#🗣️┃chill-chat
1 messages · Page 12 of 1
yeah it does a good job at melodies
A noticed some UI updates to Suno when I logged in (it's been a few days) that look suspiciously similar to Udio. I freaking love it!
i feel like it understands genres better sometimes
Have you found that you have to prompt differently between the two?
honestly Ive mostly just been using band names for Udio
You just blew my freaking mind. I did not know you could do that!
not always perfect but usually comes close enough to the style i want
Have you tried specific songs from a band, or is it just the band name?
not yet but ill try soon
Funny... because those updates where pre Udio.. just rolled out later due to some problems
Its harder for me to get the sound of music and vocals i want on Udio
That is pretty funny
I'm not musically inclined enough to know what words to use to describe it
I have the same experience, but I am going to try JimJams approach to see if it makes a difference
They were literally rolled out the day Udio launched.. so yeah 😉
I love it, it is very clean.
My position is: I vote with my money and right now I am voting for Suno but that is subject to change. Competition in this space can only do good.
well you cant give Udio money yet anyway
Hahaha, yeah fair point!
if Udio gave us longer generations id be more willing to pay
It's off topic.. 🤷♀️.. it's fine
i dont see myself paying for 30 second gens
yeah i was just joking around lol
Yeah that 30 seconds, is kinda of a strange move.
i think the 30 seconds thing is just because of their current model
Heck, even Stable Audio was 3-minutes. Granted is was 3-minutes of nonsense.
Nice chatting with ya JimJam, thanks for the tips on generations!
yea, I tried that out the other day
wow it was horrible
like, why even release
oh man yeah i forgot i tested that out LMAO it was not good
I guess the band names thing is also my issue cause like... i never like entire bands, like take Skillet for example, they have over 100 songs but i only maybe like 4 of those thats in my playlist, thats true for any band i might like 3-4 songs from the whole band, so using the band name on Udio gets me the sound of the larger majority of their music
this is my main issue with udio is it really does seem, at least to me, to directly emulate particular bands
and in this sense, ironically, I think it's less creative
but the sound quality is better
substantially so
however, I'm guessing this is part of why it's only in 30 second sections
Yesterday i was finding the sound quality worse, today its ok
it also loses voice fast
i wish I could just sandwich Suno and Udio together
so you can't have a long instrumental
probably soon
I mean, suno is still early on
and machine learning stuff is moving fast
Loses voice? as in the voice sounds like one person when you start but by the end of the song its a different singer?
def, people forget how new Suno is
they must have music2music in the works
yes
Udio, i guess, has confirmed Stems are coming
like, if you have a long break
Ahh that would definitely be a deal breaker for me
according to some of the youtubers
like this one, I wanted to have a closing bit
to end it off
but it changes voice
bc of a small guitar solo
I believe their context is literally 30 aeconds
yeah i noticed if you do an instrumental part in the middle it totally loses the voice afterwards
so this would explain why my, upon longer use, sense that the melodies are less coherent, is the case
Suno can pull some brilliant stuff together
if you set it up right
Ive just foudn it easier to get something i like in a couple generations on Suno, like it would be acceptable, i just like to keep going and see what i can get, but Udio i keep going because i dont like it at all
yeah id like it better if everything was generated separately
i must be doing something wrong cause i just cannot get anything good out of it
Udio is extremely hit or miss. The misses are huge though.
I like the way that Suno constructs songs better even though the quality is worse. With Udio it seems like the vocals dont match the instrumental of the songs sometimes, like it is not layered correctly.
Yes the vocals are clear, the audio quality is good, but ot me that voice and singing is just awful
YES that is exactly what i hear, the vocals dont fit the music most times
satire aggressive hard rock song. Listen and make your own with Suno.
The cadence in each verse, it came out great
wow i told Udio to use automated lyrics and it got flagged with a moderation error
you just flagged your own AI
but the audio quality isnt that big of a thing to me, i've never been that big on audio quality, like clearly i know when something sounds like crap but i listen to music out of regular computer speakers, i dont use fancy audio equipment or headphones with fancy features to hear every nuance in the music
Exactly. When you do enough side by side testing, this starts to show. Suno has almost magical way of composing around the words, when Udio just fiffs it there in stereo
Yes Suno did a great job of this in this song of mine, listen in the first verse when he says "We fought the fight"
https://suno.com/song/12365596-da12-4bea-afc1-f344b1d714ab
satire aggressive hard rock song. Listen and make your own with Suno.
track not found
probably one of them i deleted cause i didnt like it
well nicely done because i was actually quite interested
although the results it produces can sound good, theres a bit of a learning curve probably
Yeah, Suno makes the lyrics lean on and sometimes drop on the bass, drums etc, making it sound like it was meant that way
My bad, i guess there's no trash bin to restore from
its alright, dw about it
I mean its the same lyrics as all the others ive done, the singing just sounds... bad
Make your music
Has anyone tried ACE for vocals, or, how do you fix your vocals?
ACE?
ahh nah i havent
I let jesus i mean suno take the wheel. Bit mastering with bandlabs and audacity
udio must have a hell of a pr team
They don't show it do they?
I think thats why in my case its a little bit more difficult to get stuff in Udio, it wants some very specific musical descriptions and i have no experience in making music or have any musical inclination or talent besides being able to tap a pencil on rhythm on my desk, if i like a song i like a song, so using other software after downloading the suno file isnt really happening unless its minor adjustments in audacity
Oh ok.
$22/mo
I asked bing
I'm just really disappointed with the suno update and it's putting out bad music ever since on my end anyhow, I noticed the instrumentals are still good though but the vocals they all sound the same and melodies too sometimes
Another i did, good example of how the music fades out at 0:57 and then the drop hits
And the vocal layering is ick
you mean the update the other day or since v3 came out?
hmm i havent had much problem really, i mean the stuff im generating sounds good just i keep going to see if i can get something better
I will have to give it more time. How do you get the vocals to not layer but just be one person singing without the layering effect?
Can I hear some of your stuff?
I made a funny rant song this morning anime style. Lol
Oh cool thanks!
I always get that kinda.... slight choir sound to the vocals
but sometimes its not so bad, i think because im doing mostly rock stuff, the "static" in the voice kinda fits
or sounds like a raspy rock voice
but you can try "solo vocals" or "singer-songwriter" stuff like that, to try to indicate to the AI "I WANT ONE PERSON SINGING" but it doenst always help
It's been happening since v3, it was fine in V3a. Doesn't seem to affect all genres though
I mostly notice it on the Chorus, maybe it's confused about Chorus and Choir haha
Thank you for the tips, Xeno98
reputabletoad, lololol yep
Here's my AI rant, I was in a mood I guess. It's ironic because it's a rant against AI doing creative things, (which I use it for and enjoy) and a complaint that it's not doing the stuff I don't like doing like chores. https://suno.com/song/b834a5b5-32a6-481d-88dc-d34b9b743a54
Anime rant, emo, female vocals, candy song. Listen and make your own with Suno.
I wonder if one used [Refrain] instead if you would get a better result. Assuming the tokenization is wonky idk
I tried, it sometimes helps a bit but when it does it often lacks the energy of a proper chorus
Yeah a lot of alchemy here
Yes ive had that same thought lol
Another tip, hyphenate some things, where you have "A I" and its sounding like "uh eye" try "A-I"
Oh thank you!
Like in mine when i typed helldivers i kept getting hell-div-ers, div pronounced like the word division, so hyphenating it got it correct, same with Malevelon creek it kept saying Male-velon creek so i had to type Muh-level-on creek
Okay, I had a song where they lyric was 'necklace' and they sing it like 'naked lace' so bizarre
Interesting, the only thing I've run into is ambiguous pronunciation, wound as in wound like a string or wound as in a cut or abrasion.
Hmm to me this singing is quite expressive. There's qualities in that which are good imo
Udio definitely feels harder to steer for sure though
Sometimes its compositions are looser in ways that are intriguing and other times looser in ways that are meandering 😆
and since you can't choose where to continue from you have to leave all those parts in the track lol
I think starting with the "intro" or "start", as Suno does, makes better sense too. The way Udio starts from a section probably works against it
So, is there a visual tutorial about inserting a drafted paragraph in the middle of the lyrics as a modified version when creating a song in suno custom mode?
At the moment you can only continue forwards (you can't insert). When continuing you can start the lyrics at the point it was up to. Sometimes you might need to put a line earlier (and sometimes I put the entire section again).
