I have autism and lyrics come to me. Like its a talent I don't know. I'm a wordsmith type Ashkenazi Jewish and I guess you'd call it autism talent I'm good enough to be considered an idiot savant probably. I can write great songs in any genre like a machine. I can't do much else very successfully. Lol aspergers autism and my family died or moved away. And so it's something to do
#📚┃suno-school
1 messages · Page 92 of 1
You know how Ashkenazi have high iQ and some of us are really good with words. That's why we write most Hollywood scripts etc. We write and people throw stuff at us cause we're a hated "model minority"
wow what a lore dump, heavy on exposition but it had flavour...
dude.. if you got it... use it, why not, just don't stop being you really at the end of the day, we all equal beings on this plane
That would be really comforting, but no it's 99% guys like MattVidPro who always wanted to use ChatGPT to write a song about a neon frog wearing sunglasses and eating donuts and god love them but I feel like people who don't have autism hate those who have it. Were not good for much of anything
urgh the animal songs drive me nuts
I don't speak for all autistic people but I never learned to fit in. I'm what you call level 2 asd I guess
stay awesome dude
if someone has autism and can write songs for any genre then what are you doing here? You could make boatloads of money selling them
i think it's an interesting talking point but i'm unconvinced that autistic and or people in general on the personality divergent spectrum are any better at writing or producing than anyone else... or maybe i'm just being a horrible ablist
Mentioning animals is okay as long as you have a reason to do so. My newest song "people are animals" I don't like how everyone doubts evolution and separates themselves from bonobos. It's bs. So this Darwinian song I wrote came out great. In v4. With lines like "feed a piece of meat to a hungry ape" that's the hook... do you guys usually add hooks to your songs? Hooks are nice
yea that's not what i meant, we all know what I really meant
I was going somewhere once and listening to college radio.. there was a song about buffalos on.. and then another.. and then another.. it finally clicked in my head it's like an hour of buffalo related songs.. and it was the best thing ever
The idea of this song is just that we rose up from rodent-sized primates. When I tell people I believe in evolution and I'm pretty sure there's no afterlife people think I'm a demon
you're talking to the child in me that loved dinosaur red cassette tapes
iykyk
Afterlife beliefs are a way we self-soothe. Dreams of dead loved ones aren't real. If life taught me anything. There's not enough humanistic/primal music. But I god might be real. Can't rule anything out 100% but it seems far fetched. "Human kind lives a lie, but the animal is alive."
depends on the audience and the person, i'm agnostic and i guess i enjoy both sides of the fantasy and the unproven truths
So I think shit like that makes interesting material for songs
Christopher Hitchens was cool.
So many things to write about and Suno's new lyric writer needs work but it's actually impressive af and while it can't write good songs from start to finish, it can do better than ChatGPT and other llm? I use all llm besides grok. And these are just tools to help me write the lyrics. I end up writing and writing, but these tools make it faster and
Another successful method for writing badass songs is to just start with a song you like. But then completely change everything in that song and just take influence from a song you like to make a completely new song that's totally different. The method works. I also believe in just getting ideas and writing.
I can't seem to NOT get some sort of background film reel sound from generating in any style of song generated... am i the only one?
I don't use them in general 90% of the time i was using it today as I wanted something cheap and dirty to test V4 and I just didn't like the limits
yep, come sit in our support group for shimmer survivors, here's a badge
is that what we're calling it? shimmer? lol
yea that's what many are calling it in the discord, i'm not sold on the terminology
Anyway, why did I write an in-depth song that tells the story of Jeffrey Dahmer from birth to death? Because suno. Lol serial killers are an interesting subject for songs. I love how suno opened it up. Looking forward to future models. V5 will probably be massive next year. I feel 1 more year of working at this and suno/udio models will be so much more advanced. I'm hoping they are on the cusp of major progress and v4 is the stepping stone to v5 which will blow minds
damn the vocals are so much better now
yep i fully understand, I've stopped generating in V4 for now 3.5 is ok for a while longer
but that... "shimmer"
it's mostly limited by the other AI models since their based on those foundations
It depends what you mean "better" better fidelity but worse instincts. The results are systematically less interesting. You gotta get lucky to work with v4 and get good results. I had 3.5 and personas set up so I could make every credit count. That's been wiped away
vocals are cleaner yes, tempo is down, vocal expression is down, vocal volume is more consistent, also vocal tone and pronunciation is seeing some problems in V4
There's quite a few other issues as well that are (at least for me) too hard to ignore
Over the last few days of 3.5 I churned out insanely good stuff. But my favorite persona no longer works in v4. No longer sounds like Linkin Park. It was literally Chester Bennington sound-alike
but maybe I care too much about the quality of it all
well, it IS in beta still, i suppose
yikes, knock knock we have an injuction to serve you
but those are pretty good descriptors
I didn't use an upload feature or anything. Don't look at me
haha chill i'm just messin
I'm chill I'm just flippant and tired and hungry
we call it hangry
At 5 am. Lol.
lack of caffeine too
damn... 5am it's 1210 am (Thurs) here
anyone want to know the winning lotto numbers 😂
Been popping V4 remasters and from scratch songs since release. Not a singly one without very noticeable shimmer. They are all useless.
if I had a dollar for everytime i've seen this comment today i'd be earning the same as a doctor
I'm sorry dude, hate to reduce your struggle like that
Metal tracks are the worst. Bass and drum lines sound like they have a reverb snare drum echoing on every beat.
yes, there's been significant issues with very apparent sound artifacts in a high number of generations
i've heard some songs that have other sounds like a whirring noise instead of the clicking, there's also a bug with cymbals being way to loud or strangely offset in the pan
Thank God this is a beta. Let's keep giving the good people at Suno a head's up, so they know the most important fixes we need for the final v4. Hopefully, no weird unwanted noise, up the creativity of the vocals and make it to where it was a catchy as v3.5. Give us our epic choruses back. Make the remastering option more accurate and fix the other issues with it. Crossing my fingers.
Apparently we have to learn how to reinterprate the prompts, but all my earlier prompts that gave heavy groovy metal sounds now give very flat and bright sounding early 90's black metal feels.
check out https://discord.com/channels/1069381916492562582/1308499132356362383
There might be some interesting remarks that you could use to mitigate the issues
remastering songs at V4 make them worse
It shouldn't be adding rap styling into rock. The rock music genre got gutted in this beta v4 but hopefully it will be back in the final release. Bring back the good aspects of v3.5 and v4 is suddenly killer
voice change tone many time during the song.. the music ''scratches''..
i'm totally sticking to V3.5 for a while
same
also the music became really ''flat'' with no arrangement.. only backgorund noise
lol oops
Moderators, guides, "alpha users" (who have red names in discord) and some other regular users are responsible for testing.
There may be others who had access to the alpha, but in terms of testing, it's mostly just selected users from the community.
fair comment, thanks for clarifying also
Suno do not provide an API.
Any apps or websites that use Suno for their music generation are doing so without Suno's permission. They can (and are) being prevented from using Suno.
how is this stuff being flagged to you?
I'm gonna keep messing with v4 and put up with the issues and see if Suno issues a statement of some kind. If they could start releasing a road map that would be great. Tell us what your plans are. If plans change, keep us in the loop. How is suno going to react to negative reaction to v4 beta? I'm crossing my fingers for a delay on the final v4 version to let them fix everything people hate about the beta. Because it's not realistic to just keep going forward with a model the community has rejected.
like the questions
v4 is fine, I really like it
I've seen a few grumpy people today and I understand, the V4 rollout isn't a good first impression for people who are coming into paid sub right now, hope the team can turn V4 around, V3.5 is good for now
you enjoy the wonderful artifacts in your instrumentals eh? each to their own
I have had shimmer in one track, the rest are great
I haven't had one without it in V4
that's not including remasters... don't get me started on that as well
This is the beta v4 and the list of issues with it is long. Sure, you can be optimistic and glass half-full. The user base largely prefers v3.5 over v4. Like I said, I'm pushing through. But it's not easy. V3.5 was sweet. This one is way worse in like 10 significant ways and way better in like 3 ways. So what we need is a final v4 version that fixes all the issues in the beta. Because I saw even huge suno users with hundreds of thousands of listens on this platform and they are like "I can't do my genre with v4 so I'm waiting till they fix it" lol. Not good but I hope it can be fixed
I was grumpy when I wanted them to hurry and release it. Now I'm like: never mind.
I don't know the internal details. The team don't reveal that otherwise it makes it easy for someone using their service in an API-like way to just change something to avoid detection.
Regarding the shimmer:
ChatGPT:
If you wanted to "hear" a digital fingerprint, you might artificially isolate and amplify the embedded frequencies or patterns. This could sound like: High-frequency static or faint, rapid clicks.
This was just one example.
If the new model was trained with tracks created by free users, the digital fingerprint (if that's what the shimmer is) could have transfered over.
earlier there were a few persons who came in which had the usual complaints but were saying that if it's not addressed they will be demanding their money back and cancelling, there was a debate about it being in Beta which I had to gently shut down, as I said before I get their frustration and I hope it can be addressed soon
so far my current opinion is v4 is fantastic for harder styles with a focus on vocals. aggrotech for instance was not really a doable style previously as it uses a lot of vocal style changes and 3.5 wasnt all that dynamic and always just make the voice gravely.
4.0 can do aggrotech.
on the other hand though i think 4.0 butchers styles that use vocals in the middle layer where its sort of melds into the instraments such as retrowave/outrun. it just forces a more extreme vocal as either clear or too robot and then on top of that seems to push instraments far into the background of the mix when singing.
TLDR
4.0 good for metal, rock, screamo, aggrotech but worse for electronic styles like retrowave
whats your opinions?
The community has noticed that personas was destroyed in v4. That's what depressed me. Along with the remastering being a failure so far. So far, it's a dud. But it doesn't mean they can't fix it, just that it's currently broken. It doesn't work right. V4 is very buggy in a way earlier versions never were. This is uncharted territory
did the personas ever work though?🙃
For me, without personas working anymore it makes it way harder to get quality tracks. It used to be you just make a persona from your best song and then everything that you produce with it is almost all exactly what you want. Now it's all gone so we're starting over.
V4 simply cannot do metal. Cymbals and snare are horribly mixed. Bass lines and bass drums shimmer like crazy. Vocals are sometimes decent, but suddenly break out of the mix either going to high or too low. I have spent several hundred credits today and not a single one of them has NOT gotten a thumbs down!
Just waking up — kinda silly how some users are upset enough about the new model that they’re attacking the alpha testers & guides who helped provide feedback.
Robert! Long time, dude! Hope all is well 🙂
yeah sorry i noticed that too. i do like it for aggrotech though as it never could do that before
hey hey 🙂
I mean... I agree, no one should be attacked, but I can understand their frustration
For most persona obsessed users. Worked insanely good in v3.5 but in v4 there's not even a real reason to use it because the vocals are all over the place. Forget making an entire album with that ai singer etc
why when i try use v4, it just givevs me a 3.5
there were many complaints that it wasn't even an option for 3.5 to V4 and/or vice versa
Am I guilty of that. I dunno. I hope I didn't attack anyone. I wanted to be an alpha tester. Still do, but when I asked by email they said no
are you a pro or premium user?
yes whatever the middle one is
Pro
communist_bugs_bunny.jpeg
Our frustrations.
All good, we’re more sociable and if people want to yell at us instead of the devs, they’re allowed to misplace their emotions where they choose.
i will say in some cases depending on the genre you can simply make the album in 3.5 then use the remaster a hundred times and eat your credits to get a decent upgrade. or only use it on some songs(generally slower songs with less instaments happening during the singing parts
I think some people have a glitch where they make v4 tracks and it gets tagged 3.5 do you have a 4.0 option
have you tried restarting your PC, closing the browser/app, clearing the cache and log out/log in
Remaster works on a song with piano and singing only. Etc
worked on my witch house, dark electro swing song quite well but other than that iv only had one other remaster turn out well. on the otehr hand i took some lyrics i made before for other songs and generated from scratch and it gave some great performances
beautiful point you make, it's the collective groups frustration, forgive me
id say over all v4 is really more like a sidestep. like using a different stable diffussion model based on the same data set
i put it like one step forward and two steps back... for now
I just made a bunch of cracks about MattVidPro and how a lot of alpha testers get paid to advertise it or else they get loads of free credit they get something, so I was doubtful guys like MattVidPro were having much of an impact on the process. But no hating the guy makes fun to watch content. Pay him but my god, looking at a guy who always has 35,000 suno credits in his account is wild.
I know when you’re not privy to going’s-on, it’s easier to speculate.
