@@cereshinI’ll let you know if I came up with decent track with those specs. Luckily there are some awesome human performances of paganini played in gypsy style😃
wow, first time hearing stuff from suno. It's scary-impressive how well it hit the genre. Sadly, generated music for commercial use will mean a flood of sound-alike music. So even less streaming money for independent artists, which means less time and energy can be put into new exciting music (something AI can't do imo). But you're right, we will still keep playing and connecting through chords and melodies and stuff :D
@@jazzguitarwithandy This is real! Last week I was watching Finnish television and there I heard AI created gypsy jazz as background music. For some reason I recognize the sound pretty easily 😂
Impressive as this is it doesn't sound anything like the way Django, Fapy Lafertin etc or any of the real gypsy's would play. But it does sound quite like a lot of the other modern players. I think this illustrates how much they rely on licks and learnt phrases going around the chords and scales rather than being actual jazz musicians imho
That's fair. It's very formulaic in nature, earlier sound with lots of chords (arps) rather than chord-scales. It's played on auto-pilot like Dixieland.
As AI, it is combining lines and possible line extrapolations from existing artists, but not (yet?) actually creating. Sorta the difference between Hindustani and Carnatic improv. But Gypsy Jazz is far and away the most predictable of improv formats, even more than simple blues.
Being a guitarist, although not of this genre of which I am nevertheless a great fan, I must admit I wouldn’t be suspicious at all hearing that. It’s much much better I would expect. Yes, the guitar tone is a bit plastic, but I would rather think it’s a weird microphone or something like that than an AI generated music. I am pretty sure it’s just a matter of very short time when no one could recognise anything.
The lead guitar vocab is off for the genre (pompe is there rhythmically, if not in tone). Impressionistically, the solos sound in the ballpark, but the solos have tons of moves that sound more generic that specific to gypsy jazz. It's like it's taking bits from very different solos, and combining them into one.
Very impressive to my ears. The lines included tasteful rhythmic variation, it wasn’t just a steady stream of eighth notes. Now and then it would end a phrase on an interesting altered note. The variation in melodic direction was there too so to my ears it sounded good. I’d steal some of those lines. But I don’t want AI to take away more money out of musicians pockets which sadly is what it will end up doing.
Yes, been playing with this over the last few weeks. The improvement in the last few months is unbelievable. Captures the genre pretty well in my opinion. Also it is a synthesis of real recordings so it kind of is real players combined.
Yes I was trying to ask suna to make me some gypsy jazz backing tracks on random progressions for me to improv/noodling to but instead of backing tracks it gave me a bunch of intense fast tempo fully composed pieces, I ran out of free tokens for creation and was rather confused and a bit impressed with AI's solo phrasing.
I also haven’t been able to create backing tracks. Also tried other ai called udio. I hope someone will create a backing track generator 😁 With tech like this it should be possible.
@@niilooikari7243 I went really detailed on the description and tried my best to refine my communication with the ai but it wasn't able to get the idea of backing tracks. I somehow really resonate with some of the ai's phrasing maybe It's because neither the ai or me was ever systematically trained for hotclub lol Can't wait to play and jam with ai real time it would be really cool.
@@thelilpeanutdotgov Were you able to generate actual jammable backing tracks with this? It seems like suno just likes to give me really interestingly phrased lead playing, some of the licks it gave me actually makes me raise my eyebrow.
I love the accordion . I have no idea how to play one . Prob y i like it more as a listener than if i knew how to play the instrument. Also love the vibrato on guitar. Also love the colors the AI chose. And decided to hang out on. Crazy man
I’ve been having fun with it, but it definitely has limits. I told it to make a 6 part fugue in the style of Bach using a 12 tone row melody. It didn’t even create a fugue at all, and no 12 tone row melody.
In the Distant Future, the Year 2025 there will be Gypsy Robots traveling around the Galaxy in SpaceVardo WagonShips playing AI Generated Gypsy Jazz on the Moon, Mars, and Titan (62nd Moon of Saturn) for denizens living in Domed Colonies! Huzzah!!!
@@niilooikari7243 absolutely... music should be for personal development and intellectual exercise...but in a financial world musicians tend to show off
I think if individual tracks could be created for each instrument and worked with individually. Notes being played on and slightly before and behind the beat . Would give it a more raw sounding jam session . The rhythm is to sharp on timing.
