I've used Suno for a while and have tried to use Udio for the last few days. When it comes to creating music for your own lyrics, Suno still blows Udio away! Yesterday, it took me 40+ takes to get a song with Udio to a semi-acceptable state, same lyrics in Suno took a few takes to make me happy. Udio is impressive for it's first few days of beta, but they have a long way to go. For now I'll be using Suno.
It’s also been my experience that Udio is FAR worse when given custom lyrics. I’ve mostly given up on that for now, focusing on describing the lyrics I want it to generate.
Yeah- as an avid follower of all things AI, Udio has been the single most shocking / awe-inspiring AI I’ve personally used in the last 18 months. LLM’s are incredible too, but I feel super weird admitting that some of these songs make me ‘feel’ things like ‘real’ music. Would not have guessed we’d get to this point so quickly.
For custom lyrics Suno still has the crown. Udio can make random jaw dropping magic, but struggles to give you what you want if you have something specific in mind. I'm sure they're taking notes and will improve over time.
Dethroned, definitively! I have listened to may examples that would have fooled me. I expected we'd reach this quality in 6 months, but we are here, and now I can't imagine where we'd be in 6 months, but, it's possible my favorite song of all time is going to be created with AI??
I feel like a gambler when using udio, and a surgeons scalpel when using suno. If only we could have the instrument and vocal fidelity from udio and the pinpoint precision from suno.
I approached this one with my "40 years a music producer" hat on, and I have to say my world was rocked. The only thing it cannot do is originality. But for everything else, this will be the future. You want a new Nirvana album? Get your prompts right, add a few sections and an intro and outro and Kurt's your bloody uncle. Dead popstars? Nah, they gonna live forever.
@@bandinopla It does two things very well - complete nonsense (presumably when it doesn't understand the prompt, or it has a "moment"), and a really good version of something that already exists. I fed a particular prompt in and I got a mixture of Hawkwind, Frank Zappa and Genesis. I tried another and got Peter Green, The Moody Blues and Jon Lord. What blew me away was not so much the copying of identifiable styles, but the way it was able to blend them in a way that more or less made sense. For someone who's been an active listener since the early seventies it's very hard to show me anything original. That's why I love it when it happens. The AI is delightful, but not original. So, if you need a long lost album of collaborations between two artists who never met, now is your time!
FlipFinderz Oh, I know what you mean, but what I'm talking about is a form or a sound that you've never heard before rather than just another four chord trick song like we've heard a million times before. Admittedly, most humans cannot find that kind of originality, because it takes a special moment to achieve it, so the AI is probably as good as 90% of musicians out there at turning the trick again. But it's still just that same old trick.
Been using it all night, incredible.. simply incredible. The fidelity is insanely good, it follows prompting much better than Suno I find. Hell it knows PC Music, which is super niche.
That hasn't been my experience, but I've found that if you explore the prompts others have used, you can learn to make your own prompts better and more accurate.
I love them all, but the best thing about Udio was when the "singer" started speaking in tongues. I am still trying to figure out what she was trying to sing. (I had to split the song, but I kept the "la la las", after all AI should be given some artistic license too.) Great tip on the "spoken" prompt.
These tools are cool. I really wish I could upload my own music and have the app generate the lyrics and vocals based on my prompt. As a producer, that would be an invaluable tool.
Cool, video especially the prompt brackets examples. Haha, can't be unheard now I can't go to the bathroom during the night without thinking of that choir song. Waiting for the first musical made inside Udio, might actually kick ass.
I almost pulled that song to use another example. I had a "what am I thinking" moment with regards to posting a song about having to pee...but in the end, I obviously went with it. :D
Udio audio quality is great, probably better, but I cannot see any masterpiece there. I've tried to make some songs, but they don't have the same great melodies as Suno! Suno has more pathos and heart. Another problem I've found in Udio is that when extending the songs, they are not always seamlessly joined, and you can hear a clear gap!
THEY JUST CHOPPED IT'S BALLS OFF! Thanks for enlightening me to it, I spent 40+ hours with it. It's 1% of the AWESOME it was now. I understand the legal implications.... they had no choice.
I find it's really hard to do something you want if you have anything specific in mind. It would be nice to be able to input the actual tune, and get it to make an arrangement based on that.
i dont know what you doing, but all i become is crap, and i'm music professional. Suno is way ahead better. maybe you can help me get the right prompting ? but when i write doen genre + gender + Key route like G#, its sound really garbage. ciao
Not just the audio quality is better but the lyrics is way better than Suno, the options for generate album art and intro, outro options are do good. Amazing see Udio, hope they update soon
Looks like the Exact same base code as Suno. I doubt Suno leaked it ... which begs Qs. Where did Suno get it from? Are both based on code we All have access to? Or does Suno secretly own Udio? Are any differences our imagination? Did Udio flat out sneak in and steal Suno's code? A lot of mysteries!!