Can’t speak for all of the alpha testers, but let me know if there’s any content you want covered in the video tutorials section #1270928841614823546 Or the Prompt Experiments #1256301881966198805 (somewhat abandoned since I don’t have the ability to remove old bot posts)
I may not outright spoonfeed requests on how to create specific songs, but I’m happy to test anything that others might be able to use in their own songs and help them familiarize themselves with Suno 🙏
Re: credits — sometimes alpha users are given a small chunk of credits to test out a new feature (nothing exorbitant)
Personally, I still pay for a premiere membership, and unless I’m actively working on songs, I don’t usually use all of my monthly credits.
This means I accumulate extra credits the longer I’ve been an alpha member, but it doesn’t mean that they’re enabling me to make millions of songs with no limitations
I won't just yet criticize V4 to the moon and back. I mean it only launched like 18 h ago and it has a lot of clear bugs, like the muddy sound, lyrics go into wrong places, etc. I think in like a day or so it will already improve significantly.
I’m guessing that YouTube dude is in a VERY similar situation.
MattVidPro? lol jkjk
Love to the people with preview access and to the devs and mods. I went through a process of grief but then I made the decision to be positive and add requests for the final v4 version. It's reassuring to know the current version is a beta and there's more changes coming to the final v4
MattVidPro has 35k credits in his Suno account. Do the math that's a boatload. There's not a plan like that unless I'm missing something
But he also has 300k subscribers
Did you just read the second part of my message? I have a good chunk of credits without any YouTube subscribers
I don’t use my 10,000 personally paid credits fast enough — so the SMALL amounts of credits that the devs gift just keep adding
Premier is 10k creds... with top up option... maybe he is pre-empting or flexing... you just don't know some people's circumstances
do the personal credits get used after the monthly?
they don't roll over
Subscription credits first — THEN extra credits
So you consider 35k free credits a month a "small chunk" a good chunk" I'm just amazed by it. I'm not shit talking. I'm a fan of MattVidPro.
Who says “per month”?
im not asking if they roll over just the order
gotchya
thanks. i know some companies try to pull a fast one
Why are you being rude to me if you're a mod. Lol wild
I’m not a mod, and I’m not intending to be rude — I’m asking you outright why you assume it’s a monthly gift of credits.
Does he run out to zero then restock? Or does it just always seem to be at 35k?
If it’s the latter, I’d JUST explained how I was in that same situation by not using my monthly credits.
MattVidPro had over 30k credits in his account. I know for us regular folks we don't get access to rollover credits. However they hook him up, it's crazy.
Coincidentally, also named Matt 😉
do you know how he got them? I'm not understanding why this is a trigger?
I have 19000 credits right now
If you're in here being rude to everyone and you're not a mod, whyd they give you special access? It's interesting.
I was never triggered. I thought it was interesting
@Robert was given 10000 credits close to 3 months ago, and hasn’t used his monthly credits fast enough to use the extra ones.
correct
That’s why he currently has 19000, not because he is given 19000 every month.
I don't even use my 10,000 a month
so the "extra" is not reset monthly with the sub creds?
So the extra credits never expire like monthly credit that's nice. I wish they'd make unused monthly credits roll over
Where is the rudeness?
I'm sorry you feel like I've been attacking you. Earlier I shared links to my workflow guides and offered to provide any assistance that I could -- I'm just trying to clarify your misunderstandings about credits 🙏
Correct. You can buy extra credits on top of your sub credits. Extras cost more but don't expire.
WAY more -- always good to note that you're allowed to have multiple paid accounts, even for a single month (way better value, just remember to cancel before you're charged for any months you don't need!)
@unique stone tbh my guy I simply love your name, it elicits so many emotions in me, just tickles me the right way
I found harber a bit rude because he was condescending in the way he chatted. "Did you read what I wrote" etc. .... it's all good.
fun instrament for your prompts is to put the cowbell and flexatone. it gives really fun results especially great in electro swing, house and industrial imo
That's what they eat in the UK. Furry beans on toast.
I apologize that came out bluntly -- I had written that long message in bed, and had explained the part you seemed to misunderstand in the second half of my admittedly huge wall of text)
I had explained that simply having a large amount of credits doesn't mean that there's a regular influx of free credits -- unless we see that user had gone to zero and was refilled to that same amount -- but just having 35k as a total across a number of videos doesn't mean quite the same thing
@broken knot ya i got the v4 option
yes i saw that's an option for both pro and prem members, I wasn't sure if they disappeared or not, i bought some extra and used them pretty quick lol
but its tagged v3.5
Maybe you're getting v4 but tagged as v3.5 I heard some people have that issue
It's cool. It's fine. I love suno. But I also have autism and don't get along well with others.
there's plenty of people with that issue in https://discord.com/channels/1069381916492562582/1308499132356362383
I added some to the v4 feedback. If suno fixes what's wrong with the beta, v4 final version could be a gamechanger
such a liar straight off his fingertips... we have been discussing for over an hour without any issues
You mean I'm a liar cause I'm able to have nice conversation. I'm super empathetic. People with autism. I've been told I have no empathy many times. It's a different type of empathy but it's there.
I bet you're messing with me but I have autism and take everything literally and I can't stop
Will V4 final final come to free users@sysmai
Yes, v4 will come to free users but it's a beta and it'll be better for them to release the final v4 to free users in a month or two rather than dumping the beta on them
I'm also told I'm too blunt, have moments of no empathy etc... so perhaps i can relate slightly...
But yes, I was calling you out for using it as an excuse, it was just a simple misunderstanding between you two, as you say... early in the morning for yourself
Wake up, caffeinate, stretch and be happy❤️
AND Eat!
The earlier models took a week or two before coming to free members -- if history gives any indication, v4 will follow suit
this was very true for V3/3.5
How many songs do you guys write about dinosaurs, vampires, extinct bipedal apes, Frankenstein? I feel like this is all I write about. Do you guys try to think of weird genres just to see if Suno can do it. Trying to make stuff sounding like Disney music. It's better on V4 than v3 if you're specifically going for that Broadway/Disney Pixar sound. V4... not bad.
@charred merlin Has anyone tried to make a song utilizing the shimmering noise / lazer? 🤣
Certainly, that's a great idea. Do it
Don't want to steal your idea but it's good
someone send me their best shimmer lazer featured song
So you should definitely do that. The v4 shimmer song
If you have access to V4, have it it.
I actually saw something earlier were they had some very invoking style tags
Not sure if this is the right channel but Here is something I am wondering is do we think Suno will at some point all the selection of voices and maybe even allow a custom voice
why can't u stop telling ppl u have autism?🤔
...weaponizing AI's tools right back at them... we'd be unstoppable.
When people get mad or question about what's wrong with me. It's best to just mention it. So people understand I'm a freak
bro... i swear if this is a flame...
I have severe mental illness. That influences what I do. I didn't mean to throw it in anyone's face
u basically want them to justify ur freakiness 😅
Validate*?
idk maybe
Not at all. Just, mention it if I think I might get ganged up on or something. To ask please don't
it's cool I'm just curious, like genuinely
that's like a moth to a flame for trolls man.. i swear
Seem like V4 is not getting the drum right, it sound like the drum is broken. Does anyone have similar issue?
can you elaborate on that a little more please?
there's definately issues, but like what exactly
I made a persona, when using it to make a song it wont follow my lyric's. It uses the lyric's from the persona 😭
Or if someone responded and said " most of is have autism" that sort of comment riles me up. I become like Chris Farley in Tommy Boy. "Now the pet is my possible sale. Hello there pretty little pet, I love you. And then I stoke it, and I pet it, and I massage it. Hehe I love itthen I take my naughty pet and I go [makes ripping noises as he tears apart the dinner roll] OOOOOOHHH! I killed it!"
So, the shimmer noize/ shimmercopter /pewpew/ whatever sound. is FANTASTIC with hardstyle 🤣
common issue, even before V4
Anyone have some V4 prompting tips? Feels like it only uses the first style tag I give it and just ignores everything else
Not sure how to explain it.:
https://suno.com/song/8230d18c-6786-4f9e-a70b-540e326e80f8
Sad emotional chillstep, reflective and deep, with soft piano, clean basslines, clear synth tones, and precise soundscapes. song. Listen and make your own with Suno.
I have prompting tips. Always use all 200 letters up. Don't leave any empty space in prompts.
That's what I do
the style tags are in order, the ones you put first are prioritised more
I have always found to use a few as possible. Simplify what your looking for in 3-4. That's what I do and it usually is pretty good.
are you referring to the extra reverb?
I do prompts on udio that are paragraphs and paragraphs of prompts. Being held to 200 letters is a challenge. Make sure you mentioned if you want a baritone or soprano etc
sounds way more clean than anything i've been able to gen
Will try to use less, thanks!
But use ChatGPT or something and ask for a 200 letter prompt for exactly what you're trying to get. Then slightly edit and improve upon that prompt.
Not sure but it sound a bit weird when I listen to the drums, around 00:47 and forward when there's drum sound.
can probably master that in another app and it'll be clean, it might have the least amount of shimmering that i've heard today
Simple is one way to go. Filling up all 200 letters give me more control than putting in like 1 or 2 words. Both can work. I definitely prefer using all 200 letters
also this one start at 00:27
https://suno.com/song/873d4dab-4665-4bdc-bcba-29e2096998ac
Sad emotional chillstep, reflective and deep, with soft piano, clean basslines, clear synth tones, and precise soundscapes. song. Listen and make your own with Suno.
it's given a static sound to the entire track, maybe try a prompt to counter that like, [no static], [High quality] etc.
It doesn't sound punchy, like a loud speaker was about to break. Sorry @knotty skiff , My English is not that good man. So hard for me to describe it
all good, yea it's the reverb you're referring to, honestly it just needs mastering... the static element to the song at around 0:48 sounds intentional, overall the levels of reverb are high and might clean up nicely
Thank you. I have been trying to figure out how to prompt for clear sound but couldn't find the good words to put in negative prompt. So hard to take the washy or statics noise away from the song
no worries, I saw it recommended earlier by a guide member
it'll go to free users in a week or two if they are going to continue how they released V3/V3.5
Ohh i hope so too then
Idk why i got 1 Free V4 Remaster lol i hope i get 1 free everyday until v4 goes free i guess... Anyways. Thanks for Clearing my Doubt mate!
Can you suggest any remastering daw or app? And which option in there that I can use to fix this?
no problem
I normally try to make a new song if suno gen came out bad, I burnt 10k credits a month. lol
truthfully that's a bit complex and it's very opinion based and needs based, I recommend that you look at researching what you need, MVSEP was recommened in here earlier
Cool. Thank you.
If you want to (re)master a song. Any DAW will do it. The issue you may face is the knowledge of the tools required to do what it is your looking to do.
There are also some programs that use AI to help mix and master music. Good ones tend to come with a price tag.
Thanks Dylan. I tried bandlab, but it just make it louder, when it's louder, the statics also comes with it 🙂
Hi I need help.. V4 no longer takes my upload vocals but changes them completely so that it no longer sounds remotely like me
V4 is a bit strang for me too.... no matter what I try it is always a female voice ... is there a trick to get male voices (I put [Male_Voice] everywhere to no avail)
yes V4 dont use my Voice.. it will be female.. ord if i remaster my voice sounds not like me
Yeah same issue for me buddy. I always use bandlab but it just makes the song louder. Quality doesn't change. It's the same with LANDR and few other Mastering sites i have tried
V3.5 is awesome but only just if there is a way to increase quality..
So from me testing around with stem spitting, it seems like the "Shiny" Sound is coming from drums in generations.
My problem is I can fix it by changing the key ... but when it is a female voice it is sometimes changing it too much ... I make covers ... but all of them are female voice ... not even alto ...
Yeah man. I stop using Bandlab. I notice the original song from suno sound better with bandlab, after I listen to it carefully
Yeah.. i tried V4 Remaster Free one today. It is so awesome i wish they provide 1 Remaster everyday until they release V4 for Free Users...
But yeah anyone if you have any suggestions for best Remaster website.. please let me know..
own voice cant be persona, and if i will make a new song or remaster or cover or extend my voice is not my voice.. v4 is not on point now
I heard it in the instruments section, not the drums.
I've been saying for a bit that it seems to be triggered by cymbals / crowd noise / peaking -- it's like the AI is mistaking the ring of a chinese trash cymbal and just wants to turn EVERYTHING that gets reasonably close to peaking like they do into a cymbal in turn
Try "Male vocals" in "Style of Music" section
Does editing with Audacity work increasing quality?
Audacity is a good tool ... but I am sure there are some better ones
thanks Paslejektoja
This could be it, it seems to show up in the cymbals a lot more
Hmm..
