAmadoAlveslive recordings maybe, but not live shows
I’m not going to stop creating music, ever.
And we salute you for it Andy. Been a massive fan since Enter Sandman Jazz (crazy that was like 15 years ago) and no plans on stopping. AI or not.
When you perform in front an audience, record the session and then you can sell CDs and copies of the performance to people in the audience. They will pay knowing it is not AI generated and that they were there. And you get money for every CD/USB stick that you sell. Without having to go through a big company, you get to keep all of the profit. You probably know all of this anyway. Good luck in your career.
Real human music does not have to stop. It will be a niche product that sits alongside a voice controlled AI media centre, where the listener creates expressive music by just asking for it without having to learn the skills.
Nobody does it like you
Same.
The reason we (At least today) recognize AI is that it sounds as "tinny" fake as the auto-tune stuff that's coming out of the major music labels. The voices are already so manufactured that *no one* sounds natural.
NAILED! I discovered your comment after I wrote mine, in which I say more or less the same......
I doubt the general public cares or even notices the difference..
That can‘t be a problem for AI. Even conventional synthesizers use things like „random pitch“ to make things sound more natural and not too perfect.
garbage in, garbage out...
I was coming here to say this. The vocal tracks all have a tinny note to it. Kinda of like when we used to download a low bitrate song and it almost sounded under water or something.
*It's the vocal compression and key artifacts that give it away. Think 128kbps quality stereo mp3 vs WAV or FLAC.*
I hear it now, but the thing is, that's what I hear in every major top 40 production vocal that's not AI as well.
That's because AI will never be original or creative, just like today's top 40. If AI existed in the 80s and we stopped there, we would've never gotten grunge, hiphop and the kind of pop and R&B we've got in the 90s and 00s.
@@miked7295 Never is a long time. Tell me you wouldn't have said that the current stuff was impossible 5 years ago
@@SmileyEmoji42 I absolutely believed this was possible. But AI is trained on previously explored things, so by definition it cannot be creative.
@@miked7295 please don't be so romantic, it can and it will.
@@miked7295 you know who else was 'trained on previously explored things' ? humans. difference is that we create by taking ideas we've seen before and blending them with other ideas we haven't seen them blend with. AI will certainly be able to do that very well in the future.
The fact we're left with such impressions from AI just confirms most of the music we have online is so generic even AI can do it better. Just another motivation to strive for originality.
So true, they spent years sounding like robots and are now shocked people can't tell the difference.
Time to throw out the metronome and only listen to Hendrix for inspiration…
Within 10 years you'll be able to ask for an AI song that's 60% Dire Straits, 20% Chet Atkins, 10% Albert Lee, and 10% Bob Dylan + a clever song title and subject.
The end result will be amazing, but if it's not perfect you can record any little fixes in a kind of scratch track and ask the AI to rework it with your edits.
I agree. Top forty music is such sheit, really lacking in the creative department. Particularly in the drumming.
I was going to pretty much say the same thing.
Going back to the bed and putting the covers over my head. 64 year-old songwriter here realizing it's time to retire. Truly disturbing.
Don't believe the lies by the tech industry. Rick is way off here
@@NoDaysOff-oz2zl care to elaborate?
@@NoDaysOff-oz2zl I think you underestimate the average banality and ignorance of people. If Rick was wrong, then progressive music would be the most popular, or acid jazz, or classical. Instead, the most popular music is the simplest, easiest, most insipid tripe. It always has been. Generative AI allows crap to be churned out at a ridiculous pace, and that's really all corporate wants...it means more profits.
@@NoDaysOff-oz2zl rick is usually way off lol
@@rikk319 if you and rick dont understand why that music isnt popular you genuinely dont understand music, people or culture
This is like what I call a "Plastic Vacation." We went to Disney World for a week, and had a great time....but it was "plastic." Then went to a cabin on the shore of Lake Superior in the U.P. of Michigan. Enjoyed that too...but on a much different and deeper level. AI music can be amazing. But it could probably make us overly critical of real musicians in a jam session or coffee house environment simple because they're less than perfect. But if you want the real thing, get out to a jam or coffee house type of venue where you can experience the real thing.
