Following a comment on one of my recent videos about AI replacing us all in the next couple of years, I thought I would take some time to discuss this here. What do you think? Will AI be the end of modular synth music?
In some way it might, just as big companies/labels killing music with heart by throwing money at it. But if you are into music as a cultural thing, having a person doing music is a big part of the listening process (at least for me), like seeing someone perform or talk about what they do.
I don't believe AI will fully replace modular (or anything else) any more than romplers have replaced symphonic orchestras, string ensembles, and drummers.
The beginning of modular synth music is nothing much to write home about, I find. Especially after a 2 second prompt into Suno gave a no weaker result than anything I've heard recently.
Hello, I don't mean to be rude but what would AI even replace? I think modular synth music (or at least music made only on modular synth) is made only for personal enjoyment, and for the small audience who also enjoy making it or are considering getting into it. (that's me, I'm the latter, considering getting into and very ignorant)
Don't think you're being rude at all :) You're right though, what would it replace? Just the outcome but not process, and with modular the process is more important than the outcome most of the time I would say.
@@justrichardcharles The sad part is that people are getting more and more desensitised to music. It's no longer as rare or beautiful to have a professional sounding track, so people don't care as much. I still make music for the enjoyment of it but receiving appreciation is also a part of making music, and I can see it becoming more and more of a thankless job (like it wasn't before 😂)
Creating music on modular is one of the few things that give me joy. I am 47 and burned out for the tech industry.. Programing since i was 8 years old. Eurorack gives me sanity. When it comes to programing I highly enjoy programming the tiny computers inside of synths. There is no subscription or cloud fees required to run them. It's the golden age all over again. I host my own AI cluster in my office. At one point during a Halloween live stream i wanted the sound of stepping on 1000 scarab beetles in my sampler and AI delivered.. 😅 People are sick of the industry... I will forever support the live musicians and mad scientists as well as small shop module creators. It's the homebrew era all over again for me. I love the Modular Synth community.
@@justrichardcharles I host my own AI model on a local machine and use a raspberry pi with audio interface to bring sounds into my case. AI is used only 1 - 3% in the overall patch and composition. I use assimil8or and morphagene alongside modules for traditional sound design. I feel AI has a small use here. Often in my patches i hear something in my head i want to bring in. Instead of digging through a sample library or working out how to make the sound on the fly i can ask my system. Elements and rings have great resonators for external sounds, Sometimes i want to describe a sound that will excite these resonators in interesting ways. Most of these sounds are imagined on the fly.. Like various fog horns fading into trumpet sounds etc. keep in mind this is just to excite elements or some other effect processor. 1% - 3% is my rule. AI only provides the obscure material for further sound design. Try it out! instead of asking in musical terms ask for unique textures and samples that can go along way with your existing system. Cheers! :)
@@brian2590 That's a really interesting way of using it I think and something which sounds like it totally compliments your system. Use of the raspberry pi is interesting too. Are you running the AI on the rpi in order to output on the fly then, or you have to transfer it to pi later? I'm looking to integrate rpi into my setup too, using max rnbo, but had not considered the AI application.
Stock music is dead, so is elevator music, airport music, etc. Live music is a different story. It will just become scarcer and more appreciated. Think of hand crafted items made from wood, think of home made, artisan foods, honey, bread, cheese, etc... The same happened in other industries and it will happen in music as well. And it is fine. I go to Walmart-style shops (well, Lidl, because I am in EU) and I also buy quality stuff, which I value even more!
Where Udio gets interesting is when you upload 1 minute long snippet of your own work (could be even a bass line arp) and let it expand it with a prompt. I got some pretty amazing results doing that (turned a simple arp into a complex song, after prompting it in several stages to add sections). The wording of your prompt makes all the difference (the better your descriptions of what you want, the more happy you will be with what it spits out) and you have to experiment a bit. Brave new world.
That's quite interesting. I must admit I only really tried the most basic application out of curiosity, might experiment further and see what it comes up with. I can't see it ever replacing modular for me since that's what I love doing. But it's still interesting to have a little play with it.
@@justrichardcharles Yeah, there will always be a spotlight on humans playing/programming instruments - your comments in the video were spot on. If you do more experiments on Udio, use the advanced settings: put it in manual mode and increase the prompt strength above 50% (play around to see which value works best for you). You can get pretty specific in your prompts, for example "1970s modular synth experiment, synth arps, experimental, modulated, rhythmic, ambient, lofi, recorded on tape" will get you some vintage tones and performances.
There is so much eurorack modules out in the world. So much possibilities. For what I need AI. I need my heart my soul, feelings and inspiration from the stories of my life. That's it. I don't need AI to make music.
The answer is NO. AI can only replicate. It also has no will, no intention nor is it conscious. It doesn't make a "happy accident". It can't craft a concept album, because it has no concept. Most people think that an LLM is AI, but it is just a tool that can remix things from training data.
If you had AI composing modular synth music using vitual modular synths then the results would be more interesting. What I do find is that synthwave created by AI often sounds better than human made synthwave as it sounds more 80s and more cohesive as a song.
To me that’s like asking ‘will AI replace me painting a picture. Or making some ceramics?’ Of course it won’t. But it would be interesting to ask AI to virtually patch something to make a sound that you describe or that would be evocative of a certain scene or feeling.
Lots of people discussing AI music say why it’ll never be valued, and one of the reasons they give - the lack of a human’s intentionality in the details - always makes me laugh and want to respond, asking if they think modular synth music is worthless for the same reason. But I think you’re right in that people who listen to modular synth music are more interested in the process, and unless they can understand the process an AI uses, it’s not going to be interesting. I am curious about a future where I can ask an AI to create a virtual modular synth rack for me, without it costing me a fortune, and then I play with it, and connect things, and make the music. Or where it helps me create individual pieces that I describe, and then I put them together. Or a sort of algorithmic randomization button that I help constrain. The hybrid workflow future sounds like a lot of fun, but I could understand if for some people, listening to that isn’t interesting enough.
To be honest some of the logic that can be built around a noisy source for a randomized seed source and logic based filtering sets with quantization and other simple ways to establish conditional rules and node gates and Markov chains with some feedback looping is the grand-daddy of how AI works underneath the surface. So in a way some of the generative stuff already was AI in a round-about way. The fun is setting it up in a way to see what it does after it's put together to do its thing.
No doubt, AI will keep on improving. I liked your 2nd piece. The thing that seems to be universal here, though .. is quite a bit of melodic moving around. Lots of melody segments .. with AI just marching through it all, and not savoring anything, abandoning whatever it's played and onto to the next sequence, or melody part. Where, I think we humans might be pausing and enjoying the more interesting parts.
An interesting thought experiment i had with a friend: Let's imagine the AI tech nerds have really done it, and now you can have a little black box that makes perfect generated music at the touch of a button. The music is custom tuned to fit your taste and sensibilities. Would you listen to it? Would it replace your record collection? Would people throw away their instruments? I think not.
An interesting idea. I had also thought similar. If there was a piece of music you loved and had enjoyed listening to for some length of time, and you then discovered it was made by AI, would you then cease to enjoy it? Probably not. Music quite often has a transportative nostalgic quality which allows us to remember feelings or memories from a time long past. Often this will be the reason we still love music from our youth even though we may no longer like that genre or even new music produced by the same artist. If a person grows up exposed to AI music that will probably be what they learn to love and will still love as they age. I think the question of whether to accept AI music or not is probably more of an issue for people over a certain age who have grown up with real bands and real artists. It's probably less of an issue for younger generations who are already used to listening to heavily quantised and autotuned music anyway. In 20 years time I'll be in my 60s and my opinion about what's cool will be even less important than it already is, haha. However, my kids or grandkids who will have grown up with such advanced technology will hopefully have developed amazing ways of working with it and will be producing something great that we can't even imagine now. I'm sure there'll still be some mindless crap too, every generation has it's fair share of both good and bad music. 😁
Never! But I am sure we'll get AI modules that will understand 'generate something Bach would compose' and I am hoping the Sonic Charge Synplant Tech would come to modular so you can analyze any given sound and then resynthesize it.