I tried this but I didn't like how it put my lyrics in. Everything was rushed. I just saw Stable Audio. I don't understand it, but it sounds good so far. Suck it Suno!!!😘
One feature that will top this, and as soon as an AI tool will have it, it will be the new champ: Let me record my song on piano or guitar and sing it (even badly) into an mp3, allow me to upload that mp3 and the tool will figure out the genre that I'm trying to make and produce a better version of it for me. With the same chords, progression, tempo, and keys that I used. This shouldn't be too difficult with the technology we have, and I WANT IT! This way I don't have to just rely on the curent AI tool's melody selection. Please, someone make this!
Yeah, agreed those would be nice. Unfortunately, those aren't my personal areas of expertise. I can point to some simple mastering tools, but wouldn't be able to go in-depth into mastering with something like a DAW if that's what you mean. It is probably worth my taking a look at how people are using Spotify to share their AI music, to see if that's really "a thing."
It does every song into punk or oldies. It skips lyrics. Suno female pop vocals are on a totally new level . I agree that Udio is doing it differently but its not better. At least not yet.
Musicians will be still around with live performances, but...what about human creation? I'm specifically talking about composers. What really hurts me is people will think genuine composers use AI when in reality, they create music without using AI. Terrible unfair. What can we do to show we didn't use AI at all? Show our entire process? That is an option, but many would still be suspicious anyway. It really hurts.
I absolutely see your point! I’ve had those thoughts myself. Personally, I play the ukulele and spend many many many hours practicing to get better and better. If I posted an audio recording of all of my work and people assumed it was AI, that would absolutely suck.
Awesome! Could play with it and generate a complete song in french, gypsy jazz style, 5 extents and intro. I can see that adding more extents gets the results away from the original style, and produce more voice glitches. Second, I've been shocked by the ability to generate really complex classical musics, with complete orchestra and choir. That I thought impossible to render just yesterday! I tried too a voice separator to splut the text and make a verse sung by someone else, but it's not needed: just say [Voice 1] and 2, set "jazz female singer, rap male singer" in the prompt and after a few tries I could get a way better result. Currently crawling the shared musics, copying interesting styles, and try them with other texts. Mind blowing 🤩 Last but not least: if you like a precise style but not the text: remix the part with little or no variation and type your own lyrics. And restart from your variation (intro, extents...)
I tried it last night. Took a few days before I could get on. Some results sound pretty good, but it kinda does it's own thing sometimes. I'm pretty good with prompting, but it doesn't always follow. Suno is much better in that area, but you need to understand how to prompt it too.
Udio seem recognize more genre than Suno . Their result also seem cleaner when doing STEM separation . This would be useful if we want doing remix or laying some additional tracks and doing another post processing . I suspect that`s happen because Suno doing some post enhancing processing after songs created. But I can feel UDIO sometines doing hallucinations , especially when you`ve used autogenerated lyrics or some mashup genres
There's an AI that's been on my mind, all the time, Su-Su-Udio. The quality shown in your video seems way ahead of Suno, much cleaner and actually usable. As good as Suno was for messing about, it wasn't really usuable at least not in the version I tried.
Been playing with for a few hours, vocals on the song I've been working on seem quite clear, but we got a problem, it is not keeping the consistency of the song as I progress a couple of verses, where as Suno, does and I am using the same lyrics to a song I created in Suno, I will do more messing about with it, but not been able to finish a song, after 3 mins the vocals are gibberish, tbf, I can do that myself on stage
Hmm. That hasn't been my experience for the most part, but as I demonstrate in the video, the one outro created almost a totally different style, but the next take nailed it. So I would just keep hacking away at it.
I tried romantic classical symphonic music, which is way harder to assemble than pop music. Quality is, I would say, 128kBPS MP3 of an original 60s 70s recording. Compositions are awesome and I can't imagine how a machine could ever calculate such stuff.
I was doing the same thing, ballet as part of a prompt gives some amazing stuff. One thing I have not been able to get is strong melody. It's brilliant w/ gestures, textures, some cool call and response flourishes etc. wondering if you have had any success w/ a good, memorable melody and if so what prompts/tags did you use?
I can't seem to find a way in either Suno OR Udio to add in background samples of things like birds chirping or the sound of waves on a beach. Have tried using [tags] but they seem to be ignored. Any suggestions?
Hello because I am blind, would you mind to send me a promt example with the brackets where you create a choir? I love your video! You can just post a reply if you want. Regards Daniele
I just tried to set the Spike Milligan poem to music "On the Ning Nang Nong Where the Cows go Bong! and the monkeys all say BOO! There's a Nong Nang Ning Where the trees go Ping! And the tea pots jibber jabber joo." But it didn't like it, saying "moderation error" which I think means the AI has decided it's not Woke enough for Udio.