I concur, it only starts to click and have artifacts in background when the drum stem is starting or has started (engaged)
Im attempting to find a fix to remove these sounds for me, probably with some EQing
Hardstyle. It makes the shimmer seem apart of the song. Sadly, that means everyone wuld have to make the same genre and not have fun, lol.
old school
Gotta keep it simple :)
ok i'll stop trying to make a male version of that song ... I will use all my credits
starnge to me that the default voice is female
It's gonna be a tricky one, I feel. Occasionally I notice it during the low-to-mid-range sections, but it comes across more of a minor waving/warbling effect. THOSE I could live with, but I'd love to hear any eventual insights from your testing, even if it's just what didnt work as well as hoped 🙏
use "Male Vocals" in your style prompt works very well for me
I wish it was as simple as throwing a low pass filter on it, but without killing the high end of the cymbals 😦
Yeah, its a shame, its quite difficult to remove
I tried ... with [, without ... lol
No... not in the lyric prompt, in the style prompt
yes I get it ... I will try again one more time ...
Is 'riot grrrrl' a play on "I know this pretty rave girl"?
go fish
I am lost about how to use meta-tags for sound effects. Is there a comprehensive list? And can somebody show me an example where something like [shouting] worked. I mean does it matter where it is put in a line or stanza. How can it be more consistant?
Thank you AweClaw ... works in the song creation part ... not in the cover anymore ... but thank you so much for your hlep !!!
it's all AI based, sometimes it picks it up and runs with it, sometimes it ignores... it's really down to the generation and how you use it
I've seen others discuss this and they recommend just writing something simple about how you want it to sound in [ and ]
I've also used ( and ) the signs for back up vocal to be used for shouts or distortion
no problems
Hi! I just noticed I can't change going from V4 back to v3,5. It says v3,5 in the creator window but it creates in v4. Anyone experiencing the same issue?
it has been known to happen, copy what you have in prompt and log out/ log in or close the window all together and try again
I would say V4 has better sound quality KingRotten ... but it strange you can't go back still
Thanx a lot, I'll try that then
Error code 422 I’m sorry if this is asked before, but I have been getting this error over and over and over since last night. It just started again, but it won’t let me do anything.
ok well i found out why the ai broke. i shared a song in suno showcase and like 2 people went through every one of my public songs and liked them all without listening in the hopes i'll follow them back. i mean i think they are all good but if you havnt listened then you could be liking garbage made for like farming or something and training the ai on trash.
maybe report in the https://discord.com/channels/1069381916492562582/1082775596767117312
I won, but at what cost?
I notice that when V4 generates instrumental tracks, there is always a strange buzzing sound in the background. However, this issue does not occur when V4 generates vocal songs. Does anyone know how to solve this?
you solve it by using 3.5
so far 30 minutes... what's that? like $20?
I guess so
I'm trying to figure out how to diminish the metallic sounds but its a struggle LOL
it's a known issue and the Suno team, and others are trying to find solutions
Such as Matt here
I better stop and start working on my college Graphic design homework, I don't have time for this lol
age old procrastination... I got many an A+ turned to B+ because handed in late because I'd rather play LoL than do my Assignments on time
Lol
I'm not sure about that actually. Yeah, it's very strange and I've tried but it still creates in v4, which is very annoying.
There's no comprehensive list because nobody really knows all of the sfx that are in the training data OR how to reliably trigger them. I've done some testing with several here:
https://suno.com/playlist/49a40042-a09a-421b-9df1-3f6d976221f9
Some are fairly reliable (especially [*beep*]) but Suno can and will ignore any of them depending upon... the styles for the prompt, the surrounding lyrics, the alignment of the moon and stars, etc.
common complaint, sorry that's happened try and look at the feedback channel and bug channels for support, the voice generation has been changed significantly and you will find that across the board not just with the personas
I tried that also
Hello everyone, how are you? I'm having issues creating songs using Suno. When I ask it to create, it loads but says it couldn't generate the song. I can only remaster songs.
I'm Helton, I'm from Brazil.
Hello, Suno team, iI want to ask about a little detail about billing. I had a friend pay for me and now that it is done, i want to delete the billing information of my account. I already cancle subscription but still cant delete billing info bcause it says it cant with current plan active
Hi @full root, discord isn't an official support channel 🙏 You'll need to email billing@suno.com
is there limited use of v4? all my v4 gens are now coming out at 3.5?
ok
If the message says "Song stuck in generation, no credits have been used", we've been seeing that in relation to something in the lyrics or style prompt being blocked by the automoderator. https://help.suno.com/en/articles/3198209
We support a welcoming community for anyone to make music, but inappropriate content may get flagged before you can generate a song. At the moment, so...
Other users have reported seeing the v3.5 tag on their v4 songs, also seeing v3 tags on their remasters -- as far as I know, this is a visual glitch, but the devs are aware 🙂
I'm sure many has brought this up, but having the personas only produce result with V4 is really annoying. As the vocals are not close to the original persona as it was in 3.5. Is it meant to be like this or is it a bug that will be fixed?
thanks. but it sounds 3.5 too
yeah i noticed it ignoring those 99% of the time. no wind, thunder, rain, or aztec death whistle for me i guess T~T
is it possible to generate a baseline track only given my own audio clip?
Thank you, bro, that's exactly what happened.
Hey guys!!! 👑 Just want to congratulate the team here at SUNO for the release on V4 the song structure and over all sound is phenomenal! Only thing I have noticed is that there is a strange SKIPPING/GLITCHING noise that can be heard in the background of certain genres and intrumental versions of genres such as LO-FI or Metal etc. It is a weird glitchy sound that isnt really heard if the same song is created using 3.5. These strange artifacts that appear make it sound weird.
Sometimes it can be tricky to figure out what is being blocked, so it can be helpful to use chatgpt to paste the lyrics / style prompts & ask it "what about this would be blocked by an oversensitive moderation program", and then asking it for help to rewrite those sections 💪
Best of luck with the songs, dude!
The shimmer / cymbal noise?
yeh V4 is amazing.. some of my gens are ridiculously good
Yup!!! It is so overwhelming in chilled songs and in metal/rock songs 🤢
Hi everyone,
I've been experimenting with Suno and was wondering if there's a way to generate vocals without any backing instruments. Does anyone know if this is possible?
Its super noticable when using headphones
Yup. I've been seeing if I can just make some songs that sound good with it in the meantime.
"The Shimmer" is physically represented by;
- Jazz Hands
-OR- - Spirit Fingers
Tech-Chat decides.
For nature sounds, you'll have much better luck with v3, but you really need to prompt for nothing BUT those effects, or you'll be doing a CRAPLOAD of gens to get anything. That's where extensions shine. You can start with v3 to get rain and thunder, then extend from that and switch to v3.5 (or v4) and extend from there for your music or spoken audio.
Here's some Rain and Thunder: https://suno.com/song/31a91d5a-1222-4a13-a1fb-6c3df8ca7ccd
...and a few birds (lol): https://suno.com/song/7f8c9b03-31b1-4ced-bf61-85a8ae57002e
For both of those, I started with v3 to get the fx and then extended with v3.5 and the music prompts.
Rain Storm, Thunder. Wind, Doom Metal, Horror, Ambient Noise song. Listen and make your own with Suno.
angry birds, tuba melody song. Listen and make your own with Suno.
I spent all day yesterday remastering my songs when I was out and about and didn’t realize the laser sounds. When I got home last night and put earbuds on it was like the only thing I could hear. Now that we’re aware of it we can’t justify using these generations
I was going to say the same thing bro, these laser sounds on the background is very very bad and annoying
Suno fix these asap pls i just literaly paid just to find out this mess
And no free remasters like when they released covers
its alright ill throw it into fl studio and add effects liek that in there instead. easier and more importantly wont cost a million credits
You can slightly, overpower the shimmer noise by using genre's that are overwhelming with sounds, or with genre's that work with the noise. https://suno.com/song/83a6e8b0-e46a-42ff-9090-d58cb8063c26
electronica, electronic, hardstyle, song. Listen and make your own with Suno.
Ears be warned
If I use "cymbal" in exclude area will it lose the ever present shimmer stutter? Hmmm
It seems like Suno uses the shimmer almost like white noise to fill space.
THANK YOU to everyone making liberal use of 👍 / 👎 in these early days, and remember that you can flag particular egregious generations using ... > Report > Bugs > Poor Audio Quality to help improve future versions of the model -- this isn't likely to impact your short-term generations, but contributes to the eventual betterment of the model for everyone 💪
There was a NOTABLE difference in v4 between where it started with alpha testing to where it stands in the beta version -- I can't wait to see what the team is capable of once they parse the scale of feedback that this launch has provided them with.
At least this song I wrote, the mass laser shooting isnt too annoying... but still 😫
I tried v4 beta. 300 credits worth than stopped. I made some good songs, but nothing is usable because v4 is really in need of some dev love. After using it all night I'm giving up on Suno temporarily, waiting two weeks, hoping they release a v4 patch, then switching back to 3.5 for the rest of the month if I have to. I'm not satisfied with my v4 results. I gave it a fair chance. The weak arrangements, instrumental issues, and warble/shimmer issues are too massive. I can't just use up hundreds of credits. Especially because they will probably patch it in a couple weeks. I have until the 18th to use my credits I can't waste them on this version of v4. But I'm not bitter.
Heavy metal 2000s, melodic, hard rock, gritty screams and melodic clean vocals, double bass drumming and ghost notes. song. Listen and make your own with Suno.
It can produce laser sounds
It can add this cool warbly effect to the vocals
@native rock ☝️
Will I get WAV files if I subscribe??
so here's a fun fact: The clickign existed in 3.5 as well, just no one noticed it
I'd still love to know why personas do not follow new lyrics
Yup -- downloading .wav files is available for paid subscribers 💪
another thing for Suno to revise, new songs made with v4 only goes up to 3:14
thats a big fix needed
afaik, the devs are aware, as you're right, the limit is supposed to be 4m
i do suggest to have the ability to goes up to a max of 5minutes
That's more a matter of #💡┃suggestions-web since the current limit is 4m, but that would be nice.
This is a song from Nov 17th, so it's not a permanent thing, and likely stems from one of the late changes made to v4 -- I expect the 3:14 limit will be resolved before long.
https://cdn.discordapp.com/attachments/1270928841614823546/1270928991947063386/multiple_song_selection.mp4?ex=673f411c&is=673def9c&hm=761d795aaf7a9729ab8002594eb9ef316614432b54dae575566bb4b79ed61832& You can use desktop selection methods to highlight more than one song at a time -- @naive quail showed me that you can actually highlight multiple pages at once and delete a LARGE number of songs at once.
I'm going to call you "Cotton Eyed Joe", kind stranger -- because I have no idea here you came from, or no idea where you went
"showed me that you can actually highlight multiple pages at once and delete a LARGE number of songs at once"
Multiple pages? HOW? xD
Doing lot of replace segment this is a must have
1 sec 🙏 gotta find some pages I want to trash 😅
btw, i hope they'll change the way they take credits for replace segment, it will kill my credits in few days if it stay like it's now T_T
It's still a more tedious process than it needs to be, but it does work:
I just made a few instrumentals that got 4 mins. So either it's fixed or it's an inconsistent issue
with personas
i just did as well with persona go all up to 3:14
Oh, that makes sense! It was my persona stuff that got stuck at 3:14 🤔
@prime anvil
I read articles claiming sound effects were massively improved on in v4. Not that I can tell. Trying to trigger sound effects into my suno tracks has not worked. I'd love to be able to add rain drops, thunder, etc
You're a life saver!
ctrl+shift is the key 😄
sorry for the gif, can't hold inside the joke xD
where does one access the remaster feature?
Under a track, under the "create" tab
Click the three dots next to a song and then under the create
But go easy on the remaster feature. It's not ready for prime time don't eat up all your credits. They need to fix the feature.
I wonder.. the system is having issues at the moment.. but I made a really good song.. I wonder if I try to remaster it tomorrow or next will it fix the static
yeah, used 1k on remastering, kinda cool, but just try it once for the fun
the shimmering and some artifacts, like distorted sounds can be a pain
man V4 has some serious issues with leaving otu words, mispronunciations and now for some reason i cant get it to finish a song, it just always stops at 3 minutes 14 seconds even thoughthe song hasnt ended.
personas is active
I'd hold off a couple days and see if Suno releases a patch. V4 "remasters" don't come out right 90% of the time. The arrangement get worse, weird noisey tracks. I bet they patch it in a couple weeks or even faster. It's really bad right now.
I made this it sounded good earlier, now I notice is sounds like garbage throughout the whole song: https://www.youtube.com/watch?v=9OU12eqsa78
Classic 1980's rap song turned ska.