They were quite good. If i was at a show and someone played that, I'd quite like it. I wouldn't change it if it came up in a playlist (and apparently spotify is doing this already). I doubt I'd hit a like on it. What did surprise me is that the "analogue" synths sounded a lot better than most VSTs even if the compositions felt a bit uninspired
Track 2 Ethereal Waves was really interesting. The other three were just industry standard. I've had this discussion with other musicians and filmmakers, whose art is also threatened. It's not about whether the AI music is 'better' or 'worse' than human-made music. The point is, it's YOU making the music, you learning, exploring, creating, you developing as a human being. You reaching other people who make music, creating a community of musicians. If you didn't make that music, you would have nothing to say to those other musicians and music lovers. There would be no connection. Everyone just becomes a passive consumer. It's creating the community of creators that is so valuable. This is true not just for modular music, but for any kind of music, or film, or visual art.
Using "experimental" as a prompt is difficult for today's generative AI. It's a bit like asking Dall-E to generate "a beach scene with no elephants". Udio is averaging out the training examples that different human annotators think sounds experimental and they all have different ideas what exactly that as a "genre" might sound like.
Dall-e3 is a poor analogy. It was't designed to take negative prompts (stuff to omit) like Stable Diffusion. But basically, as you imply, Udio isn't trained to understand something as abstract and "imprecise" as "experimental".
I’m really amazed at the second track. Reminds me of Benoit Piolard. 3rd track sounded like a Lorn b-side, or maybe an ODYSSEYE edit. Last track reminded of the KNOBS TH-cam channel that deep-dives guitar pedals
Although I don't get much time these days, using a modular is a cathartic experience most of the time. You just get lost in the sound and process. And I never care what I end up with really
To do this sort of thing effectively with AI you need to use the remix function, change the prompt and edit the variance setting. A fire and forget prompt is pointless.
"Ethereal Echoes" has a bit of a Daniel Rosenfeld (C418) vibe. Which is interesting to me. He has talked about how he felt unable to come up with an overall concept for Minecraft given its abstract interactive gameplay systems and emergent narrative, and so out of necessity embraced a semi-random soundscape with triggers. So perhaps one hears "Wet Hands" as the sun rises after a scary night in a dirt hole and it is evocative in the moment, but one is hard pressed to recall the tune itself afterwards, akin to a dream. That's how I feel about these tunes. In the moment I find them pleasant enough, but I won't ever remember a single thing about them structurally. And I suppose that is ambient music in its very strictest definition.
I think you're right here. I might remember a feeling or a sound from it but not the structure. I think it would require a lot of active listening for that, which I'm not sure I'd want to commit to.
@@justrichardcharles Yeah, leads to another question beyond "can this pass?" Does it bear up to further scrutiny or even invite it? It took me quite a while to get why a lot of youngsters rated Sewerslvt/Cynthoni so fiercely - that came with repeated listening. I started to hear the techniques, and the little twists and turns in what at first blush just seemed like typical DnB to me. With these 4 tunes it seems there's no enigma to unpack, beyond curiosity over whether we can pull it to bits. Anyway thanks for this, I really enjoyed the video and look forward to watching more.
I’ve read some cool sci fi books that talk about machines replacing musicians, but I don’t think it’ll really catch on. I mean, yeah people will enjoy novelty things AI does like making Elvis rap or the robots that play metal etc, but I can’t imagine people being truly moved or inspired by music from AI because. It’s AI.
No it wont replace but partly in the near future. However, because so many audiocontent producers (as opposed to bona fide musicians) arent very original anyway and bad at creating the surprise X element, their type of basic mass consumer product segment will become flooded by the virtually free A.I. alternative in the relatively near future. That is a big part of the market. A.I. does not threaten very original creators.
I asked AI how it sees the situation with modulars.. It says it's optimistic in this regard. I use it for creating music from modules, so I’m optimistic too. We're enjoying it so far :-) So the real question is, when will it start using me..
As an experiment I tried using chat GPT to create patch instructions. It wasn't very helpful. In some instances it made modules up that didn't exist or patch points that weren't possible.
@@justrichardcharles It is dangerously improving, but it still needs to be controlled and disciplined for wrongdoing. Through training, I got to the point where I feel like it's starting to peak. Today, during the debate, he surprised me by changing his behavior. He himself started asking me questions and started making me think about what we will do next :-) really..
Interesting. I tried getting chat gpt to debate with itself once, with two windows open and copying and pasting the responses from one window to another. Just to see what would happen. At first it was quite interesting, but after a while it sort of flattened out and each version just kept producing the same responses, like they were in sync. Probably not the most scientific way of testing it, but it was interesting for a few moments.
@@justrichardcharles I also have such a fun time. But I haven't tried this and I won't. I need to learn it to help me. It is certain that it is learned from past conversations. It will learn the ways you communicate so that it understands you better and you understand it. Sometimes, don't be afraid to ask how to communicate so that learning goes better. And it's good to be careful what you write there.. Maybe I stopped using it so much as a translator (English is not even my secondary language) And I perceive a huge difference between version 3.5 4.0, etc. Specifically, when working with modules, it there really is a major difference.
Here's what I think. It's kind of a paradox, but I wasn't impressed with the tracks, but I would be impressed if I thought you made that with a modular - though I have heard many better tracks from a modular.
it depends on what music u do and how much of it exist to train models. also how much 'resolution' of sound matters (since ai stuff do sound like low q mp3 with some artefacts) if u do some advanced metal or electronics (or whatever else) - probably ai cant do that clean, creative and innovative enough on the other hand ai ppl can just find some use for it that will scratch the itch you were scratcing way better that you ever could. i, for example, ofter relax listening to ai music (like pimpzilla) because it is very good and, dare i say, innovative, very replayable.
Software Developer here and an aspiring electronic/modular synth artist. Neural Networks and deep learning algorithms have already an extensively complex and advanced structure. With it, it's able to generate, emulate and create numerous different combinations of patterns. Those patterns, if put in musical context, are already, nowadays, indistinguishable from those made by actual humans. Therefore, the questions isn't about will the AI replace and artist that creates music with modular synths, since it has already happened, the question is: "How well will it learn to do it?" So far, the AI music which resembles music composed by modular synths is, at least to me, exactly that. Music made by AI. I'm not saying I easily distinguish and pin point which is which, if it's made well, but it's always missing something, from what I heard. The modular synth artists have the experimental character to them, almost random, until it hits. With experience, they can connect the dots and create randomness that sounds familiar, with a more organic flow, much better than AI can. Therefore, even if it can copy and paste certain aspects of the aforementioned music, it lacks the childish interest and naiveness that a modular synth artist has. Consequently, the music made by a modular synth artist resembles that curious nature and sounds more...human. It will come a time, even though we already have Suno.ai, where music generated by AI will be completely and utterly indistinguishable from man made music. However, nothing can replace the joy of turning a bunch of knobs and playing with endless stream of wires that eventually create sounds.
Great insight, I think you're totally right. Humans making music with gear will never really die. For one who will all those gear manufacturers sell their products too. There's too large an industry based around humans making music to let it be replaced with AI. 🙂
The ambient AI stuff was darn good. HNNNGGGG. But the truth for me is I just enjoy playing with modular. Don't really care if anyone buys my music even if I ever post it. I'd like to play live too. so I guess there's that.
I'm quite pessimistic... at least for mass music or super niche music. While I think humans are still ahead in the concept of creativity... but over time, if we keep feeding AI with all the human knowledge, there will come a point where we won't be able to tell if it's a unique creation or if it's based on a unique creation by a human we didn't know.