Other people might have better techniques for this 😆
what update?
nah u can add anymoment intro and choose if its instrumental or not its way better, but udio has 1 problem u just cant continue from point that u want, but the quality is just killing suno
I find the suddenly proliferated idea that there was a stealth update to the model kind of odd
in terms of timing
Hmm i think I get what you're saying. Somebody else might have better advice. Sometimes when I want to force it to go to a verse I will use
[Continue]
[Verse]
Start from here
(this is in a blank lyrics box)
I already warned you not to double post, and even removed one, now to find this again.. I answered you in #📚┃suno-school
You've been timed out for 1 day for double posting, and ignoring my warning @fallow vale
If you actually want to add a part to the middle of the music you can't do this (yet)
I'm sure part of it just subjective and what i consider to be good singing but i just dont like most of the vocals i get out of udio
what I mean is that there isn't a way to start from an "Intro" and the "Intro" or "Start" often sets the mood of the music
i don't always like parts of them, but there are parts I do like. That said, you do feel way more in control with Suno lol
its random with every generation so ?
and it is better if u can generate intro as many times as u want with suno u just cant
yea, i think generating outwardly from the intro seems to work well. Something about continuing onwards from the start of the track (to begin with), makes it easier to get the type of track you want... I could be wrong. This is how it feels to me for some reason.
eventually you'll be able to use the "in-painting" feature to go over selections of music, when they add that feature
what I want from suno is stick to the freaking prompt, like i can put 10x violin and... there is none lol or skipping words 😦
well, Udio seems to stick to the prompt even less 😛
well i fell opposite everything that i put was out heh 😅
haha, maybe in some situations it can be better, im still trying Udio out
(it just hallucinated random words to that haha, but it does sound cool)
i like this... but I would've preferred if it was starting from the start of the music. Then I can work out what im doing better
but udio is in beta and well it is BIG, suno is in V3 and ekhm well i hope they will get upgrade soon!
i'm sure Suno are constantly working on things 🔥
well if not they will lose 😦
https://www.udio.com/songs/1Me335YgtsXU5AvQh52zUN
This is the most impressed I've ever been at any music generator ever. Take a listen to 1:40 minute mark, tell me if this doesn't sound like an absolute classic
Make your music
yea, I've been able to, thank to being on the opposite side of the planet, use the app pretty extensively
it's like Hey Jude 2 🤣 (that part)
and thus far I can do basically everything in suno that I can do in the new one
(whose name for some reason I am consistently unable to remember)
but not the other way
certain things are just not replicable, because it veers too hard towards known artists
and especially for instrumental
neo-anti-cello jewish niggun, hyper-anti-piano, experimental Avant-garde, darkly creeping song. Listen and make your own with Suno.
I literally cannot hey anything close to this
that part been stuck in my head for quite a while, it goes so hard Jesus
I think thats my problem with Udio it replicates a lot of popular stuff that i personally dont like
sort of udio jailbreak, the trick is to write the text in another language but make sure that it sounds phonetically the same
very moving !!! udio can kind of veer out of control in ways suno doesnt
agreed
It's beautiful.
An orchestra? I'm pretty sure Udio can do that
yea that's an awesome piece, very soundtracky
one example where I really couldn't get up to the same fidelity in Suno is
https://suno.com/song/44811e63-4a22-41fc-9605-7afc31929767
vs
acoustic neo-delta blues revival, black 1940s anti-solo soloist song. Listen and make your own with Suno.
it's apples to oranges here
but the sound is so rich
it's not comparable
What do folks use to come with the words for how to describe the music to Suno? Is there a good source where I can learn what different style combinations sound like or ideas of how to finesse a style like that one?
however, it's so niche
and the suno progression IS stronger
in particular, his voice literally becomes the guitar
at one point in a way that is so cool
it'll be the same for me now lol
and something you won't get out of the other
it's not an orchestra
it's the melody. It's just good.
experience with other LLMs
not other, as it isn't an LLM
but I mean, other models
I get that sometimes, othertimes it does come across fairly creative. Like in certain styles it comes out more original (to me) than Suno, but then it also screws up things in ways Suno wouldnt, and because you cant continue from a time stamp it limits you a lot
which helps me know where to research
I fond ratemusic to be a good start point
also, I listen to music like, idk, 8 hours a day
I think your name should be anti-Yolkhead haha
believe it or not that has been a prompt
Math-bit,anti-Yolkhead,anti-🌳 song. Listen and make your own with Suno.
this one takes a bit to hit the dirty notes
hahaha loool
it's like, one particular set that occurs around 40 seconds I think
where it gets low
that makes this one interesting to me
and then the progression
it's basically one huge buildup
for 2 minutes
but yea, udio just can't touch the weirdly intelligent method suno makes music
like, if you just make some options, it'll get it "right" at some point
like, if you know how to listen
and are willing go roll with it
Suno is killling dubstep 🤩
oh absolutely
easy
but I mean, not so much killing
as it is making more of all styles
anything purely sound based will be easiest
lyrical stuff, for now, is harder
but I see no reason why it will remain this way
but for the time being, having access to quality material is like having gold
Yeah
if you're making rock, indie, alt, metal, anything
I try to make Deathcore (Instrumental) but it allways crumble in a noisy mess... with udio however it's just clean
yea, this is an issue
you can try adding vocal solo or soloist or anti-male solist, post-male soloist
and see
solo vocalist worked here
also tags like raw, live vocals
anti-pop-punk lofi cassette, anti-sound anti-fx, anti-recording style, trap solo vocalist song. Listen and make your own with Suno.
however, note it's a woman
which is another option
as they don't have the issue
in general
it's almost always male voices that degrade into chorus
I've noticed this yeah, but being that im doing hard rock, it tends to work fine cause it just sounds like a bit of rasp in the voice
it also can just sound like feedback
i never get the same chorus out from the song when i'm extending Oo
I'm fairly sure it has to do with instrument pollution
like, something about how it does voices
is similar to, for example, the distorted guitar
but who knows
(Patsuno)
Maybe it'S because voices matching the frequenzy of the other instruments
could be. I really am not sure. I'm guessing basically it has to do with scaling. But in any case, the good thing is the next model will have better fidelity regardless. So in the meantime, just lean into it. One day, the stylistic decisions made to work around these issues may be an actual sound people like.
like all good music eras
half the style comes from limitation
you only even notice it when you're on our end
like, I've had 3 different people now ask me when I learned how to play guitar
And I'm talking people like 6 years younger than me
one of them. I have no idea how on earth they thought I oculd produce 3 albums in 2 months
but point is, it passes
I mean suno is such an amazing ai playground, it has much more potential then this, like i said in #💡┃suggestions-web for example create music and lyrics but without using voices for own singing for example.
But first they really need to fix the noisy issue and the non repeatadly Chorus. (it's bothering really, to get a consistent Chorus with a slight varaition out there)...
I Mean Music is defined by patterns, why you can't create first a pattern then you can tweak or generate the pattern in the pattern itself as ai for example? So you are able to copy the chorus and then click on (generate alternative with a slider where you can choose how different the chorus sound for example) 😄
THAT would be an industry killer :3
I mean i think we are still in a pretty infant stage of AI in general, so i mean most of what i have done with Suno has been above my expectations
Thats right... it's just the start
True, I was going to abandon it but with some tweaking on a DAW maybe I can do something with the voice problem
With a mp3? I doubt it...
oh hot damn
nice to be back
Can Suno take multiple AI generated tracks and then have the AI generate a new track based on the data of both? Like how Midjourney can take different images and come out with a new one based on the data of all of them?
on suno for a bit
How do you keep it from skipping lyrics? Sometime when I extend it skips who blocks of lyrics and jumps down or starts in the middle of a verse
are you specifying the time to continue from or just letting it pick the very last second by default?
So I tried udio, the only nit pick I have is that the 33 seconds clips are a bit problematic for certain tracks, trying to extend a clip right now and all of them so far have had blending problems.
The second
Is that a problem?
I got a bit intrigued by Udio, so I’m wondering if anyone here knows if we can use it yet?
I still think that Suno is probably just as easy. The main difference seems to be that people can apparently use artist names in prompts on Udio, which we can’t on Suno.
Sometimes that very last second of a part may be a near silent note
the audio quality is definitey good on udio, but I need more than just a 33 second clip.