I'll tell you what I think. I think the minute that egregious auto tune use was everywhere was a big indicator. Auto tune was initially used to fix bad vocal notes for the most part. It was meant to fly under the radar to grab a moment from a great performance moment and save it by a very inconspicuous move. But for some ungodly reason, Producers everywhere took that and said "lets make it totally audible". That's where it got interesting. And really stupid. Every producer / artist in L.A. was trying to out-do the other guy and that's EXACTLY what'a already started here with A.I. So yes I totally agree... this is going to explode and put many great artists and composers out of business. It will ruin the ENTIRE creative music industry at the same time ... think of The Beatles. If they had A.I.? at that time, imagine them saying "well this is easy - lets write an LP". Now you have taken the human element away, and they would have had to have written based on OTHER ARTISTS. So no Beatles. They would have been a combination of the bands they wanted to mimic, the artists they loved. They would have sounded like a combination of Little Richard and Carl Perkins or something. So if that's the mentality that people want to embrace then God help the people of this new generation who expect to hear great original works from artists who are living geniuses in the making.
God knows how many incredible artists that will never even try now to search their souls to create "I Am The Walrus". This will annihilate the art of making music. But nobody cares because of the bottom line. money. money comes before anything in the universe. Easier. Simpler. No brain activity needed? and a paycheck for doing it? And you're right Rick how the hell does anybody track this? Like the "intellectual property" and so on? They let a lion out of the cage and have no idea what will happen. Modern human behaviour never ceases to blow my mind. It's always a half-cocked race out of the gate for people with no sense of direction ( Monty Python ). Here comes the train wreck. We can either stop the train or get on it. My tiny part, just for my own sanity I'm going to do anything I can to stop it, for what it's worth. I hope there are like minded people out there.
Rick, as someone who has been thru a lot in life, music has been a great source of salvation in my life. No fan of auto-tune. The flaws, the scars, the errors, the vulnerability, are the beauty in the art. The thing that touched me and moved me about many songs and music was knowing someone else, somewhere, another human being, was feeling the exact same things and that I was not alone, no matter how much I felt that way. Discovering music that touches a nerve and describes what I feel is magical and transcendent. I dont hate computers nor am I a "Luddite". But stop messing with my music!
I think what your son is hearing, is that the vocal segment sounds like it was sampled with a very low sample rate... Which has a unique acoustic quality to it. It's as if the vocals are recorded on an AM radio...
Sometimes sounds a bit like additive synthesis artifacts, or FFT plugins that can cut out frequencies very precisely like ReaFir
Yep, this is it. I can hear it.
Yes! This describes it well
All Ai generated music tends to have that mushy audio quality because its not generating individual tracks, its generating a whole audio segment at once. That's where the artifacts come from.
If you look at Ai-drive software like SynthV, the vocals can sound much more realistic because they are isolated and clean. More importantly, it gives you control over the notes and how they are sung. So I can basically dictate to the Ai how I want it sung and it will do it. That's where Ai is headed IMO. More tools.
Exactly! To me it almost sounds like the vocal is a bit out of phase or something.
I'm totally unable to make a difference between AI and natural voice, but my wife says she 'suffers' when she hears AI voice in any media. Like they switched the news reader to AI voice in some local radio stations and I couldn't tell the difference except in rare occasions when the AI couldn't pronounce words properly. However, my wife immediately switches station when AI news begin.
I can't tell either, what it is they can hear, but sure as hell I'm deaf to it. And I totally agree, in just a few short months, neither will anyone else anymore. The development is so fast in this area.
Sadly.
Listening to AI speaking feels like chewing a meal that's made of paper. I hear it, but cannot remember anything it said.
And it's the same with AI 'text', 'music' and 'art'. It is not the real thing, and never will be.
I am a full-time piano teacher, now mainly online all over the North America. Thank you for your efforts to bring attention to this composition/ musician killer.
By robbing us of creativity AI will turn *all* humanity into two-legged cucumbers.
Initially as a 75 year old pro musician since '63, I am SO glad I lived through this "golden age" of music/sonics/gear/sound production. I can't really imagine the world not too far ahead......