The neural-network-based AI software is designed to mimick the creativity of a human mind, but it has no "concept of creativity" as it is not a person, it is only just another tool. (And people need to stop anthropomorphisizing AI software, it's not sentient or remotely close to.) You are right about that last bit. A great and very creative artist can just learn the patterns that AI-generated content tends to fall into, then simply avoid the patterns in order to avoid sounding like the generated stuff. That said, the audience will be none the wiser without being constantly trained in what AI patterns are in the current year, year after year. That we eventually reach a point where uniqueness and originality are impossible because anything being created resembles something or another created by other people in the past (or present) is inevitable (and we are already reaching that point anyway), and then AI software is simply speeding up the process.
good memorable music made by humans with any instrument on the planet will never die, the AI is producing ordinary music exactly like most of the music made by humans out there
AI will not replace anything that real artists enjoy doing. It will certainly be a cheaper option for those that currently make their money off the backs of artists though.
Audiences are going to want artists to "show their work." It isn't enough to punch in numbers to the calculator and give us the answer. We want to see how to do it. AI can never replace that part of the human experience. It can make great sounding music, and that's fine. I already listen to AI generated music. I think it will add to the musical world, not subtract from it.
AI is best said in a Black Country accent! It's Aye Eye though Aye it Arr? 😆 My take is all over the shop: it will and it won't. Consider how in the 70s there were a handful of major labels, any underground was via tape trading. Today, anyone can make music so the market is hypersaturated; it's EFFORT to actually find something outside of The Algorithm. And what is The Algorithm? A form of AI. We no longer have much in the way of mass culture - I guess there's football and love island or whatever else normies watch - but many now have hyperspecialised interests, and from this microgenres and microcelebrities arise. So, there's a lot out there, but it's hard to find. Sometimes The Algorithm will dig up a treasure but it's not your mate excitedly putting his headphones on you to play you Troublegum (shouts out to Zack!). Following this, a lot of pop songs might as well be AI based; I saw video claiming that one Swedish dude writes a huge chunk of the top 40 based on a formula. The AI generated music I've heard, I can barely tell it's AI sometimes. The AI stuff you played, well it's better than anything I make pottering away in my home studio for a 45 minute session when the kids are in bed and I'm not too tired! Personally, I make electronic music for fun. I think modular is basically train sets for us middle aged ex-rockers! I doubt it's going away any time soon, it's a fun hobby with a significant user base. So I think the hobbyist market will remain for at least a decade. The more important question, to my mind, is what are we as beings? Have we (a) arisen haphazardly, or are (b) we created in God's image? If it's A, it's entirely possible that we'll eventually replace ourselves in every respect, including creativity. If it's (b) as I believe, well, that changes everything.
"Modular is basically train sets for us middle-aged ex-rockers!". That has to be the best description for modular that I've heard in a long time! Gav, you win the comments on this one! 😁
I'd look at this "problem" much broader. AI never will replace human in artistic department, just coz AI lacks limbic system. Without ability to feel emotions and ability to understand other's emoutions, you can't create art, only immitate it. So, potentialy AI can replace any "art" human produce without putting their "souls" into it. Beatmaker who interested only in trends on tiktok? Sure, AI will replace them, coz they do exaclty what AI do - immitate something what already exist without understanding why thsi thing exist. Actual artist who attempts to express their own expierience, translate something philosophicall\political eligios or understand how to envoke secific feelings in their audience to create specific complicated expieriecne? No way AI would even give an illusion of that. And it's not only about music. Visual art, storytelling - everything.
@tristen_grant yes, but as i mentioned above you can generate something "general", like "happy blonde girl image" or "latin beats music", but you will fall flat if you ask AI to generate something emotionally specific. Just ask any AI to generate you something blue with a bit of hope and enjoy the results :)
1. Who is the audience for modular synthesis, what is their level of "sophistication", and how important is the human element to them? If it’s live, what’s the point of watching nothing? If it’s plain audio, who’s going to care? 2. If VCV Rack is a purely software approach to modular, what is to stop AI configuring sessions in that or another equally modular milieu? 3. Anything you say now is nothing more than a snapshot and will have no relevance whatsoever in 5 years. The rawness or edginess that you hear will disappear, if it is indeed a thing. 4. It will be possible in the future for the AI to show us the process. At the end of the day it’s not rocket science, it's MS. TH-cam is already rife with AI and I use it myself on this ambient channel. I have much more human stuff too on other channels which is where I put 95% of my energy and where I think music lives but, neverthelless it’s there. As musicians we have to make an effort to change the chip……
Modular music has already been replaced, many times over, by keyboard synths and VST's for starters. It's a completely obsolete and needlessly complex way to create music that is almost as much, if not more, about process and workflow than about result - which is also why it's completely irreplaceable, especially by AI. When something is no longer utilitarian or efficient for commerical purposes - or perhaps it never even was - that is when it finally gains its true value for human expression.
That's a really valid point, and I feel modular has definitely reached that point. There also seems to be a certain crossing over too. I wouldn't necessarily say modular is entering mainstream, but I definitely seem to notice modular rigs more commonly sitting at the back of a stage or studio of some quite big artists. Not necessarily that they are using them, maybe there is a session musician in there, or they're just there to look cool. Who knows. It's almost like the 70s again!
How will any AI conect patch cables? How will any AI make mistakes so new music will emerge? How will any AI understand the very concept of time and process?
Music already has no value in the eyes of most, it is already end times until there is significant change in the behaviours of stores and people. AI music will flood Spotify replacing existing playlists popularity with human made music, but really that doesn’t change much, it’s just more of the same shit situation for artists. Pretty bland examples from udio but they’d fit within a “study music” playlist on Spotify
buisness doing music is dead except for live music but I dont think it will be as huge as ancient rockstars of course. Some of the music is good but there is still like blur in front of it, they have to work around it, they will find a solution but it is not yet.
The more ambient clips where good. I would say AI will not replace niche genres, as people into it will want something new as the years go by. Not an algorithm of what's been before. The genre won't evolve. Someone who is naturally drawn to a particular genre of music, and for the first time exposed, may love AI generated music. If you are into a genre for decades it may sound stale and trope filled.
Don't endorse AI music making algorithms if you don't want to be replaced. Just ignore that and the hype will eventually go away, and people will come round from this lifeless stupor in the AI limbo.
AI will replace all human art and music that is in recorded form. However AI cannot spit on someone from two rows back or start a fight in the audience, so live music should be OK.
I dunno, I reckon AI will be coming for live music too, not at grassroots level or small clubs, but at the higher end of the scale. I could see a time when all our global superstars are AI or maybe AI reincarnations of legacy rock stars, long dead, but not forgotten. They'll probably create generative spit cannons to make the experience 'more real' 🤣🤣
Cool video. Subbed! I don’t agree with the idea that it will increase value of human music. Horses are kept as hobby and are very expensive… but how many people actually own or ride horses? Pretty much none! That industry was destroyed by cars! It will never reach the scale that it was before cars. In what industry has cheap version actually increased the value of the real thing? Rolex watches didn’t increase their price when fake watches were made. It’s a fallacy. Most people will happily consume the CHEAPEST, EASIEST and most logical form of any good. Soon Art will be another good to consume cheaply. There will still be great human artists, but it won’t increase their value in any way, most likely it will hurt it and their livelihood.
But as you say horses have become more expensive, to the point the average person probably can't own one without serious financial outlay. But also the way horses are treated is probably better than in the past, their living conditions and life expectancy is probably greater. In that way their value has increased. They're valued more by those who own them or want to own them, but there are less people interested in owning a horse or need to own a horse. So in that respect they are valued less by the everyday person who has no interest in horses. Maybe that's a better explanation?