Check my YouTube channel for other fabricated stories and songs.. Gorilla Biscuits as a pop band? Sick of It All before Lou joined the band? Samiam sad song turned happy. More to come. Suggest me some interesting songs to make.
I just made another one that's really catchy, but filled with same issues here and there
static, freezes here and there, etc.
Yes, don't try using the word "iron" v4 pronounced it "I ron" ... I ron, who you be??
we have a singer called Ron in Italy 🙂
arrgh every song is 3 minutes 14 seconds, why why why....
Hexx, you have Personas active
guess it only saves so much in my history
I can't go back to grab the beginning of that song I don't think and try to remaster later
nevermind I found it was looking in the wrong spot
Just so you are aware, you are not allowed to use Suno to make covers of songs you do not own the rights to (this goes for using other peoples lyrics, and uploading audio.)
yeah I don't care.. but anyway, they announced a new feature to make covers.. wonder what that's about
"Popular features now supercharged by v4:
🎶Covers: Reimagine originals in new styles"
I checked him out. That's chill.
You should care. You are breaking the Terms of Service by doing this. Covers is for making covers of your own music.
"covers" is the name of the feature -- doesn't give you the right to infringe copyright (or be rude to the community guides just trying to point that kinda stuff out)
Maybe they should be more clear on the feature then
Even cover bands need to get permission before doing it seriously (or pay mechanical license fees)
people use this feature because they're not musicians and haven't made their own songs lol
there's tons of people on youtube doing it
I'm not disagreeing that others do it, but that doesn't mean it's something that's allowed by Suno's ToS, and the users in Blue are just trying to help keep the place tidy
yeah I understand, but I could care less.. I'll just get rid of the service and save my money.. it doesn't work properly anyway
Thanks for stopping by 
oh that remaster feature isn't what I thought it was.. it just made the whole song different
Yes, but if you are looking to use public doman songs both the composition and the sound recording both need to be public doman. If the recording is public doman, the song might be copyrighted as a composition. Every January 1st new public domain stuff becomes available. Be careful. If it's not public domain, we don't "own" it or have right to it. Even if it's a demo. I did extensive research a few hours ago.
For example, this year, 2024 the composition became public domain so we should be able to reuse those lyrics all day long. The Annette Hanshaw recording is also public doman, now that the composition is also public doman, technically we all "own" it if we want to. Anyone can use it.
The recording is usually the kicker -- using your own public domain stuff is a "safe grey area" in regards to the ToS from my current understanding, but sometimes that's misinterpreted as anyone having the rights to any unique recordings of a public domain song
"because it's public domain, we all own it, so I'll upload a recording of someone else playing auld lang syne since they can't own that" sort of thing
✅ Warning logged for summerpeaches06. I couldn't DM them. || Please do not post the same message in every channel. This is considered spamming and is in the #📜┃rules, @cunning crater
https://suno.com/song/2aba742a-68fb-4711-b594-3138ea870a6e - I did many sound effects in this song. To be fair, making them is confusing.
In v4 I haven't been able to get any sound fx
Yes, only stuff from 1923 or early is public doman as far as recordings, for the most part. But not always. Thankfully, some Annette Hanshaw stuff is now 100% public domain as is some Al Jolson stuff. Some Gershwin is entering public doman in January. The song Rhapsody in Blue along with the 1924 record. So I'll definitely sample that.
I have a song from around that date, and seen a song from another user (one was May 18, the other May 20 I think) - this is just before v3.5 was made available. So it's possible that they did something around that time which tagged the songs with something other than "v3.5" but which is now being interpreted as "v4".
In 10 years, public doman accessible sound recordings are going to be flooding the market. As for right now, we're only going to be on 1924.
Has anyone else noticed an issue with songs since the launch of v4 yesterday? There’s a consistent background noise that disrupts the quality of the music and detracts from the otherwise great tracks.
We all noticed. I'm pausing my suno use for a couple weeks in hopes they put a fix out for v4.
Great, thanks. I have just payed 30$ yesterday morning for 10000 credits.
I don't hear it as often with acoustic folk songs but not everyone is into that genre.
The big issue is the weaker than 3.5 compositions and the weird noises and the weak instrumentals of v4. Besides that, it's great.
Yes, I love it, but ut ruins almost every second song I create, I can't use them.
Depends on the track. Sometimes the vocals sound insanely warbly on v4 even if it's a stripped down track, but you have a better shot of getting something decent. I think people are not wanting to notice the "shimmer" aka the warble. If you tell yourself its not there, it could work as a placebo.
Is there any way to remove it with another app? I'm so disappointed that I won't be able to use many great songs because of this. At first, I thought the issue was with my headphones.
It is a constant. I've tried cleaning it in any way possible inside of suno and I have not been able to pull it off. I can't control something that can start on the first beat
After I write several v4 songs, it suddenly force output v3.5 even I select v4. and have a water mark on the cover.
Thanks 😉 I just want to create something that sounds as perfect as possible. For Spotify.
You can do things like reverse flanger and Phase correction but since it goes through all frequencies it is not able to be removed.
😭
Ah came to check if anyone else was having issues with weird background noise with v4 and it seems I'm not alone. A shame, it's so good otherwise.
They will fix it
Where do they release updates? I'm so new to SUNO.
I'd suggest sticking with 3.5 for now, but v4 has some crazy vocals.
One method that works on any suno output is to "remix" on udio. You have to do it in 2 minute chunks and splice them together, but the results are impressive
The majority happens in the background
The likelihood of them patching v4 in the next two weeks is sky high.
Vocals and everything is better with v4. Except this, hope they'll fix it quickly. I sent them an email.
Yeah, this is gonna be a priority.
I'd say so, this is a glaring issue, that and pronunciation that may be tougher to fix
We can just phonetic around the pronunciation stuff though
I also noticed that when I create a track, it shows a length of 4 minutes but stops playing around the 2-minute mark. However, when I download it, it plays perfectly.
V4 is incredible. I look at it like a toy behind a window to a kid before Christmas
can official team please fix why my V4 gens are coming out as 3.5?
Same here! 😉 I discovered it on Sunday, bought the regular version, and then upgraded to the 10,000 credits version the very next day. 😄
For me it goes 1: the weaker arrangements. Worst issue. 2: the weird warble, shimmer noises 3: personas and remaster not working properly 4: The pronunciation issues. That's the 4 big things that need a quick patch ASAP. If it can be done, v4 will become the goat.
Welcome to Suno.
I’ve noticed all these issues too. When I create songs in French, some words are pronounced very strangely in one track but perfectly in another. It’s quite inconsistent
There are some strange weights going on in lyrics. I believe that is a heavy part of the problem. They focused on vocals so much that a tradeoff happened. But the vocals can go crazy in short runs. Crazy in a good way. If they get this fixed up, then I think it has the ability to overlap two separate Vocals. Like a Broadway. Two separate vocals singing two separate overlapping lyrics.
Spell them phonetically to get around that
I notice covers works better in v4, it seems to respect more the prompt and stop re-spitting the music
Version 4 has difficulties to follow melody structures when making a cover. Often bad notes
You can use chatgpt or something else to do it if you don't feel like it
3.5 was amazing at arrangements if personas was being used correctly. I could just get it to spit out stuff that was perfect. v4 is a losing battle we can't win in the current version. We just need those 4 fixes I bet 90% of us would be really happy with v4. But it's a work in progress. Don't waste credits on remaster till they patch it. I'm really excited for what it will be in a few weeks/months. Ask not what Suno can do for you. Ask Suno devs what we can do for them.
yes I had tried it one with 111 😉 it worked but not every time
Just gotta let v4 work itself out. The magic that happens behind the seens
Cool. I have had gpt rewrite an entire song phonetically. It takes more characters but allows you to make different vocal runs
I didn't try covers in v4. I have temporarily paused usage, assuming a patch will happen in the next 3 weeks and I'll save my credits for after the patch.
How do you mean the entire song?
As in, every word spelled phonetically.
But if covers can actually do covers now? That would be massive
V4 is baby...it is still learning itself. LOL. I wouldn't do much on v4 for a while (I won't be) I switched back to v3.5. I'll only use 4 if I want cleaner vocals.
What I did last night was, songs that the vocal style was good, but the audio was dirty, I had v4 redo the vox and I put it in the song. it lines up with the music so it's fine, the remastering is literally a remaster, not a redesign of the song.
v4 has been pretty awful making songs. I blew thru 100 credits last night before I realized that it is not ready. all Drum and Bass songs were rock. ALL of them.
Will remastering into v4 keep the original song?
and not good rock btw...was generic awful crap you'd hear in an elevator or dept store
I have the same issue. All of my v4 creation/remaster attempts gets v3.5 tag only
Yes, it just brings things out more. I have yet to see remaster change the songs structure for my music but...mileage may vary
Anything I made in v4 I'll either redo it from scratch once v4 is patched or I'll use the remaster feature if it's also patched. I'm just going to assume they will patch the heck out of this sucker.
I had this problem yesterday. are you making brand new songs or extending? if you are extending it will be 3.5 if you made it in 3.5
Within weeks. Even before the final v4 version drops we should expect some sort of patch or hotfox
I replaced a part in one of my songs, but then went into the newly made full song and realized I wanted to replace another part. But now the timeline doesn't match up with the lyrics displayed. What do I do?
yeah it's super early. it was nice to get in and see what it may be capable of, but it's not there yet. it's not even close. and I'm not going to bleed thru credits to get it there. I'll just wait. 3.5 is fine.
Imteresting... I generate a WAV for download and nothing happens. Anyone else getting anywhere with WAV downloads? I'm assuming it's just a traffic issue?
Im getting wav files no problem
That isn't normal you should have no issues downloading wav files
I've tried everything - starting a new song from scratch, extending a song, cover song, remastering the song, using a different browser, switching to incognito/private mode, and even using my phone's browser, but nothing seems to work.
I know. weird. Maybe a restart... or bang the screen! haha
Let's start with a reboot...lol
I wouldn't say "it's not close" it has 4 or 5 major issues. If they can fix the genre issues that are exclusive to v4, sucking the life out of tracks, dull arrangements, noise issues, then it will be awesome. Especially if they patch remaster and persona features. They do that and we're back in business.
Maybe we could get v4.5 à la v3.5 being an improved v3 🤔
Those are pretty major issues, no? I mean the songs have no life in them, the remasters suck the energy out of the songs. Oddly I have noticed that some of the stuff coming out of 3.5 is cleaner and sounds better but that may be me halucinating.
Hmmm I am at a loss, open a support ticket
We could hope but I'll take a 4.1 at the moment. They can revert to the drift that was in 3.5 and I'd be better off with that because it was controllable.
I mean @broken knot those issues are so major that If I was their producer, I would not have allowed this to release.
I mentioned "major issues" but hopefully they can figure out the root causes of these issues and patch v4, then it will be beautiful.
I may have to just extend from right before where I wanted to replace in the original track and just fix it all and generate an extension and make a new full song. As I don't think it will re-sync for some reason.
I assume there was a decent amount of pressure after partnering up with big names
no I saw you did. I'm just agreeing with you, I would have held it for a few weeks (or sprint cycle) and then put it out.
They should have held off another month, but maybe they needed the feedback from the subscribers. Better a late hotfix than never.
mmm pressure sure. but if this is your baby. your main product. you don't shit the bed to make partners happy. that's a great way to loose "users" and get a bad rap, social media news travels fast
Now they know exactly what to work on for the final version
The issue is training a model. Some of this is code, but a lot is in the model. I don't know how much training is needed for an audio model, but I would assume it is huge because of the frequency range of things like cymbals.
I already went through the 5 stages of grief. You're only halfway through. Lol
at least...I wouldn't shit the bed, I'd tell them to wait. that we got one shot at this and while we can hotfix the hell out of it. I wouldn't want to do that either. we do that now on a mobile game I work on and players are never happy when we do. and to me...the players are more important that the shareholders
training an AI model for audio is a lot of heavy lifting.
I know, that is what I mean, so a "patch" is more time consuming at the model level.
Yea, I have sent e-mail to support already.. and opened a "ticket" to #1308499132356362383
Yeah v4 really needs work! Getting a totally different style from my usual music genre prompt and im getting this really wierd warble on every song in the background like a car is idling or something. V3 and v3.5 has the same issue where in songs it sounds like the song was being played through a fan a bit but i looked past that. But yeah v4 still needs alot of work. But considering this is like a day 1 thing, im gonna give it the benefit of the doubt because all day 1 things come with bugs. So hopefully this gets fixed and improved
I used suno v4 all night and pulled an all nighter so I had time to process the shock of all this, now I'm just looking forward to the next patch and focusing on Udio until Suno gets the hotfix... I may write a song "the shimmer warble garbage arrangement v4 song" someone suggested it as an idea. . Embrace the madness
for now, I'm going to use v4 for updated vocal audio. it works like a charm. 3.5 makes better songs and with better accuracy on what I want. maybe I will check out v4 again in Jan to make new songs.