See how AI glitches can become an inspiration: yesterday I made this song with Udio th-cam.com/video/r0FXMa_mbFk/w-d-xo.html&ab_channel=EmilBiljarski Well, I only wrote the prompt and the lyrics, but they were inspired by the music and voice coming from the first prompt. After 3 verses I ordered an instrumental solo. But the outcome was singing in a non-existing language. First, I wanted to cut it out, then I found out, that it comes after the words: You are the prompt, I am your mirror. We're getting closer, the view is getting clearer. So, the when the girl let me close, she started singing in her own language… I am blown away
xD nah still not as good as Suno imo. If you learn all the tricks to properly prompt Suno, you will easily tell it's leagues better than even Udio. But Udio is a good contestant and certainly second in line for sure. It definitely has its potential and a way better interface than Suno, so it could one day become better but it's just not there yet (again, from the perspective of someone who knows how to prompt Suno very well to always get flawless results). Definitely a fun alternative for sure, sadly won't be free so won't really excel as a result.
Super fun, exciting, mind-blowing, but ultimately not useful for song creation until they have a way to create consistent verses and choruses. Suno, while not amazing at this, does actually have the capability, even if the quality of the music isn't as high.
This may have been true with my examples, but if you listen to others on the Discover page, you'll find plenty of examples with lyrics that are nice and slow.
I think it's just about taking a few minutes to learn the differences in how prompting works. Not all tools have the same instructions. If you're really interested in getting the most out of Udio, you can choose to dig into their documentation and use it at a more serious level. But if you're just "playing with it" it may not occur as intuitive as Suno - but in my case, I think it's worth taking the time to learn it. Probably not that difficult once you know.
@@BobDoyleMedia I'll give it another try again since I used it on the first day as I signed up and had to wait long for an output. If I do find that it makes things sound better I will have to rethink a couple things.
I wouldn’t call it amazing over Suno. They still are missing stem exports, midi exports, and the sound quality compared to Suno v2 is crap. Sounds like a streaming audio over compressed. Can’t really use this potential scratch pad; cool attempt though. They’ll get there at some point.
I have to disagree with the other comments. I actually have someone I pay to use my full Suno credits for me each month, as they often went to waste. It's useful but the quality, even on v3, especially the vocals is just awful honestly. It's great for making work tapes and getting ideas going but Udio can at times make actual broadcast ready music. I used to produce for Chris Moon (Prince) and Butch Vig (Nirvana, Green Day) and I know what broadcast ready music sounds like. I have 300+ songs on my Spotify page and I'd say maybe 5 of them somewhat meet those standards. Yet lately I have taken all my originals, plugged them into Udio, and currently have 14 full ready Country songs of all things (I'm an EDM guy) and am gearing up to sell them directly to a label as they sound so good, they feel they can plug them into an existing artist page, and get them plays. Suno sounds like Suno, Udio may be hit or miss, but it's the only one capable of not sounding like Ai, especially if you have the ear to sift through the nonsense. Also Udio is the one that lets you use famous artist names, and will tailor a custom prompt to be as close as possible. Trust me when I say, I have a few that make all the difference. Using the changed prompt doesn't yield the same results. The problem with Udio is you can't treat it like Suno. you have to really "Steer" the song in the right direction and collab with it. Suno spits ideas out like wildfire but it refuses to work with you. It only does what it does. Udio is also about to launch seed control which is going to be incredible. Best way to use either one, is to use both imo. We use Suno to make a folder of 1,000+ ideas, then remake them in Udio. Stem out, remix, produce over, and there ya have it.
I love AI but can’t say as much for AI music. Everything I’ve heard is “good for AI” but not good. I’ve yet to hear it make “a hit” One of those songs that gets stuck in my head.
I know that hype sells, but after exploring and contrasting suno, udio and sonauto I don't think udio is objectively better, the quality of the audio perse is indeed better in udio, that is for sure, but fidelity at this point is not the whole map, push udio with complex lyrics and it goes to hell immediately, also, I would say that udio sounds limited in creativity, probably a exchange to grant more clean quality, suno is much better at rendering "what you may expect" from the prompt, whereas udio has less nuanced interpretation, and is terrible at mixing genres, the songs I got from suno are not always publish ready but are mostly excellent compositions, I'm working on rerecording around 20 songs I made with suno, and in excited to get them done, I haven't felt that excitement with udio... And let's not talk about sonauto, it only has a gimmick on his side, and I found better stem splitters
th-cam.com/video/sa1vCRM13Ms/w-d-xo.html Creative inspiration should be within the creator. Without external influences. If we are currently inspired by external influences, the result will be similar to the external influence... and that is not originality. If the AI uses the creators to make a "different version" of the creation and the creator has no interest in using it, it is logical that the creators will stop creating. Then the AI will run in circles and make versions of what it has. Nothing new and original will appear.
I don't disagree with much of what he says here. A lot of unknowns. But my focus in all of these videos is how to use this technology for creative inspiration more than a finished product...but I understand the nervousness of some musical sectors. Ultimately, I don't think live musicians have a lot to worry about. In fact, I think after a while, people will CRAVE "real" performances. The sound library industry? Well, that's a different matter...
