#๐ฃ๏ธโchill-chat
1 messages ยท Page 10 of 1
that's your top 54 iq response brother
Dude, the mods have been actively discussing it in here too. We keep it to offtopic as per the rules.
@uneven holly #๐ฌโgeneral-chat message not vibing lol
People interested in music generation are gonna be interested in music generation stuff, that would be the same with any community. Suno could ban the word Udio if they wanted, though that wouldn't be a good look
Atleast it's better than you're 20
y'know it takes at least one person to make a disc.
therefore
it is personal ๐
anyways, if you guys think it's okay then you are braindead and there is no telling you otherwise
/me waves
Good to know, thanks!
I think even Suno team would want to know what their users think about a direct competitor. it's useful information for all parties
ao far not impressed personally
The voice got me, and the artist profile page
ofc servers ded so godawful generation time
but like...
Exactly. We're not shitting on suno, we are giving them ideas or what not lmao.
me too and I am weirdly pleased that it's not as good - maybe that's my suno bias wanting our tool to be the best. maybe it's because will.am.i is connected lol Idk
we've been promised artist profile pages at some point so big deal
Has anyone generated anything they think is exceptional with Udio? I need to see some cherrypicking
actually is quite good too lol (like the way the voice moves), now that im listening again
but i get what you're saying overall the track isnt what you want
like, the overall results arent great in this case, but there are elements that work if you get me
Make your music
udio is good and bad... but it is new and has some top guys at the healm so it'll improve in time, but for now toe public will be suitably impressed
maybe peep the staff picks for a good read
I always want the new tools to be better but it's funny to see the hype channels trying to go crazy for every trac regardless of quality
i made this, not exceptional by any means but it was easy enough to produce https://www.udio.com/songs/4Hp5kaqJebkuhMnU7XbQSp
I'd honestly be happy if this pushes Suno's release of Artist Profiles into production sooner
kinda cradle of filth vibes
does udio have profiles? i missed that
Yes
Huh... it doesn't have the overmodulated bass thing going that Suno does
haha definitely does sound like the model is leaning into that
it seems to pull a lot of the style from the lyrics over the actual prompt which is super annoying
like really annoying
Sometimes the lyrics and the music style prompt in Suno will fight
that is a thing
Cradle of Filth sound with Monster Mash lyrics. At least we can all come together in collective disappointment at both tools inability to generate decent lyrics
sometimes i dont mind it, like its like another level of guidance. But it can go both ways
is anyone having an issue with creating on Suno - I can't seem to get into my account. I keep hitting create but nothing is happening
haha the music at the end is actually quite rad. I might be so new to Udio that im not picking up on its cliche tropes that it might do, but parts of that song were quite good
edit: i mean the music, not the lyrics ๐
log in
I'm impressed by some of the things other people have made. But playing around with it it seems like udio is not really generating anything good based on my prompts. I trashed everything I've made so far. It is really straying from the genres I'm telling it. I got like a death metal song when I asked for kpop.
I know I got logged out
oh man... all the lyrics are made with gpt4 except the outro which i let udio make on the fly xD
It's not even letting me do that
i was honestly impressed myself
@twilit needle it's not giving me the option
like they have a really good llm under there
absolute garbage when trying exact same prompts for my finnish rap in suno
sec
I think Claude/PI are a it better than GPT on lyrics. GPT is a pretty derivative writer
almost sonauto level
You click sign up
gpt is awful for lyrics, i have to go through it line by line... its so cliche and boring
they're the same thing (in this case)
okay let me try again because that's what I was looking for @twilit needle
yea, it does seem pretty good, I think I'm coming around to it. Initially i was a bit weary. It's all good competition for Suno
where is the artist profile in udio?
pretty solid comedy https://www.udio.com/songs/naR7Zz2CxT1z2aSPnrhGnd
Make your music
Whoever can give us the ability to save and reproduce a sound (vocals, instrumentation) wins.
underneath the song I think?
okay I see what happened @twilit needle - i'm using my company laptop they blocked the site
that part in the track from 2:06 is so awesome. I love that lead guitar line!!!!! ๐คฃ
I'm on my personal laptop
big oof
ah, just my user account
sunos extend from anywhere is such a great feature, so many of my saved generations are just for the first 5 seconds
I know....I'm working from home @twilit needle - I guess they look at all sites as a threat now one of these companies that no one wants their info LOL
i thought more like saving the specific voice like in sonauto
Let me know if you get to where you can log in ๐
thanks for the assistance have a fab day creating โค๏ธ
Np ๐
@twilit needle I wa able to log into my personal laptop ๐ฅฐ
It's not quite the same since you already have certain elements that are going to be set by that point but that is a good approach to trying to get a consistent sound
Hey guys. Theres a super intense storm out there right now. Super windy and currently flooding..lucky i live on a hill
yea that's a good technique!!
Same here (about the hill part)
I haven't played with Sonauto yet, the quality doesn't seem to be amazing but if it can reproduce a voice, that's a great feature
lol when v4a comes out it's only gonna get worse
I think they called it sonauto to catch our attention
Yeah, it's above pretty much every other song generation tool, Suno just had such a moat prior to Udio and now we have some real competition
I stress tested Udio's 'generate backwards in time' using your clip, prompted with a classic rock-music prompt. So at time 0:33 the song transitions between my prompt and your clip.
- (Not well connected at all, but the new clip is a nice song) https://www.udio.com/songs/dijzc6BnrKoBCD13fUt6M2 (This song is now changed and replaced by something more like #2...!?)
- (still a harsh connection, but the overall vibes are closer fit) https://www.udio.com/songs/aFP5B3G4djQACA358BH3n6
Arguably this is working as intended, since I totally changed the style prompt. Though maybe I would want a smoother transition instead of one that is so abrupt ideally.
Make your music
Make your music
Yeah, love the team, but at the same time. Watch your 6
udio more like udi-slow amirite
what the... this is great !!! instantly like this first track
like the part that comes before the Whisper has very little relevance... but sounds absolutely amazing
I thought I might have heard a penguin but I'm not sure. Sounds like shoveling mayb?
It is interesting
6 mins of 1 udio gen and its still loading lol
I have to imagine Udio predicted the hug of death but there's only so much you can do to deal with it, I guess, at least the site hasn't gone down entirely
huh, what happened lol??! i first heard this awesome late 1960s track, or early 1970s track... now I cant hear that?!?
