6:59 you enter the seed number yourself, that's why you don't see anything in the window. Pick a random number, enter it (and make sure to make a note of it!), then if you like the generated result, you can reuse that seed number on future songs and generations. Of couse the results will not be 100% the same, but many features will be similar and you can potentially use it to create a personal sound/style. Note that it ONLY works in manual mode. I haven't really used it myself yet, but it's said to influence things like panning and voices. 7:23 - the edit function is primarily intended for editing the TITLE of the song and the COVER, not for the lyrics. 15:17 - context length. This has nothing to do with the length of the generations created. It's used for determining how far back in the song the ai should look for context when creating the next generation. One of the biggest problems with Udio used to be, that you couldn't get consistent verses and choruses, because the ai - so to speak - had forgotten what earlier verses/chorus sounded like, so it would more or less create a new melody for every one. The context length is - in plain terms - the memory bank of the ai. So if you set it to max, it will remember the last 2 minutes and take them into consideration when generating the new extension. If you want a change, however, like another mood/style or even just creating a bridge that doesn't sound like previous content, move the slider to a low position, and this will help the ai "think out of the box" and create something new. The "Lyrics strength" slider is primarily for giving Udio freedom to improvise with background vocals/answers/echos for example, like oohs, ahhs, yeahs, mmmhs and so on. If set on max, you're telling Udio "just do what I say, ok? No messing around with funny ideas", while on min setting, it may decide to completely disregard your lyrics if it "thinks" it can do better. Clip start is, as you also mention, a sort of "memo" to the ai, about approximately where in the song you're currently working. This means that a setting of 0% will tell the ai, that you're at the very start, so shouldn't begin "mid-sentence" so to speak, while setting it to 90% or above means, that you want an ending, so either a fade or a cold end. In between these values you help determine the "intensity" of the generation - like you'd typically have a more sparse instrumentation near the beginning, with more layers, effects and choirs being added the closer you get to the end.
Thanks for that breakdown! 🙌 What about the quality option? Does it mean we can now get cleaner / crispier outputs? If so, are we able to bump the quality of our older songs? Same with seed, are we able to generate seeds for our previous songs so we can keep the same voice and vibe?
Great information my friend, I was wondering about the seed numbers as they had stated a prompt will always yield a similar output when a seed is used.nothing about creating the seed numbers yourself. I think I was expecting a midjourny seed number. Thanks for watching and the great information
@@ChrisUnlocksAIthe good thing about this it shows wot a minefield these AI UIs are. They can be very messy and very user unfriendly. Midjourney is the same. I thought I was going mad when I tried searching a prompt in the explore menu, u could never go back , u then had to type it out again. Turns out you gotta click in a tiny area around the image to close it and return to your search??? I'm gonna sign up for this looks fun ;)
Both Udio and Suno are good; Udio has better control over song creation, and Suno writes better songs. When I use Suno and find a song I like, I keep extending the song while tweaking my lyrics. I might create 50 tracks, keeping faulty tracks that contain small parts I like. Then I cut out everything I plan to use and compile it into a single file or two using Adobe Audition. Afterward, I export what I decide to keep into a single track or more, depending on how much I keep. I then separate the stems of the track and import them into a multitrack session in Audition. Using the multitrack session, I can reuse, replace, or combine vocals and rearrange all the different parts into a song. For example, a refrain I really liked might become a bridge instead, allowing an even better refrain from one of the alternative song extensions I created in Suno to take its place. This process enables me to make the best use of all the material I generated in Suno.