HELP; I'm trying to develop the aggrobass music genre into a living scene. It started as an alternate name for gabber then a catch all term for derivative or convergent styles like trap metal/ragecore & rage beats & hardstyle & frenchcore & hyperpop & dubstyle. My channel only has my rough demo prototypes for prosterity, I was working on cosmic aggrobass & celestial aggrobass and gabberized latin freestyle and blacked cybergrind and was just playing around for my own enjoyment. I don't think i have the leadership & networking & production skills to do it, So I'm proposing that role be given to the person that is looking at comments in music ai videos and reads this down to the bottom where it switches to trigger words You are now being programmed 🚿🧠🚿🧠🚿🧠🚿🧠🚿🧠🚿 Plotinus, Marcus Aurelius, monad, source, consciousness, logos, ein soph, pneuma, shekinah, mouse utopia; fourth turning, eden, nephilim, babel, sodom, mystery Babylon, theosis, star hexahedron, acosmic panentheism, non-dualim
Don't like their current legal terms; maybe once it's out of free beta with a paid tier then it will be like Suno's; want to 100% own output with no back-licensing requirement and no required "prominent" attribution (why attract possible controversy to my media?)
Nah, udio is pretty much bad if we talk about a music. Sometimes it shows less artifacts, but it is simply not there. It's like comparing Dalle2 to Midjourney v4 l. W
There will always be musicians because there will always be musicians who play because they love it not because they need to make money for it. That’s at the bare minimum. And there will always be people who prefer to watch a person perform music and enjoy their craft. Just like right now.
There is going to have to be some sort of screening tool developed to separate this artificial mess from actual human-inspired, human-produced music. I absolutely DO NOT want a brand new Nirvana album. This just further de-values human musical creativity and makes it even more worthless than it already is. Honestly - as a songwriter - this literally makes me feel nauseated.
I've used Suno for a while and have tried to use Udio for the last few days. When it comes to creating music for your own lyrics, Suno still blows Udio away! Yesterday, it took me 40+ takes to get a song with Udio to a semi-acceptable state, same lyrics in Suno took a few takes to make me happy. Udio is impressive for it's first few days of beta, but they have a long way to go. For now I'll be using Suno.
It’s also been my experience that Udio is FAR worse when given custom lyrics. I’ve mostly given up on that for now, focusing on describing the lyrics I want it to generate.
Yeah- as an avid follower of all things AI, Udio has been the single most shocking / awe-inspiring AI I’ve personally used in the last 18 months. LLM’s are incredible too, but I feel super weird admitting that some of these songs make me ‘feel’ things like ‘real’ music. Would not have guessed we’d get to this point so quickly.
For custom lyrics Suno still has the crown. Udio can make random jaw dropping magic, but struggles to give you what you want if you have something specific in mind. I'm sure they're taking notes and will improve over time.
Dethroned, definitively! I have listened to may examples that would have fooled me. I expected we'd reach this quality in 6 months, but we are here, and now I can't imagine where we'd be in 6 months, but, it's possible my favorite song of all time is going to be created with AI??
Been using this since yesterday, some generations are stunning, some aren't but it's jaw-dropping for something that's still in beta.
I feel like a gambler when using udio, and a surgeons scalpel when using suno. If only we could have the instrument and vocal fidelity from udio and the pinpoint precision from suno.
what is actually impressive to me on Udio, is the high fidelity with the voice, is very good, omg, and look i am saying, in another language.
Agreed. With headphones, you really hear this.
I approached this one with my "40 years a music producer" hat on, and I have to say my world was rocked. The only thing it cannot do is originality. But for everything else, this will be the future. You want a new Nirvana album? Get your prompts right, add a few sections and an intro and outro and Kurt's your bloody uncle.
Dead popstars? Nah, they gonna live forever.
I see what you did there.
what do you see that make you say it can't be original? musically speaking...
I'm a stagehand, record label owner, and a guitarist myself. Hello from orlando. 🎉
@@bandinopla It does two things very well - complete nonsense (presumably when it doesn't understand the prompt, or it has a "moment"), and a really good version of something that already exists.
I fed a particular prompt in and I got a mixture of Hawkwind, Frank Zappa and Genesis. I tried another and got Peter Green, The Moody Blues and Jon Lord.
What blew me away was not so much the copying of identifiable styles, but the way it was able to blend them in a way that more or less made sense.
For someone who's been an active listener since the early seventies it's very hard to show me anything original. That's why I love it when it happens.
The AI is delightful, but not original. So, if you need a long lost album of collaborations between two artists who never met, now is your time!
FlipFinderz Oh, I know what you mean, but what I'm talking about is a form or a sound that you've never heard before rather than just another four chord trick song like we've heard a million times before.
