Jesus, I'm totally in love with Udio. I just put a whatever prompt and they gave me a song, using my beautiful lyrics, with a Daft Punk 'One More Time' vocal effect that I didn't ask for and I don't have a clue to make another music with the same effect!!!. I AM DOING THE KIND OF MUSIC I LOVE TO HEAR, CUSTOMIZED MUSIC TO MY EARS AND I'M AMAZED!
As a musician that’s exactly what I’m waiting for before diving in. I can see the value in what’s it currently does but I want to be able to pull it apart, edit and augment it.
Ok, but users should be aware that any material uploaded by them will belong to Udio FOR FREE, to be used raw by others users and for training their AI. As long as people agree on liberating Udio from all that copyright headache, and making the startup richer, it's fine for me. I will still be making my music in a way that belongs TO ME.
@@marcosazambuja When you say "Belong" you mean, they have a license to train their model on it. Probably pretty similar when using AI mastering plugins too tbh. In fact, quite a lot of the music Distribution platforms also have terms that by uploading you are granting them a right to train AI models on the stuff. Not even music, likely this very comment and yours too. That is the world we live in. As far as how people make content, that is up to the individual. 99% of my stuff has never seen an AI plugin let alone generative AI, but I can certainly see a time when I will ask for a funky vocal that compliments my original music. Nothing has changed here other than opening up new possibilities. And if you don't like those possibilities that is okay too.
@@jumpstar9000 I will like when these possibilities get free of legal infringments (otherwise we can't use it, for real - I'm not risking having a lawsuit due to the use of AI material in the soundtracks I produce). How long will it take, and will worth it? I don't know.
Well is here but it just needs some fine tuning and a few functions but this is just implementation,but the core is already here ,i can already imagine it will evolve into a DAW kind of interface( such as pro tools) for those that want full control, it will then create everything given a few samples ( including any person's voice ) and create a fully produced pro song
If you could select a audio section and delete, replace, regenerate or remove elements then it really would be a game changer. Would also be great to be able to save stems.
As a composer, I have some mixed feelings, but the audio-to-audio feature works quite interesting by inputting a single instrument track and generating audio with little variation (so the melody is not modified, e.g. a violin playing). The output can serve as an extra layer of audio that can increase the realism.
This update is definitely a big step to make more custom music and a hybrid between human and ia, however I still notice strange artifacts, the snares, cymbals and drums do not sound real, sometimes the snares sound like they are playing with brushes ... I find it a fun toy but and exciting but I will also say that I detect a leagues the songs generated with ia. Just like I detect the texts written with chat gpt... and I've been listening to songs generated with ia for months, with incredible voices, but none of them move me... does this happen to you?
Yes definitely, most time you can still tell it's ai, also keep in mind us ai music producers have heard hundreds of songs now, so we are training our ears to pick up those noises. Thanks for the great comment, and Thanks For Watching
Now I just wish they'd let you merge different tracks into one and introduce more editing tools and then it'd be perfect. A lot of the time it doesn't listen to my prompts even if I select prompt strength to be 100%
When competition is present, the user wins, you can do the same with SUNO. The sound quality maybe better with Udio but I can never get the results I get with SUNO. It will come down a preference and what you can work with better.
Looks like you can download and then edit segments of the song to re-upload it with what you like about the track to see if the AI can build on your original idea, including the small segment it wrote for you earlier.
This is exciting. I can see that very soon, we as weekend producers, could design an AI artist and then produce a full album of content around the style of the AI artist you built.
Uploaded an actual 5 minute demo song. It plays the whole song back but it will only give you that 32 sec part after playing the 5 minutes, then it kicks in with the AI stuff. Lots of experimentation to do! I can see the credits diminishing quite fast now LOL.
I am dreaming about times, when Udio will give me opportunity to open a stem, see all the plugins used, and I can only re-record my leading vocals))) Feels like heaven)))
What I'm missing is the possibility of creating a guitar solo for 4 bars for a song that I composed and recorded myself. Where you can specify the key and chords, or upload the location where the solo should be produced.
I knew this was coming, but didn’t think it would be this soon. The ability to upload your own audio changes everything when I comes to songwriting and copyright ability. The chord progressions and melodies can be yours and Udio just follows your idea. That means it can be copyrighted. This is going to get better in terms of quality and how well it follows your lead. Also coming according to their Office Hours presentation on Discord are STEMS and MIDI. There’s no argument against using this as far as I’m concerned.