Suno is popular, but the latest AI is Udio. You should try it. It's more erratic in scales and harmonies, it can sound amazing or terrible alike, but the overall quality is better.
I fully disagree. Suno is more musical, while Udio is kinda random and tuneless. I'm completely baffled at why there are always people who think the newcomer is better. It isn't.
I mean it sounds like most of the people I jam with. Technically sounds and in style but a bit dull and soulless. And yes I include myself in the dull and soulless category😅
It doesn't sound natural because the playing at the turn of phrases becomes muddled and skewed, quite the opposite of what a good player would do. Also there is no theme, no discernible melody, no leading idea, except to go on and on and just ramble some phrases without an end. In the late 1980s, we have been programming applications called "The Politician", which would assemble "speeches" out of randomly generated sentences, with simplistic rules, but the text, like this music, had no intelligent theme or subject. That was the early "AI", not different from the AI 38 years later: there is no intelligence in artificial intelligence.
What people don't understand is that tools like these don't have to offer a final song that people would come to listen to only. But can be a source of inspiration for musicians, the idea of being able to hear what you have on mind just by prompting it to AI is incredibly amazing. Ai is very limited melodicaly and harmonacly when it comes to jazz language. Looking forward to the future.
Yes. Good idea! I will put a download link into description. It's also here: drive.google.com/drive/folders/1B0B53AGGiQouYnibyxi55lncTsZ71wKl?usp=share_link
Generating something that stands out seems like the big challenge. The examples you played fit the genre a bit too well, and sounded quite generic. I am curious how much precision prompting currently allows. The more precision a prompting system permits, the more actual domain expertise may be required, in addition to being competent with the prompting system itself. My only AI music explorations so far have been in isolating Paul Chambers bass lines from the original Miles Davis tracks, and that has been really helpful.
I have been a fan of gypsy jazz for years and not bad, but too much ego. LOL Interesting sounding vibrato. Perhaps someone will come up with AI that can replace the person who made this AI. Perhaps AI should raise children.
@@niilooikari7243 I tried playing around with this in other genres, the gypsy jazz seems to be the most fun. It wasn't too good at Grateful Dead space jams.... lol
Suno is the most impressive but pretty unusable because of its 2 mins limit, lack of intro and ending, and very limited control (you can only type the genre of music and the kind of ambiance you want). I prefer Stable Audio 2.0 for the kinds of music I like : 3 mins of duration, proper intro and ending, high control (you can ask complex things like "psychedelic drum'n bass with violin, slowly increasing in intensity"), seeds control, etc.
impressive at first but i think it start sounding dull rather quickly. Something about it is just too bland. I think it's missing human error and imperfection. Other than that, that's great for "utility music" like elevator music and all. It's still better than corporate ukulele at least !
I feel like that what's missing is "big picture" coherency. Line-by-line it sounds pretty decent (dare I even say creative), but everyting gets lost in this weird synthetic meandering. It's too close for comfort though.
Be groovy to use AI for ideas around the harmony of a song and then we just add our own organgic feel to it. This inspires me hearing the lines, using it as a tool isn't scary. Yeah its impressive and picked out some juicy licks but will never replicate the feel in the the music. Be also cool to have the notes displayed for this tune. Edit: fully man noticed the single note lines and that mad walk. Like most things we've created that are controversial. "To be used and not abused" in the wrong context anyways.
I wonder how original the lines are. Is it just taking snippets of lines and mixing them together, or is it understanding the fundamentals of how to create lines.
From what I've read, it's not copying existing music, but synthesizing new music. There are familiar passages here and there, but so it is in human-made music as well. We're building on what came before.
The better the ai gets at playing gypsy jazz, the more likely it is that django reinhardt was actually a time travelling ai in the simulation earth v.12
Nah man I'm not happy about computers making this stuff. This "music" is like a car without an engine. I'm worried for the future 5 years from now when this technology is way better than it is now, and human musicians potentially being valued less and losing work because of it. I don't think this is a side of technology to be celebrated. Thank you for making the video!
I like the analogue of a car without engine. This technology sure makes one think about the future. I try to think that in the long run AI-generated music will raise the appreciation for human musicians since we will be doing music no matter how advanced AI will get. Thanks for commenting. I really admire your musicianship :)
La pompe is terrible and chorus is so well on tune It feels plastic. Not hearing any "touch" on notes either. Looks like video game music more than jazz BUT composition is quite interesting and might be a starting point for a real musician.