I wonder if people will know what real music is. I wonder if songs will be half as long for twice the price. And if it will be illegal to record your own singing voice or instrumentals. Did you know it is illegal to capture rainwater in some states? Rainwater. A free, natural resource.
I do understand you very well. I'm 69 yo.
@@velvetbeescorrect about the rainwater same in Australia
@@velvetbees Which states would those be, genuinely interested to know.
The thing is, some people may not be able to imagine the world not too far ahead, but the point is SOMEONE will have to live in it, like my son. Throwing up your hands about the future and saying, "well, at least it is someone else's problem, not mine because I'm older" is half of the reason we're in this predicament as humanity.
@@velvetbees Well, I did a simple web search, and found out collecting rainwater ISN'T "illegal" anywhere in the United States...but it is regulated in 11 states, for different reasons. Sometimes humans can be as disingenuous and deceptive as AI text generators.
Really, reading comments on TH-cam filled with nuggets of "wisdom" requires just as much skepticism as always.
Scary times ahead. The scariest times will be when no one can tell the difference. The worst thing that technology is doing(IMHO) and has done is making the "art" of creating music way too easy. Too many apps, AI, recording software etc. The REAL creators of music will be lumped in with all the non creators, ending careers and losing the soul and art of real songwriting, performing.
Yep, the worst thing that people wouldn't be inspired and interested in music, cause it's too easy to create your own song at home with just some simple prompts
The voices sound very similar to any MP3 compression - it is specifically on the vocals. It sounds like the sampling is low.
Nothing can replace the interaction an audience has with live musicians. It is a powerful force that AI can never replace. So there’s still hope…
Than why is there more DJ set these days than live show with musicians …
Tell this the Hatsune Miku fans 😅
@@hugorodrigue3131 thats exactly why there are dj sets and not just music playing from a playlist. People want to watch someone regardless of what the music is, the best bit with most DJs is they are playing other peoples songs so I don't get the appeal but it must be there as DJs appear to be the modern day rockstar.
You're forgetting that most hit songs are recorded in a studio, no live audience required. You could always tell it to make it a fake "live" recording I suppose.
....yet
Saddest thing i´ve ever seen, it´s like a nightmare I didn´t even imagine coming 15 years ago, when I was obsessed with music.
At least we have Lage.
I'm afraid for what is happening to the world and for what my grandkids and young people might have to endure. Everything is changing too fast. For music and graphic artists what will be left for them? Soon robots will be doing all the work too. Maybe we will become like the people in the Wall-E movie.
I saw it coming.. once the computer came in..
@@tralfazyall artistic endeavors will suffer severely...
@@paulgentile1024 I'm afraid so. Even in my profession.. software engineering, AI is starting to make a big impact. Not entirely yet but probably soon.
I firmly believe this is the best moment for us musicians to remind people the power of live music. We sure had a good time earning some money through selling our cds, or even streaming royalties but, for better or for worse we’re all going back to what it used to be: getting paid to let an audience hear, connect, let loose and feel live music.
AI has been changing the world a while ago and it ain’t stopping and it ain’t gonna feel sorry for anyone… it’s better to get ahead of the curve, learn how you can use it creatively too and maybe even make some money while at it too, why not?
I can see people dancing to a salsa band. Or New Orleans zydeco music. Or bluegrass with fiddle and mandolin. Or Celtic pennywhistle. Can AI create the in the moment of that? It never will! But where will these musicians be if they can't make any money in the music biz because the biz replaced them with computer music? Playing on a street corner?
I am sad to predict that only those old enough to remember hearing live musicians will care or know the difference.
Yea, there is a group in our country who call themselves a sequencer group…they have a laptop playing all the instruments and they sing karaoke style in their concerts.. one of them pretend to play guitar.. and on closer listening, their voices are digitally processed. Kind of “live AI”😅😅
KISS just started to destroy your thoughts about "power of live music" with their "digital avatars" thing... I feel bad about all of this, honestly, and can't seem to find a solution where PEOPLE will still be making money out of their own creativity (except those who get creative at prompting, maybe? Which also won't make a lot in the long run...)
The terrifying thing though is that ai music can be performed live too.