Right, v3.5 has controllable drift.
Udio is hot garbage.
oh you got the warble too LOL first 5 times I was like "what in the F is that?"
that's when I knew to back off.
It sounds mostly like a phase issue. The timing is getting misaligned off the jump
No, it's not. Not if you're an expert at using it. It can do amazing things. It's difficult to prompt unless you're an expert. You can make prompts 5 pages long
For example, you can remaster Suno tracks in Udio and it actually works. That's a great feature.
I'm not under the impression that's a "support ticket" -- just another way to share feedback
3000 character is the lyric limit on suno. 200 letters is the genre style of music box. It's sniffling. But I write insanely long prompts in Udio and get authentic styles Suno can't do including legit sounding 60s music, wall of sound, it can do perfect Seatle grunge. It can do so-cal 80s punk. Suno can't do any of that.
Udio has no limitations on prompt size so I write prompts that are several pages long
And the stuff I produce is astonishing.
Well now, if you do say so yourself...haha
Also, it's udio v1.6 no telling how massive v2 will be
As far as it's accuracy to specific subgenres, it's astonishing. The fact that their data set is trained on stuff like The Germs makes a big difference when you want to capture authentic so-cal sound
Sounds a lot like a conversation for #🗣️┃chill-chat
It's not even a conversation I'm done mentioning it.
then I maybe suck at it. i've had 4 good usable songs come out of it. most times it spits out noise or the audio sounds a bit off. like it has noise (especially the cymbals or high hats)
show me how
Does anyone know how future generations might be affected by songs that have previously been generated in my library? I'm wondering if I need to periodically nuke my library to get a fresh start to avoid material going down the wrong direction. I just noticed that I wrote a silly song about cats because I sometimes like to hear those. Then I shifted work to a completely different creation and the thumbnail had cats in it even though it wasn't appropriate at all.
The 32 second model is way better than the 2 minute model on udio, it's also difficult to prompt
Maybe you used the 2 minutes model and didn't tweek your prompt enough
that may be it. but I also don't want to go 32 seconds at a time to make a song man, that's a pain in the ass. or at least...sounds like one. I might as well just do it all myself using Native Instruments and my DAW if i'm going to work at that pace.
as that is the same thing...actually I work faster using NI and my DAW
@sunomusic Suno V4 drops soon and is set to drastically improve audio quality.
But why wait when you can enhance your Suno creations now?
Here’s my secret workflow for how to upscale your favorite Suno tracks today:
That's the tutorial to do a proper Suno remaster
Udio is an entirely different implementation and likely doesn't handle prompts in the same way. It is also not Suno, which is what this channel is for, so discussion regarding it (or workflows involving it) should take place in #🗣️┃chill-chat
I'm holding off on premium udio until v2.0 but I don't want to bother anyone, so no more udio talk in here. But I'm a huge fan of both.
Suno does not intentionally remember anything about your previous creations when you generate something new. It is possible that something could be stored in your browser's memory that is not being cleared out when you switch to making a new song. If you find this happening continuously, see if doing a complete page refresh or using an incognito/private window stops it.
This is how I feel Suno is designed too...but without any specific documentation, who knows?!?
I asked one of the devs about this previously, and their response was along the lines of "I wish it could do that".
Thanks. I hadn't thought about it being a browser cache.
Negative prompt and lyrics also seem to be "seen" even if you switch off exclusions or go to instrumental... I had a "lyrics too long" for something that had the instrumental toggle on recently.
I've also noticed things like where I specifically ask for a trumpet solo...I won't get it. Then low and behold a few generations later, there's a trumpet solo.
How can we know if any dev know and work on our bug https://discord.com/channels/1069381916492562582/1308503194938118164 ?
I honestly feel like library gets biased while I'm working through an idea where when I try to shift gears drastically, there feels like there are residual artifacts from previous prompts...but I could just be finding patterns in the clouds. It's so hard to tell with music what I'm perceiving with a high degree of objectivity.
Pretty common practice to not reveal trade secrets for closed-source projects 🤷♂️
The devs typically don't respond directly to reported bugs and are likely looking at the large amount of feedback in response to v4 so they will likely see this and look into it. I think I've seen this mentioned a few times in different places on this discord so most likely it's a known thing.
Hi guys, i'm new to discord. Is there a way to download the entire library i've created with lyrics and mp3 or wav instead of one at a time? chatgpt allows it. anyone can help?
Suno is messing up again!
In so many ways! ONE) Tons of song rendered ages ago, now turn down, in the actual file and no upping volume helps, and distorted.
two) non of the music tags render correctly.
If Suno does have residual memory, I'd really like to be able to launch up a new instance, maybe something called "Song Space" that had a fresh start. It would also help to create a kind of "folder" to keep material being generated for a specific song or project. The "Song Space" could have global settings with a separate prompt to do unique tweaks to each generation.
That's not like a "major" bug ? Paying customers cannot select a model, the server auto fallback on v3.5 whatever we select (and it's not a browser bug, it's serverside) and it's cost credit anyway
The amount of work and finetuning for a hidden feature
Suno is messing up again!
In so many ways! ONE) Tons of song rendered ages ago, now are turn down, in the actual file and no upping volume helps, and distorted.
two) non of the music tags render correctly. (Like I say I want Goth rock, it will give me blue grass or rap...or just render whatever the heck it wants!)
If it's been reported in the feedback or bug report channel, the devs will be aware of it. I don't know what's being prioritised etc. Here I was just stating something that was mentioned a while ago about whether Suno remembers what people have prompted for.
There is no way to revert to 3.5 when you use custom beat then extend song. Fix That. I Want to Make song with 3.5. Becuse v4 is is just weak
Yep. V4 is a bad dream.
Yes i'm sorry to bother you, but there is no way to know. I need to spend credit every X hours to see if it's fixed so i took the chance to ask you
Sorry I don't have an answer 😦
I agree! I cant use V4 am poor. Yet I notice downgrade in old songs made in the past (from me and other users) using 3.5 and new ones rendered with 3.5 sound like crap! And dont even follow the tags!
okey this remastering feature is neat. but it is a bit of hit and miss still
i understand, np
I notice lot upset upon 4V don't crap renders too,.
It's almost all misses and very few hits. I'm waiting for a hotfix patch
the vocals are much clearer. but the gender is still confusing. been mainly using female voices so when i try remastering i get a male instead
Suno is glitching again, and in so many ways!
1. A lot of songs that were rendered a while ago are now coming through distorted and at a much lower volume, and turning it up doesn’t help at all.
- The music tags are totally off – when I request a specific genre like Goth rock, it either gives me something completely random like bluegrass or rap, or just renders whatever it feels like.
I upload my custom beat and suno on v4 Makes crap Out of It, its Not using the beat i have uploaded. Just Like faint residue of a beat. Suno v4 is Not following Simple orders. And you cant revert to v3.5 when using custom audio
Can you elaborate on this? Are you just hypothesizing?
hello, I'm struggling to get Suno to generate a duet. Any tips would be greatly appreciated
takes a bit to start generating a style and tone, but that may make sense remember some early stuff i did tha ti am no longer able to redo even i fi use the same tags
the rattling in every song is very annoying
Thanks for the explanation. I do have to admit that I am very curious about how Suno works under the hood.
I know this isn't the answer you want to hear, but we all have been there. What I have tried.
is
[Verse 1]
[female vocals]
whatever the verse is..blalalalalal
[verse 2]
[male vocals]
blhahahaha whatever lryics
100% speculation -- @unique stone asked one of the devs if it learns in the shortterm from your voting, and they replied "I wish it worked that way"
They can't even fix some old bugs and you think they doing all of that 
Anyone else noticing old songs, that you have created months ago, now coming through distorted and at a much lower volume, and turning it up doesn’t help at all.
Even if I select 3.5, the song has v4 tag
Everybody loves a conspiracy 🤷♂️
Same bug than this: https://discord.com/channels/1069381916492562582/1308503194938118164
All im doing is generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash, generate, move to trash,
So it's not possible to make v3.5 at this point? Luckily I'm running out of credits. No need renew the subscription before this mess gets resolved. 😦
Agreed.
That's been my v3.5 workflow for months. It's how my trash got to 1700 pages. 
so its not me...3.5 is being a turd?
Ok I'll expand on this a bit more then, since it wasn't clear. They wish it would work that way, but as of now, it does not. It definitely should not be remembering previous prompts.
Nope. The v3.5 sounds like v4 crap
POO! I feel like this scam some how by SUNO! To force us join with money. Their killing off 3.5. By choice!
The songs don't get changed after being generated. It's possible the website may be playing them at a different volume or a browser update changed something. Download the MP3 or WAV of a song you've previously downloaded and compare them. They should be the same.
Lol i Just used My persona and It gives different Voices
Here is a screenshot of where I had generated a few covers of a silly song for my cats (It's kind of a parody of Fairytale of New York, so I was trying to get it to sound like the Pogues)...then I resumed work on a completely different song...The thumbnail was cats for a a couple generations even though the song and prompt had nothing to do with cats. Perhaps it is a browser cache thing...but it did get me thinking about how future generations may be influenced by songs I had in my library.
BUT ANYONE! Please riddle me this...
Has any one been noticing old songs, that you have created months ago (and ones your very please with, that met your standers) are now coming through distorted and at a much lower volume, and turning it up doesn’t help at all.
That without a shadow of a doubt, that you would bet your very life on, the audio has changed it self, from how you remembered it?
Download an older song again that you previously downloaded. Compare the two.
Think the Best service are getting Free users, to lure them to subscription
Gosh...this would create an aweful lot of overhead on the servers if true!
If you're into that kind of thing, generate a hash of both files (if they're both MP3, or both WAV). This will conclusively determine if there are actual differences with the file.
Woild you like me to test a song I downloaded months ago, and redownload the same file, so we can test if there's any difference?
I can make a quick video and share my findings
This is most likely. I hear certain things in v3.5 that are simply clearer in v4 but now I know they're there I noticed them
That would not working. What am implying is the original file has already changed. Within the site itself. If that make sense?
...I swear I still hear shimmering on Weezer's Hashpipe in that one sections 😂
I write 808 (beat) and It gives me violin without any bass drop
Does "Replace Section" just sometimes not work? I've spent over 200 credits trying to replace the words "My baby" with "This boy" and it never does it. All the lyrics are correctly highlighted, etc...
Well, I've made my suggestion regarding testing whether a file is the same as a previously downloaded one but feel free to not take it
Not quite sure I understand-- I downloaded the tracks when I FIRST made them, so they're on my computer as they originally sounded.
If I download that same file nowadays, you're saying the file has changed, yes?
I've only used it once and was successful with it, but there have been lots of people reporting issues with it
the chosen one!
Ya, I was trying to replace two phrases in a chorus. One worked 1 out of 10 tries, this one never works.
well maybe not for you? if your got lucky? As there so 1,000+ users. so if this some bug/glitch am sure not all are effected. But I bet my very life my old files have got messed some how...ones left on the site.
I'm not sure what you mean "not for you" -- are you saying that the songs sound different than when they were originally made? If so, how are you testing this?
For the record, I'm not saying you're wrong, I'm just trying to understand so I can investigate & try to report findings to the devs 🙏
well it's find of weird that personas in v4 just produce the same song over and over instead of varying the generations... even if you change the prompts.
Yes. The songs sound different than when they were originally made.
It sounds twisted and distorted, and volume is turned down. This is within the very site itself
Good day! I'm new here and I've come because I was unable to find a specific information. I asked this in general-suno, but that place is too crowded and in 5 minutes, 70+ messages were ahead. I deleted there and I think this chat might be the correct one.
My 'issue' is about Suno's melody. I wanted to know if we have means to control it, I mean the notes, tempo, time, D minor, major, things like that, or if we cannot really influence that part. If we can, can someone point me to where we learn how to write these parameters and compositions in a way that Suno understands? Thanks in advance.
ya of course. Not saying you are..am having hard time explaining
It might not impact every song, but have you downloaded any of the songs so you can compare how they originally sounded & how they sound now?
Or is it primarily just KNOWING that they sounded different? Like, a matter of having listened to it a bunch of times, then listening later & it not sounding the same
it didnt work 😦
As far as I know there is no way to do that. At least there is no good way. You can tell it a spefic key but chances are it will ignore it.