Stems and MIDI would completely change the game. Udio is making me hate this tech a lot less by making it far more controllable. Sure, the AI is existentialy threatening, but how can you be a songwriter and not find all of this exhilaratingly fun, too?
This is the part where it helps to know about music production. 1. It helps if the audio sample is on beat. 2. Giving it looped files certainty helps. 3. if you are using a long audio file, it helps to extend from a good spot, anything short of that will yield weird results.
Suno's singing definitely needs a lot of work. the quality of Udio vocals is much better, but it tends to layer the vocals in a way that doesn't always sound natural.
It’s great to finish songs that were not finished by the original artist before they died Or to take a song that you like and capture the essence of it continue the song and then cut out the original material
Hey Brother. I love what your doing with Udio here. I work a lot with Suno. I wonder what we could create. By combining some of our music together. Keep up the great work My Friend. I did Like and Subscribe Buddy.
Well full music production already here but it just needs some fine tuning and a few functions but this is just implementation,but the core is already here ,i can already imagine it will evolve into a DAW kind of interface( such as pro tools) for those that want full control, it will then create everything given a few samples ( including any person's voice ) and create a fully produced pro song
What are the t&c's with this feature? I assume they keep your uploaded data and use it to learn the AI for other users to create similar tracks from your sounds
I've tried it. It's not a bad concept but the fidelity just isn't very good. Certainly not there yet. And you can't "Inpaint" over your audio yet. It's a good roadmap idea, but for well realised mixed 24bit demos that are uploaded, the quality lets it down ATM. Give it a year...
I must admit it is dope but that 11 Lab sounds so real. I don’t know how they get the individual instruments to sound so distinctive, where here you can tell it’s more a flatter mix when quality should instead require dynamics. Suno is definitely for the culture though #hiphop #NC #atl
I saw your face and I was like hey, that guy looks like someone I used to go to middle school with, I think his name was Chris. Color me shocked when I exited fullscreen and looked at your username. You ever live in Maryland?
I testing it aleardy and this is amaizing. At the moment UDIO beat SUNO, but we wait for SUNO v4. Another part of cake is generate start from Suno and upload to Udio to extend it and add vocal. You know what i mind? This is creazy.
@@drmarioschannel Nobody need Suno beat after that. Now you create your short midi and Udio make from this awesome music. This resolved Udio rythm problem.
Yes you can, if you add the lyrics in the custom box. You could only do verse 1 then do verse 2 ect. You could add all the lyrics at once in Suno hope that helps. Thanks for watching
I say do it for a month and see if you like it. It's easy to cancel. And I'm not affiliated with Udio in any way. Hope that helps. Thanks for watching.
With this do you own the copyright? im thinking maybe no because the ai generated additions use stolen copywrited training data. I don't have a subscription to know, what does it say about this?
Great question, if you create lyrics or a melody you own the rights to that, you can not copyright a chord progression as per the Ed Sheeran lawsuit. Hope this helps. Thanks For Watching
Of the pieces generated so far my favourite is the one in the following link. It has its issues: Udio all but muted the right channel throughout one section, and I still need to properly fix it when I have the time. But I like the flow of the lyrics which have a rather complex rhytmic structure (the lyrics are by a Polish poet, written in the 1920s). The title translates to "Hangman's Ballad": th-cam.com/video/3Vw0MkH53yk/w-d-xo.html
Meh. the problem is you gave it almost nothing to work with. You have to give it more than a couple notes. It needs at least 30 seconds of starting data.
If you are paid subscriber you can upload it anywhere. So you can upload it to Spotify. Now can you make money from Spotify? That's a whole other question😂 Thanks For Watching
Not true. Didn’t you notice this allows audio uploads? You can use it to produce your own music. And I’m sorry that you’re in such denial, just because someone uses AI doesn’t make the entire thing copyright free. You think anyone using Adobe’s AI in Photoshop can’t copyright anything they make? Youre in for a shock. And you’re in for an even greater shock in the future because they’ll continue to make tools that let you use Ai in music production, which means it will be everywhere and there’s no legal justification to say someone who extends their own music and works with it back and forth with it in highly detailed ways doesn’t own their own track at the end
@@Edbrad Of course everybody can generate tracks thru Udio, Suno and the other services. Copyrightless tracks that belongs to nobody (yes, no copyright for the user). Until you publish them on that client's project and all of a sudden receives some emails regarding (others) copyright infringment. Why? Because these startups uses other artist's phonograms for training their models without consent. And your client will never work with you again.