Yes! This is exactly how I like to do it, you may go more in depth than I do, but we both tear into the song and try to make it what we want. Send me some links I would love to hear them. Thanks For Watching
@@ChrisUnlocksAI I put a lot of effort into this one, and there's a buildup through the singing as the song progresses. My next step will be to replace everything with real instruments and vocals. Still in Love with You fRDkwepb_zg This is a work in progress, and it contains at least four different verse progressions. The instrumental parts at the beginning and in the middle are from a totally different song. Purple Eye JkMSXmSo8d8 I put a lot of effort into the lyrics (English is my second language). Sledgehammer lBJG2QmTK8g
For the Seed maybe you have to enter the seed before you create? then it sticks to the song? What they need to do is have separate the vocals from the audio so you can change vocals without harming the music
Perhaps it says a prompt will always yield a similar output when a seed is used.i assumed a seed would be generated like midjourny. However that's not the caes
I like the new update on Suno. It's doing a lot better with my songs as far as structuring them. I like that the censorship seems to have decreased as well. The only thing I don't like is that now that there are 4 min generations the AI seems to think it has to fill that entire time even if it has to replay lyrics I don't want it to. So I've gotten several songs I've liked but haven't kept cause it'll pick a previous verse, replay it and it gets cut off while singing. I know I could extend but I'd prefer being able to do like on Udio and clip the end to just get the song I wanted out of it.
@@ChrisUnlocksAI Yeah, but sometimes I don't want to extend. I'd rather truncate and have the song end where I wanted it to. Suno has a habit of repeating verses to fill the entire 4 min generation. It would be nice if suno gave an option to cut the end of a song to have it stop sooner.
Suno has this wavy white noise problem that I wish they'd address before adding new bells and whistles, typically when all the instruments come in, it just sounds like fluctuating white noise which ruins an otherwise good track. Alot of times when the track fades out you can catch a good earfull of that garbage noise fade out in a delay, almost like you're listening to audio of Ocean waves that are turned up way past peak volume. 🤦♂
I use suno to give the basic idea and then I rebuild it using my instruments and additions. The vocal can be then separated also too which you can either tune or drop into another AI site in order to reproduce the vocals with a better singer algorithm.
Hi, thanks for all your tips, but I need to say something, I do not have a clue when the matter is music, I'm an old Brazilian guy that loves Rock and Roll, so I look your videos to see if I can learn something, but about Udio and Suno as I use the free account (has no reason to pay as I'm not musician) Udio I just get 30 seconds and extend 1.05, and in Suno I I get more than one minute yesterday I work with an AI we chat and together (yes because I always use COT when I talk to chatbot) and I get a music letter I send it to Suno I get a 4 minutes song, very good. I hope I can learn something with you, thanks
@@teebodk3917 Hummm, really I do not know, but if you are saying I believe and I will look for how this can be done. Thank you very much for your attention and answer
Suno, has extended the fist generation to up to 4 minutes. Udio says it has, however it has not yet officially. I hope that helps. If you are looking to create more music perhaps subscribe to either one and you will have more control over the content. Hope this helps, thanks for watching!
Being an older guy I generate a lot of 60s/70s/80s sounding stuff. Udio excels at this. Suno is horrible. I tried 70's yacht rock and it created the most generic sounding country track with little variation. It all sounded the same. Udio is all over the place but once you get the sound you want it's easy to extend and get a quality song. I have the PRO plans for both but I keep going back to Udio.
@@j.spencer2635 Udio has only rejected what I already knew they might so to me that's not a flaw, Suno is way more strict to the point as a paying sub there, I haven't done anything since March.
This is a fair point, at the moment audio gives far more control of the overall song. I hope you get some great results from both of them, and thanks for the fantastic comment!
Suno should at least lower their prices if they want to stay competitive and survive. Once consumers pick a favorite it's hard to make them change their mind, and they are getting left quite behind, like Leonardo or Jasper. On the other hand, Udio's ToS suck with that whole having to give them attribution thing. Suno should leverage that along with a more competitive price.
You make some valid points, I do think Udio Changed the terms so you no longer have to give them credit. I will have to double check that however I think I read it in an announcement a while back. Thanks for watching.
I don't like the vocals generated in Suno at all, this is the reason i prefer Udio. However I do like the simplicity of Suno. It's like you stated, Suno and Udio need to make a love child.