I get it now! The person who produced my CD put some software on my voice and I didn’t like it. I know that sound. Thank you 💜
I hate this is happening, HATE IT!! I’m not a musician, just a lover of music and I’m in awe of musicians! The dedication to learn an instrument is inspiring and I admire anyone who learns an instrument and I am moved to tears sometimes by the masters!
We're done as a species.
When computers started beating the best humans at chess, did we all stop playing chess? Why play chess at all if a computer is so good and can beat any player on the planet? Somehow chess is more popular than it has ever been in history. Why do we want to see people play chess instead of computer programs duke it out? The human to human connection music gives us can never be replicated artificially, just like seeing two humans square off across a chessboard is an authentic experience. Like any technological revolution, this is an opportunity to embrace new tools. It does introduce new challenges, of course. However, just as you stated, people are still going to be looking for that human connection. Perhaps we will come to value authentic performance even more than we do now.
I'm sure that there will be a holographic band AI tour some day, like there was for Hatsune Miku. Those will be fun in their own way. However, we'll still want to see live bands play the songs we love. Maybe the impact of AI is that live in-person music is in for a major revival.
I’m actually pretty happy about this, as AI will only replace the soulless pop that has become som mainstream now.
My hope is that AI will make mediocre music abundant, and people will go back to making proper music
As an old old (now retired) engineer, I spent much of my career trying to persuade people that "Just because the technology exists, you don't have to use it."
You're a class act, @brianpateman2666
Well, what technology? Where do you draw the arbitrary line Brian? No, don't use computers to edit songs because it's way too easy. Yes, you can only edit by manually splicing tape, this is clearly the better way to do things, right? Or wait, maybe it's just the use of electricity at all? Or are you referring to quantization? Tuning vocals? Wherever the line is, it's probably arbitrary, and you're probably a dino.
@@travisyee7278Smart Fridges that can post on Twitter
AI Doorbells that tell you who's at the door
AI in your computer mice that does...idk??
NFTs
Cryptocurrency
Blockchain
Those stupid swearing filter boxes from the 80s
"You got baggage, I got room" is a beautiful line as well
AI knows how cringe humans are
You can keep your baggage, I don't want it.
thank u
Sung by a woman who wants to “fix” someone. This AI hasn’t dealt with its own issues, until the troubles begin.
Probably plagiarized from a human. AI plagiarized whole paragraphs from the NYT, so why not steal from songwriters too ¯\_(ツ)_/¯
My fear is that the world will be pumped full of this AI stuff and the real creative will be lost in the crowd. I'm a little older than you and your channel has caused me to hit the record shops again, getting all the albums I couldn't afford when I was young. Also love your artist interviews. We live in a wonderful time when I can hear stories directly from band members. You're a treasure. I'm now addicted to your channels and Tim Pierce.
Rock bands need to record albums like they did in the early days. Everybody playing in the same room and leave the little mistakes. Good old fashioned Human Intelligence. No more auto-tune, edits, or quantizing.
Yeah but then you may have people saying that they don't want to pay big $$s to hear a garage band.
Like George Harrison's All Things Must Pass album.
there will be like 300 listeners per month lmao that's about 40 cents
The last song by slash sounds like it. Pro players with pro equipment having a blast (though is a generic blues cover).
Then the producers will AutoTune, pitch and pocket it all away unless the band is already big enough to not have that done.
I've always been strongly critical of autotune and all other technologies made to "correct" singers' voices. The imperfections in our voice have a unique quality, a real kind of beauty that shouldn't be tampered with. And if a singer depends on autotune to sound decent then they shouldn't be singing in the first place, they either lack effort or talent (or both).
Exactly right. The reason it's hard to tell is because everything today is over processed. The only antidote is real music with real voices.
Exactly my thoughts. I guess it would be much easier recognize something AI did if it was told to create something that would sound as real and raw as possible. At least at its current stage of development.
I agree partially. I disagree with the antidote part. The only way to know for sure a voice is "real" is having the person directly singing to your ears. With no mic, no PA, no overdub, etc. And my second point is: soon enough no one will be able to tell the difference. Like with a professionally photoshoped image.
It will take time but, over time you will not be able to tell difference.