What do you mean? You can use it now in beta
Yeah its definitely hard to work with that
Ah okay
I don't mind the 33 second clip thing at all, not really. The bigger issue is coherence, it forgets musical style, sometimes voices but just up and forgets the chorus. Might be due to the 33 second thing
I mean even Udio isnt totally allowing "artist names" its just pulling a description of that artist and converting it, which you could go find and type in yourself on Suno
For like, a 'value' idea or whatever, so what 33 seconds. I use Suno like that, lopping of tons of time to get what I'm after
It's different though, very much so, than Suno
If you want to make weird stuff Udio is better
Weird as in strange genre mixes and noises and things
I said it a bunch last night but it's really compelling
the 320 quality thing is nice too
I have an orchestra metal clip right now, intro meshed fine with the first clip but trying to mesh anything after the 33 second clip either clips or just plays something that doesn't flow in a good direction, it's a bit of a hassle.
The issue with 33 seconds is that what you're limited in the time of what youre trying to achieve, for example in Suno i can let the AI generate 3 stanzas, sometimes those three stanzas span over 1 minute, sometimes a minute and 33 seconds, sometimes a minute and 40, sometimes it repeats the stanzas twice over 2 minutes, theres more variation chance and stuff that could happen between them
What would that description be?
So it’s not possible to replicate or get a very similar style of music to any artist on Udio, I’m guessing? If so, it makes sense.
If we put in tempo does it recognize that, anyone notice?
Yeah but the other side of that is you can cram a whole bunch of words in Uido and watch as it makes the speaker/singer speak faster in order to try to fit it in the 33 seconds 
I mean, depends, hang on i'll find it, it literally says when you type it in "replacing artist name with:" and gives you the words
Ah, I see.
Saw that yeserday when someone tried Whitney Houston. Udio states right there (changing artists name with this: (words)
So it might describe their style
it's like playing chess with chatgpt, first few moves are great, but then either it's or your pieces start hopping all over the board in the mid to end game is the best analogy I can offer right now.
yeah like i did Red Hot Chili Peppers and said replacing artist name with "alternative rock, male vocalist, smooth vocals, melodic" etc etc
Udio rules completely over Suno without question in being able to generate before whatever you are working on though
You never fight trying to get an intro with Udio
Just add it whenever you want
The downside is you kinda always start in the middle of the song
What if you’d name a well known album in the prompt on Udio?
Would that have any effect, if you know?
Since it won't, probably, give you your intro to start
gosh imagine having ai generation but with the interface of a track editor.
I started a spoken intro on udo, and on every extension it would not start singing... 😒 suggestions?
That does seem very useful. But they both have their strengths and weaknesses, of course.
"Oh I want to delete this section, and I want to copy this section here." And then have the AI do the blending.
The other thing about Udio I've found
It's great to be weird
But it will always be weird
How do people use Udio? It says there’s a waitlist
It sticks to whatever the genreation is and will. not. step. outside.
wait list? I just created an account
idk I made account yesterday
Might be new and due to so many people jumping on it
Wait, I seem to be able to make stuff now
Probably. It's working properly now
quite quick as well
remember that Suno had to upgrade their servers because they got hugged
Is this the correct screen I should be seeing?
george, the dev here for Suno (hi gerog) has talked a few times about infilling, aka inpainting, for v4 of Suno
That would be most of the way you are talking about
is that the app?
dude that woudl be rad, cause I love using Midjourney because of that
That's the other other issue with Udio. You can't cut anything at all. It's 33 seconds every time, no changing that\
phone or what? on my pc it looks kinda different
Can’t you extend it?
yes thats annoying too
Midjourney lets you take two images and then the AI creates a new interpretation using that data
You can sure, but exactly 33 seconds
You can't pick your time
its nice until 38 seconds ^^
Exactly
Ah, okay. Well, it’ll just get better and better, I’m sure.
All songs for Udio are easily recognizable because of the lengths.. 0:33, 1:06, 1:38, 2:11
yep
Extensions also become a bit distorded as well
guys how can i pay suno ? i dont have credit card..only paypal but i cant find this
I feel it is better for just for fun songs for now.. serious songs with guidance are really hard there
asylum of reverie 😵💫
This is about the only thing from Udio i kinda think sounds good, its not my style of music entirely that i would try to make, but still is something i consider good singing
https://www.udio.com/songs/o9rPsKosJevyndyVPjWZQP
Make your music
Wonder if the music stuff will settle to how the image generators are. You have your Dalle one, your Midjourney one, and your Stable Diffusion. People use Dalle because it's 'free' and follows prompts well, Midjourney is the prettiest and quite full featured, and different, and then Stable Diffusion which is whatever the heck you want it to me, unmatched features and depth if you dive in. Plus open source and your own hardware
Stable Audio has that aspect of covered for the audio side already, it will get there.
Oh well.
Midjourney is more for dummies imho.. everybody can make pretty pictures even if the prompts are bad
if you're gonna do AI images go for stable diffusion, if your PC can handle it
Stable Diffusion is king for me because yeah, my own hardware and there are a billion +1 different things for the creatives, also run it right in Photoshop
segmentation and Ella and control nets and and and
Yep.. I have it too.. Automatic 1111 and ComfyUI
is there somewhere on udio i can actually see the lyrics?
I switched to Forge vs. Auto1111
Forge is simply not-shit Auto1111
Literally the same interface but lots of stuff built-in and sidesteps all the weirdness
I like ComfyUI.. it can make videos 😄
ComfyUI is probably the best since it gets all the new stuff faster
might look into forge at some point, i had auto1111 installed but did a windows reinstall from 10 to 11 and never set it back up
But I haven't took the time to get used to the UI because
boy
it's weird
Stable-Diffusion-WebUI-Forge is just plain better than Automatic1111. There. I said it :p Why? Well, performance and resource usage to start with. Even better still, some of the oldest bugs related to model loading have gone, plus you get loads of stuff just built right in! All that complete with the familiar A1111 WebUI gradio interface. Sure s...
I can't overstate how drop-in replacement for auto1111
All the plugins and everything
Betteridge's law of headlines is an adage that states: "Any headline that ends in a question mark can be answered by the word no." It is named after Ian Betteridge, a British technology journalist who wrote about it in 2009, although the principle is much older. It is based on the assumption that if the publishers were confident that the answer ...
See with Udio... i like cant even understand the lyrics in the intro i added
https://www.udio.com/songs/37q73TmYGr6MMnNN6iaFvZ
Make your music
I let it auto generate
exceptions are cool
and I can think of a few of them
In this case Forge is better than Auto1111 since it's the same, but works properly
But if you don't care about that then keep using auto1111, it's still there
It's Udio
Guys do you know alternative suno? Udio dosen’t work
Yeah, but the difference with doing that in Suno is you choose where it cuts off!
It's Suno and Udio, those are the only ones even close to each other
Yeah the being locked to multiples of 33 seconds is the main deal breaker for me
Sonauto isn't bad but I don't know if it's accessible
Hm, seems like studies show the opposite, the answer seems to be 'yes' most of the time though I think the reasoning behind this 'law' is a sound one
Oh right that one, man I really don't like sonauto
It's certainly not as good as Suno or Udio but it's the only other one that gets close
The way you interact with it doesn't jive with my sensibilities
You're right though
Raw audio is a lot better in Udio. But making a coherent song is significantly easier in Suno.
im still using it, but yea that can be quite annoying, like when it's trying to base the lyrics pacing off the 33 secs. I think... or speculate at least... when you add stuff BEFORE and it has to make that first section meetup at 33 seconds... I wonder if that is what can lead to funny transitions. Ironically sometimes they can be clever musically, it's just they happen so often
True dat
Guess it's a question of which team can address their weaknesses faster. Cheers for the consumers!
Make your music
i agree about the audio vs ease of use stuff
I mean I hope both are successful and stronger for it
That quality really comes through with certian genres
Udio shines with stuff that would have distortion
lol.. Suno peeps are motivated.. I can tell you that 😉
Since it seems like the quality advantage it has means that distortion stays disctinct
Instead of just mashing into noise
Suno team does a good job at listening to their community
Good. I'm paid for the year. LOL
This will hopefully lead to both services improving a lot over time. Competition is always good
Me too.. twice.. premium
The community is worlds better with SUno
This discord, the discourse, interaction, that stuff is not even a competition
It's a lot harder to build that
Than to just 'make better model'
this is why Competition is a W
true dat come on we have a helpful tree. How they gonna compete with that
W?
LOL Udio's definitely got the double whammy of being new and being huge. It does not make for as cohesive community feel.
we have a great community here, only time i ever have problems with people they seem to have this next to their name
lmao
and they seem to disappear pretty fast too
What I really don't like with udio is the 33 seconds which leads to much less consistency throughout the song. Also I had way less success with genre switching.. but at least it can sing lorem ipsum well
Suno's major pro's ATM are:
- Easier to make a longform coherent song
- Has paid plans and priority
- Doesn't Require Credit for Commercializing on the Paid Plans.