Yea. Its mostly on a hope and pray. I keep playing with it, I get lucky sometimes.
Am sorry ❤️
You can also try [Male voice] [Male vocals] or [male artist]
Nope. It has been tested extensively. Suno responds to the content and formatting of the lyrics. As a matter of fact, often one of the outputs (two versions of your requested song) is in major and one is in major, one is faster and one is slower etc. This is often random, but again depends on the lyrics. If you want to steer it, try to write paranthesis after each verse that conveys the emotion you want, this is by far the best practice to lure the music output into the outcome you want.
you aren't going to have luck with this unless you generate like 50+ times and you may finally get lucky. I think it finally worked for me once out of many tries when I gave them names, like Bob and Jill but again, had to generate lots of times
I don't believe that anyone has found anything that works consistently. People swear that they've mastered these things but when I listen to their tracks, the tempo will be different, the chords are different, etc.
Hmmm, is there an issue with extending clips? Anyone know a way to keep it from sounding muffled at the start?
they all sound washed out (I know that is style) bad, sounds like crappy you went to live show, and tried to film it with your phone sound...and post it online..when they didn't before.
I see. Thanks for your replies. I'm not a musician, but I love composing things in my head. I was testing a composition of a musician I like to see if it would translate in the same way, but Suno made something really bizarre instead.
Extending, persona and cover has switched over to V4 tech even if you want 3.5. It is completely broken as of now. I am speaking as a user with V4 access, it may be different to others.
Yeah am trying to extend a V4 Persona song.
I'm not a musician, I just love making music
hate to break it to you
I expect that we'll be able to do all of those things and much more in the future.
And without fail this layout I entered always worked for me...but now am just getting bloody rap...(or shade to people who like it)
https://suno.com/song/ede987e5-fa0b-4ae4-9479-489f71d7887b
Ya I want this sound back again...but when I use this prompt it turns into bloody rap!
female vocals, gothic, neoclassical synthpop, darkwave, Cinematic, 80s Film, ethereal , baroque, symphonic song. Listen and make your own with Suno.
everyone facing same problem. 3.5 is good one and everyone want this
Yeah, my beautiful operatic duet prompts now suddenly give drums, even when using 3.5.
while the vocals are clearer in v4 i keep thinking the singing was superior in 3.5 i am putting together another music album. and there is no heart in the v4 for some reason
Yeah, if we can control at least the melody, I wouldn't mind that the rest was AI-controlled. Sometimes, it will just make some odd things, with a nice chorus, and vice-versa.
ya it took me months to find that good 80s sound..for what if cult flim sound I wanted..
not clear at all singing different pronounses
what you mean?
Having clarity is of no use if the pronunciation isn't clear
got a point, being able to hear the vocals doesn't not aways help
Good to see that I'm not the only one. Who saved their certain prompts. When their looking for certain sound.
i have a entire playlist of songs with prompts that i go back to when i am looking for a certain sound i liked
Hmmmz, think I got the Extend to work.
I added [Fade Out] [Spoken] (Text) [Fade In] and it seems to be working favorably. Could just be random chance.
Version 3.5 was really good; there were no issues with it. I can’t tell if this is an upgrade or a downgrade
it also feels like [theses] dont work anymore either..
also i wish there started to exist a library of theese little command prompts we can use like how to do duets, backup , spoken and so on
hey duet is one of the big problem here
100% Sure not all the rolls I did met what I wanted..But at least they matched what I put in!
you basically would at last get what you asked for even if it wasn't neccesarily the right variant of it right?
0% output 100% credits out
yep. Like am looking certain sound. And though it matched 100% its like "Nope not that one"
like say neverending story was my suno song..maybe I wanted a female singer not a male
yet sadly I would have render it all over again and all but ya
I just heard this preview..sounds pretty crappy...in my humble view.
#📢┃announcements message
#1308499132356362383 has been set up for general feedback about v4 🙏
is there a way to get female voice and not male voice when I indicate it in the styles?
style: female vocals , exclude : Male vocals
thank, does it work for you? I tried it, but mostly it didn't work for me
99.9% of the time
hm, thanks, I will stick to it and let's hope I'll see more good results
I have 2 saved characters. is it possible to have one voice sing verse 1 in one song and another sing the second verse? or is it impossible to do this at the moment
Does anyone know an AI website where I can learn and test different genres to make styles more precise to my liking?
gpt is useful, describe what you want to hear and it can recommend styles
Hey, anyone already figured out if you can extend a v4 Song with an v3.5 Song and reversed?
I don’t know if a non subscribtion user can collab with a user with a sub like this.
Yeah -- there's just a disclaimer in red that it's not optimal quality -- v4 plays best with v4 (or remastered v3x tracks)
Yes, I use but, I would like to actually listen to it as I don't know every genres
👇 extending a v4 track with v3.5 (still lets you do it, just warns you)
👇 extending a v3x track with v4
Are you sure? I got the same thing but the section was labeled with v4
Duets are fun, tho! https://suno.com/song/bb6224f1-4e69-4232-93b6-accf13faaf84 🧍♀️ 🧀 🧍♂️
female vocals, gothic, neoclassical synthpop, darkwave, Cinematic, 80s Film, ethereal , baroque, symphonic song. Listen and make your own with Suno.
Okay, thank you. Seems like there are some quality tests needed for this.
does anybody else have problem to enter the Libary ?
Such as? Mine loads fine right now. Yesterday, there was a period of time where the library wasn't showing ANY of my old files but would show new files. That got fixed.
Application error: a client-side exception has occurred (see the browser console for more information). its only on the libary , the others are working fine
doh!
try restarting the browser and/or clearing your cache?
did already , also relogged in doesnt work
"Client-side issue" means a problem displaying the site on your end
ik , but it was only the site of the libary, thats why i was wondering
now it works somehow
its good to note exclude is only a paid for freature.
Client side exception means that the problem occurred because there is a difference in the rendering between what came from the server and what Javascript decided to render. In simple words - the developer f*** up 😛
This also means that these are most often errors about displaying some element and not about some functionality or logic.
I have a scrolling issue, when I scroll the library turns empty
The developer made the site, but the issue is presenting on the client-side -- when people say "is anyone else experiencing a client-side issue right now?", there might be others experiencing isolated similar incidents, but it's not a system-side issue. It's a number of isolated incidents -- yes, caused by the developers, since they programmed the thing.
When someone asks about client-side issues here, the only thing we can advise is basic tech housekeeping -- "check that nothing on your end is preventing the script from loading properly"
For 1 week at least they should offer free credits to ease the transition and all these issues.
That's the only thing I know cause I am a web developer, so believe me that this is typical React hydration error about render difference between server and client side.
I can agree that there is nothing that either you or the person with an error can do.
Fair, I was unsure if it's not been public yet
I want to use it so bad! but am so poor..and my mother will not allow me to used it...since the site doesn't offer paypal. I wish they could offer a gift feature...so I could repay back a friend through paypal...
get a preloaded debit card, that would solve that issue, no?
well am autistic and my mother is my keeper, and can only do so much..of she says no..she says no.
I have tried to ask for my own debit card, shes no. am 36 years old. but then she ask nicely she makes me feel bad.
specifically a preloaded card, which you can load with $X and it can't pull more
yep told her that but my mom hates getting new information..and somehow she attacks me emtionally, like am spoiled...so I better just not say anything at all ❤️
is there a place I can voice to suno they should offer a gift sub feature
have remastered a song like 10 times and every time it says "re-collections" instead of "rec-o-llections"
When I klick on cover song it no longer says beta but takes me to the manage subscription page. What's wrong? On another note I feel like the last couple of hours there's been much much less of annoying artifacts in my songs
I figured out something, that might be useful... Extend - Remaster (Keep the ReMaster as part of the Extend) = Which it currently is not
why we cant extend on version v3.5 too, when the track remastered its changed to another vocal or the music it self sound different so we need both available in the replace section !
#💡┃suggestions-web is the discord channel they've set up for ideas from users, or you could email support@suno.com -- keep in mind that neither guarantee that the idea will be followed up on, but if people don't let them know, they can't understand how important it is to them 🙏 
I'm not sure what you're asking -- you are able to use any model to extend any model.
now we can`t replace section on v3.5, we need both available on replace section
I'm confused -- we were talking about "extend", right? I recognize there's separate issues with Replace Section, but just making sure that I keep everything straight
no iam talking about the replace section, its only available on v4, not 3.5 too
its was typo mistake at first i mean replace section not extend
You're right -- Replace Section, Cover, & Personas seem to all use v4 by default now. I genuinely want the ability to swap between models as well.
Unfortunately it's not something that other users are able to assist with, so #1308499132356362383 or #💡┃suggestions-web seems like the place to direct that info to the devs 🙏
Is there any workaround or solution to when your persona origin song contains lyrics that are similar to the ones you're using to generate a new song, and it just reuses the old lyrics?
This "can" happen, I have found, if I fill it up with enough lyrics, it won't do it
But it's usually only 1-2 generations that does this, so I just ignore them mostly
it's pretty consistent for me
Is it like the old song that you based the persona on?
Yeah and the new lyrics are somewhat similar to the old ones, and it just reuses the old ones. Even if I change the new lyrics significantly, it sometimes replaces them with gibberish
Hi everyone, is it me or the V4 struggles more for the extend and replace parts ? Like my song is totally different from Part 1 and 2
Well, there's your culprit
at least let people have a choice to use v3.5 in all features like replace section, then by time the new version will learn more and be more effective than 3.5
I agree -- I hope that you share those opinions in #💡┃suggestions-web or #1308499132356362383 where the devs are reading those kinds of ideas
Good afternoon, please tell me, in SUNO, when creating songs, an extraneous text sounds at the beginning, what does this mean? How can this be fixed, this has never happened before
Hi @flint imp, is the message at the beginning the sound of someone reading out messages suggesting that you visit suno.com, and to not use fake sites, etc?
Something like the name of the site, but someone reads it out
If you are using https://suno.com, or the official iOS app (https://suno.com/ios), you should not be receiving these messages -- if you email support@suno.com with an example of a song that features this voice, they will be able to investigate the matter properly to help resolve this 🙏
If you are using a different app or website to access Suno, these messages are intentional
are the devs aware of this shimmering issue? its actually unbearable
Thank you
Did the weird audio clicking/popping from yesterday get addressed?
from yesterday
😅
I've had a lot of overloaded basslines yesterday, not as much today?
It definitely feels like something changed from yesterday to today 👀 Not 100% yet, but there's notable improvements in the frequecy of hits vs misses I'm getting in my tests
I kind of got a solution. If you want to create for example an instrumental use meta tags with lyrics enabled and then the shimmer isn't there
I swear it only appears for me when the instrumental is checked in
Hey! I just signed up for a paid account following yesterday’s launch but haven’t been able to generate a single one, getting A Clip got stuck while generating every time. Is this normal even for paid users?
Okay never mind metal will always shimmer
I think you've just had some lucky gens. I never click the Instrumental Box and I do mostly instrumentals.
No that’s not normal
Have you tried to sign out and sign back in?
Yeah, I’ve tried that on multiple devices
Yo metal lovers here
Yea I think my luck ran out I was getting like 10 gens back to back with no issues. Well I mean maybe that means devs are working on it
It's been occuring frequently for many people lately. Check this post: #💬┃general-chat message
Uhm.. the fact you have tried on multiple devices tells me it has nothing to do with bugs or caches. But issue could be the account. When did you move to pro account , today?
Oh thanks! Good to know I’m not alone 🙂 and that was yesterday!
Revo, https://help.suno.com/en/articles/3198209 "stuck in generation" has recently been related to something in the lyrics or style prompt being blocked for moderation.
We support a welcoming community for anyone to make music, but inappropriate content may get flagged before you can generate a song. At the moment, so...
Hi, is this the right channel to discuss a new issue I'm seeing consistently with v4?
Tried and contact the support@suno.com. but as the link suggests multiple users has been experiencing same issue. Fault is on the system migration . Give it another 24 hours perhaps
Yes
Not really sure what goes where atm 🤷♂️
#🪲┃bugs-web for bugs using the site
#1308499132356362383 for comments about v4
Here we can help give advice about how to use Suno
Oh it seems like there is a lot of critical feedback already in #1308499132356362383
OK, well specifically I was hearing a lot of weird reverberation on my songs.
I haven't gone back to mess with it more since this morning since I was kind of disappointed after being hyped up
lol
All I really care about is, does the dev team know about the feedback and plan on taking some action
Aw I thought it was funny
Worst case can just use 3.5 for now
😛 
Just to pre-position -- covers, personas & replace section are locked to v4 😅
oh. hmm
I understand different people have varied experiences but so far I am loving v4
I find that for example trance and electronic tracks that don't have the usual analog drums and analog guitars and Distortion there isn't any artifacts. But when I do something heavier it does.