Just like the v2 days
Wait a minute my clips just changed... weird. Now they are both in the same style I described number 2? Maybe a website bug
prefer to wait six minutes and hear a SOLO vocal for a change
okay so definitely a weird universal glitch
the first thing was crazy arghhh, shouldve downloaded that ๐คฃ
i genuinely wanted to hear it again
Udio finally finished my first track, some reggae and it did consistently get the patois accent. The music itself is alright https://www.udio.com/songs/odtsE7DzQFuXqmtw1QRetQ
Have y'all experimented with music in different languages on Udio? THat's half of what I do with Suno since the lyrics suck
I'm glad you listened to it, I thought I was going crazy for a second
Don't gotta worry about neon lights in Cambodian because I don't know what that word is
if i heard that anywhere I would've looked to see who the track was by so I could listen to it later lol
like it was super convincing to me
Hmm. It doesn't seem to be adhering to my prompts as well as I'd like. I tried a few random things like I've tried in Suno: "glitch music box", "sitar edm", "sitar dubstep", "banjo hip-hop" and it seems to pick one part of the prompt and ignore the other
Sooo
The first track (before the edit) sounded like the sort of song they'd make up for a TV show that had to use non-commercial beetles like music
After playing with Udio, I feel like the voices are more stable - but they don't actually know how to generate stables songs. Suno usually gives me a melody that works with almost any lyrics I feed it. But udio actually sounds like I actually would in real life ๐
Yesterday i tried generating a full song with Suno
It's the AI "That Thing You Do"
everything done on just Suno
k
with only small edits for like , ending part , since most songs at the end just cut out
there was a Beatles feel yea, though it did have some nice progressive vibes or something to it... i hope it magically comes back somehow
asked for kpop and got this https://www.udio.com/songs/ncwAj1A5UMweThFpQ849R2 nothing like kpop and sounds terrible. I think I will just wait a few days and try it out.
I have been liking the vocal quality in Udio. Nothing has really hooked me yet in terms of my enjoyment of the song but there are some areas where the quality is better.
I don't think we have any idea of how Udio works but it might be a load vs inference time thing where there are so many people so the model is cutting corners.
same
https://www.youtube.com/watch?v=KxEbfVmI67o stems are planned according to this vid
Amazing News today! Udio, the remarkable AI Music Generator that we looked at just a few days ago has surfaced, and even better: It is FREE during this open beta!
I had a call with the Udio team after my recent video, and they offered me access, so naturally I decided that I would make a tutorial for you as well!
So this is everything you ne...
so thats cool
definitely not kpop, but it would be a great intro for a different genre
were you using manual mode?
Stems would be a huge deal, then we could potentially layer the best takes of multiple generation, though I guess inconsistent rhythm and BPM is still an issue there
Still, just having an isolated vocal track would be amazing to use it on its own or remove it for a clean instrumental
Instrumental versions of things, now that'd really be somethin'
now i have 2 generations that are just stuck in forever loading, meaning i cant even try to make more because thats the limit i guess
It took me 10+ minutes to generate but it got there. Udio's just hammered right now
Seems like today will be a good day to use Suno, if everyone's flooding Udio instead
I wouldn't be surprised if there were more people using Suno today to compare and contrast with Udio than there would be on normal day.
hmm maybe, though most ppl probably wont compare lol
Perhaps. But then it'll be 50/50 split and with Udio taking a long time to generate... That might slow usage down
actually oh yea good point
Happy thoughts
There'll also be a slurry of people assuming it'll be better for whatever reason
I just think interest in music generation in general is going to see a peak when a new tool like this comes out. SUno is generally pretty quick with generating most of the time so I'm not sure you'll see a lot of gains but maybe.
All of those youtubers mentioning Udio in relation to Suno are going to get people playing with both, though most will focus more energy on the shiny new toy.
yea that is true. I forget that not everyone knows about Suno even
people have gotta make some songs about Suno vs Udio if it isn't happening already
I think that example was ones I did auto generate but I've also tried manual mode.
Most of the people talking about it only have viewers that are already in the AI bubble so I don't think it will be a huge day for Suno but I imagine they're still getting some traffic.
it's a good bubble ๐
so, what do we all think about Udio? What's the gist?
It's different.
today right now it's 100% trash
maybe next week, tomorrow who knows
I do wish there were some more critical thinkers in the AI youtuber world. There are some good channels but most like to hype the new thing regardless of whether it's actually any better.
wonder what the costs will be like
it might inform future costs for Suno, and Suno will inform pricing for it, vice versa
and i think there's also Adobe and Nvidia music generators coming out some point too
so will be interesting how that all affects pricing
summarized the above few dozen messages with chatgpt, in case anyone's interested:
-
Vocal Quality Appreciation: There's a nod towards Udio for having some quality aspects, particularly in the vocals, though it hasn't completely won everyone's heart... yet. Some folks are warming up to it, suggesting there's potential if not an outright win.
-
Technical Teething Issues: The chat hints at possible technical limitations or growing pains for Udio, like load vs. inference time impacted by high user demand. This has led to slow generations and even stuck processes, which seems to frustrate users a bit.
-
Features and Future Possibilities: The mention of "stems" planning (separating tracks like vocals and instruments) is generating excitement. This could allow users to mix and match parts of different generations, enhancing creative freedom despite concerns about consistency.
-
Comparative Curiosity: There's an undercurrent of comparison with Suno, suggesting that users are toggling between the two platforms to gauge which serves their needs better. Udio's newness appears to be its current draw, but the efficiency and output of Suno are still valued.
Matthew Berman is pretty good but Wes Roth and even MattVidPro tend to just force the hype
6 mins and my song still hasn't generated on udio
pretty quality info
good deal
That is a good summary of the recent conversation. From what I've heard I'd agree, not really better or worse clearly, just another option
the stems would be interesting
MattWolfe tends to be okay
how long does the generation take on udio?
good for summaries on the new stuff
๐
perhaps overloaded from all the buzz
I mean, I like all of these channels, I just try to take the hype around a new release with a grain of salt.