Yea it’s pretty good! I think it has more realistic sounds. I made a video where I test it: th-cam.com/video/Zng6w1hE-ao/w-d-xo.htmlsi=RTVXJFNCOf7lb5WD
Yes, it is unsurprisingly technically accurate, and to large extent good. But this not the same as experiencing a human connection playing together including the interaction of a viewing audience. Emotion is a funny thing and I think ai will be a while before it can compete with flesh and blood - and a bit of alcohol and a broken heart. 😊
It's crap. Not more impressive than what Band in a box was doing 17 years ago (except for the sound samples). We should not promote this because this will threat human musicians. You can hear it is AI for there is no logic in phrasing. Sounds like it's been trained with 8 bats of Adrien Moignard or othe rmodern player. Thanks for mentionning this Ai gypsy jazz. I've been talking about it for months, especially the horrid playlists on Deezer or Spotify that invade the platform and can deceive people about what gypsy jazz really is.
Love Adrien's playing. 8 bars from him includes lots of data. 😄🔥 Yeah, there is no soul in AI's playing because it's a machine after all. It can't have unique human experience no matter how complicated programming. Still it's somewhat fascinating. I think those spotify and deezer playlists are often made by AI. I prefer albums or my own playlists. For this style of music live performances are the best! :)
When technology advances it will become better and better at imitating the pros. If you say its crap why are you so afraid that its gonna take your job/glory? Why don't you just try to do something that cannot be replicated by a propability machine. I agree it sounds bit boring but It definetly isn't bad. Especially the accordion. But you can still remember how bad AI generated images were two years ago compared to now.
Ai is NOT making music - it is compiling data from artists and using it to mimic those artists. And, if it is not giving credit to those artists, it is essentially ripping them off. The technology is fascinating and has amazing possibilities, but the source artists from which the code mimicked needs to be recognized.
Yeah. Thats tricky with generative AI. It can use so many source artists it might be hard to do, but you are right. Im also concerned about: did they have a permission from every artist they used to train these models? Probably not…
After a while of using Suno or Udio, you will notice that they are using licensed artists for their models. Each genre has a limited set of model sources. And the source artists have been paid for training their model.
@@BluesInSeattle Are they paid for training the machine or are they paid a royalty each time one of these "songs" is played? Is it the musician's composition of the programer's or the machine's?
What do you think about AI-generated music?
Scares me. Now, all that´s left for mankind is to work in the quarry, since machines do the creative things in life.
dude i sub if you get an ai to generate a song with django reinhardt and niccolo paganini! all the ai tools i used were not up to the task
@@Hrldryeah it’s scary. But on the other hand: what a time to be alive to witness this. I want to believe that new tech leads to good things 🤖
@@cereshinI’ll let you know if I came up with decent track with those specs. Luckily there are some awesome human performances of paganini played in gypsy style😃
@@niilooikari7243 Would be the first time for technology not to be abused. But hope always dies last.
wow, first time hearing stuff from suno. It's scary-impressive how well it hit the genre.
Sadly, generated music for commercial use will mean a flood of sound-alike music. So even less streaming money for independent artists, which means less time and energy can be put into new exciting music (something AI can't do imo).
But you're right, we will still keep playing and connecting through chords and melodies and stuff :D
Yeah we keep the music going!🎶
It's a touch soulless, but honestly, pretty damn impressive. But, can it drink and play around a campfire?!
Sounds like Django to me.
Can it set fire to its caravan and come back from the dead?
@@DonQuickZote that's the true measure right there.
Soul less means limitations of humans...
AI is just awesome...much better than humans
Well that's all of the background music sorted for UK TV 😂
😅
@@jazzguitarwithandy This is real! Last week I was watching Finnish television and there I heard AI created gypsy jazz as background music. For some reason I recognize the sound pretty easily 😂
Impressive as this is it doesn't sound anything like the way Django, Fapy Lafertin etc or any of the real gypsy's would play. But it does sound quite like a lot of the other modern players. I think this illustrates how much they rely on licks and learnt phrases going around the chords and scales rather than being actual jazz musicians imho
That's fair. It's very formulaic in nature, earlier sound with lots of chords (arps) rather than chord-scales.
It's played on auto-pilot like Dixieland.
As AI, it is combining lines and possible line extrapolations from existing artists, but not (yet?) actually creating. Sorta the difference between Hindustani and Carnatic improv. But Gypsy Jazz is far and away the most predictable of improv formats, even more than simple blues.