I'm sure there are more, but these are the big 3 for me.
Suno's community - well I do wish there were some more channels to split everyone into, and I'd love if the song sharing was limited in General. But I do feel at home here in a way I sure don't elsewhere.
Yeah, Udio community is a little rough being so new
I'm decently jaded so I don't expect it to improve
i dont really understand why Udio didnt have discord setup before launching the beta. i thought that was REALLY weird
I generated some cool tune if you want to use it https://suno.com/song/9581d99b-2828-4187-bbea-8665ba41cf5a
acoustic guitar, dark, anti-melodic song. Listen and make your own with Suno.
Ah come on how important can a community be ehh 😄
Do you think they went like "Ah shit I knew we forgot about something"?
it looks like that lol
Too busy trying to be stealth mode.
when i joined it was pretty clear they had JUST created the discord server
Well that's definitely something
All that is why I don't have confidence Udio's community focus will improve
hi
yeah seems like the discord/community was an after thought
I hope you are wrong but I can see where you are coming from yeah...
i mean im pretty sure they banned Julie Wong almost right away so thats gotta mean something
shes a notorious troll lol
I mean you can see how important that 'community' is to Suno just in how it's handled. Moderators promoted from the community, guides, and they are in here seeing what we are all doing
That's effort nobody puts in
Not sure they have the best vetting for who they are putting as Community Leader's though
IMO
yeah the thumbs up thumbs down system? lol
Most things like this fumble on the community side
just screams "oh shit we didnt plan for any of this, you guys decide!"
no effort
Yeah, the guy that is currently Community Leader was getting ratioed in the chat for stupid takes just 12 hrs earlier
So then chumps are going to roll in like reddit and then there are weird in-groups and feifdoms and ugh
Yeah no thanks.
Yeah, that was a serious red flag for me
lol really?!
yeah, I can pull receipts, but probably have to put in Off topic
oh, wait, we in here
lol
good thing thats where we are!

i wonder if its the guy i blocked
yup i blocked him for dropping into my DMs just sending links to his songs without even saying anything
What a sweetheart
a...degree in AI
dude straight up spamming me
Yeah, and he is a Community Leader in their Discord
i have an accounting degree, you dont see me bragging!
Man you guys are far more trusting than I am. I join a new discord, the whole thing gets muted immediately. Also ain't nobody DMin' me first without going through whatever discord's thing is
i also mute immediately
Hey isn't that guy in here
sup
i saw the message request, saw it was just a song link, blocked
Yeah lol
so we really talking trash now
Yes
I can't see anything I don't know what you are talking about 
i started seeing "blocked message" pop up so im guessing hes in here now
They are yeah
The call is coming from inside the house
im just impatient when it comes to people DMing me
I use 'block' as a mute button for people I'm not interested in, very easy and quick to use it
yup
Okay, okay, I've played around with the new beast. It still lacks prompt understanding and isn't quite at the point where I would rely on it as an additional tool. But, the clarity and vocal models are pretty amazing. I think this is the main reason why everyone is talking about UDIO. I've been with Suno from the very beginning, so sit tight and let's see how it evolves next.
if were gonna talk thats fine, but people just send me links or try to trash talk ai or whatever and i just dont have the patience
do t know what your problem is
Yep.
I sent you a Friend Req, but not a DM
that was a couple days ago
There are literally billions of people on this planet, and many thousands I can interact with directly, right now
on this account? I had to switch back to this one recently
Scrubbing one scrub (heh) from your view ain't shit
oh no, the jimjam3 I think
yeahhhh it got banned LMAO
JormJeames
i made a joke about being 12 and discord took it VERY seriously
accepted it! lol
maybe I misunderstood your take, re-reading it. I think I agree, but there will be a lot of, "nobody's".
I’ve written 310 song title ideas so far today.
nice
I like a lot of them
Lol
Really long list though
A personal record or something
Just spontaneously
I rarely get more than 1-2 good ideas per week. I'm not sure what I would do with 300 lol
I have a few ideas queued up and a some poems i need to convert over still
been chilling for a bit, have been applying for jobs among other things
Good luck and happy hunting!
i’m unemployed and bored out of my mind
made a lot on the stock market a couple months ago tho so i’m vibing
Interesting, yeah, never messed with Stocks, have some money in crypto, but a long term holder type of guy
I don't invest in volatile crypto, long term bets
XRP for example
the job market in general is not great unless you are trying to find work that pays less than 50k a year, and even that isn't great.
Speaking for the USA at least
I just tried Udio, and I’m not too impressed. Lol. Maybe I didn’t do it properly, but I asked for an instrumental in the style of Michael Jackson, and it made a short song with a voice in the start that actually sounded like a robotic and distorted version of Michael Jackson’s voice.
Listen for yourself. https://www.udio.com/songs/vWmojPGC3yYrv2bP14TaMX
Make your music
But it had lyrics and singing
So not an instrumental
Although the words weren’t very audible
It sounded like lyrics lol
just sounded like a bunch of discombulated sounds IMO
Well, that’s a pretty accurate description lol
How were you able to use it? I just signed up today to see what it was about and I can't even use it. It says they're at capacity 😑
I was able to use it now a bit, but it does say that it currently has server issues
The Udio interface/UI is so far pretty hard to navigate. At least on mobile.
Like, I click on buttons and it either glitches or nothing happens.
I have tried refreshing
I made a whole song and I thought it came out pretty decent. The sound quality is amazing. But the creation process is painful. https://www.udio.com/songs/fhZLFZTyCbejkMjTEaVZ5Y
Make your music
Pretty decent actually
Although I think the voices on Suno are even less robotic
In some cases
I’ve gotten very smooth, nuanced, human-like vocals on Suno
the trending songs on udio are unbelievably good. havent been able to try creating
Less robotic, but there is a static in the Suno ones that is not in the Udio ones, even when I'm not battling the Suno reverb. I tried to make the same song in Suno and going back and forth directly, you definitely hear that Suno's voice quality is lower.
But Suno has a much more dynamic delivery. To me the pacing on the voices seems the same on everything from shoegaze to speed metal in Udio.
the word that comes to mind is "coherence"- it seems to better stick to the traditional structure of a song
whereas suno sometimes its not clear where the chorus actually starts
And Udio completely ignores previous choruses. It's just roll the dice until you get something sounds close enough that it might be the same song.
but again i only am seeing songs that have made it to trending so not a good sample size
ah, gotcha
no complaints about suno mind you its all amazing lol
I still prefer Suno. Here I am chatting after all. But I can't say I'd be disappointed if competition prods them to their A-plus game. 🙂
Make your music
I agree, the quality of sound is better on Udio, but I won't trade that for the overall package that Suno offers at the moment
udio is pretty dope aswell https://www.youtube.com/watch?v=LKrxCBqLQno a 90s rap song - which can tell the vocals are more clear
I also made udio and sunoai songs in bluegrass style -- the 1st half is udio and 2nd half is sunoai which is fun, https://youtu.be/-noe7wFlUY8?si=iD-dpJGXV3NHgJ79
Making AI music is pretty dang cool!
The first half of the album is udio and the second half is sunoai.
I wanted to try to make some bluegrass folk songs!
#aimusic #bluegrass #countrymusic #bluegrassmusic #aibluegrass #aicountry #alienmusic #ufomusic #fullminialbum #fullalbum
So far, Udio has been much worse than Suno, in my experience at least.
I'm just going to say that healthy competition is good for everyone. LLMs have it (and even an arena where they battles other LLMs!!), image gen has it, video gen has it, voice gen has it. I think this is a good development,competition in image gen led to improvements all across the industry like now every tool has a "region" or "spot" modify.
at first udio was not as smooth as suno, but I think it because sunoai does alot of favors in backend to mix up lyrics and structure? it seems to have higher success rate
my fav song i made it did a good job mixing up the lyrics to flush out the song
Chill Lo-fi hip-hop with female vocals song. Listen and make your own with Suno.
of course the competition is good, I agree.
Shit goddamn it literally one of the big examples of where suno has room for improvement
and udio didn't do it
udio only has 33 seconds of context so 🤷♂️
Yeah, sadly Suno's continue feature is the best we've got for making fine changes
Not sure what the context is. It only gens 33 seconds at a time, but it could be listening to 60. But I suspect it's not. No, @runic depot any problems you think Suno has with consistency are seriously 10x worse in Udio.
I think that feature was built in assuming that people were too dumb to write proper song structures
it was supposed to be idiot proof
And guess what the ❤️ back on 🙂
but turns out to just be a pain in the butt
But it does sound effing amazing.