I even watch Wes Roth, we need to support him with his struggles in being SHOCKED by everything
snowflake on steroids
haha i dont know who he is, but that sounds funny lol ๐คฃ
It's probably better at certain genre's worse at others. From the limited amount i've seen, it seems better at country for example
tried to have it correct the grammar in one of my EMAILS and it refused because it was related to a criminal matter I was discussing....like hello it can't understand context yet?
Epic Orchestral Music is on another level on Udio +320kbps
Well, that's not just him. Seems like half the AI videos have SHOCKED or STUNNED somewhere in the title
probably works too well and they keep doing it lol
it has crystal clear voices when it comes to country or smooth songs but for example in greek language its cutting down on lyrics for me at least. Also considering how much hype was built upon you would have expected with the people backing it up to have much better infrastructure at release.
nice!! i love the chord progressions its using. Not too cliche, the right amount!!
udio server all but down lol
barely usable now
I'm getting Backend Processing Error for basically every generation
I love it too!
lol i think im ready to pay for Udio now and ive barely even used it, but im getting more convinced
SOME Udio vocals have soul and character. They're mostly high quality but there are a lot of generic vocals on those tracks
suno is much better atm just because you can do 2min songs and if you are familiar with prompting on it it gives fire results. But once udio reaches the point of 2min generations and fixes their server issues i can see it being close. I am gonna get paid sub on both anyway once the time comes
Managed to pull the old version out of an unrefreshed browser tab by screen recording. weird bug
that's wild
god damn @hazy oxide
it even sounds like it's old degraded footage
isn't it, that if we have 100 of these apps that they can't expect to make money with it themselves?
it generates songs that sound like they are ripped right off a record
yea does seem like that at times lol
btw reminder that suno song vocals sound fire when mastered with bandlab or similar tools. The difference is day and night. Probably the final output on udio goes through mastering or something
is udio good
yes!!!!! I swear that chord progression reminds me of a track... it'll come to me
Edit: its a Rush song its reminding me of... which one is it lol
i mean if I heard that on spotify i wouldn't think twice it was from a blues album from back in the day
Ah,so it knows Leadbelly? This is good, Suno isn't great at this folksy vocals
need to have it generate me some old neil young classics ๐
lhttps://youtu.be/NdgrQoZHnNY?si=qdQgVorLy4klHCp9
Delta blues great Eddie "Son" House performs his classic "Death Letter Blues." From the Vestapol DVD "Legends of Country Blues Guitar, Vol. One."
udio works but my renders r stuck haha
How long does udio take to generate ?
I mean wait 5 months and every 7 year old can make deepfakes and mad soundtracks on their parents phone better than wiping their noses...
REally? butis open to everyone?
yea
does udio do edm
I'll try agin then
do u know why udio is stuck
is udio connected with suno?
Besides been stuck do you like it?
no
it looks promising
atm it's not generating anything for me. been 40 mins
Suno vs Udio is like ex-Meta vs ex-Google
Same as Suno yestterday ๐
Have y'all tried doing much with meta tags? It would be interesting to see how its tag adherence compares
yupp like ten minutes. 15 mins ago
only for more info
ok, thanks
The progression reminds me of this haha (I know its different, but thats what I was thinking of)
https://youtu.be/gymNRI66BBk?si=dn2ZxmlsTejAqveq&t=49
https://www.udio.com/songs/5qKstBE7j7qQ9vYCNZAPAK
so far not impress servers are Fried , and completing and extended track is nie impossible.:(
Make your music
yeah im hoping the long generations are just from server issues
quality is great
cant wait for Suno to get this good
yeah quality is great. i left the server for like 40 mins spinning.still zero
now they got some competition ๐ฅ
also i asked that youtuber if he knew any info about their copyright stuff (he had an NDA with them previously!)
Yup, that's definitely Basquiat
yup i thought so too
same here...
sounds pretty good!
my first one is still generating, been 30something mins, but my second finished around 15 mins
It's probably just gonna keep getting worse today until most of the US is asleep
too early to tell, quality on first generation was very good, 2nd generation is taking forever
Been interesting to play with it for longer gens.
https://www.udio.com/songs/14aeFxSTbZ7fBcTqK7pQun
Make your music
Voice is better than suno but still not perfect
yeap, get off the servers goto bed america :0
but its only 2PM!
what that got to do with it, it 7 pm here
i still gotta use up my credits for today i guess back to suno since its taking 15+ mins for generations on Udio LMAO
if you fancy watching 2 hours of content a day about AI, he's your man
https://www.udio.com/songs/2rtaoZ4HwecA93b3PeHHeB definitely sounds like cobain
Those are rookie numbers
i have moments like that, so i will keep him in mind... even if its over the top
i didn't even look at the title of this i just played it and my first thought was.....sounds a lot like cobain. Now I see why ^_^
almost sounds like staley/cobain mash up
Wes's actual content is reasonable, it's just the clickbait titles and hype for whatever the shiny new thing is but that's pretty much all AI youtubers
i see i see
following the competition! ๐ lol
Yeah, I get a lot more Staley from this. I'll need to play around with some grunge.
you going to get banned ๐ by suno ๐ญ
nah the mods are advertising the competitor also so i think they are good.
54 mins of generating and still not done
i generated a song with the prompt "in the style of neil young" please be good please be good
oh damn im only at 5 mins
mine did it in 15 mins
I generated 27 seconds of silence with Suno one time lol
Which mode? I had the same problem with "Instrumental mode"
i checked lyrics bubble and didn't use manual mode
I think Floridians top them
i used custom lyrics
dark-jazz synthwave samples fx modal modal modal synthwave industrial rock-feel rock-feel modal song. Listen and make your own with Suno.
i'm also using hip hop style
listen to this
Ok, I cannot help you with that one. I know that if you try instrumental mode it doesnt work or least didnt work for like 4h
looks like my suno membership ended... im supposed to get 3 free months from their social media stuff. wonder when that kicks in
Epic
shit is just straight using other artists voice
i want my credits back lol
Subo imo still wins on instrumentals
I've been doing some experimentation with extending songs from a point, and then extending the extended versions of them.
Fun stuff.
like it's no contest
email support sooner than later if you don't want to find out after the fact haha
You did it on purpose lol
I smell jail time ๐
you don't get it back
well i didnt actually the style tags were to generate a normal track lol, but even still its all part of random generation
you think people will go to jail over something like this? lol
i was like dammmn
this: https://www.udio.com/songs/vXH8UY8zeKyKrcAJZKLrRM
vs:
https://suno.com/song/3d5e05e0-2d1d-4589-8899-f9fa94b23010
post-anti-folk post-hyper-cello anti-revival, anti-instrumental song. Listen and make your own with Suno.