Being a guitarist, although not of this genre of which I am nevertheless a great fan, I must admit I wouldn’t be suspicious at all hearing that. It’s much much better I would expect. Yes, the guitar tone is a bit plastic, but I would rather think it’s a weird microphone or something like that than an AI generated music. I am pretty sure it’s just a matter of very short time when no one could recognise anything.
so it's good enough to get a AFM card
@@DSAK55 maybe, but I have no idea what is an AFM card:-(
@@unwrought9757 American Federation of Musicians
@@benjaminholt6640 okay, thanks for explaining me, I didn’t know.
The lead guitar vocab is off for the genre (pompe is there rhythmically, if not in tone). Impressionistically, the solos sound in the ballpark, but the solos have tons of moves that sound more generic that specific to gypsy jazz. It's like it's taking bits from very different solos, and combining them into one.
Very impressive to my ears. The lines included tasteful rhythmic variation, it wasn’t just a steady stream of eighth notes. Now and then it would end a phrase on an interesting altered note. The variation in melodic direction was there too so to my ears it sounded good. I’d steal some of those lines. But I don’t want AI to take away more money out of musicians pockets which sadly is what it will end up doing.
This is incredible and scary. But I would like to see a blind test: AI vs human. Then let people decide their favourite
A few dynamics (vibrato ..bends) missing. But otherwise, very cool ! Thank you for this experiment !
Wow… it swings well enough to dance to for sure.
Your videos are really funny (your sense of humor) but educative
Yes, been playing with this over the last few weeks. The improvement in the last few months is unbelievable. Captures the genre pretty well in my opinion. Also it is a synthesis of real recordings so it kind of is real players combined.
Yes I was trying to ask suna to make me some gypsy jazz backing tracks on random progressions for me to improv/noodling to but instead of backing tracks it gave me a bunch of intense fast tempo fully composed pieces, I ran out of free tokens for creation and was rather confused and a bit impressed with AI's solo phrasing.
I also haven’t been able to create backing tracks. Also tried other ai called udio. I hope someone will create a backing track generator 😁 With tech like this it should be possible.
@@niilooikari7243 I went really detailed on the description and tried my best to refine my communication with the ai but it wasn't able to get the idea of backing tracks. I somehow really resonate with some of the ai's phrasing maybe It's because neither the ai or me was ever systematically trained for hotclub lol Can't wait to play and jam with ai real time it would be really cool.
something helpful i learned: sometimes less is more in the ‘style’ box
@@thelilpeanutdotgov Were you able to generate actual jammable backing tracks with this? It seems like suno just likes to give me really interestingly phrased lead playing, some of the licks it gave me actually makes me raise my eyebrow.
I love the accordion . I have no idea how to play one . Prob y i like it more as a listener than if i knew how to play the instrument. Also love the vibrato on guitar. Also love the colors the AI chose. And decided to hang out on. Crazy man
I’ve been having fun with it, but it definitely has limits. I told it to make a 6 part fugue in the style of Bach using a 12 tone row melody. It didn’t even create a fugue at all, and no 12 tone row melody.
It's funny that you think you could detect it
Suno has very limited text comprehension unfortunately, you can only give it a musical genre and an ambiance.
In the Distant Future, the Year 2025 there will be Gypsy Robots traveling around the Galaxy in SpaceVardo WagonShips playing AI Generated Gypsy Jazz on the Moon, Mars, and Titan (62nd Moon of Saturn) for denizens living in Domed Colonies! Huzzah!!!
Just give it a few more years....... I already lost significant amount of client on my Fiverr music gig ....
Months?
Wait I thought jazz already had autofill what do we need an AI for?
Thanks man - interesting.
Unbeatable....
Humans need great practice to play this
@@clearbrain It’s getting even crazier. Luckily for us humans music is not a competition 😄
@@niilooikari7243 absolutely... music should be for personal development and intellectual exercise...but in a financial world musicians tend to show off
I think if individual tracks could be created for each instrument and worked with individually.
Notes being played on and slightly before and behind the beat .
Would give it a more raw sounding jam session .
The rhythm is to sharp on timing.
Yeah it's all on a metronome
Suno is popular, but the latest AI is Udio. You should try it. It's more erratic in scales and harmonies, it can sound amazing or terrible alike, but the overall quality is better.
Yes I have heard about it. Sounds fun. Haven’t tried it yet. Been sparing it if I make another reaction video or something.