If you add lyrics or melody parts you 100% own the rights to that. There is no grey area there. This helps make that happen. Hope that helps, and Thanks For Watching
@@ChrisUnlocksAI Nope. If the AI was trained using 1 million tracks (which is the case), then you gotta SPLIT your ownership with all this people and companies. Oh boy. You need to read Udio/Suno's terms of service, they almost laugh at your face. Anyway, copyright laws doesn't work the way you painted. There's no Alice in this wonderland. Plus: Udio and Suno used phonograms WITHOUT PERMISSION to train their models. Which means that, if you generate a track and decide to show it to the world via the internet, you could be notified of copyright infringment. If it's a risk you want to take, it's up to you. But US Supreme Court decided last november that any AI generated material CANNOT BE COPYRIGHTED. I guess this ends the question.
Jesus, I'm totally in love with Udio. I just put a whatever prompt and they gave me a song, using my beautiful lyrics, with a Daft Punk 'One More Time' vocal effect that I didn't ask for and I don't have a clue to make another music with the same effect!!!.
I AM DOING THE KIND OF MUSIC I LOVE TO HEAR, CUSTOMIZED MUSIC TO MY EARS AND I'M AMAZED!
If Udio gets Stems and we could directly regenerate single stems, that would be another game changer.
It's coming!! Thanks For Watching
I hope, and say to udio the tonality etc.
As a musician that’s exactly what I’m waiting for before diving in. I can see the value in what’s it currently does but I want to be able to pull it apart, edit and augment it.
Wow. That is pretty cool man. Now it really starts to become a real production tool. Incredible. Thanks a lot for the headsup!
Ok, but users should be aware that any material uploaded by them will belong to Udio FOR FREE, to be used raw by others users and for training their AI. As long as people agree on liberating Udio from all that copyright headache, and making the startup richer, it's fine for me. I will still be making my music in a way that belongs TO ME.
@@marcosazambuja When you say "Belong" you mean, they have a license to train their model on it. Probably pretty similar when using AI mastering plugins too tbh. In fact, quite a lot of the music Distribution platforms also have terms that by uploading you are granting them a right to train AI models on the stuff. Not even music, likely this very comment and yours too. That is the world we live in.
As far as how people make content, that is up to the individual. 99% of my stuff has never seen an AI plugin let alone generative AI, but I can certainly see a time when I will ask for a funky vocal that compliments my original music. Nothing has changed here other than opening up new possibilities. And if you don't like those possibilities that is okay too.
@@jumpstar9000 I will like when these possibilities get free of legal infringments (otherwise we can't use it, for real - I'm not risking having a lawsuit due to the use of AI material in the soundtracks I produce). How long will it take, and will worth it? I don't know.
Honestly udio is top right now and if you are already a musician or artist you can take your music to places you never thought
Yes, wait till you try Suno's it's insane! Thanks For Watching
Holy crap! By next year this will just be a virtual producer!!
Yes sir, a Brave new world😊
What do you mean by next year? It already is.
Well is here but it just needs some fine tuning and a few functions but this is just implementation,but the core is already here ,i can already imagine it will evolve into a DAW kind of interface( such as pro tools) for those that want full control, it will then create everything given a few samples ( including any person's voice ) and create a fully produced pro song
If you could select a audio section and delete, replace, regenerate or remove elements then it really would be a game changer. Would also be great to be able to save stems.
Dope! your my new guy.
keep it up bro.
This is unreal. like a dream. Is this real or am I living in a simulation.
Been waiting my whole life for this.
Haha probably both!! Thanks so much!
As a composer, I have some mixed feelings, but the audio-to-audio feature works quite interesting by inputting a single instrument track and generating audio with little variation (so the melody is not modified, e.g. a violin playing). The output can serve as an extra layer of audio that can increase the realism.
This update is definitely a big step to make more custom music and a hybrid between human and ia, however I still notice strange artifacts, the snares, cymbals and drums do not sound real, sometimes the snares sound like they are playing with brushes ... I find it a fun toy but and exciting but I will also say that I detect a leagues the songs generated with ia. Just like I detect the texts written with chat gpt... and I've been listening to songs generated with ia for months, with incredible voices, but none of them move me... does this happen to you?