True dat but it even loses things in the first part so all we know it could even be 15 seconds context.. or could 2 hours context but dumb af 🤣
like listen how often the voice and stuff changes https://www.udio.com/songs/p66uVGEgifEBLdoR5Ttyue
still good though
Yup to both the lack of consistency and to it sounding good.
did udio fix the weird songs changing after gen a few times yet? that is kind of a hilarious issue. You be listening the result and then next time you play.. different song.. next time yet again different
Yes
Seems like some people in this server just expect Udio to be perfect right off the bat. I mean, it's their V1 and IMO it's better than Suno, even though it's flawed (just like any AI model atm)
@queen dove Sonauto and Udio are the only other 2 worth mentioning and Sonauto is kinda crap
Or like because it's not perfect, it's not a serious contender. It's definitely real competition. But I've no skin in the game. I'm happy to use the best tools available to me. And it might turn out that for some songs that's Suno and for others it's Udio and for still others it's some startup in stealth mode. I don't care who the winner is. and hopefully there isn't one winner.
Until Udio makes a Paid tier, it could just be a Data Harvesting tool lol
LOL, it could be. blanchon, is that you?
Hell no lol
I meant Udio might be blanchon, but thanks for putting any doubts I had about you to rest!
No, I have no interest in stealing anything and I'm not smart enough to make my own AI
?
Just checked out this week's track drops on @GoGalaMusic! 🎶
The lineup is incredible, especially excited for the "RZR" soundtrack NFTs.
The fusion of music and blockchain is changing the game
I mean that is kind of the reason for relatively generous free usage right? Get much "real world" usage data fast.
can we share songs here
i keep forgetting the rules
its kinda busy here so i dont stick around for long usually
how is everyone
well, I more so meant it could hypothetically just pop up for a week and then disappear forver, but the incentive for that would be odd. I was mostly joking
I think so, but the #🐣┃suno-showcase is for Suno stuff
But I have seen people posting Udio songs here
what is udio
Different Music AI
suno was first right
Yes
yeah ill keep investing in them then lol
i have been struggling to get my song heard for 2 decades
buddy shows me this
7.3k views in 28 days
i cant even believe it
ai is insane
Ah well could be. But yeah would be odd. I misunderstood the message then 😅
hey thats ok
misunderstandings happen lol
we are human
anyone else having good luck with youtube channel views
or what are your goals??
Not as good as yours
what are you making
I just want to monetize eventually
No.
I, Robot
You need 3-4k watch hours in the last 365 days and 1000 Subs to run Ads
on youtube
what do you know about distrokid
i litterally just word vomit into suno, and generate what i like to run to then throw it up.
heres one if your interested
for a reddit user from the junk yard
Sincerely yours,
The Sandwich Maker
and i am trying distrokid
Interesting methodology
I make music from love. Paying for my Suno sub is the dream. But I'd settle for paying for my DistroKid sub at this point LOL.
Audio-quality wise, yes it's better
yeah, I'm not trying to game the system
want to actually like what I put out for the most part, although it's hard to make a perfect song especially with AI
fair
... and you can replace one word and keep the melody. Take that, Suno devs LOL
unless it's at the end of the song right, then you can't remix because it's too big
but only on first gen right? After that it doesn't work anymore? Or is there a trick I overlooked?
Yeah, can only remix the first segment in a song. But remix means remix, and you can swap a word out.
Me registran
The number of times I've accepted one wrong word in a Suno gen because it was the first segment and it was the one that owned the melody is ... non-zero.
So anybody tried a manual mode lyric and know how to continue before it start with gibberish at Udio?
Fill it with more lyrics than it will fit into 33 seconds was my attempt which seemed to work
Mira no te entiendo nada me lo puede alguien pasar. A español
Yeah but if you like the current song?
english here please
Que
you want me to give you a time out for mocking me?
Throwing in more lyrics? it's tough, if you're lucky it will humm or something instead, it seems to just fill the space with whatever it wants
So no way to keep the same song... I have to start again
You can probably keep genning and get a good one, but as far as I can tell there's no direct control really
Hoy can speak spanish
He just said he didn't understand you, chill
Somente inglés porfavor, amigo
I don't need your help, thanks..
You can Speak spanish
Ah so you already have a gen and like it but it ends in gibberish?
You really are one of the worst MODs out there Jesus. Enough with that attitude, you're here to help people
There is an spanish chanel.. but if you want to talk here, you will have to talk in english so we can all udnerstand
No entender very mucho
https://discord.com/channels/1069381916492562582/1154779774858833940 Here is the spanish chanel
oh, did I post the wrong one
You can regen it and set the similarity high. Maybe alter the lyrics a bit. I'm no expert but there is a similarity slider on the regen screen.
isn't spanish and espanol the same thing?
Thank you
Hi Vito, due to your recent message that didn't align with our community guidelines on respect, you've been temporarily timed out for 24 hours. Let's take this time to reflect on how we can interact more positively in the future.
yes 😄
But there are different kind of spanish.. You've got Mexican who speak spanish too..
They do understand each other though
gotcha gotcha
Yeah.. but it is the first part.. and I was just testing but I got this gibbrisch taking over.. just hoped something like a time stamp 😦
this example. Was just a test: https://www.udio.com/songs/qAcTotujQ3BMy7rHravwNL
Make your music
Oh if it's first part try remix for sure
it's proper remix it is what @rocky meadow meant with the similarity slider
True dat.. But that is not my point.. What if you REALLY like the first outcome.. but just not the last 15 seconds..
Hmm I can't test right now but when i used it earlier remix was proper remix and let you correct small mistakes maybe try with similarity cranked up and just add lyrics at the end? In theory it should come out very very similar
Interesting.. will try.. thanks 🌼
Lol.. totally different.. but okay.. It's fine.. Just exploring 😄
Ah well unlucky then. I think in the udio discord someone figured out something with a seed. But i need to check that out myself if the code snippet he posted is safe. Will let you know tomorrow
Thanks.. I would love to hear that!
ideogram.ai is good at making ads ngl
"waj hte bekt lhe i/trua tihe vor nonyith sheny yor, and frost so inspiring
Udio seems to do really well with weird stuff
dude we sahre the topic ^^
I really enjoyed that
-_- clickable
wtf its 2 mins ^^
Make your music
sry Idk better
udio is terrible for me ;x
this took quite a while, next on suno as well like 20 outputs
udio can do magic stuff it only takes forever, and often it's very odd like mine as well
The vocals are way better and soundquality often, too. But for me Suno is more like music while a lot of udio is totally off or really weird. and than it doesnt go anywhere.. the style remains, no real chorus, the lines are off, etc.
ah and with suno I'm annoyed with the intros. Ihave a spoken one and half of all generations are totally muffled and buhu ^^
Suno loves the word "clear" on vocals if you want the balance of singer versus instruments to tip to the singer. "clear female vocals" in the style prompt is my go to.
ok thank you.. I'll tryon the next generation.. I'm at 18 now and most of them are musicwise ok but I can't hear the intros -.-
https://suno.com/song/b2e890e6-ed46-438e-838d-bd21db0ae32a this is a "good" one 😄
theatrical New Wave psychedelia, psychedelic haunted illbientwave, aggressive carnival, dark lute haunted song. Listen and make your own with Suno.
well it's clearer now, but it's not spoken anymore
idk it's only diddling around anyway
Hey 3Daizy, reported your attitude to the Staff :) thanks for the time out though, very friendly of you. Hope you either change or lose your role as a MOD
Honestly hope criticism is still allowed in this server :)
Threats against moderators are serious breaches of our guidelines. You've been given a 24-hour timeout to reflect on the importance of respectful interaction.
Threats are different than criticism.
Genuine question, where was the threat? Deleted or something?
Oh, can you share the tags?
Vito’s messages, while reflecting criticism, also border on disrespect and indirect threats, which are against our community guidelines.
I'm impressed with how it articulated all of this. Took a few tries for it to start with something creepy.
Prompt: raspy demon, reverberating, breathy, intimidating, uneasy
RIP suno
I don't think so.
Long live the king Udio!
I tried to make a song using Udio with some lyrics I made for a song I made using Suno and it butchered those lyrics and couldn't get the feel like Suno did. So I think Suno is still got it until Udio team gets better. I won't say Suno doesn't have some competition though. Udio definitely has the potential to become better.
COMEPLETELY random question, anybody here like Digimon or play the card game?
You have to give Credit with Udio, so no, not really.
Who can we tell that the quality has dropped significantly and I am using credits for junk that is not my fault. If I have to keep generating until it's not static or lyrics skipping or music sounding muddled, how is that our fault and why should we pay for that?