Yeah - like when the song gets "busy" theres a lot of issues
I'm finding that anything that causes "peaking" seems to trigger that shimmering effect -- crowd noise, cymbal hits, general loudness -- so some genres like metal are hit harder than others
Doesn't mean metal is impossible, and it genuinely seems like there's been improvements made today (speculative on my part, but I'm getting notably different generations to the later days of alpha testing)
Honestly, I think if they just slapped "beta" onto the name, people would be less angery
They did, but only on some social channels 😅
and then people would be more willing to test it out. But I guess if you slap beta on it, then people expect x number of free songs (like with covers)
(not expressly mentioned in all communications, though)
Hi, guys. Do somebody know how to get chirp model as library for python ?
Chirp has been closed-sourced & spun off into Suno.com -- the repos aren't available 🤷♂️
They open-sourced their text-to-speech model, Bark, which was developed at the same time.
I see - I found out via email:
Do somehow possible to use via API ?
there is no public API
@all i need help. how can i cover a song with my uploaded voice? the v4 changed my voice so hard. so it's not mine..
API isn't made public -- using it violates TOS 😅
If you want to use Suno, it needs to be through the website or the official app 🤷♂️
Suno doesn't offer vocal cloning in the same way companies like Elevenlabs do, but it can attempt to continue from the clip you've uploaded -- so if you sing a beat, it can try to continue the beat with the same sound, but it won't understand implicitly how to say different things accurately in your voice
😞 thank you for helping
yeah but with remastering or cover my voice is complete different
v4 is not good now for me i think
tbf -- v3 didn't offer vocal cloning either. But by all means, if Suno doesn't offer what you're looking for, you should reconsider your subscription.
No V3.5 was such that it expanded my vocals. Certainly not 1:1 like my own voice but it sounded VERY VERY similar.
u know what i mean?
Or please explain to me why it worked before to expand my texts with my uploaded vocals and to have them with almost the same vocals as I uploaded or recorded them?
i want to know how distributors identify whether we are using pro or premium
my friend bought the suno subscription
and she wanna know how she distribute
They don't, they can only tell if it's made on a paid or free account.
I think I understand -- it's a subjective matter, but if you feel like the covers were an accurate representation of your voice, it's not my place to argue. From personal experience, I've had success in recreating specific sung phrases in refrain, or mimicing a SIMILAR vocal timbre, but it clearly wasn't my voice, so it's in the ear of the beholder, I suppose 🤷♂️
like, with metadata or something like this?
ok
There might be an inaudible audio watermark embedded in the song (nothing that humans can hear, and nothing that would impact the song, but something that another computer would pick up on -- not the same shimmering kinda stuff we're hearing lately)
But also, any tracks made on a free account are still technically owned by suno -- who knows what kind of setup they have to prevent people from generating free songs & stealing them for monetization? Suffice to say, as long as your friend only distributes songs made on a paid membership, not extending from anyone else's clips, they should have a straightforward time with distribution partners 🙂
Anyone who managed to make some good house (or tech house) with v4? Let me know
I don't understand why my voice was almost the same as mine before.
let me try to make some
Really??? I almost bought a subscription. I remastered one of my old hebrew songs and it ended up way better. I guess it's not that good when you generate from scratch? can you send an example?
Alright let me know! I've been struggling to make it work lately. Stuff like jazz and samba-rock are killing it with v4, but when I try house it is sooo bad
this has been happening all day- my song won't generate and just loads until it says 'A clip got stuck while generating. No credits were used and you can retry.'
are you doing it with or without lyrics?
I'm not saying that you're wrong -- if you feel like v3.5 was reliably able to replicate your vocal timbre, even after covers, that's awesome & I hope that they bring back v3.5 covers.
Beyond that, I'd say that you're incredibly lucky to have the audio upload + extend + cover workflow chain give you similar results to vocal cloning that requires a large training dataset & an amount of time to train on -- so I understand why being forced to swap to a different setup would stink.
🙏 hoping it gets set up in a way that you're comfortable with again
Can anyone help me with something?
Depends;
-
Need technical help with Suno? We can probably point you in the right direction.
-
Need help moving apartments? No promises.
So at 0:12, I really like the effect but I don't know what its called because I want to make the whole song with that effect, song: https://suno.com/song/9b845bc2-766b-492e-8184-0c753bbcdd62
edm, dance-pop, electro, future bass, vocoder, trap, synthwave, auto-tune, robotic vocals, airy, confident song. Listen and make your own with Suno.
Seeing that more frequently in place of "content blocked for moderation" -- might be something in your lyrics or style prompts that's causing it to get blocked?
https://help.suno.com/en/articles/3198209
We support a welcoming community for anyone to make music, but inappropriate content may get flagged before you can generate a song. At the moment, so...
Ah, i see what was wrong- thank you tho ❤️
@potent pelican I did tech house instrumental.
first example: I noticed a couple things? But otherwise seems clean.
https://suno.com/song/bcfe3d2b-0d51-4aa0-a2f0-dc3edcb4f1ca
Second example: detecting shimmer in some quiet parts I think. Yeah, definitely. Which is a bummer cause this song is actually nice lol
https://suno.com/song/59154876-4992-4e5e-85df-4168b6d66151
^
Like that rolling glitch effect on the first word?
Yes on the "We roll up" part
Not sure what I'd call that, like a distorted glitch -- admittedly it might be difficult to prompt that with specificity through the use of prompting (not impossible, but might not be a matter of just asking for "distorted glitchy vocals" if those kinds of vocals are determined by the genre seemingly at random)
Might be easier to add a distortion effect to the vocal stem in post, provided the song "living on suno.com" isn't a requirement
So should I just do a cover song and add in distorted glitchy vocals?
@charred merlin
I need it explained briefly again. What is the purpose of uploading the vocals?
Does covering silence still yield better results in V4?
Both, but I would say mostly instrumental
Yeah this is exactly what I mean... now go on Spotify and type in 'tech house' and play the first song that pops up... you'll then realize that it is nothing like tech house. I think the Suno model doesn't know what it is
ahh I gotcha
Good evening, might be me but the new v4 songs have some weird background noise (sounds like humming)
maybe techno house might get you better results?
I'm not sure but it's worth a try
Yes I tried this, techno house makes it do weird stuff. House, techno does work sometimes (1/25 or something) and then is mostly not good still
https://suno.com/song/0ea2b283-d5f2-4447-a916-48d14b50237f Here's a remaster I did of a 3.5 song, and I feel like I hear echos on the snare and some other things
tribal celtic shanty song. Listen and make your own with Suno.
I've managed to get 1 banger tho using House, Techno, Funky. But it took me 1300 credits
Are you getting bad results like the music is just not good, or is the genre/style feel like it's not working at all - like you get generic EDM rather than what you asked for?
Not sure the specifics why Harber recommended it to you but, generally speaking, many people use Audio Upload to help set tempos or a particular type of beat.
It could be used to try to get a certain vocal property (e.g. the tone) but it tends to be a dice roll from what I remember
Yes! Thats actually what I was looking for, yeah I mostly get generic EDM and it just doesnt sound good
Just as a side note, when you upload a clip with vocals, it gets restricted -- You can't make it public, and others can't extend from it. This is the first UX clue that audio upload isn't primarily designed for vocals -- not a comment on whether it works or not for vocals, but if you'll notice all of the original demonstration clips of audio uploads showed people making music with instruments -- relatively simple patterns that the AI could analyze in bite-size and continue from.
When you extend from a track, Suno doesn't implicitly know what it sounds like, it just analyzes the patterns in the audio waveforms that it sees leading up to it, and mathematically predicts how those patterns would be likely to continue (this is a wild over simplification, but gets the point across)
When you upload your voice to Elevenlabs, or the vocal cloning tool built into iPhones, you give a LOT of training data, and the system does those same kind of pattern recognitions, but for a LOT longer and trained specifically ONLY on your voice -- this way it understands the little inflections and timbres of your voice, so it can try to accurately mimic content that it's never heard you say in the training data.
When you upload your voice to Suno, it doesn't have that same luxury of training time, so it just looks at the previous patterns and tries to continue it -- if you're just singing "doo wop, doo wop", it will have an easier time analyzing that as a "bite sized" clip to repeat over and over. If you have a lot of variety to your vocals, it doesn't have as much to recognize as a pattern, and while it might understand certain pitches, it's not capable of accurately representing content that it's never heard you say originally.
</wall of text>
Personally, I stopped try/using credits for now because of glitchy audio artifacts I noticed
Was waiting for some word on fix plans/etc
Okay, in that case, try this advice. I wrote this when V3.5 came out about V3, and I'm not sure it will help, but at the very least it's worth a try. Essentially it's checking if maybe a V4 style word is weaker than it was in V3.5, so you need to turn up the knob a bit, in the only way we can in Suno.
I wrote this post when V3.5 came out for things to try if your V3 style prompt isn't working the same (wrong or too generic) in V3.5 and I think these may be worth a try with V3.5 -> V4 if your problem in V4 is not getting the right style of music:
My V3.5 style in V4 is giving me something more generic than the specific genre I asked for. What can I do?
(I didn't edit model versions in the rest of the text.)
- Repeat Style Tags - **Try the simplest thing first: copy and paste whole style prompt two or three times into the style field. ** Use a longer style prompt in 3.5 than the style you used in 3.0. Adding unique style words is best (too much simple repetition risks making the music worse, in my personal experience), but simply repeating important or weaker style words also helps. If the style
shoegazeisn't giving you shoegaze, it could be as simple as tryingshoegaze, shoegaze, shoegaze, shoegazeinstead. - Replace any spaces with dashes if you used any compound words. 'Acid House' -> 'Acid-House'. (this may be specific to V3.5, but worth trying in V4).
- Consider reducing the number of very broad genre terms in your prompt. Repeating words is great, I'm specifically referring to repeating broad terms like
countryormetal. For example, if you're looking for a specific genre of 'metal', don't overuse the word 'metal' too many times in the style prompt. (I'm less certain about this one, let me know if it seems like it doesn't help.)
3b. Try your prompt without[verse]and[chorus]tags, this often shakes things up, gives Suno more space to deviate. - Try copying some of your style words in the lyrics. In the simplest case, just copy a few style words into [brackets] in front of your first verse. Better is to integrate styles and instruments you want throughout the song - but just doing that can be enough. As an example, you can get live music by writing [live concert] in the lyrics with a blank style prompt https://suno.com/song/2e22d633-c076-466d-841b-993f0d633ffa
- (Power Prompters): If you are using very unusual style prompts (not musical words, not real words at all) then you may be the exception and your 3.0 prompt may actually just mean something different in 3.5.
- Consider the effect of the content in the lyrics on style. Check what style your lyrics might be creating with a blank style prompt. <https://suno.com/song/ce64b2b4-d90d-4c22-bb42-ef410d4bb58a If the lyrics are giving you the opposite style you want, you may have to start your song with "fakeout lyrics" and then Extend From those with your real lyrics.
A song. Listen and make your own with Suno.
...my closing tag didn't work. my wall of text seems to be leaking -- sorry to the downstairs neighbours.
Lyrics can affect style?
https://suno.com/song/65a2e520-e54a-4def-be67-d483a08a49eb
no style prompts used -- just the lyrics "I'm so angry right now"
A song. Listen and make your own with Suno.
Welcome to Level 2.
Shimmer on v3 
Hah
Woah, thanks man, I'm gonna try all these things immediately and I'll let you know!
harber, i'd link you my song that repeats the f bomb for the whole song to childrens music, but that would probably violate server rules 😉
hmm, this song seems to have fluctuating tempo
I thought you said flatulating tempo figured that seemed appropriate.
The first 2 tips actually have some impact, I know no longer get the generic edm you were tslking about. But it is still not tech house unfortunately.
I'm not sure how Suno works, but I think Suno doesnt have the right training/instruments for tech house
Well pushing it off generic EDM is something, at least, thanks for reporting back.
Tech-House, Tech-House works, bit the quality is really bad
If there any subgenres to tech house that use unique words, I would try spamming them and avoiding broader terms. I didn't mention this but I personally think there is a little bit of a quality tradeoff when you have to resort to too-much-spam to get the style to work. I get better results, musical results, if I use a little spam (2x, 3x) but also a lot of diverse other musical words that are not repeated.
Hi everyone. Is Suno aware of the lazering everyone is talking about?
You can also [fill your lyrics with evocative phrases] which can be a huge placebo effect, but it is absolutely the case that Suno is influenced by lyrics in a broad sense, and adding literally anything in brackets increases diversity of songs generally.