Admittedly, most humans cannot find that kind of originality, because it takes a special moment to achieve it, so the AI is probably as good as 90% of musicians out there at turning the trick again. But it's still just that same old trick.
Been using it all night, incredible.. simply incredible. The fidelity is insanely good, it follows prompting much better than Suno I find. Hell it knows PC Music, which is super niche.
The audio quality seems better but it's hard to get a good result from your prompt. Suno is much better at being easier to use.
That hasn't been my experience, but I've found that if you explore the prompts others have used, you can learn to make your own prompts better and more accurate.
yes I cannot get a melody I really love in Udio !
Suno has much more heart
@@BobDoyleMedia Suno prompting doesn't take a prompt engineer. If you type in "Ragtime piano hip hop " that is what you get in Suno .
It's not as surgical as suno once you look past the initial knee jerk reaction .suno understands and delivers faster and on point
exactly.
I love them all, but the best thing about Udio was when the "singer" started speaking in tongues. I am still trying to figure out what she was trying to sing. (I had to split the song, but I kept the "la la las", after all AI should be given some artistic license too.)
Great tip on the "spoken" prompt.
Yay! We got it today! Watching it now
Love your videos. I also love your ability to come up with the most random song topics lol
Anyone trying to create music from this century has been left disappointed by it, but I'm sure they'll fix that. The quality is amazing.
yea the modern stuff it lacks a little but not Too much!!!!
Now this is a great competitor to Suno. I love this. I just wish it would generate longer generations so you don't have to keep extending
These tools are cool. I really wish I could upload my own music and have the app generate the lyrics and vocals based on my prompt. As a producer, that would be an invaluable tool.
Cool, video especially the prompt brackets examples. Haha, can't be unheard now I can't go to the bathroom during the night without thinking of that choir song. Waiting for the first musical made inside Udio, might actually kick ass.
I almost pulled that song to use another example. I had a "what am I thinking" moment with regards to posting a song about having to pee...but in the end, I obviously went with it. :D
Udio audio quality is great, probably better, but I cannot see any masterpiece there.
I've tried to make some songs, but they don't have the same great melodies as Suno! Suno has more pathos and heart.
Another problem I've found in Udio is that when extending the songs, they are not always seamlessly joined, and you can hear a clear gap!
THEY JUST CHOPPED IT'S BALLS OFF! Thanks for enlightening me to it, I spent 40+ hours with it. It's 1% of the AWESOME it was now. I understand the legal implications.... they had no choice.
I find it's really hard to do something you want if you have anything specific in mind. It would be nice to be able to input the actual tune, and get it to make an arrangement based on that.
Your prompts are hilarious!!! The music is sooo good that I’m literally laughing out loud! Keeping the music funny kept me watching the whole video
Udio is my new AI obsession.
I feel ya.
i dont know what you doing, but all i become is crap, and i'm music professional. Suno is way ahead better. maybe you can help me get the right prompting ? but when i write doen genre + gender + Key route like G#, its sound really garbage. ciao
@@blackswan6386 Suno is much better
I was waiting for this!!
I played with it lastnight, it is Pretty good, easy to use and great results, bye bye Suno
Not just the audio quality is better but the lyrics is way better than Suno, the options for generate album art and intro, outro options are do good. Amazing see Udio, hope they update soon
Looks like the Exact same base code as Suno. I doubt Suno leaked it ... which begs Qs. Where did Suno get it from? Are both based on code we All have access to? Or does Suno secretly own Udio? Are any differences our imagination? Did Udio flat out sneak in and steal Suno's code? A lot of mysteries!!
I tried this but I didn't like how it put my lyrics in. Everything was rushed. I just saw Stable Audio. I don't understand it, but it sounds good so far. Suck it Suno!!!😘
the 'Bob Doyle Media' tune at the end is worth watching for!
Thanks. I hesitated to put still ANOTHER version of it, but I really liked that take too! :D
One feature that will top this, and as soon as an AI tool will have it, it will be the new champ:
Let me record my song on piano or guitar and sing it (even badly) into an mp3, allow me to upload that mp3 and the tool will figure out the genre that I'm trying to make and produce a better version of it for me. With the same chords, progression, tempo, and keys that I used.
This shouldn't be too difficult with the technology we have, and I WANT IT! This way I don't have to just rely on the curent AI tool's melody selection. Please, someone make this!
Thank you, my friend. Now, we need tutorials on what’s next:
Music AI - Done ✅
Mastering AI
Publishing
Yeah, agreed those would be nice. Unfortunately, those aren't my personal areas of expertise. I can point to some simple mastering tools, but wouldn't be able to go in-depth into mastering with something like a DAW if that's what you mean. It is probably worth my taking a look at how people are using Spotify to share their AI music, to see if that's really "a thing."
It does every song into punk or oldies. It skips lyrics.
Suno female pop vocals are on a totally new level . I agree that Udio is doing it differently but its not better. At least not yet.