I like Digimon but haven't played the card game even though I have some cards. lol Honestly the series kinda died off for me after a while when they started getting crazy with the story.
some generations are just going to be stuck in that loop
I have something for you then!
this is like dalle vs midjourney, but in music generation
You can try #💡┃suggestions-web or #📚┃suno-school though
Thank you
Pretty much. That's what I was thinking when I seen some YouTubers covering them both. lol
Wow this is really great. Nice one.
If you're ever bored I recommend just taking some of your favorite cards, their titles, and card text then tossing them in
Thanks been working on it for hours. Hard one to finish but I’ll finish it and release it. Probably gonna do a full 80s glam hair metal album
You are definitely on the right track. I really like the vocalist; they're very authentic. Just so you know, I figured out that Udio allows you to access the seed (unlike suno), but it's not exposed in the UI. However, you can use it as a parameter when making the generation request. I'm not sure if you're technical or not, but I hope that makes sense. I thought that might be useful to you if you want to create an album with the same vocalist for each song, maybe.
Everything I tried to "port" from Suno to Udio I was disappointed with. You really need to start from scratch. We've all kind of learned the rules for Suno and believe they are THE rules engraved on tablets from on high. But Udio does behave according to its own edicts and it makes beautiful (though not necessarily consistent) music.
Divorce Papers on Christmas - the best christmas present!? 🎄
Also I'm so surprised the female voice came in in the later part of the song, didn't expect that but it turned out pretty well I think 😄
Interesting. I know about the whole seed thing but not with Udio. I think maybe I’m going for various artists on each song and just release on an album so it’s all the same genre but different variety of singers. Seems like a little extra work trying to get the same singer. But that’s good to know.
Yeah I figured it would have its own edict but I was hoping it might get the feel from lyrics we make of our own that wasn't AI generated. -shrugs- Everything I've generated using Udio hasn't been great even if it wasn't a port from Suno so I'm kinda leaning towards staying with Suno for now.
Only thing I don’t like is when Udio makes a song it starts from like halfway or something so auto generated hasn’t worked out too good for me yet. It’s very hard to complete the song with auto generated keeps mixing up choruses and verses. A bit messy. Might have to make my own lyrics in the future.
Brutal, lmao. Such a dissonance between the lyrics and the tone of the cheerful Christmas song.
I'm still here. But damn is the quality there on Udio. Just the voices, the instruments. So clear. The vocal delivery is stilted (to my ear) and it tends to forget from one segment of the song to the next what it is supposed to sound like. And I just like the interface on Suno better. But it''s a decent product. I made one whole song and I was happy with the result.
I did kinda find a little trick to making sure songs keep the same beat/feel when extending them on Suno though. If you get so far and the next extension sounds like its starting to deviate from what you wanted it to do, then get the full song before attempting to continue extending the song. So far it has worked for both me and my hubby. I thought it was a fluke at first but then we kept getting consistent results last night with extending songs.
yeah it forgets how the chorus is supposed to sound, but it often works out kinda nicely lol
Agreed, for their first model, Udio is really astounding, considering where Suno started. The clarity and range in vocals are what really get me. It's just next level.
Agree. The clean quality is the big one. They need to work on the UI a bit and workflow. It’s a bit tricky. It’s also annoying the songs auto play and it’s a bit hard to remove the ones you don’t like and like.
Quality is simply astounding. For a Beta that's insane
Udio’s beta is like Suno v4 if they had it lol
The way I see it, it's just another tool in the toolbox. Like how I use both Claude Opus and GPT-4 as LLMs, each has its pros and cons, and I use them where appropriate. In the same way, I plan to use Udio and Suno, as they both have their strengths. It's clear Suno is trained on way more material and genres.
V3 Suno sounds like 128kbps. It ruined so many of my good songs.
That's funny, I felt the same thing when I first tried it. I felt like I was using what I imagined Chirp/Suno V4 to be.
I also noticed guitars and guitar solos sound very realistic with Udio. On Suno it was terrible and rarely sounded real.
I've had some pretty awesome guitar solos on v3.
I’ve done a bunch of Rock on Suno and some are pretty decent but when I tried it on Udio it just sounded cleaner and more realistic? I dunno. I just like it more on Udio.
My main complaint about Suno is the reverb. I know that isn't the #1 of the community, but my overall song quality has fallen because since I started hearing the reverb I can't unhear it and I'll choose the no-reverb clip over one I would have liked the delivery or creativity from otherwise.
https://suno.com/song/943b1cb8-8146-4f2d-a343-adcc0bba571a
Here is one I did with v3. You are right compared to udio its not as clear.
electric guitar power ballad solo song. Listen and make your own with Suno.
If Udio increases their 33 sec track (at one time) limit (which I'm sure they will soon enough) and somehow makes the AI remember and replicate the way verses and choruses are sung in previous sections - they'll be unstoppable
My biggest isssue was almost all song had that artifact sound on both the music and vocals. I also feel the vocals are stronger and cleaner on Udio. I’ve heard some powerful singing on Udio as well as Suno but it’s just more cleaner and realistic on Udio.
That’s not bad but yeah you can hear that artifact somewhere in there.
You can actually tell the difference without someone telling you which is Suno or Udio. Suno has this signature sound lol.
Yes, the 33-second context window is the biggest issue with Audio compared to Suno right now. After that, it would be the bugs and issues with their system, as well as the restrictions on generations and all the errors in their process. If they can work on those it would be pretty big.
And make gens faster. Suno is like 10-15 seconds gen time.
The singing is cleaner on Udio. But I swear there's more emotional range (from my limited testing) in Suno. Suno can make me feel longing or despair. I have not achieved that in Udio yet. It's pretty, but pretty like a statue.
OK, now here is one I did with Udio. You can hear the guitar come in around 1:10. Definitely a difference. It really sounds like a capture of someone playing.
They also need to fix the UI and options for better workflow. I like Sunos work flow. When you don’t like a track you just dislike it and it disappears making it easier to know what’s good and bad. And I like how it shows where the track continued from.
Also the timestamps is good.
Yeah that’s nice man
I know exactly what you mean. I tried to replicate the specific emotion and feel of longing/despair with udio, and couldn't do it. The MP3 attached is from V3 alpha and really captures the feeling.
Here’s a sample of some strong solo vocals. I found it better than Suno too.
The quality has gone down since a couple of days ago and the reverb is ruining the music. It's too much and off center. Also, it skips lyrics, hops and the music is generally more muddled. I hope they can fix it but it's not Midjourney. They don't tell us anything here like they do over there where we at least have office hours and can discuss with them directly problems and solutions. I am contemplating cancelling my subscription for now because without communication, How are we supposed to know if they even care about the problems. They are still getting paid and in fact we are spending good credits on bad output. I hope they notice and are working on fixing things.
I miss v3a soooooo much!
I agree, but compared to the very first web UI Suno did with V2, Udio's inaugural offering is really quite good.
There’s a guitar solo in this too
any recommendations to switch to Udio? Nothing in the past 2 days since the changes they made is sounding remotely good.
Not the genre I'd expect to have an electric guitar solo, but it works, lol. Sounds good.
The changes in suno or udio?
The changes in Suno.
I know this proves your point, but what changes? There's nothing in announcements.
Was there any purple name actually saying what was changed?
There was a message that they were doing something with their systems a few days ago and it Suno was offline. After that, the quality has been degraded significantly.
It was a banner at the bottom of the page I believe, orange.
Or red
I thought they just made some DB changes to improve capacity, no actual model changes.
My gens today sound about the same as last week
Something happened unless it's Mercury in retrograde, lol
Because it's not just me, my friend who also uses Suno played something for me from before the change and same prompt after Suno went back online and the difference is striking. One is great the other horrible
And today for me it's garbling lyrics now
I dunno nothin'. The general channel moves so fast now-a-days I can't keep up. So I've been on the AI music showcase discord instead of here until Udio came out and made off-topic a decent place to hang out, LOL.
Ah okay is Udio better?
Better... for what? How you answer that question will determine the answer to the question, lol
It has its strengths and weaknesses
Yeah, I mean less muddled sounding music and more clear vocals, not garbled, and not over processed
Sound quality is amaze-balls. But workflow and consistency are less fun.
Udio is "better" in that case. However less range of genres I believe.
You can only do one gen at a time and there is a significant chance any extending of a song gives you something seems like it was made for a completely different song than the one you are extending
I will have to explore. I like that with Suno you can mix stuff up or at least i used to be able to now I get muddled sounding noise, literally noise, lol,
Oh that's so funny
Also, FYI, Suno absolutely did not change the model on the backend over the last few days; it was strictly database work.