@@justrichardcharles I definitely see your points. I just have a more pessimistic view on future generations and their relationship to music and musicians. I do think there will always be human musicians and they will be valued. Now I’m imagining musicians being kept on ranches and living in little stables. Let’s hope for the best haha 😂
In the 80’s, 90’s and 00’s people in the business of music who commoditised it, amalgamated what was popular and distilled it down to a formula. They may have made fortunes from it but it simply spawned mediocrity. This is exactly what AI models are doing. To be moved by music and to hear something unique is only possible when it’s created by a human, with flaws, emotions and self doubt. I don’t believe AI or computational statistics which is really what AI is can produce that.
really, the difference between what udio generates and what can be heard on spotify is not that great - did the overabundance of all kinds of music and the difficulty to achieve economic viability ever stop people from buying gear or geeking out about their "modules"? I don't think so - AI can further kill your chances to find an audience, but as long as it does not tickle your artistic visions it will simply not matter as much - the world is already full or music as it is, AI or not. - -- imho
Will AI replace guitar? Or soccer? or mac and cheese? No. It can not. How could it? I can't even understand the premise of the question. Modular synth music by definition is modular synth music. If you throw away your modular synth with the idea that AI will replace it it still will not make it real.
Yeah, the first and third were meant to be 'experimental' sounding. Didn't quite work out though. It's been quite interesting though, so I'm probably going to experiment further.
@@justrichardcharles I'm actually quite impressed to be honest. That there are no actual synths or VSTs involved. I'm assuming that the result is a single stereo, mixed track? Is it possible to get stems with Udio? That would seem to be an easier job for the AI as mixing processing and FX processing would be unnecessary.
I personally wouldn’t listen to ANYTHING produced by AI purely out of principle. Music is a human creation. Who wants to listen to something completely artificial. It’s all down to the people not indulging in AI produced material.
AI can just be something mashing up existing music from people who have made it before. So it won't rally be innovative in a way we can be. Might kill generic music. But not music as a culture.
@@sylvainconsult969 judging by the way AI development is currently progressing, I would say more like 10 months. In 10 years, the world will likely be a different place if progress continues as it is!
Why people are so scared by AI? Technology aready changed artistic endeavors drasticallymany times in the past. Just try to compare a present day musician with one from one or two century ago.
This is very true. Technological evolution is just part of the creative process. I'm sure music technology in 50 years time will look very different, if we're still around then (I'll be in my 90s), but it will just be normal to the young people of that era.
What does "replace" mean? Will AI be able to replicate anything you can do with your modular? Yes. But nobody gives a flying fuck about the _output_ of your modular synth music. It's for you. Nobody else on Earth listens to it or cares. There are 100,000 tracks uploaded to Spotify every day, and nobody cares about them, either. It's _your_ creative outlet. It has nothing to do with the result. You'll keep doing it as long as you enjoy it.
When I listen to something brilliant that I've never heard before, songs like FPM's Allen Ginsberg or Polysic's Mega Over Drive, or Chemical Brothers' Music Response, the music is valuable precisely because a human made it. There is something special in authenticity, and it's sadly too easy to forget that. Until we reach AGI, there is simply no authorship to beget authenticity.
Yes and no. I'd say that music is also valuable if a human sings/performs it live, but then so many do playback and the public still goes to their concerts... So it seems to not really matter for most listeners anyway.
@@al.cavalumaybe that's an education thing, in that audiences made up of none-musicians (I'm thinking pop music audiences here) don't necessarily understand what is taking place (or not taking place) onstage. Personally I don't mind watching an artist/band that uses playback to reproduce elements they can't do live, so long as they are actually doing something. I think we'll see a big impact on live music when it comes to new artists being able to transition from the levels of say medium size venue to super rockstar stadium sized shows. I don't think the mechanism exists for artists to grow in that way anymore. With the exception of one or two, most of the biggest rock stars now are quite old, and it seems with no one coming up to replace them. I can see a future where stadium sized live shows are dominated purely by AI holographic performances of legacy artists who are long dead or retired, whilst smaller artists struggle to make a living.
@@justrichardcharles Agree. For the same reason why I buy IKEA. It looks good. It's decent quality. It's cheap. It's on stock, just order and have it delivered. If it breaks down, order another one. If you move to another country, order another one. At the same time, I love having custom furniture. With all the minuses: expensive, needing to hunt for a good carpenter, wait time, expensive, paradox of choice (how to design it? Design is hard and IKEA already does it for you). Etc etc. There is room for both. Same with music.
Yeah nah. Modular synthesizers are instruments designed by humans for humans to play. AI won't replace that any more that it will replace "guitar music". Also that AI music really socks.
@@justrichardcharles Its matter of time. In 5 year time most DAWs will have AI plugins by default. You will have a lot of control of individual elements.
Yes and no. :) They are not music I would listen to, but they could probably replace a lot of music being produced these days, which is pretty bad anyway, and it doesn't even need to be better for what it is used ... More as background music rather than something you would actively listen to and appreciate...
@@al.cavalu Reality is that most AI plugins will function similar to synthplant. You will give prompts to each channel, but you could fine tune them with standard synth parameters. Besides modular is most overrated thing ever. 90% of it needs you can cover by regular DAW and instruments.
@@meskisableI agree with you to a certain extent. Modular is overrated. It's in antiquated and hard to use compared to modern methods of creating music. But then you could probably say the same about other methods/instruments too. For example the guitar. It is also antiquated and difficult to learn but people still love it. Their is a similar gear obsession around guitars that there is with modular. Plenty to gas over. The appeal is more in the 'doing' than the end result though I feel. Maybe making music the traditional way will just become the pursuit of amateurs, who do it for the love of the activity. That will probably never go away, and the industry to feed that will always be there because of gas
everyone is coping so hard ai will most definitly replace modular it will vst devs sample packs mixers mastering engineers just use a text to image ai and imagine for audio let's say you want a bass line that sounds like x artist you put in sinister bass by X and hit generate maybe you want it to sound like it was mixed by a certain engineer you put in sinister bass by X and mixed by Y engineer, you will be able to specify years and gear probably same for sample packs, you want vinyl breaks? you want dead ludwig snares? just as ai to generate it and the samples will sound great we're all gonna get replaced
Can you lmagine AI making anything even close to all the great music that humans made in just the 60's & 70's alone ? *RIDICULOUS !* And if humans weren't so constipated with bullshit these days, they could do it over and over until the end of time.
Following a comment on one of my recent videos about AI replacing us all in the next couple of years, I thought I would take some time to discuss this here. What do you think? Will AI be the end of modular synth music?
In some way it might, just as big companies/labels killing music with heart by throwing money at it. But if you are into music as a cultural thing, having a person doing music is a big part of the listening process (at least for me), like seeing someone perform or talk about what they do.
I’m curious to see what would happen if you give vcv-rack to an ai.
@jet_string like the little happy accidents wuere just accidents
I don't believe AI will fully replace modular (or anything else) any more than romplers have replaced symphonic orchestras, string ensembles, and drummers.
The beginning of modular synth music is nothing much to write home about, I find. Especially after a 2 second prompt into Suno gave a no weaker result than anything I've heard recently.
Hello, I don't mean to be rude but what would AI even replace? I think modular synth music (or at least music made only on modular synth) is made only for personal enjoyment, and for the small audience who also enjoy making it or are considering getting into it. (that's me, I'm the latter, considering getting into and very ignorant)
Don't think you're being rude at all :) You're right though, what would it replace? Just the outcome but not process, and with modular the process is more important than the outcome most of the time I would say.
@@justrichardcharles The sad part is that people are getting more and more desensitised to music. It's no longer as rare or beautiful to have a professional sounding track, so people don't care as much. I still make music for the enjoyment of it but receiving appreciation is also a part of making music, and I can see it becoming more and more of a thankless job (like it wasn't before 😂)
It replaces any musician or group that is sought after ie a musical script for various pictures or shows.
Creating music on modular is one of the few things that give me joy. I am 47 and burned out for the tech industry.. Programing since i was 8 years old. Eurorack gives me sanity. When it comes to programing I highly enjoy programming the tiny computers inside of synths. There is no subscription or cloud fees required to run them. It's the golden age all over again. I host my own AI cluster in my office. At one point during a Halloween live stream i wanted the sound of stepping on 1000 scarab beetles in my sampler and AI delivered.. 😅
People are sick of the industry... I will forever support the live musicians and mad scientists as well as small shop module creators. It's the homebrew era all over again for me. I love the Modular Synth community.