Vocals from Udio are really really good! The way how Suno make it are a catastrophe. At least should provide to the user the capability to extract the Stems to process in a better way the vocals with other software. Sometime is even impossible to extrat them directly.
Detailed moderation error is not implemented. I went to try the new features, and got locked out of a generation path with a moderation error. Using original, g-rated lyrics, and still no details on why the moderation was triggered. These are the lyrics in question: Thinking 'bout the people that I meet, The places that I wander with my feet, Life's a puzzle piece by piece, Every day I get cold feet
I'm not sure. It said Thanks, it doesn't clarify that, just says says a prompt will always yield a similar output when a seed is used. Nothing about writing your own seed?? Thanks for watching.
Wow, it almost seems as though Udio has become far worse than it was before the update. Previously the voices were incredibly good, but even more important, the emotional expression, intonation and phrasing was excellent, almost scary. But in the examples in this video, the intonation and phrasing is so basic, robotically following the most basic beat without any emotion or intonation whatsoever. I'm wondering whether this might be due to Udio having been accused of using copyrighted music in its initial data set, where songs would sound almost exactly like famous pop stars ... maybe they had to remove this content from their data set ? Not sure, but all of these examples sound so elementary, generic, basic and uninspiring compared to their prior version.
Logic pro 11 and Audacity have both added that feature play a few chords sing a lala etc and you can generate a song from that ith of course all controls in own hands There are a few more DAW'S editors that say hey have this feature to. So how does one stand for example i like one of those songs you played even yourself now if i know the chords etc that are playing i just play and record myself Now when as a songwriter you copyright a song or whatever you feel is good and maybe worth releasing etc could i legally copyright entire song in my own name or would i have to give co write credit to AI machine whichever used to generate original structure of song or maybe you use entire generated Then again how would anyone know or is their some sort of tracking that the melody could be identified somehow. Also sampling i need a four bar bassline in style of whoever have you now got a copyright free sample or U know legality issues strange business
Great points, chord structure can not be copyrighted, so if you do do that it's cool. It's only the melody that can be copyrighted so as long as you have paid subscription you can gain control over the melody. Hope this helps. Thanks for watching
Suno has better beats and modern instrumentals but udio has amazing vocals but they sound is always 80s or 90s sound but I made some amazing songs with udio but took forever but I love sunos instrumental I like hip hop and udio doesn’t do good with that
Thank you so much for letting me know. I went and bought a pop filter and some denoising apps to help with this. I had no idea. Hope this helps, and I hope you stick around with less mouth noise 😂🙇♂️
6:59 you enter the seed number yourself, that's why you don't see anything in the window. Pick a random number, enter it (and make sure to make a note of it!), then if you like the generated result, you can reuse that seed number on future songs and generations. Of couse the results will not be 100% the same, but many features will be similar and you can potentially use it to create a personal sound/style. Note that it ONLY works in manual mode. I haven't really used it myself yet, but it's said to influence things like panning and voices.
7:23 - the edit function is primarily intended for editing the TITLE of the song and the COVER, not for the lyrics.
15:17 - context length. This has nothing to do with the length of the generations created. It's used for determining how far back in the song the ai should look for context when creating the next generation. One of the biggest problems with Udio used to be, that you couldn't get consistent verses and choruses, because the ai - so to speak - had forgotten what earlier verses/chorus sounded like, so it would more or less create a new melody for every one. The context length is - in plain terms - the memory bank of the ai. So if you set it to max, it will remember the last 2 minutes and take them into consideration when generating the new extension. If you want a change, however, like another mood/style or even just creating a bridge that doesn't sound like previous content, move the slider to a low position, and this will help the ai "think out of the box" and create something new.
The "Lyrics strength" slider is primarily for giving Udio freedom to improvise with background vocals/answers/echos for example, like oohs, ahhs, yeahs, mmmhs and so on. If set on max, you're telling Udio "just do what I say, ok? No messing around with funny ideas", while on min setting, it may decide to completely disregard your lyrics if it "thinks" it can do better.