Been 100% blown away with Udio so far
It's Little Richard meets Otis Redding basically
It does oldies really good.
Suno, def better
it's really just vocals
and quality of the tracks
they might slow it down but it can't be stopped anyway. as cartman put it: your wiener my balls they're all publicdomain now ๐
I prefer Udio on this one
eh, no
Less energetic but good nonetheless
i like the composition on Suno, but the sound on Udio
no contest lol
intense breakbeat climax (dark synth) japanese instruments song. Listen and make your own with Suno.
that's a given
Wdym
I was just being cheeky
ah xd
theyre two different styles too, like they'd suit different situations
but in this case, the udio one isn't really even making a melody
it's aimless
but not in a goodcway
if they were music for a movie or the background to something
I think it works better with lyrics
i could hear it in a movie lol
idk it sounds pretty good to me lmao
just audio quality wise in this example i think the audio is "clearer" for udio but yeah suno is better in everything else
it's cinematic
yea i hear ya haha. I could see some guy wandering aimlessly around a city, its very tragic
did you just copy paste the logs or do you have a script/tool to do that for you?
copy and paste and ask to summarize
i could see this being used like a snippet tool for an audacity-similar program in the future
reminder: suno initial output if put through mastering (you can use bandlab or even paid mastering apps) the mastered output is much better. Udio vocals are more clear probably run through a mastering plugin
Also Suno sometimes makes the vocals a bit autotuned
What we need is "AI music bot arena" feed in a prompt and you get the Suno and Udio output and you pick your favorite
That's a great idea
are so casiotone for the painfully alone
someone make a rap battle between suno and udio lol
They're just great
wow those frequencies are going crazy in my headphones
I did not consent to that my lawyers will be in contact with you shortly
the high ones lol
Also how does Udio generate their lyrics? I found them to be much less generic than the usual LLM lyrics
Suno's songs usually start with "In the world of..." and use the word intertwine a lot
they likely have a fine tune
ask on the udio discord ๐
they def used way more non-produced
and indie
which will have much less "middle-america" overtones
Suno offtopic = Udio discord
(i'm aware there is a proper one too lol)
Udio's lyrics are maybe better than Suno/GPT's but I'd still give the edge to Claude
Lol, look what I did (suno ai)
https://suno.com/song/96661372-da61-4179-817a-4160a497f063
Vocaloid, Miku Hatsune voice, speak fast, hyperspeed angry dubstep, math rock, experimental funk, j-pop song. Listen and make your own with Suno.
PRESENTING SUNO!!!!
i guess thats one way to keep tabs on your competitors
Jesus lmao
in the context of the conversation this track is very funny ๐
Maybe this will push suno to make their voices better
i know, lol
my testing with Udio seems to not be great for vocals, are any of you guys actually getting anything close to V3 quality? only examples i've seen arguably close (or actually much better) are country songs on the featured page, everythign else seems closer to V2
i cant make a final answer since its takes like 20+ mins per generation
Suno vs Udio is kind of like GPT vs Claude. If Suno is convinced V4 will be better than Udio, then they don't mind them generating more customers who will then switch over to Suno V4.
It's very volatile atm, some songs sound like v2, others sound like v5 to me
lol its a funny track
Yeah, it's pretty inconsistent so just keep generating at 1 hour per generation and you'll get a hit eventually
hmm ill keep trying then but it's definitely weird
max song length of 1m30 if I understood correctly, seems unusable too, but then again there's no monetizing yet anyway
this is the way
same
Thats not true tho, there are songs over 3 mins
Some genres Udio seems to really nail that Suno can't like those really folksy blues vocals
I was just going by their faq on how continues work, but I haven't been able to test it yet
Also their old songs are just amazing
i haven't been able to get it to generate any rap whatsoever. with custom lyrics. just generates forever
https://www.udio.com/songs/gajKkTvivm1nJx6oNQjFk1
This one is over 3 minutes
OK, so.... yaaaaa
I guess I should've put more effort into these lyrics:
https://www.udio.com/songs/gANcowiNkqmBvG5T2961yL
https://www.udio.com/songs/52Tz1DuVPdHrghUMrsuBKo
2024 is the year of LLM music I guess
gotta make a song for that art ๐คฃ
Ok welp. To many people are on it.
Just play the official song of Minnesota or whatever
that is so damn scary
right?
and also impressive
I'm going to make some really weird shit once I can generate enough to "dial in"
right now it's basically ally useless
I'm going to see if I can get the AI to summon Cthulu
plz
My song got flagged I am going to prison
depends. are you in a woke country?
this was the sister track
Like Suno, it doesn't seem to do animal noises or things like that. Testing with a standup comedy spoken word thing next
Udio can get ultra weird, yeah
you will die in 7 days
BEcause Suno does notify the worldwide woke police
no
This one gives me the creeps wtf
then you'll probably just be euthaniezd
lol!!!!! wtf ahahahahahaha
Damn I guess the devs sold their soul to devil or something lol
Welp what i've seen of udio is ultra compelling
You ask it for weird shit it goes 'yeeee'
I've had only one piece of music I'm interested in out of it so far
I've only got a single set of generations out of it
But I want to get it wierd
Have to wait till it stabilizes I guess
How come they have a voice chat in their discord but we don't
for me the sound has been worse than suno and the generations are 33 seconds long each
No doubt, but it seems to be willing to mix things that Suno struggles with
or just give you widly experimental stuff, like Yolkhead posted
Suno always tries to sing
Which is good, on the whole, since that's what people really want
I'd like to play with it in a more experimental sense
that'd be nice but basic stuff like orchestral music sounds like it was recorded by a high school band
idk most of what people post here using Suno also sounds like garbage so I'd be a bad judge of song quality since I'm so high on my own supply 
agreed, but suno does really well with classical/orchestral compared to stuff shared on discord. the sound quality is much better
the cherry picked Udio examples I've heard so far had much better quality, especially in vocals, but everything I try to generate is crap
even though I'm generating really basic stuff trying to get a feel for how classical genres map to their training data
Yeah and like I said, I've just got one through and it was vaporwave country and it pretty much nailed what I imagine that would sound like
It is just a couple of 33 second clips but
The blend
This is precisely whats happening to me
Yeah the quality seems good, the vocals seem to have a bit more expression/life to them but that's not useful if I don't find the music interesting
the other thing with Udio atm is its so damn slow to generate, so if you could generate more in a row you might be able to pick the ones you like better
I imagine you raised your pinkie for this
Yeah my Udio stuff sounds like an ameteur recording studio, or like some kid in his basement
like a true afficianado
very much so
yeah i'd like to throw money at them so i can get priority queue like Suno but... they launched free only (wut)
but with a fairly decent-ish recording setup
this is the real thing though
It's so slow to generate that it breaks the flow
I don't know about you guys but typically how it works for me is I'll be messing around then all of a sudden you 'catch' something and end up trying to chase that down
the reverb on anything ive generated is even worse. it's like they're recording in grand central station
hard to do that when you can only generate 1 set at a time and it's many minutes before the next
yea that quick generation is everything ๐ฏ
Some of them maybe
breaks the flow and you wander away
true, is a bit hit and miss
I'm sure it has it's 'feel' like Stable Diffusion or Dall-E or MidJourney
Those all make images but they speak their own language
I imagine it's going to be the same for any other generative AI systems including music
True yeah midjourney definitely does have its distinctness compared to stable diffusion
What was that other s-named music LLM
sunauto?