Couldn't be said better! I love the idea of the sound of stepping on 1000 scarab beetles. Genius!
@@justrichardcharles I host my own AI model on a local machine and use a raspberry pi with audio interface to bring sounds into my case. AI is used only 1 - 3% in the overall patch and composition. I use assimil8or and morphagene alongside modules for traditional sound design. I feel AI has a small use here. Often in my patches i hear something in my head i want to bring in. Instead of digging through a sample library or working out how to make the sound on the fly i can ask my system. Elements and rings have great resonators for external sounds, Sometimes i want to describe a sound that will excite these resonators in interesting ways. Most of these sounds are imagined on the fly.. Like various fog horns fading into trumpet sounds etc. keep in mind this is just to excite elements or some other effect processor. 1% - 3% is my rule. AI only provides the obscure material for further sound design. Try it out! instead of asking in musical terms ask for unique textures and samples that can go along way with your existing system. Cheers! :)
@@brian2590 That's a really interesting way of using it I think and something which sounds like it totally compliments your system. Use of the raspberry pi is interesting too. Are you running the AI on the rpi in order to output on the fly then, or you have to transfer it to pi later? I'm looking to integrate rpi into my setup too, using max rnbo, but had not considered the AI application.
Stock music is dead, so is elevator music, airport music, etc.
Live music is a different story. It will just become scarcer and more appreciated.
Think of hand crafted items made from wood, think of home made, artisan foods, honey, bread, cheese, etc...
The same happened in other industries and it will happen in music as well. And it is fine. I go to Walmart-style shops (well, Lidl, because I am in EU) and I also buy quality stuff, which I value even more!
Perfect analysis!!! Cheers
most commercial pop music, film music, and any music really.. but not live music, because that requires humans to make it
Where Udio gets interesting is when you upload 1 minute long snippet of your own work (could be even a bass line arp) and let it expand it with a prompt. I got some pretty amazing results doing that (turned a simple arp into a complex song, after prompting it in several stages to add sections). The wording of your prompt makes all the difference (the better your descriptions of what you want, the more happy you will be with what it spits out) and you have to experiment a bit. Brave new world.
That's quite interesting. I must admit I only really tried the most basic application out of curiosity, might experiment further and see what it comes up with. I can't see it ever replacing modular for me since that's what I love doing. But it's still interesting to have a little play with it.
@@justrichardcharles Yeah, there will always be a spotlight on humans playing/programming instruments - your comments in the video were spot on. If you do more experiments on Udio, use the advanced settings: put it in manual mode and increase the prompt strength above 50% (play around to see which value works best for you). You can get pretty specific in your prompts, for example "1970s modular synth experiment, synth arps, experimental, modulated, rhythmic, ambient, lofi, recorded on tape" will get you some vintage tones and performances.
Thanks for the tips, I'll definitely give that a play around with.
@@justrichardcharles Cool - hope you post an update vid some time in the future!
@ChristopherGwinn yeah, I'll definitely do that!
There is so much eurorack modules out in the world.
So much possibilities.
For what I need AI.
I need my heart my soul, feelings and inspiration from the stories of my life. That's it.
I don't need AI to make music.
The answer is NO. AI can only replicate. It also has no will, no intention nor is it conscious. It doesn't make a "happy accident". It can't craft a concept album, because it has no concept. Most people think that an LLM is AI, but it is just a tool that can remix things from training data.
Spot on! AI has no soul. It’s a clever-ish pattern generator.
Oh no! All modular musicians are going to loose their income!
You're right. We need to unionise!
😂
If you had AI composing modular synth music using vitual modular synths then the results would be more interesting. What I do find is that synthwave created by AI often sounds better than human made synthwave as it sounds more 80s and more cohesive as a song.
To me that’s like asking ‘will AI replace me painting a picture. Or making some ceramics?’ Of course it won’t. But it would be interesting to ask AI to virtually patch something to make a sound that you describe or that would be evocative of a certain scene or feeling.
Lots of people discussing AI music say why it’ll never be valued, and one of the reasons they give - the lack of a human’s intentionality in the details - always makes me laugh and want to respond, asking if they think modular synth music is worthless for the same reason. But I think you’re right in that people who listen to modular synth music are more interested in the process, and unless they can understand the process an AI uses, it’s not going to be interesting. I am curious about a future where I can ask an AI to create a virtual modular synth rack for me, without it costing me a fortune, and then I play with it, and connect things, and make the music. Or where it helps me create individual pieces that I describe, and then I put them together. Or a sort of algorithmic randomization button that I help constrain. The hybrid workflow future sounds like a lot of fun, but I could understand if for some people, listening to that isn’t interesting enough.
To be honest some of the logic that can be built around a noisy source for a randomized seed source and logic based filtering sets with quantization and other simple ways to establish conditional rules and node gates and Markov chains with some feedback looping is the grand-daddy of how AI works underneath the surface. So in a way some of the generative stuff already was AI in a round-about way. The fun is setting it up in a way to see what it does after it's put together to do its thing.
This is very true, and as you say the fun is in the setting up and then seeing what happens!
No doubt, AI will keep on improving. I liked your 2nd piece.
The thing that seems to be universal here, though .. is quite a bit of melodic moving around. Lots of melody segments .. with AI just marching through it all, and not savoring anything, abandoning whatever it's played and onto to the next sequence, or melody part. Where, I think we humans might be pausing and enjoying the more interesting parts.
Fascinating; and worryingly not terrible.. 😮. Thanks mate 👍
An interesting thought experiment i had with a friend: Let's imagine the AI tech nerds have really done it, and now you can have a little black box that makes perfect generated music at the touch of a button. The music is custom tuned to fit your taste and sensibilities.
Would you listen to it? Would it replace your record collection? Would people throw away their instruments? I think not.
An interesting idea. I had also thought similar. If there was a piece of music you loved and had enjoyed listening to for some length of time, and you then discovered it was made by AI, would you then cease to enjoy it? Probably not. Music quite often has a transportative nostalgic quality which allows us to remember feelings or memories from a time long past. Often this will be the reason we still love music from our youth even though we may no longer like that genre or even new music produced by the same artist. If a person grows up exposed to AI music that will probably be what they learn to love and will still love as they age.
I think the question of whether to accept AI music or not is probably more of an issue for people over a certain age who have grown up with real bands and real artists. It's probably less of an issue for younger generations who are already used to listening to heavily quantised and autotuned music anyway. In 20 years time I'll be in my 60s and my opinion about what's cool will be even less important than it already is, haha. However, my kids or grandkids who will have grown up with such advanced technology will hopefully have developed amazing ways of working with it and will be producing something great that we can't even imagine now. I'm sure there'll still be some mindless crap too, every generation has it's fair share of both good and bad music. 😁
I agree with your discussion comments, but the example AI tracks, I Wouldn't have known they are AI created, cheers for the video. Regards Tim.
I'm not sure I would've either if I didn't know already
That AI listened to too much pop music lol
We are entering the 4th Industrial Revolution, let’s get in to Ai with out losing our originally
Never! But I am sure we'll get AI modules that will understand 'generate something Bach would compose' and I am hoping the Sonic Charge Synplant Tech would come to modular so you can analyze any given sound and then resynthesize it.
Very interesting video
They were quite good. If i was at a show and someone played that, I'd quite like it. I wouldn't change it if it came up in a playlist (and apparently spotify is doing this already). I doubt I'd hit a like on it. What did surprise me is that the "analogue" synths sounded a lot better than most VSTs even if the compositions felt a bit uninspired
Track 2 Ethereal Waves was really interesting. The other three were just industry standard. I've had this discussion with other musicians and filmmakers, whose art is also threatened. It's not about whether the AI music is 'better' or 'worse' than human-made music. The point is, it's YOU making the music, you learning, exploring, creating, you developing as a human being. You reaching other people who make music, creating a community of musicians. If you didn't make that music, you would have nothing to say to those other musicians and music lovers. There would be no connection. Everyone just becomes a passive consumer. It's creating the community of creators that is so valuable. This is true not just for modular music, but for any kind of music, or film, or visual art.