I fully disagree. Suno is more musical, while Udio is kinda random and tuneless. I'm completely baffled at why there are always people who think the newcomer is better. It isn't.
I mean it sounds like most of the people I jam with. Technically sounds and in style but a bit dull and soulless. And yes I include myself in the dull and soulless category😅
a bit remarkable really, ok the guitar sound reveals the fakery but it does swing and the lines flow and are dead on for the genre. a bit scary.
scary indeed
It doesn't sound natural because the playing at the turn of phrases becomes muddled and skewed, quite the opposite of what a good player would do. Also there is no theme, no discernible melody, no leading idea, except to go on and on and just ramble some phrases without an end. In the late 1980s, we have been programming applications called "The Politician", which would assemble "speeches" out of randomly generated sentences, with simplistic rules, but the text, like this music, had no intelligent theme or subject. That was the early "AI", not different from the AI 38 years later: there is no intelligence in artificial intelligence.
What people don't understand is that tools like these don't have to offer a final song that people would come to listen to only. But can be a source of inspiration for musicians, the idea of being able to hear what you have on mind just by prompting it to AI is incredibly amazing.
Ai is very limited melodicaly and harmonacly when it comes to jazz language. Looking forward to the future.
Huh! Itelle menis täydestä tämä.😮
Hello Niilo, please can you share those AI tracks if possible? They are really cool, I wanna have them in my playlist.
Yes. Good idea! I will put a download link into description. It's also here: drive.google.com/drive/folders/1B0B53AGGiQouYnibyxi55lncTsZ71wKl?usp=share_link
Generating something that stands out seems like the big challenge. The examples you played fit the genre a bit too well, and sounded quite generic. I am curious how much precision prompting currently allows. The more precision a prompting system permits, the more actual domain expertise may be required, in addition to being competent with the prompting system itself. My only AI music explorations so far have been in isolating Paul Chambers bass lines from the original Miles Davis tracks, and that has been really helpful.
I want to hear a new solo over an existing song. Can it do a chord solo or double stops solo?
Im not sure but I think that's not possible with tools that we currently have, but in the near future I think that can be done.
I have been a fan of gypsy jazz for years and not bad, but too much ego. LOL Interesting sounding vibrato. Perhaps someone will come up with AI that can replace the person who made this AI. Perhaps AI should raise children.
Scary. But actually inspired me to practice. Can it do backing tracks without the guitar?
Im glad to hear that it inspires you! Same for me. I haven't tested it for making backing tracks, but I guess it's not possible yet...
@@niilooikari7243 I tried playing around with this in other genres, the gypsy jazz seems to be the most fun. It wasn't too good at Grateful Dead space jams.... lol
Suno is the most impressive but pretty unusable because of its 2 mins limit, lack of intro and ending, and very limited control (you can only type the genre of music and the kind of ambiance you want).
I prefer Stable Audio 2.0 for the kinds of music I like : 3 mins of duration, proper intro and ending, high control (you can ask complex things like "psychedelic drum'n bass with violin, slowly increasing in intensity"), seeds control, etc.
0:27 Lelo Nika can play it !
impressive at first but i think it start sounding dull rather quickly. Something about it is just too bland. I think it's missing human error and imperfection. Other than that, that's great for "utility music" like elevator music and all. It's still better than corporate ukulele at least !
In 10 years we'll probably have Neurolink apps that create music based on what it learns produces the most enjoyment in your head.
Sounds horrifying ;D
I feel like that what's missing is "big picture" coherency. Line-by-line it sounds pretty decent (dare I even say creative), but everyting gets lost in this weird synthetic meandering. It's too close for comfort though.
Be groovy to use AI for ideas around the harmony of a song and then we just add our own organgic feel to it. This inspires me hearing the lines, using it as a tool isn't scary. Yeah its impressive and picked out some juicy licks but will never replicate the feel in the the music. Be also cool to have the notes displayed for this tune.
Edit: fully man noticed the single note lines and that mad walk. Like most things we've created that are controversial. "To be used and not abused" in the wrong context anyways.
Now what’s the chord on your intro? ☺️very human
It’s F#dim7 -> G -> Gmaj9 ☺️
I wonder how original the lines are. Is it just taking snippets of lines and mixing them together, or is it understanding the fundamentals of how to create lines.
From what I've read, it's not copying existing music, but synthesizing new music. There are familiar passages here and there, but so it is in human-made music as well. We're building on what came before.