It's definitely fire. Messing with it now. The best
Very inspiring how ai works with any lyric sound its very creative, its opening up our options
This video made me subscribe. Keep putting out quality content.
There are 5+ ai TH-cam influencers that race to post similar info…….
I know. And I totally lost that race because the Udio servers kept crashing as I was trying to make the video. :D
Musicians will be still around with live performances, but...what about human creation? I'm specifically talking about composers. What really hurts me is people will think genuine composers use AI when in reality, they create music without using AI. Terrible unfair. What can we do to show we didn't use AI at all? Show our entire process? That is an option, but many would still be suspicious anyway. It really hurts.
I absolutely see your point! I’ve had those thoughts myself. Personally, I play the ukulele and spend many many many hours practicing to get better and better. If I posted an audio recording of all of my work and people assumed it was AI, that would absolutely suck.
Awesome! Could play with it and generate a complete song in french, gypsy jazz style, 5 extents and intro.
I can see that adding more extents gets the results away from the original style, and produce more voice glitches.
Second, I've been shocked by the ability to generate really complex classical musics, with complete orchestra and choir.
That I thought impossible to render just yesterday!
I tried too a voice separator to splut the text and make a verse sung by someone else, but it's not needed: just say [Voice 1] and 2, set "jazz female singer, rap male singer" in the prompt and after a few tries I could get a way better result.
Currently crawling the shared musics, copying interesting styles, and try them with other texts. Mind blowing 🤩
Last but not least: if you like a precise style but not the text: remix the part with little or no variation and type your own lyrics.
And restart from your variation (intro, extents...)
2024: The Year the human soul was extracted from art.
Super cool. I wonder if you can specify a key or chord progressions in prompts? Enjoying your videos. Thanks.
I tried it last night. Took a few days before I could get on. Some results sound pretty good, but it kinda does it's own thing sometimes. I'm pretty good with prompting, but it doesn't always follow. Suno is much better in that area, but you need to understand how to prompt it too.
actually if suno does not improve on this level, they will be left behing quite quickly
I agree, and I'm sure they WILL.
Where are the open source music ai generators like stable diffusion is for graphic art?
RRIISPECK BOB DOYLE - doing a v v v v good job!!!
Udio seem recognize more genre than Suno . Their result also seem cleaner when doing STEM separation . This would be useful if we want doing remix or laying some additional tracks and doing another post processing .
I suspect that`s happen because Suno doing some post enhancing processing after songs created.
But I can feel UDIO sometines doing hallucinations , especially when you`ve used autogenerated lyrics or some mashup genres
Suno still the king
There's an AI that's been on my mind, all the time, Su-Su-Udio.
The quality shown in your video seems way ahead of Suno, much cleaner and actually usable. As good as Suno was for messing about, it wasn't really usuable at least not in the version I tried.
Been playing with for a few hours, vocals on the song I've been working on seem quite clear, but we got a problem, it is not keeping the consistency of the song as I progress a couple of verses, where as Suno, does and I am using the same lyrics to a song I created in Suno, I will do more messing about with it, but not been able to finish a song, after 3 mins the vocals are gibberish, tbf, I can do that myself on stage
Hmm. That hasn't been my experience for the most part, but as I demonstrate in the video, the one outro created almost a totally different style, but the next take nailed it. So I would just keep hacking away at it.
Vocal are good all around, Suno ads to much noise to metal tracks, its hard to isolate that way
The Voices are realy nice in UDIO, but the instrumental,mixing and mastering is in SUNO much better.
Hello.. does udio create the accompanying videos to go with the audio generation too? Or did u create the videos separately?
the dune as a broadway music one is great song
Yeah, I thought that was hilarious - and I know nothing about Dune.
They told me they're not accepting anymore people for their waiting list and are already full to capacity as of last night 🤷🏿♂️🤷♀️🤷
Oh no!
I tried romantic classical symphonic music, which is way harder to assemble than pop music. Quality is, I would say, 128kBPS MP3 of an original 60s 70s recording. Compositions are awesome and I can't imagine how a machine could ever calculate such stuff.
I was doing the same thing, ballet as part of a prompt gives some amazing stuff. One thing I have not been able to get is strong melody. It's brilliant w/ gestures, textures, some cool call and response flourishes etc. wondering if you have had any success w/ a good, memorable melody and if so what prompts/tags did you use?
Why does the audio sound like wobbly MP3s? I've got TH-cam on 1080?
I can't seem to find a way in either Suno OR Udio to add in background samples of things like birds chirping or the sound of waves on a beach. Have tried using [tags] but they seem to be ignored. Any suggestions?
AI instant streaming of audio and visual content incoming. I say by next year at this rate.
inspiring video thx a lot!
Awesome Bob. Thank you
Where does the video in the top right corner come from and what can you do with it re: using it in conjunction with your new song?