Clip start is, as you also mention, a sort of "memo" to the ai, about approximately where in the song you're currently working. This means that a setting of 0% will tell the ai, that you're at the very start, so shouldn't begin "mid-sentence" so to speak, while setting it to 90% or above means, that you want an ending, so either a fade or a cold end. In between these values you help determine the "intensity" of the generation - like you'd typically have a more sparse instrumentation near the beginning, with more layers, effects and choirs being added the closer you get to the end.
Thanks for that breakdown! 🙌
What about the quality option? Does it mean we can now get cleaner / crispier outputs?
If so, are we able to bump the quality of our older songs?
Same with seed, are we able to generate seeds for our previous songs so we can keep the same voice and vibe?
Brilliant, the clip start and lyrics had me scratching….. until now. Thank you.
Great information my friend, I was wondering about the seed numbers as they had stated a prompt will always yield a similar output when a seed is used.nothing about creating the seed numbers yourself. I think I was expecting a midjourny seed number. Thanks for watching and the great information
@@ChrisUnlocksAIthe good thing about this it shows wot a minefield these AI UIs are. They can be very messy and very user unfriendly.
Midjourney is the same. I thought I was going mad when I tried searching a prompt in the explore menu, u could never go back , u then had to type it out again. Turns out you gotta click in a tiny area around the image to close it and return to your search???
I'm gonna sign up for this looks fun ;)
Both Udio and Suno are good; Udio has better control over song creation, and Suno writes better songs.
When I use Suno and find a song I like, I keep extending the song while tweaking my lyrics.
I might create 50 tracks, keeping faulty tracks that contain small parts I like. Then I cut out everything I plan to use and compile it into a single file or two using Adobe Audition.
Afterward, I export what I decide to keep into a single track or more, depending on how much I keep. I then separate the stems of the track and import them into a multitrack session in Audition.
Using the multitrack session, I can reuse, replace, or combine vocals and rearrange all the different parts into a song.
For example, a refrain I really liked might become a bridge instead, allowing an even better refrain from one of the alternative song extensions I created in Suno to take its place.
This process enables me to make the best use of all the material I generated in Suno.
Yes! This is exactly how I like to do it, you may go more in depth than I do, but we both tear into the song and try to make it what we want. Send me some links I would love to hear them. Thanks For Watching
@@ChrisUnlocksAI I put a lot of effort into this one, and there's a buildup through the singing as the song progresses. My next step will be to replace everything with real instruments and vocals.
Still in Love with You
fRDkwepb_zg
This is a work in progress, and it contains at least four different verse progressions. The instrumental parts at the beginning and in the middle are from a totally different song.
Purple Eye
JkMSXmSo8d8
I put a lot of effort into the lyrics (English is my second language).
Sledgehammer
lBJG2QmTK8g
For the Seed maybe you have to enter the seed before you create? then it sticks to the song? What they need to do is have separate the vocals from the audio so you can change vocals without harming the music
Yep you need to give it a seed before generation so that it attaches it to the results. '-1' is commonly used as random seed
I would love the option to be able to separate the vocals from the audio.
Perhaps it says a prompt will always yield a similar output when a seed is used.i assumed a seed would be generated like midjourny. However that's not the caes
I like the new update on Suno. It's doing a lot better with my songs as far as structuring them. I like that the censorship seems to have decreased as well. The only thing I don't like is that now that there are 4 min generations the AI seems to think it has to fill that entire time even if it has to replay lyrics I don't want it to. So I've gotten several songs I've liked but haven't kept cause it'll pick a previous verse, replay it and it gets cut off while singing. I know I could extend but I'd prefer being able to do like on Udio and clip the end to just get the song I wanted out of it.
You can always pick the exact spot to extend from in both Suno and Udio, hope this helps, and thanks for watching!!