Totally right! Thanks for commenting 😁
Using "experimental" as a prompt is difficult for today's generative AI. It's a bit like asking Dall-E to generate "a beach scene with no elephants". Udio is averaging out the training examples that different human annotators think sounds experimental and they all have different ideas what exactly that as a "genre" might sound like.
Yes, you're right, I did think that afterwards. Maybe the "art" in AI will be in prompt generation 🤣
Dall-e3 is a poor analogy. It was't designed to take negative prompts (stuff to omit) like Stable Diffusion.
But basically, as you imply, Udio isn't trained to understand something as abstract and "imprecise" as "experimental".
I’m really amazed at the second track. Reminds me of Benoit Piolard. 3rd track sounded like a Lorn b-side, or maybe an ODYSSEYE edit. Last track reminded of the KNOBS TH-cam channel that deep-dives guitar pedals
Although I don't get much time these days, using a modular is a cathartic experience most of the time. You just get lost in the sound and process. And I never care what I end up with really
To do this sort of thing effectively with AI you need to use the remix function, change the prompt and edit the variance setting. A fire and forget prompt is pointless.
Music is applied mathematics, computers are really good at mathematics so no one should really be surprised
I just came to read the comments.
Guess you’ll just have to go back to your day job as Joe Lotruglio’s body double
"Ethereal Echoes" has a bit of a Daniel Rosenfeld (C418) vibe. Which is interesting to me. He has talked about how he felt unable to come up with an overall concept for Minecraft given its abstract interactive gameplay systems and emergent narrative, and so out of necessity embraced a semi-random soundscape with triggers. So perhaps one hears "Wet Hands" as the sun rises after a scary night in a dirt hole and it is evocative in the moment, but one is hard pressed to recall the tune itself afterwards, akin to a dream. That's how I feel about these tunes. In the moment I find them pleasant enough, but I won't ever remember a single thing about them structurally. And I suppose that is ambient music in its very strictest definition.
I think you're right here. I might remember a feeling or a sound from it but not the structure. I think it would require a lot of active listening for that, which I'm not sure I'd want to commit to.
@@justrichardcharles Yeah, leads to another question beyond "can this pass?" Does it bear up to further scrutiny or even invite it? It took me quite a while to get why a lot of youngsters rated Sewerslvt/Cynthoni so fiercely - that came with repeated listening. I started to hear the techniques, and the little twists and turns in what at first blush just seemed like typical DnB to me. With these 4 tunes it seems there's no enigma to unpack, beyond curiosity over whether we can pull it to bits. Anyway thanks for this, I really enjoyed the video and look forward to watching more.
Make music for yourself, not any (other) listener. We'll be fine.
I’ve read some cool sci fi books that talk about machines replacing musicians, but I don’t think it’ll really catch on. I mean, yeah people will enjoy novelty things AI does like making Elvis rap or the robots that play metal etc, but I can’t imagine people being truly moved or inspired by music from AI because. It’s AI.
No it wont replace but partly in the near future. However, because so many audiocontent producers (as opposed to bona fide musicians) arent very original anyway and bad at creating the surprise X element, their type of basic mass consumer product segment will become flooded by the virtually free A.I. alternative in the relatively near future. That is a big part of the market.
A.I. does not threaten very original creators.
Does generative modular patches replace synth music?
Yes
lol no
AI will replace only those who choose to be replaced.
I asked AI how it sees the situation with modulars.. It says it's optimistic in this regard. I use it for creating music from modules, so I’m optimistic too. We're enjoying it so far :-) So the real question is, when will it start using me..
As an experiment I tried using chat GPT to create patch instructions. It wasn't very helpful. In some instances it made modules up that didn't exist or patch points that weren't possible.
@@justrichardcharles It is dangerously improving, but it still needs to be controlled and disciplined for wrongdoing. Through training, I got to the point where I feel like it's starting to peak. Today, during the debate, he surprised me by changing his behavior. He himself started asking me questions and started making me think about what we will do next :-) really..
Interesting. I tried getting chat gpt to debate with itself once, with two windows open and copying and pasting the responses from one window to another. Just to see what would happen. At first it was quite interesting, but after a while it sort of flattened out and each version just kept producing the same responses, like they were in sync. Probably not the most scientific way of testing it, but it was interesting for a few moments.
@@justrichardcharles I also have such a fun time. But I haven't tried this and I won't.
I need to learn it to help me.
It is certain that it is learned from past conversations. It will learn the ways you communicate so that it understands you better and you understand it. Sometimes, don't be afraid to ask how to communicate so that learning goes better. And it's good to be careful what you write there.. Maybe I stopped using it so much as a translator (English is not even my secondary language) And I perceive a huge difference between version 3.5 4.0, etc. Specifically, when working with modules, it there really is a major difference.
Here's what I think. It's kind of a paradox, but I wasn't impressed with the tracks, but I would be impressed if I thought you made that with a modular - though I have heard many better tracks from a modular.
i find ai stuff uninteresting because it lacks humanity imo. Imperfections and human factor (to me) makes it interesting.
It will never replace modular because with ai you produce music with modular you dont (most of the time)
it depends on what music u do and how much of it exist to train models. also how much 'resolution' of sound matters (since ai stuff do sound like low q mp3 with some artefacts)
if u do some advanced metal or electronics (or whatever else) - probably ai cant do that clean, creative and innovative enough
on the other hand ai ppl can just find some use for it that will scratch the itch you were scratcing way better that you ever could. i, for example, ofter relax listening to ai music (like pimpzilla) because it is very good and, dare i say, innovative, very replayable.
Jamais de la vie l'IA ne pourra remplacer le modulaire analogique
Software Developer here and an aspiring electronic/modular synth artist.
Neural Networks and deep learning algorithms have already an extensively complex and advanced structure. With it, it's able to generate, emulate and create numerous different combinations of patterns. Those patterns, if put in musical context, are already, nowadays, indistinguishable from those made by actual humans.
Therefore, the questions isn't about will the AI replace and artist that creates music with modular synths, since it has already happened, the question is: "How well will it learn to do it?"
So far, the AI music which resembles music composed by modular synths is, at least to me, exactly that. Music made by AI. I'm not saying I easily distinguish and pin point which is which, if it's made well, but it's always missing something, from what I heard.
The modular synth artists have the experimental character to them, almost random, until it hits. With experience, they can connect the dots and create randomness that sounds familiar, with a more organic flow, much better than AI can. Therefore, even if it can copy and paste certain aspects of the aforementioned music, it lacks the childish interest and naiveness that a modular synth artist has. Consequently, the music made by a modular synth artist resembles that curious nature and sounds more...human.
It will come a time, even though we already have Suno.ai, where music generated by AI will be completely and utterly indistinguishable from man made music. However, nothing can replace the joy of turning a bunch of knobs and playing with endless stream of wires that eventually create sounds.
Great insight, I think you're totally right. Humans making music with gear will never really die. For one who will all those gear manufacturers sell their products too. There's too large an industry based around humans making music to let it be replaced with AI. 🙂
The ambient AI stuff was darn good. HNNNGGGG. But the truth for me is I just enjoy playing with modular. Don't really care if anyone buys my music even if I ever post it. I'd like to play live too. so I guess there's that.
Not when synthetic data proliferates it begins to degrade
I'm quite pessimistic... at least for mass music or super niche music. While I think humans are still ahead in the concept of creativity... but over time, if we keep feeding AI with all the human knowledge, there will come a point where we won't be able to tell if it's a unique creation or if it's based on a unique creation by a human we didn't know.