it is, but at the same time its probably like made by one person hahaha
'# Suno Is Superior
Honestly, elevenlabs should make something like suno, if someone can do it better it would be them
It reminds me of that one image generator when all the image ones popped off
Do you guys remember? The image one?
close!!!
Honestly the only complaint i have about suno really is that choir vocals at times, but i know thats something the devs are working on so once that gets sorted out its golden
It made crap images but they had their charm
can we also note that the cover gens are um...
https://www.udio.com/songs/wgK43ohStJj6SpRhCJ8yDN
accurate
and bruh my udio generation is going on 10 minutes now
I'll check Udio more at like 3 am EST
I'm asleep then lmao
Since when did people around here sleep lol
16 minutes
17 mins and nothing
yea, I mean, I think it's a general issue with some nuddiness. You can lean into it and get some amazing music
but this does open up some possibilities
i love gen AI every time you think you know shit something new pops up
I mean yeah sometimes it does work well exactly how you want it
I hope it's just like this now
but at other times its not desired at all
every week I'm like getting my mind blown
Exactly. It's really cool to be here for this right on the edge of it all
Whats amazing to me is like my first full song on suno is still my best one ive done, like i did a few generations to play aroudn with it but i ended up getting one that was just amazing and nothing ive made since ive liked as much, maybe its a sense of pride in that being my first, but still it just sounded absolutely perfect, nailed the lyrics on its own, everything
time out for 24 hours for not keeping it PG13
I give up for now, it's impossible to experiment with udio at 1 song per half an hour
Yeah mines still going
Make your music
cursed creep polyphonic, weird, paradoxical, complex, dynamic time signatures, backroom, catchy song. Listen and make your own with Suno.
Ok it seems to handle stand-up comedy performance pretty well: https://www.udio.com/songs/6s4M7f5CcRPGjpDPKnMCTg
Impressed that it picked up the standing one leg / covering an eye part
i dont wanna laugh at those jokes ๐คฃ , but the voice clarity and delivery is very good
Now that's really good
Yeah it's meant to be really dumb stuff nobody would actually laugh at
Love his raspy voice
when the peepee poopoo am I right
So Udio stand up comedy
on Suno you might have to do quite a few generations before you get it right. Like im guessing you did this first shot
Suno version for comparison, which turns into a musical at the end: https://suno.com/song/171895f2-32c0-463f-b4e7-94b1885f746e/
standup comedy, live, chatter, laughter song. Listen and make your own with Suno.
I.... uh oh
I blame Eric Andre
and Aqua Teen Hunger Force
Udio is superior for stand up comedy
that is what we have learned today
definitely definitely seems that way
I mean Suno's version gave me plenty of soundboard material such as
even crazier considering you would've done this first take
I dunno man, it seems AI will never understand human comedy
oh crap the AI understands human comedy
FUNNYBOT
that got me good
this is just... ๐จ
Both the Udio and Suno versions were the first generation.
ahh nice to know!!
I'm going to have to write a comedy now
I was running out of credits at the time xD
literally dadaist
V3 will turn into a song but V2 will generally stick with the comedy, is one way they are pretty different:
https://suno.com/song/33ba057f-8fe6-4769-8d0e-c6252e6e0ddc/
standup comedy, live, chatter, laughter song. Listen and make your own with Suno.
Sounds like Foo Fighters combined with Nirvana:
https://www.udio.com/songs/mi3V5kdiRDQYf68DhQXV55
(just copied lyrics from one of my Suno songs)
Has more laughter in it as well. v3 really wants to sing
So my issue when it comes to the style of music im looking for, Udio makes it too hard of rock, too metal sounding
Not bad
Sooo.. what's up with the "liked songs".. did anybody check that with Udio?
yea it really sounds like Nirvana for some reason lol
Make your music
satire aggressive hard rock song. Listen and make your own with Suno.
like the melody from "Comes as you Are"
I keep pressing the play button
competitive intelligence 8)
"Mother I declare I declare myself to be a hat stand!"
32mins to create 33sec output. 1min per second damn kekw
I can't
about the same for me
the weirdness, genre mixing
Do you also see songs that you didn't "like"?
i think they just run the generations through mastering program before the final output. i mean some of them are similar to what i have mastered from suno outputs. I guess v4 on suno could implement something like that potentially for paid subs
the ones ive heard all have good sound quality but all sound very generic to their genre
What is up with that?
i havent seen anyone come up with anything that has character on udio
ITS CALLED ART MOM SO SHUT UP
yea no meme songs just generic
hear it now... its like 1970s
for me it seems to be an endlessly scrollable list of songs I've never even seen
and everything i try to generate myself sounds really bad, like recorded in a concrete box bad
Seed control would be nice. I could experiment with the same seed to see what sort of outcomes I get from certain style tags.
no idea how this bug is even in the system lol
that is not what I mean.. if you click "liked songs".. it is not your liked songs. nor other liked songs
do you remember the tags used to generate this? or the link to the track on the site
We're the same boat... the same concrete box boat... that sunk...
i have no problem with this, my other generations i wasnt as fond of overall, but this is good
ah sorry, i havent looked at that. i dont care for those features
the amazing thing with that is how clear it is, honestly sounds "right"
"liked" songs..
