It will only be a matter of months before there are infinite music ‘radio’ stations where ppl can listen to ‘all ai music, all new, all the time’ based on their exact preference mix.
I want a radio station that plays music about me and my life all day long! All generes, it doesn't matter. I'm fascinated by music made for me. If I could somehow allow it to see my emails and stuff, it'd be amazing.....
@@erikprestonTV vs human/corporate generated slop? lol Yeah, the backstreetboys, and the like were genuinely great performances by real artistic people....
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.
That is exactly my impression. If you are happy with whatever random stuff Udio comes up with, maybe you will like Udio better as it may have slightly better sound quality, but when it comes materializing a certain musical idea, I had no luck with Udio, while I can still coax Suno to come up with what I had in my mind, after trying often enough.
Udio is generating lyrics, music, vocals, instrumentals... and PRODUCTION. I've managed to get Polkas about crossfit (not a cult), Eminem-style raps about Babylon 5... and more, and I'm just constantly amazed. Also, if your prompt is just "cat" - well, I think Udio put out a song that sounds EXACTLY like was created by a cat. My three kitties agreed.
Just wanted to thank you for the videos. It's been a while since I have done. Played around with Udio for about ½ a day yesterday and ended up with an EP. I'm not huge into music, but being able to make what actually moves me... Been playing it on repeat enough that I might end up giving my Saturday to making more. Edit: Since having posted this I did spend that Saturday and a couple other days to end up making an album as well (Called Triguṇa on my channel if you want to have a listen). I'm even more satisfied with it than the EP! 😄
This is great. Me too. I don't see anything wrong with being able to make your own music. Seems to me that only people out to sell something are upset about it. But I think deep down everyone knows that nobody owns art and all talent is God given which doesn't mean any of us have the right to profit off beauty. We make things to make us smile or even better, somebody else smile.
@@zatoichi1 It'll take time for the old economic mentality to catch up to tech. In that regard we're well past the technological event horizon heading fast toward the singularity.
@@zatoichi1 All talent isn't God given to you, since you have to use an automated process to create the a whole song. It is not about selling something, but the value of effort, dedication, discipline and imagination that is being comodified and industrialized into a product that you think no one has an economic interest in. You are feeding the rich that have no sense of beauty in art and crafts. They only want your money and gave you a toy to play with.
At the 30sec mark, Udio is king. The frustrations come after when you try to link more parts. Suno allows you to determine the time cut-offs before linking to next clip. Suno also does better at filling in transitions between clips while Udio seems to more literal so if goes silent at end of first clip, the silence remains while Suno will fill that in.
Matt, PLEASE use a low cut filter on your mic. I keep thinking the neighbours are banging watching on my home cinema setup. You are "popping" the mic alot and the bass explosions are very noticeable on good quality speakers.
No one will ever produce a masterpiece on these ai softwares . Real People produce enough crap music already. These things will just flood the market even quicker. Aided by people like you. Enjoy.
As a musician I think Suno is vastly superior to all the others when it comes to its ability to actually make music that sounds "good" as a song. Vocals and how the AI phrases it is easily best on Suno. The potential for really nice flow in the vocals, even with lines that are totally unsuited for songs, is extremely impressive in Suno. Any lyric I write it will just churn out catchy stuff. Feels like magic.
Hi Jimmy! I'm a musician too, and I'm curious to know how other musicians are dealing with AI tecnology, for what purposes do you use Suno (or any other system)...
I love love love this app [Udio]. I made an ambient downtempo track that was almost spot on like something Carbon Based Lifeforms would produce, a prog rock Song that was very Pink Floyd/Early Genesis and an Icelandic Post-Rock piece that was very much like Bjork Or Sigur Ros. I'm fully impressed.
The cat song comparison is the reason why Suno V4 or V5 will dethrone Udio, and why Udio probably will never really get past it without changing their entire method. Suno builds the music around the lyrics, each word and even the sentences. When generating aggressive rap/hiphop for example, you can hear how the words literally become part of the music, like the words are leaning and dropping on the beats. Udio has better sound quality as of now, it has better quality of life for controlling intros and outros etc, but Suno can and will get all of those too.