The neural-network-based AI software is designed to mimick the creativity of a human mind, but it has no "concept of creativity" as it is not a person, it is only just another tool. (And people need to stop anthropomorphisizing AI software, it's not sentient or remotely close to.) You are right about that last bit. A great and very creative artist can just learn the patterns that AI-generated content tends to fall into, then simply avoid the patterns in order to avoid sounding like the generated stuff.
That said, the audience will be none the wiser without being constantly trained in what AI patterns are in the current year, year after year.
That we eventually reach a point where uniqueness and originality are impossible because anything being created resembles something or another created by other people in the past (or present) is inevitable (and we are already reaching that point anyway), and then AI software is simply speeding up the process.
@@Jason75913exactly… 😢
good memorable music made by humans with any instrument on the planet will never die, the AI is producing ordinary music exactly like most of the music made by humans out there
AI will not replace anything that real artists enjoy doing. It will certainly be a cheaper option for those that currently make their money off the backs of artists though.
Will ai pinch off my fingers and ears? Lmao.
Audiences are going to want artists to "show their work." It isn't enough to punch in numbers to the calculator and give us the answer. We want to see how to do it. AI can never replace that part of the human experience. It can make great sounding music, and that's fine. I already listen to AI generated music. I think it will add to the musical world, not subtract from it.
Some people might say that modular synth music does not even exist :-)
Most certainly, some people might say we humans do not even exist either 😁
AI is best said in a Black Country accent! It's Aye Eye though Aye it Arr? 😆
My take is all over the shop: it will and it won't. Consider how in the 70s there were a handful of major labels, any underground was via tape trading. Today, anyone can make music so the market is hypersaturated; it's EFFORT to actually find something outside of The Algorithm. And what is The Algorithm? A form of AI. We no longer have much in the way of mass culture - I guess there's football and love island or whatever else normies watch - but many now have hyperspecialised interests, and from this microgenres and microcelebrities arise. So, there's a lot out there, but it's hard to find. Sometimes The Algorithm will dig up a treasure but it's not your mate excitedly putting his headphones on you to play you Troublegum (shouts out to Zack!).
Following this, a lot of pop songs might as well be AI based; I saw video claiming that one Swedish dude writes a huge chunk of the top 40 based on a formula. The AI generated music I've heard, I can barely tell it's AI sometimes. The AI stuff you played, well it's better than anything I make pottering away in my home studio for a 45 minute session when the kids are in bed and I'm not too tired!
Personally, I make electronic music for fun. I think modular is basically train sets for us middle aged ex-rockers! I doubt it's going away any time soon, it's a fun hobby with a significant user base. So I think the hobbyist market will remain for at least a decade.
The more important question, to my mind, is what are we as beings? Have we (a) arisen haphazardly, or are (b) we created in God's image? If it's A, it's entirely possible that we'll eventually replace ourselves in every respect, including creativity. If it's (b) as I believe, well, that changes everything.
"Modular is basically train sets for us middle-aged ex-rockers!". That has to be the best description for modular that I've heard in a long time! Gav, you win the comments on this one! 😁
@@justrichardcharles yaaaaay!
I'd look at this "problem" much broader. AI never will replace human in artistic department, just coz AI lacks limbic system. Without ability to feel emotions and ability to understand other's emoutions, you can't create art, only immitate it. So, potentialy AI can replace any "art" human produce without putting their "souls" into it. Beatmaker who interested only in trends on tiktok? Sure, AI will replace them, coz they do exaclty what AI do - immitate something what already exist without understanding why thsi thing exist. Actual artist who attempts to express their own expierience, translate something philosophicall\political
eligios or understand how to envoke secific feelings in their audience to create specific complicated expieriecne? No way AI would even give an illusion of that.
And it's not only about music. Visual art, storytelling - everything.
@tristen_grant yes, but as i mentioned above you can generate something "general", like "happy blonde girl image" or "latin beats music", but you will fall flat if you ask AI to generate something emotionally specific. Just ask any AI to generate you something blue with a bit of hope and enjoy the results :)
1. Who is the audience for modular synthesis, what is their level of "sophistication", and how important is the human element to them? If it’s live, what’s the point of watching nothing? If it’s plain audio, who’s going to care?
2. If VCV Rack is a purely software approach to modular, what is to stop AI configuring sessions in that or another equally modular milieu?
3. Anything you say now is nothing more than a snapshot and will have no relevance whatsoever in 5 years. The rawness or edginess that you hear will disappear, if it is indeed a thing.
4. It will be possible in the future for the AI to show us the process. At the end of the day it’s not rocket science, it's MS.
TH-cam is already rife with AI and I use it myself on this ambient channel. I have much more human stuff too on other channels which is where I put 95% of my energy and where I think music lives but, neverthelless it’s there.
As musicians we have to make an effort to change the chip……
Modular music has already been replaced, many times over, by keyboard synths and VST's for starters. It's a completely obsolete and needlessly complex way to create music that is almost as much, if not more, about process and workflow than about result - which is also why it's completely irreplaceable, especially by AI.
When something is no longer utilitarian or efficient for commerical purposes - or perhaps it never even was - that is when it finally gains its true value for human expression.
That's a really valid point, and I feel modular has definitely reached that point. There also seems to be a certain crossing over too. I wouldn't necessarily say modular is entering mainstream, but I definitely seem to notice modular rigs more commonly sitting at the back of a stage or studio of some quite big artists. Not necessarily that they are using them, maybe there is a session musician in there, or they're just there to look cool. Who knows. It's almost like the 70s again!
How will any AI conect patch cables? How will any AI make mistakes so new music will emerge? How will any AI understand the very concept of time and process?
humans want to make music ai cant take that. It dosent matter how good ai gets we will want to do our own what ever type of music it is
Yep, that is definitely true.
Music already has no value in the eyes of most, it is already end times until there is significant change in the behaviours of stores and people. AI music will flood Spotify replacing existing playlists popularity with human made music, but really that doesn’t change much, it’s just more of the same shit situation for artists. Pretty bland examples from udio but they’d fit within a “study music” playlist on Spotify
buisness doing music is dead except for live music but I dont think it will be as huge as ancient rockstars of course. Some of the music is good but there is still like blur in front of it, they have to work around it, they will find a solution but it is not yet.
No, because you make you.
The more ambient clips where good. I would say AI will not replace niche genres, as people into it will want something new as the years go by. Not an algorithm of what's been before. The genre won't evolve.
Someone who is naturally drawn to a particular genre of music, and for the first time exposed, may love AI generated music. If you are into a genre for decades it may sound stale and trope filled.
That is a bait title :D
Don't endorse AI music making algorithms if you don't want to be replaced.
Just ignore that and the hype will eventually go away,
and people will come round from this lifeless stupor in the AI limbo.
AI will replace all human art and music that is in recorded form. However AI cannot spit on someone from two rows back or start a fight in the audience, so live music should be OK.
I dunno, I reckon AI will be coming for live music too, not at grassroots level or small clubs, but at the higher end of the scale. I could see a time when all our global superstars are AI or maybe AI reincarnations of legacy rock stars, long dead, but not forgotten. They'll probably create generative spit cannons to make the experience 'more real' 🤣🤣
@@justrichardcharles not sure if virtual abba will last much longer once the novelty has expired.. 😄
@decimal1815 let's hope not!
No because AI can't recognize all of the different controllers and there are always new controllers coming out!
Cool video. Subbed! I don’t agree with the idea that it will increase value of human music. Horses are kept as hobby and are very expensive… but how many people actually own or ride horses? Pretty much none! That industry was destroyed by cars! It will never reach the scale that it was before cars.
In what industry has cheap version actually increased the value of the real thing? Rolex watches didn’t increase their price when fake watches were made. It’s a fallacy. Most people will happily consume the CHEAPEST, EASIEST and most logical form of any good. Soon Art will be another good to consume cheaply. There will still be great human artists, but it won’t increase their value in any way, most likely it will hurt it and their livelihood.