AI still needs that human touch, without it there is something missing. I doubt you would notice if you only listen to mainstream music.
Yes definitely, most time you can still tell it's ai, also keep in mind us ai music producers have heard hundreds of songs now, so we are training our ears to pick up those noises. Thanks for the great comment, and Thanks For Watching
The pace of Udio's development is impressive. I can't wait to try the new features :)
It's awesome! Thanks For Watching
Now I just wish they'd let you merge different tracks into one and introduce more editing tools and then it'd be perfect. A lot of the time it doesn't listen to my prompts even if I select prompt strength to be 100%
Yes I agree, I believe it's coming, they definitely are hearing what artists are wanting.
When competition is present, the user wins, you can do the same with SUNO. The sound quality maybe better with Udio but I can never get the results I get with SUNO. It will come down a preference and what you can work with better.
That’s cool! Thanks for this video!!! It’s making much easier for composing from scratch)😊🎉
Definitely, Sumo has one coming out as well I got to try it, it's so awesome. I'm going to do an A&B on them tomorrow. Thanks For Watching
Looks like you can download and then edit segments of the song to re-upload it with what you like about the track to see if the AI can build on your original idea, including the small segment it wrote for you earlier.
Wow! Excellent stuff bro! 👍👍👍BTW This opens up a whole new world of song creation for me as a bass player.
As a fellow bass player I 100% agree with you. I got to try Suno's version today and it's insane!! You are going to live it! Thanks For Watching
This is exciting. I can see that very soon, we as weekend producers, could design an AI artist and then produce a full album of content around the style of the AI artist you built.
Yeah. And who will be the audience for your AI artist?
AI audience
@@thedevilsadvocate5210 AIDIENCE :D
@@marcosazambuja billions of humans on Earth 🌎 🌍 🌏 + AI matchmaking + AI scoring.
Uploaded an actual 5 minute demo song. It plays the whole song back but it will only give you that 32 sec part after playing the 5 minutes, then it kicks in with the AI stuff. Lots of experimentation to do! I can see the credits diminishing quite fast now LOL.
You have to crop 😅
Yes, crop, then extend.
I am dreaming about times, when Udio will give me opportunity to open a stem, see all the plugins used, and I can only re-record my leading vocals))) Feels like heaven)))
I agree, it's coming!! Thanks For Watching
It’s right around the corner🙏🏽💎
What I'm missing is the possibility of creating a guitar solo for 4 bars for a song that I composed and recorded myself. Where you can specify the key and chords, or upload the location where the solo should be produced.
I don't think we are that far off from that. Thanks For Watching
I knew this was coming, but didn’t think it would be this soon. The ability to upload your own audio changes everything when I comes to songwriting and copyright ability. The chord progressions and melodies can be yours and Udio just follows your idea. That means it can be copyrighted. This is going to get better in terms of quality and how well it follows your lead.
Also coming according to their Office Hours presentation on Discord are STEMS and MIDI. There’s no argument against using this as far as I’m concerned.
Stems and MIDI would completely change the game. Udio is making me hate this tech a lot less by making it far more controllable. Sure, the AI is existentialy threatening, but how can you be a songwriter and not find all of this exhilaratingly fun, too?
Fun fact, Udio will be having Stems soon. And Suno has said it's on the roadmap. So hopefully even more control.
if you put your own music there and adding Udio orchestration , parts etc you don't have copyright.
Anyone can use your song without limits
This is the part where it helps to know about music production. 1. It helps if the audio sample is on beat. 2. Giving it looped files certainty helps. 3. if you are using a long audio file, it helps to extend from a good spot, anything short of that will yield weird results.
Suno is coming out with something similar, I like them both but Suno a bit more. Can't wait for you to review that
Yes! I got my hands on it today and it's incredible! Thanks For Watching
@@ChrisUnlocksAI Have you tried uploading Suno audio into Udio to improve the sound quality? How does it cope with vocals?
Suno's singing definitely needs a lot of work. the quality of Udio vocals is much better, but it tends to layer the vocals in a way that doesn't always sound natural.
It’s great to finish songs that were not finished by the original artist before they died
Or to take a song that you like and capture the essence of it continue the song and then cut out the original material
Hey Brother. I love what your doing with Udio here. I work a lot with Suno. I wonder what we could create. By combining some of our music together. Keep up the great work My Friend. I did Like and Subscribe Buddy.