I feel like suno can do any genre, any accent, any language, any instrument very well. It can even do things like sing in russian with an indonesian accent, japanese with a spanish accent etc. It's good at mixing up genres and making really good music with them, even if they're completely different genres (for example I can write in the genres something like: uk grime, balkan folk, arabic style flute + drums, 8-bit electronic, soviet war anthem, traditional chinese style, balalaika solo) and it literally does exactly that, with elements from all the things I wrote. And I specifically wrote the weirdest genres to try and confuse/break it, but it actually makes really good and unique stuff every time. I don't think udio is able to do that at all. It's good at making good rock/pop songs and the voices seem more emotional and real, but you can't really get crazily creative with it like you can with suno
I've been playing with it since lastnight and the one thing that really annoys me is the hard limit that I didn't know it had and then had to completely rewrite the song to fit within the 4 minutes 20 seconds limit otherwise I freaking love it! ❤
@michellemonet4358 that's what I encountered a couple days ago my song was exactly 4:22 seconds long and when I tried to extend I got a message saying it was too long or something to that effect
Use it friend, don't get left behind. AI music can't replace a live musician yet, and if we ever reach that point as a species, we all are culturally dead.
What will really impress us is if an AI program can reproduce the voice and singing style of a good opera singer, including a medium-rate, full-length-of-the-note, operatic singing vibrato. Maybe showcase this in a future RU-vid video for both a male voice and a female voice. Just show us which AI software you used and the prompts, etc. you specified to get a good, realistic performance.
Is there any way to use ANY app to upload my own music and lyrics, and have it generate ONLY a vocal track that fits the music? Basically cause I can't sing, lol.
Thanks for this, Matt! I'm just getting back into composing (New Age/Neo-Classical) after many years and am looking forward to integrating AI music tools with that! 👍✌
Suno's audio tracks are laden with distortion, and the lyrics don't flow together with the song as well as music from Udio does. Udio is the clear choice here.
Great Channel and great show as always. When I first saw Udio, I thought it was a play on audio, without the A. I would have never thought it was STudio without the ST...
Matt, thanks for the reply. I'm a musician, and this stuff fascinates me. I live in downtown Toronto and pretty much walk everywhere I need to go. There is certainly a vibrant artistic community here, and they are all petrified about AI. I live near a number of large art galleries, including AGO, which is the Art Gallery of Ontario. There have been a few protests about AI by various subsections of the artistic community, fearing they will lose their livelihood. Some of them make a half-hearted attempt to try to convince people that AI art won't be as good as human art, but I'm not sure that's the case. I tried to have a conversation with one of the protesters to see if I could convince any of them to take my approach, which is to use it, understand it, learn it better than anybody else and be better at it than anyone else. They were not impressed. I will absolutely use AI to make my music better, just as I used all of the latest and greatest computer and internet technology to make it better. All this, from a guy who is retired and in his mid-60s. Again, this stuff fascinates me.
One thing I look for (and ChatGPT does this too) is a song including "no regrets" at the end of a line. It likes to rhyme that with "set", "met", "bet", and "yet." It's a little funny to see how often those pop up, but it's something I look for now and tell AI to do some rewriting when it invariably pops up.
What a joy to have AI replace a lot of the boring parts of making music. For me making music has always been about finding those nuggets by listening for something cool. Same process with AI music creation but way faster. I'm having a blast right now!
A collaboration between your channel @mreflow and that of the RU-vidr @DotCSV would be great. Just Yesterday He Showed Us Some Of This. Greetings Matt and Really Both Channels Would Benefit.
I still want to be able to create music, not just generate AI slop that's not in any kind of proper key. I'd like to be able to input the melodies for sections of a song.
@@erikprestonTV maybe though. Example: “create a groove with tabla drums” instead of using Software like EZ Drummer for example. “Create a walking bass groove in the key of E, in the style of Jaco Pastorius” and capture that into a stem on your bass track…
Just checked out the Udio website and all the songs on there and damn my mind is blown. AI fully replacing human creativity is amazing. Truly impressive what AI can create nowadays.