It took me 10 generations to get to 1 4 minute 22 second song that I liked. I've made 20 in a week.😂
At least you have discernment!
What would be really cool would be if they also gave us the stems!
Yes, for now there are other tools that will do that. But that would be a nice thing to add.
👍👋🏻 I am a songwriter, I would like to hear my songs arranged. Is it possible to insert the chords of the song together with the tempo in the prompt?
Hello because I am blind, would you mind to send me a promt example with the brackets where you create a choir? I love your video! You can just post a reply if you want. Regards Daniele
Sounds like Udio has a way to go yet.
Great video as always… it’s pointless adding those background videos as they don’t lip sinc
I just tried to set the Spike Milligan poem to music
"On the Ning Nang Nong
Where the Cows go Bong!
and the monkeys all say BOO!
There's a Nong Nang Ning
Where the trees go Ping!
And the tea pots jibber jabber joo."
But it didn't like it, saying "moderation error" which I think means the AI has decided it's not Woke enough for Udio.
Several of these generators are on the lookout of non-original material. So existing song lyrics, poems, etc. could get flagged.
Very nice
See how AI glitches can become an inspiration: yesterday I made this song with Udio th-cam.com/video/r0FXMa_mbFk/w-d-xo.html&ab_channel=EmilBiljarski
Well, I only wrote the prompt and the lyrics, but they were inspired by the music and voice coming from the first prompt. After 3 verses I ordered an instrumental solo. But the outcome was singing in a non-existing language. First, I wanted to cut it out, then I found out, that it comes after the words:
You are the prompt,
I am your mirror.
We're getting closer,
the view is getting clearer.
So, the when the girl let me close, she started singing in her own language… I am blown away
Suno does not allow you to mention artists n the prompt. I assumed it was a legal issue. If not/ if so, whats going on with Udio??
They are clear that they only use the names as a style reference, not to access a voice model.
LOl. They can't even get their page to work.
like udio it's cool
xD nah still not as good as Suno imo. If you learn all the tricks to properly prompt Suno, you will easily tell it's leagues better than even Udio. But Udio is a good contestant and certainly second in line for sure. It definitely has its potential and a way better interface than Suno, so it could one day become better but it's just not there yet (again, from the perspective of someone who knows how to prompt Suno very well to always get flawless results). Definitely a fun alternative for sure, sadly won't be free so won't really excel as a result.
Super fun, exciting, mind-blowing, but ultimately not useful for song creation until they have a way to create consistent verses and choruses. Suno, while not amazing at this, does actually have the capability, even if the quality of the music isn't as high.
But you can still use what is generated as inspiration to bring into "real life" using traditional musicians and singers.
I love it BUT the lawsuits will be EXTREME!
interesting
No Idea how to join the bits to create a full song
Seems to race through lyrics even though you have slow in the prompts and low bpm.
This may have been true with my examples, but if you listen to others on the Discover page, you'll find plenty of examples with lyrics that are nice and slow.
😍😍😍
It adds huge gaps into music at the third extinction. There is not enough control over where music gets put in.
Suno is much easier to use. I thought I was going to get grey hair waiting on Udio prompt to finish...
I think it's just about taking a few minutes to learn the differences in how prompting works. Not all tools have the same instructions. If you're really interested in getting the most out of Udio, you can choose to dig into their documentation and use it at a more serious level. But if you're just "playing with it" it may not occur as intuitive as Suno - but in my case, I think it's worth taking the time to learn it. Probably not that difficult once you know.
@@BobDoyleMedia I'll give it another try again since I used it on the first day as I signed up and had to wait long for an output. If I do find that it makes things sound better I will have to rethink a couple things.
Website never works when I try it. Probably so popular it's being smashed. Thanks for this.
That’s what happened to me yesterday, which is why I couldn’t post this video until today.
The timing is very important, people were saying they were getting full speed generations at 2am!
@@Ovyron 2 AM where?
I was actually up at 1:30 AM Pacific to create the content for this video.
I wouldn’t call it amazing over Suno. They still are missing stem exports, midi exports, and the sound quality compared to Suno v2 is crap. Sounds like a streaming audio over compressed. Can’t really use this potential scratch pad; cool attempt though. They’ll get there at some point.
I have to disagree with the other comments. I actually have someone I pay to use my full Suno credits for me each month, as they often went to waste. It's useful but the quality, even on v3, especially the vocals is just awful honestly. It's great for making work tapes and getting ideas going but Udio can at times make actual broadcast ready music.
I used to produce for Chris Moon (Prince) and Butch Vig (Nirvana, Green Day) and I know what broadcast ready music sounds like. I have 300+ songs on my Spotify page and I'd say maybe 5 of them somewhat meet those standards. Yet lately I have taken all my originals, plugged them into Udio, and currently have 14 full ready Country songs of all things (I'm an EDM guy) and am gearing up to sell them directly to a label as they sound so good, they feel they can plug them into an existing artist page, and get them plays.