Hey, I'm interested email me and we can try to do something fun!
I love Udio and all these updates!
Well full music production already here but it just needs some fine tuning and a few functions but this is just implementation,but the core is already here ,i can already imagine it will evolve into a DAW kind of interface( such as pro tools) for those that want full control, it will then create everything given a few samples ( including any person's voice ) and create a fully produced pro song
Writers block is a thing of the past
YES! Thanks For Watching
Greate tutorial. Now one can spend long hours playing with it.
What are the t&c's with this feature? I assume they keep your uploaded data and use it to learn the AI for other users to create similar tracks from your sounds
I've tried it. It's not a bad concept but the fidelity just isn't very good. Certainly not there yet. And you can't "Inpaint" over your audio yet.
It's a good roadmap idea, but for well realised mixed 24bit demos that are uploaded, the quality lets it down ATM. Give it a year...
As a fast UDIO convert I figured this update would take a year (not a month!!) and I have probably 300 dumb little song ideas it can finish for me. 😮
I agree! I did not see this coming. Thanks For Watching
I must admit it is dope but that 11 Lab sounds so real. I don’t know how they get the individual instruments to sound so distinctive, where here you can tell it’s more a flatter mix when quality should instead require dynamics. Suno is definitely for the culture though #hiphop #NC #atl
Great new feature you can even take the suno instrumentals and generate music quicker now, because i like suno generations more and not the vocals
Suno's version is also insane, got to try it today! Thanks For Watching
Nice work bro...
Thanks so much!! And Thanks For Watching
I saw your face and I was like hey, that guy looks like someone I used to go to middle school with, I think his name was Chris. Color me shocked when I exited fullscreen and looked at your username. You ever live in Maryland?
Hey, sorry born and raised in Kentucky. If you ever find my doppelgänger tell him I said hey!! (Spider man points finger)
Is it me or the quality of the songs in udio improved right now
Yes even better now!
I have had some bad outputs with this. I hope Suno is much better
Do they give subscribers daily credits or is it like suno that you only get a fixed amount if you subscribe
Yes it's daily credits at the moment. Hope this helps.
Thanks For Watching
What is the max length of audio u can uplaod?
Great question, it doesn't say. I had no issues with uploading a little over a minute.
Thanks for watching
Maybe you need to adjust the prompt strength
Yeah I agree. I'm going to dive deep into this when I have time and try to actually create some songs. Thanks for the advice and watching.
Great video! 🙏🏼
Thank you 🙌
I testing it aleardy and this is amaizing. At the moment UDIO beat SUNO, but we wait for SUNO v4. Another part of cake is generate start from Suno and upload to Udio to extend it and add vocal. You know what i mind? This is creazy.
udio sound quality has suno beat.
@@drmarioschannel Nobody need Suno beat after that. Now you create your short midi and Udio make from this awesome music. This resolved Udio rythm problem.
Could you upload just a lead vocal for a whole song then have udio put music to it? Or wouldn't that work? Cheers.
Yes you can, if you add the lyrics in the custom box. You could only do verse 1 then do verse 2 ect. You could add all the lyrics at once in Suno hope that helps.
Thanks for watching
@@ChrisUnlocksAI Brilliant, thank you for taking the time to respond. Take care, all the best.
tempted to buy a subscription now
I say do it for a month and see if you like it. It's easy to cancel. And I'm not affiliated with Udio in any way. Hope that helps. Thanks for watching.
Can the music promised using Udio be monetize of TH-cam?
Hello, yes it can. Thanks for watching.
I click DOWNLOAD, but it don't work!
It's not rolled out yet. Hopefully very soon. Thanks For Watching
So, ... I can upload long lyrics, and Audio creates a song up to 15 minutes?
It
Ah not on the first roll but yes, it you want to keep feeding it verses and choruses it will go to 15 minutes
@@ChrisUnlocksAI Thank you, now I see.
Awesome review
agree 100% - i am close to buy a paid version:)
Suno was doing this and udio released it beforethem lol
You use it they way I use it... iterate through it until I get it.
What if they eventually add a good DAW ⚡️Look out !
Yes!!!
With this do you own the copyright? im thinking maybe no because the ai generated additions use stolen copywrited training data. I don't have a subscription to know, what does it say about this?