Presses random buttons https://www.udio.com/songs/dr8NMfV7tZSG192vHB4n4F
they even shadow reduced the generations that werer initially 1200 to 600 or at least tahts what it says. 600 monthly generations. unless the other 600 are not mentioned and are slow. but if 600 priority are 30mins long imagine the non priority xd
the clarity given the randomness really intrigues me with Udio. Like its all over the place... yet I can hear exactly what is happening clearly
For me in my own language, I keep hearing the same male voice.. I even recognize the male singer as "Stef Bos" because he has a very distinctive sound and pronunciation..
i havent listened enough to start noticing similar vocals
so thats definitely a thing
4 dutch songs.. and 4 same voice.. have maybe done 20 songs top and 4 in dutch ๐คทโโ๏ธ
haha maybe he's the only singer in the Dutch training data
now that I think about it a few songs have sounded like Nirvana and Pearl Jam
Idk.. but this is super weird
so maybe I have noticed the voice thing, i just havent listened to enough tracks
You can even add artist name in there
yea to have 4/4 is weird
Same with Russian. If you ask for Russian ska, there's a certain iconic voice that often comes out
I am not trying to diss them.. I swear lol
mine keep changing too
I had it earlier when I heard my husband try something and he said. I hear flavor flave and public enemy.
this is so weird
woah yeah what
so random lol!!!!! i wonder what in the world is going on under the hood to even make that happen
mine too
mmmmmmmmmmm
Ok, you win this model
you people are saying like you listen to the same generation later and its changed?
yea, its a weird bug
thats odd
1970s prog or something
https://www.udio.com/songs/ietvC7Mu5L9kVhZTisi8dp
Make your music
Why would it like, revisit the generation?
I reported some extra details of the song changing bug to them, they are clearly working on it, just launch day things
It must be making a whole bunch of them they like, reviewing them somehow?
Why would it revist that prompt?
20 minutes or so later? So weird
Especially unprompted
If we just had quick enough gens
for variations

I get the feeling these would be spectacular
Maybe thats why its taking so long to generate stuff? Its actually generating like 15 different versions and then is getting confused on which to display to the user so there's a lot of processing power getting wasted
bc we're all sharing first gens
Kinda maybe feels like that?
Something along those lines, extra generational load we can't see
Otherwise I have no idea, practically, why generations would change
i do not like that haha
Or maybe it is generating more than 33 seconds and it clips? That doesn't make much sense either
its a blessing and curse
boy I'm confused
yeah its definitely weird
but, they are new, i get it, i see potential there but if im trying to make songs for my gaming playlist im sticking to suno
https://www.udio.com/songs/17NVGh2LZtZrhKKiuiyPVL bruh 33mins+ for this kekw
Make your music
Udio gave backend errors when I tried continuing my song
heh, looking at all the staff pics and 'top tracks' is just like Suno
predictable dreck 
lmao I'm such a prick
anyone already tried things Suno does not have.. like digeridoo and bagpipes?
Lol, my first creation on Udio was an instrumental and they titled it "Neon Labrynth" I was like...here we go.
im getting this now on a new song
edit: working now
Make your music
N E O N
I'm just trying to get it to make weird synth
haha it can do the Suno thing where it sings the prompt
https://www.udio.com/songs/gX76ZydoZnsn1ZxfipNwxy
Make your music
My operatic profanity song has failed to generate multiple times and now has a moderation error
udio sounds better and is free it seems.
have you tried putting that into a prompt?
Hahaha, not just me. Long live Neon!
If I could generate consistently then yeah I might
LLMs love Neon for sure
I bet if they stepped into they real world theyโd be mad that everything wasnโt neon
ayy is udio any good?
maybe
Pretty good
One of the first thing I did is trying to see if it responds to songs that is too graphic.. and figure out quickly it didn't want to generate my song..
give it me straight - suno or udio
Suno
no bias
Oh no bias? Then udio
Suno since I'm familiar and it's quick and you can use it right now
yeah i thought suno would be better
I am compelled by udio though
I am glad they have competition
did u see what paper it is using lol
I want to mess with a new thing
its 3 year old research
This is supposedly "sitar, edm" - https://www.udio.com/songs/888dxiCY9W5qh2abbKKWwu
Make your music
suno is sota whilst they using some old ass stuff
I keep getting errors generating anything on udio now. frustrating.
But who would pay for suno when udio is free? That is the question. And you get commercial usage on the free generations. And they give many free gens per month
Its a problem for Suno
Sounds good but WHAR SITAR
suno devs after udio releases
The percussion at the start bothers me
this wet sloppy hihat noise
The expiermental stuff like this is totally my jam
I am guessing they are going to paywall it pretty quickly for commercial use. Right now they are just drawing people in
Aaah.. I give up for today.. I keeps giving me backend proccessing errors
Not me
Yes its been giving backend processing errors for me as well, they are overwhelmed
well congrats suno team - you survived another day
It sounds like this is only during their beta testing period. I imagine they will limit it to paid users when they release the final version.
I doesn't like me then ๐ญ
Another thunder outside.. and geez itโs close
anyway this new udio thing - doesnt have a 3daisy moderator so who cares
So we will see what their pricing looks like.
Ok, so I think the reason the lyrical gen is better
im getting that too, i tried simplifying lyrics etc... still getting it
is bc its shorter
That is happening to me as well
Make your music
so far suno has better quality of life features, better UI, is much quicker, and seems to follow my prompts more closely. but the sound quality of udio is hard to ignore..
It depends on how it handles prompting and the interaction
this is much much much much much better

their llm is a step above cause it runs the final output through mastering. if you master the tracks on suno its pretty similar
like, I'm just telling you, it straight is. I didn't make this, it was an accident
better UI, really? I mean udios UI is pretty barebones as well, but atleast you can search your generations.