@@ChrisUnlocksAI Yeah, but sometimes I don't want to extend. I'd rather truncate and have the song end where I wanted it to. Suno has a habit of repeating verses to fill the entire 4 min generation. It would be nice if suno gave an option to cut the end of a song to have it stop sooner.
Great video appreciate you give us info. I'm definitely excited about the new features on Udio. Can't wait to try them out.
Definitely some cool features. I love the direction Udio is going in! Thanks for watching, and the great comment!
Suno has this wavy white noise problem that I wish they'd address before adding new bells and whistles, typically when all the instruments come in, it just sounds like fluctuating white noise which ruins an otherwise good track. Alot of times when the track fades out you can catch a good earfull of that garbage noise fade out in a delay, almost like you're listening to audio of Ocean waves that are turned up way past peak volume. 🤦♂
I agree it's not always perfect. I have gotten some quality tracks it does take a lot of work though. Thanks for sharing and watching.
You create the seed number and save it for future use. Also with the edit button that's to edit the title of the track.
Thanks, it doesn't clarify that, just says says a prompt will always yield a similar output when a seed is used.
Chris makes bangers 🎶🎶🎶💯💯💯💯
Hahaha thanks!!!
I use suno to give the basic idea and then I rebuild it using my instruments and additions.
The vocal can be then separated also too which you can either tune or drop into another AI site in order to reproduce the vocals with a better singer algorithm.
Love this, using ai as a tool is exactly what I am hoping for the future. Thanks for sharing, and watching!
Yes it's getting interesting
I'm pretty pumped!! Thanks for watching!
Thanks, great videos! 🙏🏼
Hi, thanks for all your tips, but I need to say something, I do not have a clue when the matter is music, I'm an old Brazilian guy that loves Rock and Roll, so I look your videos to see if I can learn something, but about Udio and Suno as I use the free account (has no reason to pay as I'm not musician) Udio I just get 30 seconds and extend 1.05, and in Suno I I get more than one minute yesterday I work with an AI we chat and together (yes because I always use COT when I talk to chatbot) and I get a music letter I send it to Suno I get a 4 minutes song, very good. I hope I can learn something with you, thanks
Udio tracks can be extended up to around 15 minutes (could be up to 18, I don't remember exactly) - you probably know, but just in case.
@@teebodk3917 Hummm, really I do not know, but if you are saying I believe and I will look for how this can be done. Thank you very much for your attention and answer
Suno, has extended the fist generation to up to 4 minutes. Udio says it has, however it has not yet officially. I hope that helps. If you are looking to create more music perhaps subscribe to either one and you will have more control over the content. Hope this helps, thanks for watching!
Being an older guy I generate a lot of 60s/70s/80s sounding stuff. Udio excels at this. Suno is horrible. I tried 70's yacht rock and it created the most generic sounding country track with little variation. It all sounded the same. Udio is all over the place but once you get the sound you want it's easy to extend and get a quality song. I have the PRO plans for both but I keep going back to Udio.
Suno was good but they reject my original lyrics far too often
@@jwdurbin Yeah. I've never had Udio reject a lyric.
@@j.spencer2635 Udio has only rejected what I already knew they might so to me that's not a flaw, Suno is way more strict to the point as a paying sub there, I haven't done anything since March.
This is a fair point, at the moment audio gives far more control of the overall song. I hope you get some great results from both of them, and thanks for the fantastic comment!
Suno should at least lower their prices if they want to stay competitive and survive. Once consumers pick a favorite it's hard to make them change their mind, and they are getting left quite behind, like Leonardo or Jasper. On the other hand, Udio's ToS suck with that whole having to give them attribution thing. Suno should leverage that along with a more competitive price.
You make some valid points, I do think Udio Changed the terms so you no longer have to give them credit. I will have to double check that however I think I read it in an announcement a while back. Thanks for watching.