Suno sounds like Suno, Udio may be hit or miss, but it's the only one capable of not sounding like Ai, especially if you have the ear to sift through the nonsense. Also Udio is the one that lets you use famous artist names, and will tailor a custom prompt to be as close as possible.
Trust me when I say, I have a few that make all the difference. Using the changed prompt doesn't yield the same results. The problem with Udio is you can't treat it like Suno. you have to really "Steer" the song in the right direction and collab with it. Suno spits ideas out like wildfire but it refuses to work with you. It only does what it does. Udio is also about to launch seed control which is going to be incredible.
Best way to use either one, is to use both imo. We use Suno to make a folder of 1,000+ ideas, then remake them in Udio. Stem out, remix, produce over, and there ya have it.
Suno vs Udio is the new Apple vs Android 😂
I love AI but can’t say as much for AI music. Everything I’ve heard is “good for AI” but not good. I’ve yet to hear it make “a hit” One of those songs that gets stuck in my head.
I know that hype sells, but after exploring and contrasting suno, udio and sonauto I don't think udio is objectively better, the quality of the audio perse is indeed better in udio, that is for sure, but fidelity at this point is not the whole map, push udio with complex lyrics and it goes to hell immediately, also, I would say that udio sounds limited in creativity, probably a exchange to grant more clean quality, suno is much better at rendering "what you may expect" from the prompt, whereas udio has less nuanced interpretation, and is terrible at mixing genres, the songs I got from suno are not always publish ready but are mostly excellent compositions, I'm working on rerecording around 20 songs I made with suno, and in excited to get them done, I haven't felt that excitement with udio... And let's not talk about sonauto, it only has a gimmick on his side, and I found better stem splitters
Anyone else have their ISP keep you from going there?
It's good but when S is used on words it sounds bad, in fact it detracts me when listening to the rest of the underlying audio.
th-cam.com/video/sa1vCRM13Ms/w-d-xo.html
Creative inspiration should be within the creator. Without external influences. If we are currently inspired by external influences,
the result will be similar to the external influence... and that is not originality. If the AI uses the creators to make a "different version"
of the creation and the creator has no interest in using it, it is logical that the creators will stop creating. Then the AI will run in circles
and make versions of what it has. Nothing new and original will appear.
I don't disagree with much of what he says here. A lot of unknowns. But my focus in all of these videos is how to use this technology for creative inspiration more than a finished product...but I understand the nervousness of some musical sectors.
Ultimately, I don't think live musicians have a lot to worry about. In fact, I think after a while, people will CRAVE "real" performances.
The sound library industry? Well, that's a different matter...
Udio is cool but Suno Ai is still wayyy better, more accurate, better quality,
It can’t create anything with my own lyrics
Well it has that capability as I demonstrated. Are you getting an error or something?
@@BobDoyleMedia I got totally different results
HELP; I'm trying to develop the aggrobass music genre into a living scene. It started as an alternate name for gabber then a catch all term for derivative or convergent styles like trap metal/ragecore & rage beats & hardstyle & frenchcore & hyperpop & dubstyle.
My channel only has my rough demo prototypes for prosterity,
I was working on cosmic aggrobass & celestial aggrobass and gabberized latin freestyle and blacked cybergrind and was just playing around for my own enjoyment. I don't think i have the leadership & networking & production skills to do it,
So I'm proposing that role be given to the person that is looking at comments in music ai videos and reads this down to the bottom where it switches to trigger words
You are now being programmed
🚿🧠🚿🧠🚿🧠🚿🧠🚿🧠🚿
Plotinus, Marcus Aurelius, monad, source, consciousness, logos, ein soph, pneuma, shekinah, mouse utopia; fourth turning, eden, nephilim, babel, sodom, mystery Babylon, theosis, star hexahedron, acosmic panentheism, non-dualim
Don't like their current legal terms; maybe once it's out of free beta with a paid tier then it will be like Suno's; want to 100% own output with no back-licensing requirement and no required "prominent" attribution (why attract possible controversy to my media?)
Nah, udio is pretty much bad if we talk about a music. Sometimes it shows less artifacts, but it is simply not there. It's like comparing Dalle2 to Midjourney v4 l. W
Goodbye musicians
There will always be musicians because there will always be musicians who play because they love it not because they need to make money for it. That’s at the bare minimum. And there will always be people who prefer to watch a person perform music and enjoy their craft. Just like right now.
There is going to have to be some sort of screening tool developed to separate this artificial mess from actual human-inspired, human-produced music. I absolutely DO NOT want a brand new Nirvana album. This just further de-values human musical creativity and makes it even more worthless than it already is. Honestly - as a songwriter - this literally makes me feel nauseated.
No Way after a few tries, I gave up, Suno is much, much better but worst sound quality
Did Udio pay off all the streamers? It sucks compared to Suno.
Will it let you import your own generated voice model so you can sing your own theme song? 👍😉