Okay, well, then I was really lucky 3-4 days ago and now I'm not lol
Sometimes that works in your favor though and gives a nice unique sound to that part of the song.
Sometimes the slot machine hates you
Hehe, I dunno do you want to hear the garbled mess I just got?
Udio is also completely free right now if you can sign up. 600 prioority gens for the month, and unlimited total gens. But you will need quite a few to finish a song.
So no down side to trying it
Okay I just spent almost $300 for Suno. I think I will cancel for a while and try Unio
That's right. But the one thing... there have been some reports that there could be more recent bug specifically in song continues, and also there are a suspiciously high number of songs ending in :58 seconds, that combined are keeping my on the lookout for a glitch there somewhere. But I haven't dug it and tested myself. (Even if that is a bug it's likely not in the model itself.)
Also the model is different betwen V3 Alpha and V3, to be clear. But that happened much more than a few days ago.
Judge for yourself. I did, and I'm hanging out here chatting with the Suno peeps. 🙂
I really liked Suno but I also want the music to work again. Maybe they will work it out in a few days. Who knows.
I'll still make some more Udio songs, but I do find it frustrating, more so than Suno.
Good to know. Jonathan here probably has more time on every version of Chirp than anyone else on this Discord, so he would know. 😄
But when Udio works it's magic.
That's cool, I'm going to try it now.
Some of the songs I made on Suno several days ago (just joined) are sweet! But now, it's been just me wasting my credits.
Just checking - are y'all getting your style descriptions just inserted as lyrics over the last day or two?
Some days it be like that
keeps happening on my end unless I super-simplify the description
Can you put in your own lyrics on Udio?
Yes
Yes.
Cool thanks!
FYI, the key to not getting garbled or garbage outputs is your style tags. Don't put too many; less is more here. Start with the base genre only and work your way up from there.
ok thank god I'm not going crazy here. and poking around I see other people noticing the much much more washed out hissy generations all of a sudden
ugh
Okay, that's a good tip
Yeah, lol
Also, V3 is particularly sensitive to lyrics having an influence on the song itself compared to earlier versions. If the lyrics don't match the genre in your style tags, you will get bad results.
I'm actually trying the Udio "optimized" tag lists in Suno and you know, they often work super great in Suno too. Udio seems to know a lot about good prompting! The optimized Udio tags are typically so long you can't even fit them all in Suno. Udio sometimes only shows user tags, but you can see the full list in the webpage source, which is what I copied into the Suno clip here.
For example if you ask Udio to make a song like the Beatles:
https://suno.com/song/165e38b7-312f-4a5a-a25f-634f66370e66/
https://www.udio.com/songs/032e5a41-e272-4d4a-8d30-86d3b1a03508
male vocalist, merseybeat, pop rock, beat music, rock, rock & roll, british rhythm & blues, love, melodic, romantic song. Listen and make your own with Suno.
My friend too. They turned me onto Suno a few days ago because their music was amazing. But then this happened, hope they fix it . I think it must be complicated though.
Here's my comparison, Udio to Suno. The Udio is a complete song because I wanted the experience of making one, while the Suno is the best first segment of about 20.. But make for yourself and judge. I don't have any reason to prefer one over the other.
Make your music
Female vocalist, Rock, Alternative rock, Post-grunge, Melodic, Melancholic, Grunge, Energetic, british indie rock pop song. Listen and make your own with Suno.
Follow my tips and try again. I've been around here since day one; it's probably an issue related to what I mentioned. They haven't changed the model since v3 release. You can always back down to v2 as well.
Oh that's interesting. I am trying to do a rant in anime style, because I heard some anime songs recently that sounded happy but she was ranting and I thought I was funny so I decided to try and make one with some ranty lyrics. I don't think the ai liked it. It just will not do the song justice and it even put this creepy sounding voice in there right tin the middle, like wot? Lol I'm still freaked out by it.
Honestly Udio does a great creating a tag list to sound like a specific artist, even for Suno, much better than the various websites or asking ChatGPT to do this. It's a fun extra use case for their site. Ask udio for a song about X, then you get a nice list of tags that may also do very well in Suno. By well, it may or may not sound like the artist, but it usually sounds better than a boring simple prompt at the very least.
Sooooo many tags. When I tried to do my comparison, I had to remove quite a few of them to fit in Suno and then put them in the structure prompts lol.
I noticed as well, but to get it to do the artist replacement tags, it seems you need to actually submit a generation POST. Or am I wrong? Too bad it's not a separate API endpoint like lyrics generation in Suno.
Yeah. I'm constantly recommending to users, and even complaining to Suno, that the default tag lists are too short and not diverse enough, and I think Udio is squeezing more of their model by doing better on that front.
V3 is better at this but too many style tags or the wrong ones can really harm the output though, in my experience.
I don't think there is a way to do it without doing a POST, but you can also just look at other user's Udio songs and see the tag list and if they asked for "A song in the style of artist X' originally. Sometimes only in web page source, not sure why it is not always visible, because I do see it on main page on some songs.
I know I tried Dolly Parton, and it really nailed the tags and got that "Nashville sound."
Udio can do didgeridoo, that's a plus.
What is interesting is the same problems it has with Suno:
- degradation of voice/song as you extend it
- "forgetting" the song/melody in each extension
- style changing in each extension
- vocal chorus effect
- hallucinating or making up lyrics
someone explain Udio's "optimized prompts" to me.
From https://www.udio.com/pricing: Our product is free for the duration of the beta program. In this period, you can make up to 1200 songs / month.
In non manual mode, it takes your tags and rewrites the list, generally adding a lot more. Occasionally making a big mistake like also removing 'bagpipes' from your list, but in general, helps a lot.
The forgetting makes sense, because "remembering" is hugely expensive and gradually becomes questionable. Even text AIs forget stuff pretty quickly. The more interesting thing is what causes the vocal chorus and degradation over time.
That I couldn't even hazard a guess at.
https://github.com/jasonppy/VoiceCraft new model and tool dropped
Ah, so that's all those extra tags that don't appear in manual mode.
im glad to see my opinion on udio is similar to others, because the quality is really good but suno seems to be better at getting usable songs if you know what you're doing.. still testing though!
all I saw was news of "suno killer" and I don't agree with that, they both seem to have advantages
This is one song I got to completion. At each phase it didn't go where I hoped it would, but I'm not disappointed either.
https://www.udio.com/songs/iMg1e2fCxjJZ6zJwekxWSB
Make your music
They do say that their long term plans are for granular control and musician-oriented features like providing references, so my hopes are high for udio
it's just... all I'm saying is... udio's capabilities are leaps and bounds ahead of suno, not just in ways like "it sounds good" and "the voices are more natural"
In one of my first gens I got an audio panning effect. There are cowboy accents and Broadway acting on front page songs. None of that is in suno's data, ever
might have something to do with the "undetectable" watermark
I agree with the quality, effects, range of instruments, but I haven't heard many good songs come out of it yet is all -- maybe I'm beyond the "wow" phase after suno came out idk
Also I'm willing to bet that with some experimentation, I'll find that udio has different hangups about what's allowed to go with what.
hi
Each tool has its own subset of pros and cons.
Udio didn't seem to understand klezmer at all and that is something Suno does really well.
V2 in some ways has less signature than V3, though a lot of tracks are overall not remarkable, kind of flag. I do get plenty of Stereo in V3 instrumentals (but not much popular sung genres), but it might be because I use very strange prompts that push the model into more atypical places, which is easier to do in instrumentals.
https://suno.com/song/71a6edfe-c4f3-4944-9db5-6b723703ce57/
",",",",",",",",",",",",",",",",",",",",",",",",",",",",",",",",",", song. Listen and make your own with Suno.
whoa
feels to me there is so much still buried poential in existing Suno Models that just can't be prompted for easily or at all, but maybe some tiny tweak could change it entirely. just wild uninformed speculation though
also bark is like this, where it can sound like total garbage, but also perfectly clear and so human it's scary, so it's in there somewhere. but you gotta get in into that bark code with a sledgehammer and start smashing things to change it up.
Yeah, all kinds of stuff in suno that is there but can't be prompted. And yet, occasionally someone figures out that you can prompt something as specific as "arpeggio" or "vocal chops"
Could be that whatever "tagging system" made the association between music and words was less comprehensive than we'd like.
I've certainly gotten broadway acting in Suno, though in V3 typically it only stays as pure acting for a little while. inevitably they burst in to song. so you have to keep re-continuing it from before that happens.