I feel like some genres will be immune to this and might go back a little in terms of production mentally
THIS
I'm using a different program that allows me to use my lyrics instead of just prompts. It has allowed me to bring my file full of lyrics to life. I've also figured out that a lot of they lyrics I wrote are garbage lol. Garbage in, garbage out, I just can't get some of them to work. The one's that do work, however, have come out pretty amazing. Some sound much like I heard them in my head when writing the lyrics, some a lot better! It has reignited my passion for song writing. Will I ever make a dime? Probably not. I do enjoy sharing them with others and having them ask me to send more
The scary part isn't how it sounds now, it is a brand new technology. The scary part is how quickly it will develop and improve. Give it a few years and it will be incredibly hard to tell them apart from the real thing.
Exactly what I’ve been saying! It’s easy to laugh at how bad some of the AI creations are, but to think how fast they have gotten better!
We are still at the beginning of AI use in arts. But still it has developed faster than almost anything IT-/computer related technology ever.
We used to measure technology development in years, but with AI it’s in months…
6 months is more realistic
Live is the future. Not arenas or mass venues where it’s easy to manipulate the vocals, but seeing bands in small clubs where you can actually see them play, feel the energy and ambiance of that particular show at that time.
Rock will devolve to the level jazz was in the 50’s; BUT the players, shows and performances will become more unique and momentous
People already all but refuse to pay for music at small clubs and small clubs don’t pay a guarantee. What do you mean by rock devolving to where jazz was in the 1950’s? That was peak popularity for jazz music, just as the 1960’s were peak popularity for rock music.
IF there are any clubs left to play.
Musicians would rather have a youtube channel than start a band.
@@swperry04039Yeah. My favorite jazz music comes from the 50s and early 60s: Parker, Monk, Davis, Mingus, Coltrane, Coleman, Brubeck, Getz, etc. Jazz didn't really begin to slip into irrelevance until the jazz fusion period of the 70s, followed by smooth jazz.
I'm over here practicing 4-part, 3-octave arpeggios, all qualities, all keys, hours per day. I know people can hear the sweat that went into that. Hours agonising over one chord choice. I hear someone say they "wrote" a song with AI and I just break out with hives. I wish more people would acknowledge the spectacle of humanity in musical performance, whatever the reason, but I'm still confident that they can feel it on an intuitive level. Although I loathe things like autotune, it's not just about computers, there is so much amazing computer music created by beautiful minds out there.
I'm glad the kids can hear the difference in a real performance. I'm sure they can appreciate the effort and dedication that goes into that too. We have to make sure all of us can.
100%!!!
Can AI replicate callouses on fingers?
I upped my string gauge for that. Try that, AI.
egomaniac
Really Liked the 'Just For Me' track - it's frightening that i like it, but I do. As a composer for screen...Am I worried?... hmmmm.
"...and the Grammy goes to Studio AI for their stellar hit love song, "I will find you!" sung by Liam N." [Canned applause]
Ah yes, lead single from the hit album “A Very Particular Set of Skills”.
I play Irish sessions with my friends. What you see is what you get, good or not so good, all accoustic, sink or swim. Loving it.
I played in sessions for years. So awesome!!!!! Wish there were some around here. :)
My biggest problem with AI is that the people who made the music (books, paintings, etc.) are not being compensated in any way for their work being used by a company who is making money off of their programs. Sam Altman has said that LLMs can't work if they soak up as much info as they can, copyrighted or not. To that I say...so? If you're going to use someone's copyrighted IP to make money, the person who owns the IP deserves compensation.
By the same logic I should demand from you compensation for reading this comment. After all, I wrote it, and it's my IP, how dare you peruse my property without my permission? And who knows, maybe you will reuse the golden thoughts contained herein to produce your own comments, and make money from it.
@@clray123 Are you really equating a TH-cam comment (which you wrote presumably being aware of Google's terms of service) with a song that a professional artist has released under a recording label? You really think that would stand up in a court?
We need to be careful what we consider 'property' here.
@@thennickeYou’re missing the point of his comment. Every work of art, across all mediums, has been inspired by other works of art. This obviously comes about through “experiencing” these works, i.e. watching, listening, etc. The question is whether a computer experiencing art is equivalent to humans doing the same. There are valid arguments on both sides here.