I don't like the vocals generated in Suno at all, this is the reason i prefer Udio. However I do like the simplicity of Suno. It's like you stated, Suno and Udio need to make a love child.
Perhaps one day they will meet in a lonely jazz bar! Thanks for watching
@@ChrisUnlocksAI 😄
Which software do you prefer ???
Suno.
Udio
Too complicated
Very simple but messes up lyrics a lot
Udio better
Vocals from Udio are really really good! The way how Suno make it are a catastrophe. At least should provide to the user the capability to extract the Stems to process in a better way the vocals with other software. Sometime is even impossible to extrat them directly.
Detailed moderation error is not implemented. I went to try the new features, and got locked out of a generation path with a moderation error. Using original, g-rated lyrics, and still no details on why the moderation was triggered.
These are the lyrics in question:
Thinking 'bout the people that I meet,
The places that I wander with my feet,
Life's a puzzle piece by piece,
Every day I get cold feet
Oh wow that must have been a glitch, I tried the same lyrics and didn't have any issues. Hope you get this resolved. Thanks for watching
Very exciting stuff
I agree! Thanks for watching!
Next epic AI battle has begun
Yes!
In the past the seed was the end of the URL. So MAYBE the part past song/
I'm not sure. It said Thanks, it doesn't clarify that, just says says a prompt will always yield a similar output when a seed is used. Nothing about writing your own seed?? Thanks for watching.
Oddly Suno ver 3 beta was the best for me, I made some great songs with that, but after?? just not that good.
I'm
With you. Hopefully it gets better! Thanks for watching!
Wow, it almost seems as though Udio has become far worse than it was before the update. Previously the voices were incredibly good, but even more important, the emotional expression, intonation and phrasing was excellent, almost scary. But in the examples in this video, the intonation and phrasing is so basic, robotically following the most basic beat without any emotion or intonation whatsoever. I'm wondering whether this might be due to Udio having been accused of using copyrighted music in its initial data set, where songs would sound almost exactly like famous pop stars ... maybe they had to remove this content from their data set ? Not sure, but all of these examples sound so elementary, generic, basic and uninspiring compared to their prior version.
These are really great points, I'm not 100% sure, I still feel like Udio's Vocals are far better than Suno. Thanks for watching and the great comment!
Logic pro 11 and Audacity have both added that feature play a few chords sing a lala etc and you can generate a song from that ith of course all controls in own hands There are a few more DAW'S editors that say hey have this feature to. So how does one stand for example i like one of those songs you played even yourself now if i know the chords etc that are playing i just play and record myself Now when as a songwriter you copyright a song or whatever you feel is good and maybe worth releasing etc could i legally copyright entire song in my own name or would i have to give co write credit to AI machine whichever used to generate original structure of song or maybe you use entire generated Then again how would anyone know or is their some sort of tracking that the melody could be identified somehow. Also sampling i need a four bar bassline in style of whoever have you now got a copyright free sample or U know legality issues strange business
Great points, chord structure can not be copyrighted, so if you do do that it's cool. It's only the melody that can be copyrighted so as long as you have paid subscription you can gain control over the melody. Hope this helps. Thanks for watching
Think I'll hold-off with AI music for now, waiting for improvements.
learn the songs , then reproduce it in your DAW
Yep! Can be an awesome tool for learning music, song writing or a tool to help craft better songs. Thanks For Watching
Suno has better beats and modern instrumentals but udio has amazing vocals but they sound is always 80s or 90s sound but I made some amazing songs with udio but took forever but I love sunos instrumental I like hip hop and udio doesn’t do good with that
I'm 💯 on board. I wish Suno and Udio would have a love child. There I said it!!
Not sure if anyone’s told you this but your heavy breathing and swallowing makes it difficult for ocd weirdos like myself to listen to
Thank you so much for letting me know. I went and bought a pop filter and some denoising apps to help with this. I had no idea. Hope this helps, and I hope you stick around with less mouth noise 😂🙇♂️