We'll see when I can play with it more
but they nailed slacker
they use a paper from 3 years ago which was a terrible ViT based model
damn
prompt better on suno and you ll get fire here too
The other one tries too hard to make things like, 5th century prose
that might be true, but the results are pretty good, the more i listen the more i'm convinced
(i'm not saying better or worse... just that it does appear to work well)
lol. No response.
thats not because their model i sbetter - its because their dataset is better
well both UIs are pretty bad, but I still prefer suno's create menu
Yeah like don't get hung up on 'stats' since that shit does not matter
The resaerch can be from a butt and from 200 years ago, it's the results that matter
For the record, I love Suno
Saying 'prompt better' is lol
but there's a reason I use my own lyrics
just go see the music on all the dubstep tracks. pretty similar
Moreover, I am subscribed to Suno
you can prompt better on sunno tbh.
I have no dog in this race
does seem that way... however... its hard to tell since generating on this is so slow
its hard to get feedback
It is
I am indeed aware
Also we aren't uh, exactly new to Suno my guy
ok but
It's ok if someone is swooned by a competing music LLM
you make it look like suno wont get v4 anyway ๐
this isn't a console war thing
no, it def will
nah i am gucci anyway just saying that udio at its current state is much better is false thats all
Like all things, I'll use both
Look as soon as georg gives us infilling/inpainting that's a whole different deal
You don't get to tell people their opinions are false though, cmon now you know better
XD
We're now approaching the feature angle
that's not it, the biggest problem with suno is the artifacts and distortion, which is not fixable with mastering. I tried using izotope's tool to fix it even and still doesn't give good results
Since yeah, the model matters less if I can inpaint
suno has much better stuff going on at the current state. saying that udio is better atm is just plain wrong. but if you like it then yea no worries
Suno is clearly worlds ahead of udio in features and UI and ease of use, no doubt
they might be further ahead in certain areas but the quality is pretty unmatched
Yolk is clearly talking about the interface of whatever is in their head and how udio reacts to the text they slap in
Strip allllllll the UI stuff away and get raw into the model interface
idk i ve had fire quality on suno too. llms like to get treated better and prompted. but true the generic outputs on udio all have better quality
There is a lot to be interested in with how this other thing is handling that
Well something is happening
It seems to be in tune to what I'm asking
Love this added it to my subreddit ๐
Just spinning here
Mine is still genning
100 monthlies is pretty good
It's been growing way more than I expected
i need to hurry up and get stuff up haha
I kept thinking it would stop but it doubles every two days. But Idk!
fun fact I am the first person to mention the word WAQF on this server
No rush! Honestly, better to make it good, I just happen to have spent a LOT of time on generative AI broadly, and also lyrics. If I was working from scratch? No freaking way I could have pumped these out that fast.
And maintained quality of any kind
aight yall, I gotta go finish some work crap, my gen is stalled anyway
I'll be back bc this is just too fun, want to see what yall make
https://www.udio.com/songs/skEnkk1wkp78UKLoBkqnem
It just deals so well with weird stuf
oh yeah i suck at lyrics
Anyone's udio working? Mines been trying to generate a track for 3 hours now ๐ฅฒ๐
Maybe it's a super complex track? ๐
yeah the gens don't finish for me either
๐ญ๐ญ
but from what I'm hearing, Suno can close down now.
there's no way they're gonna catch up to this
I just had a track completed. It was crap though. Not due to Udio in itself, just bad RNG
i dont know how udio works i put in my lyrics but somehow it starts somehwere else xD
You should try with lyrics instead /s
๐
It bothers me that I can't name my songs in Udio
you can change the title
yeah you can.. click it open and then choose edit
I don't think that's how competition works
They made a huge thing from nothing, nothing keeps them from making it huger
It made this way creepier than Suno did
Also, the competition is sloooow ๐
I'd be surprised if v4 isn't better, it already seems better in many ways
lol... Nah.. it's good to have competition.. they sound good but also still see a lot of problems too,
And.. Udio songs are copyrighted to Udio.
I've always said that Suno is 80% "there", and I really can't say if Udio is more "there" yet.
Udio has audio quality and better adherence to some genres but it isn't a consistent improvement musically
Udio only beats Suno in spoken voice for stand up comedy
as far as I've seen
not mine but HOLY CRAP
https://www.udio.com/songs/gajKkTvivm1nJx6oNQjFk1
Let alone enough improvement to say Suno could never compete with it, that's just nonsense
Behold and listen to my BAGPIPES.. https://www.udio.com/songs/vfHrYsBcjemFqewCh3KjHa
That is genuinely amazing
dammit i was so excited
LOL
haha those bagpipes totally sound like violins
i thought you finally did it
So it doesn't do sitar, bagpipes, music box or vocaloid it seems
That's a violin???
YEAH ๐
lol what happened
The Udio model doesn't know bagpipes either.. (just like Suno)
ouch
I have gotten Suno to do sitar so that's possible
Digeridoo is next up my list..
nice one!
I seem to have lost this and it's been replaced with a different generation now
For some reason i just feel like the vocals soudn disassociated from the music https://www.udio.com/songs/oEdAssEymrdh4v2xuzr769
Whoever can do throat singing will win the generator war
you're still wanting to get these Helldiver lyrics in a track (its a game?)
i think cause theyre AI lyrics it can lead it to be more generic, you know what you're saying about the vocals being disassociated from the music
the track itself has a pretty cool sound regardless
listening now ๐ง
Bark still wins here, it can be just as creepy sonically but even do it when prompted with just normal lyrics. the call is coming from inside the model! #๐ถโbark-showcase message
Note that the Malevelon Creek song is not the final version, i edited some spoken words from the game over that long instrumental part at the end, but cant really do that with suno, i dont have the mp3 file on my at work currently but its these words https://www.youtube.com/shorts/PfICvgHvqE4?app=desktop
Get ready to dive into the fiery depths of battle with the highly anticipated release of Helldivers 2! In this action-packed sequel, you'll join the elite squad of intergalactic soldiers known as the Helldivers as they fight to protect humanity from the relentless onslaught of alien threats.
Synopsis:
Set in a dystopian future where humanity is...
Yeah what to say without coming off biased.. ๐ฌ I hear it too
cant get the epic amazing grace sound though
what program are you talking about that gives you replicas of Mars etc?
I tried it many times but wouldn't do that for me!