But as you say horses have become more expensive, to the point the average person probably can't own one without serious financial outlay. But also the way horses are treated is probably better than in the past, their living conditions and life expectancy is probably greater. In that way their value has increased. They're valued more by those who own them or want to own them, but there are less people interested in owning a horse or need to own a horse. So in that respect they are valued less by the everyday person who has no interest in horses. Maybe that's a better explanation?
Thanks for the sub btw 🙂
@@justrichardcharles I definitely see your points. I just have a more pessimistic view on future generations and their relationship to music and musicians.
I do think there will always be human musicians and they will be valued. Now I’m imagining musicians being kept on ranches and living in little stables. Let’s hope for the best haha 😂
Haha, that would be strange! Maybe the ranch owners would bid for the best musicians at musician auctions. I'm taking it too far now!
no because people use modulars for the experience of using them. Same with analogue photography. We don't need no AI amalgamated funk soup.
In the 80’s, 90’s and 00’s people in the business of music who commoditised it, amalgamated what was popular and distilled it down to a formula. They may have made fortunes from it but it simply spawned mediocrity. This is exactly what AI models are doing. To be moved by music and to hear something unique is only possible when it’s created by a human, with flaws, emotions and self doubt. I don’t believe AI or computational statistics which is really what AI is can produce that.
really, the difference between what udio generates and what can be heard on spotify is not that great - did the overabundance of all kinds of music and the difficulty to achieve economic viability ever stop people from buying gear or geeking out about their "modules"? I don't think so - AI can further kill your chances to find an audience, but as long as it does not tickle your artistic visions it will simply not matter as much - the world is already full or music as it is, AI or not. - -- imho
Will AI replace guitar? Or soccer? or mac and cheese? No. It can not. How could it? I can't even understand the premise of the question. Modular synth music by definition is modular synth music. If you throw away your modular synth with the idea that AI will replace it it still will not make it real.
yes youre all going to get replaced by AI
The 1st and 3rd of the AI generated music is too "poppy" and structured. The 2nd and 4th tracks sound more like it.
Yeah, the first and third were meant to be 'experimental' sounding. Didn't quite work out though. It's been quite interesting though, so I'm probably going to experiment further.
@@justrichardcharles I'm actually quite impressed to be honest. That there are no actual synths or VSTs involved. I'm assuming that the result is a single stereo, mixed track? Is it possible to get stems with Udio? That would seem to be an easier job for the AI as mixing processing and FX processing would be unnecessary.
@TheDavidPoole yeah it's just the stereo MP3 file that's output. You could download wav stereo file or stems but that required paying 😁
I personally wouldn’t listen to ANYTHING produced by AI purely out of principle. Music is a human creation. Who wants to listen to something completely artificial. It’s all down to the people not indulging in AI produced material.
True. I don't want to listen to Taylor Swift either.
It s a question of time. AI will grow up very fast, it s a very young tech. Just wait 10 years...
AI can just be something mashing up existing music from people who have made it before. So it won't rally be innovative in a way we can be. Might kill generic music. But not music as a culture.
@@sylvainconsult969 judging by the way AI development is currently progressing, I would say more like 10 months. In 10 years, the world will likely be a different place if progress continues as it is!
Why people are so scared by AI? Technology aready changed artistic endeavors drasticallymany times in the past. Just try to compare a present day musician with one from one or two century ago.
This is very true. Technological evolution is just part of the creative process. I'm sure music technology in 50 years time will look very different, if we're still around then (I'll be in my 90s), but it will just be normal to the young people of that era.
No. Next question.
What does "replace" mean?
Will AI be able to replicate anything you can do with your modular? Yes. But nobody gives a flying fuck about the _output_ of your modular synth music. It's for you. Nobody else on Earth listens to it or cares. There are 100,000 tracks uploaded to Spotify every day, and nobody cares about them, either.
It's _your_ creative outlet. It has nothing to do with the result. You'll keep doing it as long as you enjoy it.
no.
No it won't
When I listen to something brilliant that I've never heard before, songs like FPM's Allen Ginsberg or Polysic's Mega Over Drive, or Chemical Brothers' Music Response, the music is valuable precisely because a human made it. There is something special in authenticity, and it's sadly too easy to forget that. Until we reach AGI, there is simply no authorship to beget authenticity.
Yes, totally agree. I don't know much about AGI, but I guess we're a long way from that yet.
Yes and no. I'd say that music is also valuable if a human sings/performs it live, but then so many do playback and the public still goes to their concerts... So it seems to not really matter for most listeners anyway.
@@al.cavalumaybe that's an education thing, in that audiences made up of none-musicians (I'm thinking pop music audiences here) don't necessarily understand what is taking place (or not taking place) onstage. Personally I don't mind watching an artist/band that uses playback to reproduce elements they can't do live, so long as they are actually doing something.
I think we'll see a big impact on live music when it comes to new artists being able to transition from the levels of say medium size venue to super rockstar stadium sized shows. I don't think the mechanism exists for artists to grow in that way anymore. With the exception of one or two, most of the biggest rock stars now are quite old, and it seems with no one coming up to replace them. I can see a future where stadium sized live shows are dominated purely by AI holographic performances of legacy artists who are long dead or retired, whilst smaller artists struggle to make a living.
@@justrichardcharles Agree.
For the same reason why I buy IKEA. It looks good. It's decent quality. It's cheap. It's on stock, just order and have it delivered. If it breaks down, order another one. If you move to another country, order another one.
At the same time, I love having custom furniture. With all the minuses: expensive, needing to hunt for a good carpenter, wait time, expensive, paradox of choice (how to design it? Design is hard and IKEA already does it for you). Etc etc.
There is room for both.
Same with music.
yes, AI will replace everything. pity we havn't made an AI yet
Clickbait af
This is a joke right?
Yeah nah. Modular synthesizers are instruments designed by humans for humans to play. AI won't replace that any more that it will replace "guitar music". Also that AI music really socks.
I agree with your views
I did not like the tracks There was something missing
Glad it wasn't just me that thought that! 😊
@@justrichardcharles Its matter of time. In 5 year time most DAWs will have AI plugins by default. You will have a lot of control of individual elements.
Yes and no. :) They are not music I would listen to, but they could probably replace a lot of music being produced these days, which is pretty bad anyway, and it doesn't even need to be better for what it is used ... More as background music rather than something you would actively listen to and appreciate...
@@al.cavalu Reality is that most AI plugins will function similar to synthplant. You will give prompts to each channel, but you could fine tune them with standard synth parameters. Besides modular is most overrated thing ever. 90% of it needs you can cover by regular DAW and instruments.
@@meskisableI agree with you to a certain extent. Modular is overrated. It's in antiquated and hard to use compared to modern methods of creating music. But then you could probably say the same about other methods/instruments too. For example the guitar. It is also antiquated and difficult to learn but people still love it. Their is a similar gear obsession around guitars that there is with modular. Plenty to gas over. The appeal is more in the 'doing' than the end result though I feel. Maybe making music the traditional way will just become the pursuit of amateurs, who do it for the love of the activity. That will probably never go away, and the industry to feed that will always be there because of gas
AI will be way better at making Popular music than modular
everyone is coping so hard ai will most definitly replace modular
it will vst devs sample packs mixers mastering engineers
just use a text to image ai and imagine for audio
let's say you want a bass line that sounds like x artist
you put in sinister bass by X and hit generate
maybe you want it to sound like it was mixed by a certain engineer
you put in sinister bass by X and mixed by Y engineer, you will be able to specify years and gear probably
same for sample packs, you want vinyl breaks? you want dead ludwig snares? just as ai to generate it and the samples will sound great
we're all gonna get replaced
Can you lmagine AI making anything
even close to all the great music that
humans made in just the 60's & 70's
alone ? *RIDICULOUS !*
And if humans weren't so constipated
with bullshit these days, they could do
it over and over until the end of time.