@@SirCutRyexactly 👆👆
The better the ai gets at playing gypsy jazz, the more likely it is that django reinhardt was actually a time travelling ai in the simulation earth v.12
can it play crisis?
Nah man I'm not happy about computers making this stuff. This "music" is like a car without an engine. I'm worried for the future 5 years from now when this technology is way better than it is now, and human musicians potentially being valued less and losing work because of it. I don't think this is a side of technology to be celebrated. Thank you for making the video!
I like the analogue of a car without engine. This technology sure makes one think about the future. I try to think that in the long run AI-generated music will raise the appreciation for human musicians since we will be doing music no matter how advanced AI will get. Thanks for commenting. I really admire your musicianship :)
@@niilooikari7243 thank you! I sure hope that it turns out that way in the end for the musicians.
how is he dividing up the credits or is he ai ?
??????????????????
La pompe is terrible and chorus is so well on tune It feels plastic. Not hearing any "touch" on notes either. Looks like video game music more than jazz BUT composition is quite interesting and might be a starting point for a real musician.
You should really try Udio, its already much better than Suno.
Yea it’s pretty good! I think it has more realistic sounds. I made a video where I test it: th-cam.com/video/Zng6w1hE-ao/w-d-xo.htmlsi=RTVXJFNCOf7lb5WD
I don't like how AI generated music is pointless. Like it isn't self-aware enough to fake that it has some feeling to put into the outcome
Now we know how shit sounds! ;-)
Yes, it is unsurprisingly technically accurate, and to large extent good. But this not the same as experiencing a human connection playing together including the interaction of a viewing audience. Emotion is a funny thing and I think ai will be a while before it can compete with flesh and blood - and a bit of alcohol and a broken heart. 😊
Human beings are able to channel the spirit of music. That was shopping center toaster noise.
it just kind of sounds like tricked out elevator music
It's crap. Not more impressive than what Band in a box was doing 17 years ago (except for the sound samples). We should not promote this because this will threat human musicians. You can hear it is AI for there is no logic in phrasing. Sounds like it's been trained with 8 bats of Adrien Moignard or othe rmodern player. Thanks for mentionning this Ai gypsy jazz. I've been talking about it for months, especially the horrid playlists on Deezer or Spotify that invade the platform and can deceive people about what gypsy jazz really is.
Love Adrien's playing. 8 bars from him includes lots of data. 😄🔥 Yeah, there is no soul in AI's playing because it's a machine after all. It can't have unique human experience no matter how complicated programming. Still it's somewhat fascinating. I think those spotify and deezer playlists are often made by AI. I prefer albums or my own playlists. For this style of music live performances are the best! :)
When technology advances it will become better and better at imitating the pros. If you say its crap why are you so afraid that its gonna take your job/glory? Why don't you just try to do something that cannot be replicated by a propability machine. I agree it sounds bit boring but It definetly isn't bad. Especially the accordion. But you can still remember how bad AI generated images were two years ago compared to now.
@@rq3849because if it’s one thing I’ve learned the hard way, it’s that normies are huge consumers of pure crap
Ok, let us hear your great tunes then 😊
@@rq3849 You may not have noticed, the majority of the world is strangely into crap.
It sounds good, but the musical phrases are devoid of emotion....
It’s boring “music”.
Terrible. Pointless noodling.
Impressive on first listen but the creations of ai will quickly become bland and recognizable
this is too neet andfinally so boring that I cant' listen to this for 3 minutes.
Ai is NOT making music - it is compiling data from artists and using it to mimic those artists. And, if it is not giving credit to those artists, it is essentially ripping them off. The technology is fascinating and has amazing possibilities, but the source artists from which the code mimicked needs to be recognized.
Yeah. Thats tricky with generative AI. It can use so many source artists it might be hard to do, but you are right. Im also concerned about: did they have a permission from every artist they used to train these models? Probably not…
Normies don't care
After a while of using Suno or Udio, you will notice that they are using licensed artists for their models. Each genre has a limited set of model sources.
And the source artists have been paid for training their model.
@@BluesInSeattle Are they paid for training the machine or are they paid a royalty each time one of these "songs" is played? Is it the musician's composition of the programer's or the machine's?
I would argue this is the same way humans make music.
You’ll know this is the end when you’ll start transcribing AI tunes
Well if it can help a guitarist improve his playing or come up with new lines.