@@MrPlavik Exactly, the main question here is whether the copycat's production is "transformative" enough with respect to the original work. After all, genuine artists get inspiration from other works of art all the time. That, and AI companies are careful to train their models on material that they have already licensed for this purpose (even if the licensing terms are as bullshit and sneaky as Facebook's "opt out" preferences). In other words, if you think your copyrighted work was stolen AND reproduced unchanged by an AI, by all means sue that company and you will probably win in court. If your copyrighted work was just stolen (perceived) by an AI, but not reproduced, your chances may not be so good.
I agree that I can hear an "echo" or "vibrato" in AI singing that makes the artificiality apparent in styles of music where that's not "appropriate" (e.g., it's very noticeable in the country song; it's not something that stands out in the chill track because I expect electronic modulation on a singer's voice in that genre).
More than anything, this is a perfect reflection of how mundane pop music has become. Your average listener does not care enough about music to care about who created it.Personally, I don't mind if that crap is offered up for free or even takes the place of mundane movies and advertising. We are already so inundated by crap that humans make , if noone knows the difference, it deserves to die. Hopefully great musicians can realize the opportunity to create something so indivually powerful that AI cant recreate it and it would not seem profitable to even try. Money will no longer be a driving reason for creating great art. Human spirit will not die with this. Only an old capitolist definition of music will die and it's probably about time.
100% agreed with all of your points
you mean capitalism has become?
Musicians also need to pay the bills…how many people can have the luxury of creating something great while working at an Amazon’s warehouse… that’s the problem right there…
There's a current Mexican miracle called The Warning. Check them out.
I’ve noticed that too among some people. And if you Play something from last year they’re like “oh that’s so old” I’m like yeah so what, it’s great and never gets old to me. It’s seems like there’s only a few that study music or actually LOVE ❤️ music - u know, like actually listen to a song with headphones on dozens of times. Sometimes I get in the car and go cruising just to listen to tunes while I drive.
I grew up with the Beatles and later played guitar and keyboards in a band. The last twenty years or so, I've been studying orchestration on my own and learning the intricacies of writing for the orchestra. I have many gigabytes of sound samples and think I've gotten pretty good at producing original orchestral music. I've long had a dream of writing a movie soundtrack, even on an obscure, small budget, independent film. But I've come to realize lately that I may never get such an opportunity, as filmmakers can just cut out the composer altogether and have it done to order with AI. The earliest AI attempts at orchestral music I heard were kind of lame, but now they are getting good. Soon they will be even better. I fear for the future of the musician. But I will never stop writing music. It's my passion.
You have to keep feeding in new music otherwise, it can never grow or morph the way real music does. Basically, AI simply imitates by comparing its output to real music to see what fits and what doesnt. If no new music comes in the AI stuff will go stale, always be stuck at some point. But that probably won't happen. Too much new stuff out there.
Imagine AI training on AI. I bet that goes bad quickly.
I'm surprised you can't hear what your kids are hearing... it's hard to describe, it's almost like a subtle phaser effect over the entire track, and it shifts from left to right. It's more of a overall quality issue, and I'm sure they will improve it dramatically in the upcoming versions. My guess is that the AI deconstructs the (stolen) songs it was trained on and reproduces them with artifacts in the sample rate.
Copyright law doesn’t require an artist’s permission to study their works. Unless we convince Congress to change copyright law, these works aren’t “stolen”, no matter how unhappy we may be about it.
it just sounds like double tracked vocals to me
It sounds like aliasing to me, like old, low bitrate mp3s downloaded from limewire
That’s what I hear too!!
And your description of it is fitting.
👍
I do not like any of this A.I. junk.
I like real instruments played by real, warm blooded people. With all of the nuances and complexities only a real being can make!!
@@JoeyFTL Didn't know aliasing was a thing in music as well but that's exactly my mental picture, like a low-res image with no antialiasing that is upscaled, everything is there but it somehow feels inorganic. But it won't be there in 6-12 months, or it will be imperceptible.
I'm 12 and I've been playing guitar for a few years and I want to learn about production. It would be my dream job if I could go into the music industry, but I honestly have no idea where it will be in 10 years.