I'm not even convinced that I can prompt "double kick drum" let alone "blast beat" but somehow "touhou" is valid yes I'm serious
"Liverpool" is a stronger live music style prompt than literally "live music" so I'm not surprised lol.
What prompt gets you an actual sea shanty?
The quality of instruments is night and day. It's like going from LCD to OLED. Airpods to wired over-ear headphones.
Actually in general my experience is that any prompt with multiple words like "double kick drum" is weak if it even works. If there a single word name for a specific type of that drum or something like that it may work better. Or maybe using dashes between the terms.
"blast beat" has less chance of being misinterpreted but that doesn't work either
What is a blast beat? Let me google quick
Kick drum at inhuman speed that can only be done with two pedals
Or, you know, a synthesizer.
I have had trouble getting any "double normal music speed" stuff from V3, also in EDM genres
I think I have heard it but never prompted it
I got some pretty sick speedcore but it was still well within the realm of normal music BPM
Maybe it was in a chiptune actually, so kind of cheating
You know what, let me try right now to prompt the most psychotically high energy song possible
and then just to see, I'll do the same in udio
When I test stuff I full spam mode, "maniacal, maniacal, maniacal, maniacal" and then maybe variants using ";" or " " between terms.
What is the name of a well known artist in that genre? I'll ask Udio for a song about it, then try the tag list on Suno
I think "hitech trance" was the edm genre i could not get fast enough from Suno
maybe I'll try that
There's a genre where the thing is having 1000+ BPM, so that the beat becomes a droning sound
like the artists challenging themselves to do that and still make it recognizably "music"
Two presses of the button, three fast but only reasonably fast songs, one absolute clusterfuck
Speedcore, nightcore, frantic, rapid, speed metal, touhou song. Listen and make your own with Suno.
Here's my favorite out of the others
Speedcore, nightcore, frantic, rapid, speed metal, touhou song. Listen and make your own with Suno.
I don't know that the "nightcore" tag works, but it felt right to use it
This failed epically but is actually epic in its own way https://www.udio.com/songs/ipXRzpKhBgKEkXrr3nXTvA
Make your music
"Manic" is a promising style word maybe
I love suno a ton. But this generation on udio carried me away. Couldn't believe it https://www.udio.com/songs/rMx3agX3s55U7eiYq1HU97
Make your music
Lmao that's hilarious
I use "frantic"
What the duck
It got it so wrong that it came out the other side so right
So here's what made me say that udio has stereo panning. This isn't random like yours, it's "intentional"
Make your music
I'm sure I'm not typical, but getting something like that madness from V3 is way too hard, love seeing it just pop out of the model like that. Good sign in my book.
I got one really interesting thing and one disaster
Make your music
this is weird, no idea what happened here
and the other... wait... udio actually PROMPTED for "fucked up headphones warning noise explosion"
Make your music
HEADPHONE WARNING
Omg 😅
it took my words and said "understood, here are some additional prompts to achieve the most horrible possible earblaster"
I tried using this Udio prompt in Suno. The longer "optimized" prompt in Suno, not the shorter one on top. The Udio optimizations not bad all in Suno. The firs one I removed orchestral to make room in the prompt but I added it back for second:
https://suno.com/song/d445486f-26da-4b94-8c56-345877e6dcf9/
Soothing, Opera, Atmospheric, Ethereal, Classical crossover, Neoclassical new age, Melodic, Lush, Mellow, Calm, Peaceful song. Listen and make your own with Suno.
I did get too many choruses so that was picked 1 of 4, the others were more like:
https://suno.com/song/e9e0c8e0-4091-4f8b-b321-562995bb99cc/
Orchestral, Opera, Atmospheric, Ethereal, Classical crossover, Neoclassical new age, Aria, Lush, Mellow, Calm, Peaceful song. Listen and make your own with Suno.
Wow!
the "optimized prompts" are genius really
It strips out meaningless words and adds better ones
What exactly are optimized prompts?
Udio has the same thing going on as simple mode and custom mode, and in its equivalent of simple mode, it adds/removes keywords and you get to see all the words it went with
Green is your prompt, red are the tags Udio actually used. On some song pages the optimized version is not displayed and is only visible in the web page source code.
Ahh okay
Presumably Udio knows some tags have little or no effect and ignores/replaces them, and others are used to add diversity/randomness
Using manual mode I got two winners.
Well, "winners," they sound terrible but they sound terrible in the ways we were hoping for.
Make your music
Make your music
Now, through all my Udio experiments I've mostly made unlistenable messes
Time to try making something more normal.
My normal testing behavior is to prompt the same exact thing 100 times, change one thing, prompt it 100 more times. Since I can't do that in Udio I am mostly browsing around and looking at other people's songs and prompts to get a feel for it.
Let's see what happens when I crank a bunch of contradictory genres into udio
how about emo, dubstep, and some less heavy EDM
I think to do this right, take two different optimized prompts and merge them yourself. I suspect if you just blindly try and guess all the tags you might get uncluky and one of the genres ends up much stronger than the other by RNG.
I want to see what happens to the auto-keywords when I do that
Oh of course, yeah, just ask for the hybrid to start. I was thinking you were jumping straight to manual.
the male voice is too close to Josh Groban and that is a problem there... you can get sued
https://www.udio.com/songs/uQ9WczQ4Hj4cZ89UvzXzVo
This one was too easy. Let's really set udio up to fail with contradictory genres that suno would flat refuse to do.
Make your music
Classical, death metal, country, and... vocaloid
Then sue me
lol
Not doing anything with the song
yeah udio seems very willing to create disasters whereas if you ask for the impossible from suno it'll just ignore half the tags
I think you linked the wrong song that seems just EDM genres
yeah but I got an emo singer
Oh is that what udio did when you asked for death metal country? In that case it really punted!
This is where all the udio mfs are at? yall not talking in the server ? 💀
I'm been chatting in udio server too, but it's a lot of new people, not much technical prompting discussion yet
ahh okay okay i getchu
no that's "emo, dubstep." What I have next is an actual demonstration of what happens when I ask for the impossible
Off topic question! What moderators in this server do you adore the most, and get the most help from/are really kind and nice. And what moderators do you hope to dismiss/unmod?
So when you ask for the impossible from udio, it just ignores half of what you said and goes with something sensible. Look at what it went with for "classical, death metal, country, vocaloid"
And also LISTEN TO THAT FUCKING SCREAM! ACTUAL SUNO KILLER
Make your music
WTFFFFFF
The other one from this gen was, as you say, a total punt
Make your music
LIKE. THE FACE I MADE WHEN THAT HAPPENED.
SHE ISNT EVEN REAL BUT SHE COULD CHOKE ME OUT
SHE PUT HER WHOLE ROBUSSY INTO IT 😭
Actually I remember for some reason trying to make Death Metal / Cute Anime Singer hybrid in Suno V1. It didn't realy work (more like the death metal voice took turns with the cute anime voice, instead of being a hybrid) but interesting here
udio is good with rap
Udio has a similar problem with ignoring requests for the gender of the vocalist it seems
@grim idol I think that I've a problem, I bougt the pro version but I think is linked with another email that is not the same of this discord account, I can´t try all features here, I don't know if is the same in the website
Complete failure. Let's see about manual mode.
For billing changes, you need email billing@suno.ai and they can update your account or link it properly. But it may not matter much actually because the Discord version of the music AI is not working so not having credits in Discord does not really matter right now. And even when Discord was working it was only version V1, and website is version V3.
Manual mode has been my go to.
Okay listen Udio, I realize that AIs are inherently not equipped to create truly novel things, and I understand that what makes sense to me may not line up with what's possible within your process,
but what the fuck.
Make your music
I... what the hell
this isn't rap this is the opening to a horrible r&b song
thank you
Well, udio hasn't impressed me in the field of doing better than suno when asked to do the impossible.
It doesn't meld random things together well. It also takes (my theory) strength from the first keywords and lower importance to the one's further down.
How is that mocking you? Lmao
Next, how about we try niche stuff that suno doesn't always understand the instructions for. Anyone got a niche genre that suno is too "safe" with?
How about... math rock.
Ooh, how about some screamo
I think someone already said they can't prompt for any vocaloids besides Miku, which is to be expected
best part of udio is the ability to add sections before and after a part... if suno adopt it too, and adding expanding from timestamp, or even replacing A-B section, that would completely be the killer one
"math rock" is a real winner
Make your music
You made a Suno og leave thanks to you're toxicity. Thanks
I'm coming up empty, what's another highly distinctive sounding genre to test udio with
Udio what the fuck
Instruments combos too. Banjo + genres where you would never find one or whatever. Stuff like that.
I see mentions of lanterns and dancing