Great question, if you create lyrics or a melody you own the rights to that, you can not copyright a chord progression as per the Ed Sheeran lawsuit. Hope this helps. Thanks For Watching
Echoes of melody rewrite words to cut city neon etc junk and go for it
Take the gap out!!!
You got me, what gap?
I think Audio needs to improve sound quality compared to Suno
Of the pieces generated so far my favourite is the one in the following link. It has its issues: Udio all but muted the right channel throughout one section, and I still need to properly fix it when I have the time. But I like the flow of the lyrics which have a rather complex rhytmic structure (the lyrics are by a Polish poet, written in the 1920s). The title translates to "Hangman's Ballad":
th-cam.com/video/3Vw0MkH53yk/w-d-xo.html
WOW!!
Sad its only for subscription peoples
Meh. the problem is you gave it almost nothing to work with. You have to give it more than a couple notes. It needs at least 30 seconds of starting data.
Your audio levels are too low my brother. I have to pump it up, which makes the next video in my queue too loud.
Ah man, that sucks, I will work on it. Thanks for letting me know. And for watching.
@@ChrisUnlocksAI The problem happens only when you speak, when you play some music the level is ok.
Impressive. Unfortunately, you can't copyright anything you generate from it.
I get why you say that, however, if you write lyrics or melodies you own the rights to it, that's black and white. Thanks for watching.
Of course you can. It's literally using your own audio.
I. My experience, including just yesterday, Suno is much better and the songs I actually use for self and clients but never one from Uno
So, can you now upload your song to Spotify and earn revenue from it? Or because the AI finished the song for you, you can't?
well, for what I understand it is a gray area still, but I believe you can (as of now)
You can't make money at
.004 cents a spin.
Spotify just upped their price did they give the musicians a raise?
not really related to the discussion here.
If you are paid subscriber you can upload it anywhere. So you can upload it to Spotify. Now can you make money from Spotify? That's a whole other question😂 Thanks For Watching
I MADE THIS WHOLE ALBUM ON UDIO th-cam.com/video/w12xe7KsDBs/w-d-xo.htmlsi=Gm_NJzXn4Vyy_PL6
Nice! Sounds great!! Thanks For Watching
Looks like it needs some work
Yeah, it's still pretty awesome though. Thanks for watching.
It doesn't need more work. It works great. He's just no cropping the audio before extending which is obviously not yielding good results.
@@mecd4167 show me something good then, im not a paid user (yet)
Thank you, but your voice is very low compared to music
Thanks I appreciate the feedback, I worked to improving it in my newer videos. Thanks For Watching
WTF
I know right!!
King of WHAT? The music generated by Udio belongs to ANYONE, so the users will be kings of an empty kingdom with no population...
This is my kingdom, come 😔
Not true. Didn’t you notice this allows audio uploads? You can use it to produce your own music. And I’m sorry that you’re in such denial, just because someone uses AI doesn’t make the entire thing copyright free. You think anyone using Adobe’s AI in Photoshop can’t copyright anything they make? Youre in for a shock. And you’re in for an even greater shock in the future because they’ll continue to make tools that let you use Ai in music production, which means it will be everywhere and there’s no legal justification to say someone who extends their own music and works with it back and forth with it in highly detailed ways doesn’t own their own track at the end
@@Edbrad Of course everybody can generate tracks thru Udio, Suno and the other services. Copyrightless tracks that belongs to nobody (yes, no copyright for the user).
Until you publish them on that client's project and all of a sudden receives some emails regarding (others) copyright infringment.
Why? Because these startups uses other artist's phonograms for training their models without consent.
And your client will never work with you again.
If you add lyrics or melody parts you 100% own the rights to that. There is no grey area there. This helps make that happen. Hope that helps, and Thanks For Watching
@@ChrisUnlocksAI Nope. If the AI was trained using 1 million tracks (which is the case), then you gotta SPLIT your ownership with all this people and companies. Oh boy.
You need to read Udio/Suno's terms of service, they almost laugh at your face.
Anyway, copyright laws doesn't work the way you painted. There's no Alice in this wonderland.
Plus: Udio and Suno used phonograms WITHOUT PERMISSION to train their models. Which means that, if you generate a track and decide to show it to the world via the internet, you could be notified of copyright infringment. If it's a risk you want to take, it's up to you.
But US Supreme Court decided last november that any AI generated material CANNOT BE COPYRIGHTED. I guess this ends the question.