@@vectoralphaSec is there anything you would like to do that will more than likely be taken over by a program??? I dont understand your desire to remove human effort from everything I completely understanding greatly reducing physical and mental labor which is done for $ to survive however I dont understand eliminating people from creation aside from a fucking prompt...
great coverage Matt, But with all the Ai music generators I've never heard such polished crap in my life. It may be slick but is still pop shlock. I'll pass.
I've made quite a few Suno AI songs lately, and it blows my mind. Trying to perfect the prompts to fit the lyrics can be a pain in the a**, but its really fun to play around with.
the problem I find with Udio is that it doesnt follow the lyrics you write in an orderly manner while Suno does the job better by sticking with the order of the lyrics most of the time yet perform those lyrics more in more varieties than Udio and also better use of sound variations from genres. In this case, you get more positive results and more catchy songs and that actually pop from Suno V3 compared to Udio.
Hey, Matt Udio functions similarly to Sonauto when entering an artist's name. It automatically adjusts the prompt to reflect the artist's genre, style, etc. In rare instances, it even modifies the lead vocals to sound similar to the artist I've mentioned in the prompt...
It's that, I've been successfully amazed and shocked before by language models! You waited your artistic models, and a music/song AI is a typical vector for artist. I'm more of an analyst / functionalist thinker! Language models easily do so! We all eventually do, albeit differently through adopting a persona! of what we later would be identifie by as, develop through and act, a persona! ❤ 😊
I had a listen to a few udios on the site, it is rather good. My only complaint would be how they seem to have a habit of cramming lyrics.. no singer would ever manage or want to put that many words together.. there's simply no time to draw breath. However, I did note that the suno example you played actually had the singer drawing a breath, which was a nice touch.
Really fun presentation .I think it’s a good thing that a variety of Ai music modules exist ….it is a completion but also not so in this area ….i personally prefer the udio platform…it is superb for creating orchestral music with operatic vocals and choir sections..also piano pieces ..the best for these genera’s I’m my opinion ..I have tried a few.
at first I tried it and a few generations in I felt like it was pretty subpar to suno V3 by a LOT.. but after trying better prompt, coming up with lyrics myself and giving it a few shots.. MAN this is impressive.. I ended up with some tracks that I would ACTUALLY listen to (and I am) in my playlist.. that's absolutely insane it's not good at everything, far from it, but you can manage to make a 2:30/3min song with impressive consistancy, giving you try for a bit) and it's mind blowing to think where this will be in like 6 months or a year.. Suno is dead in my eyes!
I can understand people raving about the superior vocal-quality of Udio, but creating a full song in Udio that sounds nice and catchy... that task eludes me. However, I've done it many times in Suno. All Suno needs to do is improve vocal quality and the overall audio quality (plus some music-theory tweaks). Both systems could use a much better language model for the lyrics too.
You turned me onto Suno, and I've been making songs with my 6yo for us to listen to in the car. Time to play with Udio! Edit and a few hours later: I've not yet made a single thing that's remotely decent. Suno is 1000x better. Yes, Udio can sometimes come up with something that sounds more "real", but nothing it catchy or close to anything I prompt. The very first song I made in Suno was a banger and my son and I listen to it daily in the car. In Udio it's taken 20, literally 20, tries before I got something that even came close to what I wanted. But it was lackluster and boring so it got deleted as well. I wonder if because of all the influx of visitors if they just made it dumb to be able to handle all the processing power.
The pianola, the mellotron, the drum machine, the sequencer, the sampler, the pitch corrector, and now Udio. It is amazingly awesome technology but also very very sad for the human condition.
Why try and create new crappy pop music? Let the AI try to create some new incredible realistic sounding classical music? at least then we can see what it is capable of doing.
Where Udio falls behind Suno is with modern contemporary generations. E.g. Pop and dance. I’ve found, at the minute Udio is only good at ‘dated’ styles/genres. If they solved this I’d cancel my suno subscription 😊