Wow that's good. The devs are ready for the challenge tho ๐
but to be fair.. it is not fully bagpipe.. but close
does this mean the devs are gonna retrain their model?
Made this on Udio
Yeah Suno embraces the competition.. Which is the best approach
the tracks are nice!!! you should try writing lyrics without AI, though the AI ones still work well, I think you'll get even better results if you write them yourself
muddy and kinda blurred almost
almost as if the singer is mumbling
yeah it was in cursive for a bit
oh wait was this in response to Udio or bagpipes ๐คฃ??
i thought it was in response to Udio and that got me really excited ๐ฅ
It is ๐คฃ
awesome!!!
yeah in one of the udio leaks they said "I can't believe we nerfed the suno"
Let's see what they can do
I personally like both
(bad compared to real music - mind you!)
real music ๐ฎ
you mean...
this isn't real ๐
jks jks, i know what you mean
but some AI music is very good
its incredible considering it took a minute to gen
and a random joe just threw some words into a text box
I like the no definition it sounds kind of like you forgot the lyrics but it's good from my ear
yeaa
if you compare any random single generation to real music, sure, but if you compare multiple generations and continuations etc you can get pretty good AI stuff
yeah
maybe but there is papers ive seen that are on par with real music
AI music makes mistakes but you can always make your little tweaks in a DAW
and you can make your own lyrics
It sounds like Mitski
So I just tried "digeridoo" with "instrumental"
https://www.udio.com/songs/hgQEYGngeF7FbaB3vFg7Fo
(i think if you listen carefully there is one in the background)
I had to step away, but how does it handle multiple generations?
may be one of the last times i mention it but if you haven't checked out RIPX for breaking into stems and repairing each stem, you are missing out
How does it fall apart?
Suuuure.. ๐น
even without the didgeridoo its an awesome sounding generation
Reminds me of 70s
๐ฎ .. gonna try some scat
scoobity boo bop
a cappella, Brazilian bossa nova and jazz song. Listen and make your own with Suno.
I can't find it anymore but I saw some mongolian throat singing on Udio
So it can definitely do it
make it tibetan throat singing and im sold
LOL
Udio can master almost every style.
Suno masters every style tho
it doesn't even know a bunch of styles
So, it's like 50/50 for me.
there is a crap ton of instruments and genres it's clueless about still
lol.. Not really.. but neither can suno
Pretty sure suno doesn't understand "Eurobeat"
yeah
doesn't understand krush either
i put eurobeat in the prompt and it made slapping sounds
or jersey club
Eurobeat is fairly electronic music in the 132-170 bmp range with little to no bass or drums, with fairly particular melodies
Based on italo disco from the 80s
Hell it took me forever to get proper synthwave on Suno
guess they took the euro and put parent and made a live audio of a somebody getting slapped by their parent
Suno does darksynth far better than synthwave though
LOL
yeah and not that good with spoken word either
or with like cuphead boss battle music either
Well theres a lot of voices ending up in a choir during songs, its fine for me since I make gospel, but not always ideal
mine is going on 13 minutes
Yeah 15 minutes per gen is a little long
I had one 17 minutes
worth the wait for me tho !!
still love sharing my best attempt at decent voice for comedy
https://suno.com/song/a27cde81-9da7-467f-bf1d-410473553567
standup comedy, chatting, talking, laughing, audience, solo male voice song. Listen and make your own with Suno.
15 minutes to generate a 32 second segment, you don't even know if its good, and you want a 3 minute song, do the math
it's like a full 3 minute set
I am trying things that Suno don't have.. to find out they haven't either.. so worh it.. No.. but good knowledge
the end will make you laugh if you laugh at stupid sounds lol
soloist{dynamtechnicalt math dance-pop post-anti-indietronica, midwest emo lofi, sample anti-fx, post-jumpscare revival} song. Listen and make your own with Suno.
would you believe I lost a regular suno gen
and just found it
Ima see what it looks like in Udio
I think I vaguely remember now that this one needs a better end
I like this
its rich
https://www.udio.com/songs/isCi9usUAcMYqZoYD5KBUG finished my first udio test song. This is country bluegrass.
Make your music
How many days did that take?
That's actually really good
Sounds really clean to me
Nice song, would believe
I have noticed the music off suno sounds kinda tiny in my car and on my bluetooth speaker, gonna have to learn to master that shit
Well you know what they say, competition is good for business
Welp if Udio ever gets some faster gen I'd love to see about the actual quality of the finished product
For now suno has my business, because it actually works
It's their first day, their servers are loaded. It makes sense that their gens are slower
Kinda rhymed
that might be.. but it did not deliver me digeridoo pouts
I prompted 3 times, and got what I prompted for essentially zero times
No celtic, no synthwave, no, lute, no cello, not upbeat
The only thing I got is weird music that doesn't fix with my lyrics being sung
The vocals are great though, but they're not what I'm asking for
So basically I got this (which some of you already heard)
https://www.udio.com/songs/fVYxjxinShhUgkrjg9dPXe
And then this
But its not what I want
problem is at this pace, its impossible to explore
the space
Udio seem to throw out good quality stuff, just not what I'm asking for at all
and figure out its "language"
Nothing really ever gives what you ask for, you have to kind of figure out its own connections
If I ask suno for celtic, I get celtic. If I ask for lute, I get lute. Same with cello
it knows you're not aboriginal ๐
waiting. waiting
It's probably selecting from a small list of identified trainings with that information, so you'll likely need a solid genre before you start specifying extra weights
25m oof
refresh the page
you would have expected they would have pushed an update by now
Well Udio has "eurobeat" as a tag
Lets give this a shot, see if it can make better eurobeat gospel
I think i mentioned this, but while your waiting for udio samples, the super long "udio optimized" tag lists work pretty well as Suno style prompts. a lot of really chunky varied words, probably more than most users tend to use.
Udio optimized?
Where?
Looking forward to Suno v4, because Udio is spanking Sunos butt right now
On some pages it's just the very long list displayed on screen but on some other pages the list is in the metadata as "optimized prompt", not sure why they are different. But for example the "song like the beatles" tag list was "Male vocalist, Merseybeat, Pop rock, Beat music, Rock, Rock & roll, British rhythm & blues, Love, Melodic, Romantic, Energetic, Happy, Bittersweet, Uplifting, Playful, Longing, Warm, British invasion," as the actual Udio tags