Yes, most common term has been "shimmer"
Is it on purpose? There's a conspiracy theory going around that it's there because of the lawsuit.
But how would I apply it to the whole song?
Mods/Guide are not Suno employees so we wouldn't know, but personally that sounds extremely implausible to me, technically.
I figured as much myself, but you know how people are.
lazering?
I thought you were just talking about the vocal effect -- you can add whatever you want in post processing
People on reddit are calling the Shmmering effect "Lasering". Though I don't know why.
Made a Guideline with a template and now gpt knows how to prompt on the v4
https://suno.com/song/b82c718e-2c44-4d4e-8753-f40ff825e985
[Boom Bap: glitchy beats with arcade sound effects] + [Hard Rap: mocking tone with sharp delivery] + [Humorous Battle Rap: playful yet biting rhythms] song. Listen and make your own with Suno.
I was talking about the vocal effect, the "We roll up" part
Shimmer was jut the random word that stuck here, lasering makes as much sense to me.
The same reason why people call zombies different things in the walking dead -- we all came up with different names based on whatever was first suggested & it stuck
The Walking Shimmer
Seriously how would I use that effect for the whole song?
Anybody have a anything they use in post-production for that high-pitched whistling sound that comes out with rock music? Tried damn near everything with Audacity and RX 11. No matter what I do it either doesn't take out the whistle, or it extends the whistle through the whole song.
I'm genuinely confused by what you mean -- the only effect I heard was a "rolling glitch" effect on the first word -- you could find a similar effect plugin and add that to the vocal track -- or if you wanted the entire song to sound like a constant rolling glitch, you wouldn't strip the vocal stem
Yes the rolling glitch
I want the entire song to sound like a constant rolling glitch
So I would get the instrumental stem then put in the styles of music "distorted glitchy vocals"?
would a song with the word "Oreos" in it cause a copyright/whatever violation that prevents song generation? It's spinning but I can't listen to it
Suno is glitching a bit now
Sometimes when I have the spinning circle, after a few seconds, I click submit again and it works
that worked but cost me credits doh
i mean whatever its not much
Try refreshing the page. It may look like it deducted a lot but that is because of the current counter displayed is not dynamic (unless somehow it made it count as two generations)
Every now and then I get all warm and fuzzy thinking about the fact that this year I made friends with a robot who uses science fiction magic to make songs for me.
So how would I get the whole song turned into "distorted glitchy vocals"?
Plat: I tried this and it didnt work.
Hmm maybe you should see Mr.Tree work, his styles tends to be known that area
v4 remastered issues: might have already been posted... but seems like early version songs can't be remastered (v1) or fail with error (stuck while generating... v2)
I think it would be in the Styles of music box
Hmm try extending it into 3.0+ and Remaster the extended whole song.
as of what ive found out so far with the v4 atleast is that you really cant define it to exactly do what you want, you can give it a hand and a cue on what feeling it's suposed to have and point it at the right direction, (from prompting only)
the only thing that actually listens to you 100% is the [Break], no sorry i lied, v4 have a tendency of addapting breaks for fit the flow
Hey, does anyone know if personas are one of the priorities to fix?
Try starting with something weird, and covering it into something weirder, in particular prompt a lot of "vocals, vocals" when the cover is not singing so the voices end up real weird. Example: https://suno.com/song/59e2071e-ebd5-4daf-834d-9f0891722e7c
You are testing it in V4 or v3.5? V4 seems to ''clean'' the distorsions. imo, it's more easy to produce this type of effects in 3.5
Oh alr
there litteraly are no lyrics in your text except "Begins as a pure instrumental, but there is a Title card drop about half way through that is worth the wait"
It's just at 0:12 the "We roll up" part, I want to use the rolling glitch effect for the whole song: https://suno.com/song/9b845bc2-766b-492e-8184-0c753bbcdd62
edm, dance-pop, electro, future bass, vocoder, trap, synthwave, auto-tune, robotic vocals, airy, confident song. Listen and make your own with Suno.
It's a cover of this song, where that text was just me describing the song for listeners. But I forgot to remove when I covered it so... it became part of the new version 😄 https://suno.com/song/ec0e201b-5fb0-45ba-8f4f-f636c7357a67 fixed link
oooooh thats your way of commenting it?
cool
Though that song has hallucinted vocals. For this purpose, hallucinated lyrics are just as good. Just need "human voice singing"
Yeah, Suno needs a liner notes feature, for real
good idea. but, when I try to put in any time number, it resets to 00:00 ... strange. Tried different things: crop, replace... those don't function. Tried persona - won't work. Ugh.
suno needs a deeper prompting capabileties
i'd love to be able to set the seed myself.
Well .. it's a very accurate detail. I dont know how to make that. sorry, I wanted to help you but i cant 🥲
I believe one of the devs touched this point before and mentioned that the seed will not act the way you think from what people had seen with image AI.
Basically, when they trsted setting seeds, apparently it gave pretty bad results. Which makes sense since that seed may affect a few variables but there are a ton more
why when searching up a song you've made, all 'full song' versions are excluded from the search (with no visible way to include them)? This is really confusing
makes trying to find a certain song that's buried harder
Hi, I want to see my private personas in the Personas Tab. Thanks
sure are but if i want to replicate something setting the seed would be needed.
well i'm just homan with smol brain, but comfyui is funny to play around in :3
So here's my next question: What's the magic formula for getting Suno to create complex harmonies?
styles and a guideline
There are many different ways, none that are super easy, but the simplest one is to start with complex harmonies in an Upload
I've tried a number of things. [Complex Harmonies] as a prompt. Nothing. I've tried bracketing in the lyrics where I want the harmonies, still no.
Best result I got was with an 80's Glam song called "Girl Candy".
v4 messes up my drop crescendo instead of "AWAAYYYYYYYYY" it makes "AWAayyyyy" it loses power the entire song breaks
And sadly, "Girl Candy" is one of my best songs along with "Diggy, Diggy Doo" which is an ICP style jam.
as far as i know is that you have to have a concistent theme and "feel" throughout the song to get somewhat what you want, Vocals over all though is ASS and you can only guide it some by the "feel/vibes"
[Ethereal trap + cinematic hardstyle] song. Listen and make your own with Suno.
a test i did earlier
How are you all feeling about V4? I've been seeing a lot of hate on Reddit and was curious what folks on here think.
lol
cant really comment on that ive only been using suno for a week or two so XD
i didnt edit or manipulate .. that is straight from suno
I don't get the rampant criticism. The tech is out of this world
i also tested this:
https://suno.com/song/c867e03a-b7c1-4c04-bc34-587399626581
[Euphoric hardstyle + reflective orchestral transitions] song. Listen and make your own with Suno.
Anyone having songs being made, where the lyrics are being flat out ignored? (V4)
I got that, I got unfinished songs that just drop off. We need 5 minutes and I drool over the day we can write Dream Theater length songs.
The dreaded 3:14
OMG yes. What is causing that? It's either that or 4:00 which also just stops.
Just extend it, if it's a good extension
A 15 second clip that's standalone hasn't been helpful. Not sure how that whole thing works.
Is there in future a plan to make it?
I had a Christmas song I was working on that just stopped halfway through. I was so upset.
I've been working on one song for like 2000 iterations over the last 24 hours.... I think I should just learn fl studio at this point lol
when you want a masterpiece and rng is not on your side
FL Studio is a Charlie Foxtrot - at least the last version I used about 8 years ago was. So much going on, and I understood none of it.
and oh this tests another thing aswell: https://suno.com/song/24048dac-c1af-4b68-a502-98294288819b
[Melodic trap + cinematic orchestral builds]
song. Listen and make your own with Suno.
You can just extend it
And continue the same version
When I do that, I get a clip that starts at the leave off point and lasts 15 seconds. How do you stitch them together?
It does not always only last 15seconds, they can go up to 2 minutes
yeah but extensions just mangle the melody
Incorrect
^
If you prompt it properly it does not
oh ok then
Use meta-tags when you are prompting, that way it will retain the same "seed"
meta tags?
I'm so shocked not more know about meta-tags
do tell
From the menu, Create>Get Whole Song:
...on the last extension
Check out this tutorial, it's really good:
https://www.wokewaves.com/posts/suno-ai-guide-tips-tricks-prompts
It talks about the usage of meta-tags, for example...
oh that's just guidelines
You are making a rock song, but you want to have [Subtle Rhythmic Pads] included into the track
You add this into the Lyrics part
The same sentiment applies unfortunately
right these are common proctices, you can still mangle the melody during an extension. What I do suggest is to create a persona from a large portion of a cropped clip then generate using it with complete lyrics. it will retain the style, melody, etc, but continue
Yes, this is the better approach, but I don't think you can add a Persona on Extend. Yes, you cannot add a Persona on Extend.
soo youre saying we could be able to check the metadata and use meta tags for what as if so, are you talking about the guideline cues inisde of []?
Yeah but you can basically meta-extend by creating a persona from a crop and then putting in the same lyrics. it will create the same song as the persona was based on + whatever you wrote after
I'm testing some prompts.
Bonus is that iy retains the styles way better than extend does
Will it be possible to remaster songs on v4 that are not mine?
Not particulary no, but from my thurough testing since May, I have found that meta-tags retains the same style.
define your "meta tags" please.
You can only remaster tracks from your own library 🙏
Speaking of "meta tags"... it does try to semantically parse the lyric prompts, but it's not 100%. I had this [drop] in my prompt and it literally said it and I was so upset
These are the ones I use in my generations:
[Clear Audio] [Clear Vocals] [EQ Mixed] [Panning Audio Left & Right] [Reverberation] [Echoes] [Angelic Flow] [Calming Flow] [Hard Opening] [Soft spoken lyrics] [Melodic Lead] [Smooth Vocals] [Intro with calm tones] [Indie] [2020's] [Whimsical] [Poetic] [Rhythmic Synth Pads] [Rhythmic Bass Pads] [Lo-Fi Drums] [Calm Drums] [Lo-Fi Beat] [Complicated Drum Pattern] [Soft Voice] [Breathing] [Soft Spoken] [Ancestral Voice] [Intimate Singing] [Closely Singing] [ASMR]
But this is the genre and style I always use
And they've proven very useful, I have another one for Duets as well
And Is it possible to add other people's songs to my library?
I use [best quality, uhd, hd, trending on artstation, 4k, 1girl, score_9, hifi 320kbps uncompressed master]
jk but only somewhat. Prompt making is not something that's set in stone
Yea, very funny, I laughed not only once
okok gotcha
HAHA thats a typical adult ai image prompt
Yes, I know, and this is not animegenerator nor novelai
Suno does not function in the same way
I did some testing and will be doing more testing.
GUYS
take your v3 songs and get the Instrument and Vocals stems. Remaster them separately, and you should get better results.
If you know how to use a DAW, you can then place the two on their own track. So far for me this sounds better than the standard v4 Remaster
nope but it has similarities, and if you know how to prompt you can probably get exactly what you want
Exactly, which is why I'm helping now
But if he doesn't want help 
Good luck, you know
true ^^
Not directly to your own library, but you CAN add them to a playlist
Audacity is the vim of DAWs
well well ill continue to work on my little gpt guidelines and maybe i can contribute to something nice in the future aswell :), so far it's able to generate what i belive is proper prompting by taking in lyrics and formats it correctly as an output
I think I got the harmonies to work. Any feedback would be doper than 3 leaves. https://suno.com/song/86436362-8ee6-4c94-bc33-8dd19d0f69d0
Classic Rock, 80s, Catchy song. Listen and make your own with Suno.
I have a track, I upload it, and when I try to extend it, it asks me for a prompt which.. seems to be throwing off Suno because the extended tracks are only like 2 seconds long and sound bad. (using v4).
help how do i give suno instructions without it singing them?
[LYRICS]
and I FLYYYYYYYYYYY
[INSTRUCTIONS]
- Stretch "YYYYYYYY" with echo and reverb.
- Gradually morph into synth lead.
- Sync morph to the drop.
haha we were just talkinb about this.
Put things in []
Still not guaranteed
That problem seems to be random.
All non-singing stuff should be [] or #. [] works better, but still sometimes doesnt
Yeah vocals usually is a pain in the butt, and the testing ive been doing seems like you can only prompt your way of the "feeling/vibe" not actual hey sing thissssssssssss or play that brrrrrrr
[INSTRUCTIONS]
[Stretch "YYYYYYYY" with echo and reverb.]
[Gradually morph into synth lead.]
[Sync morph to the drop.]
it won't do what you said exactly and not all the time but it will help
ok thanks guys how can i give instruction like [DROP] which seams to work + Voice effect like OHHhhhhh
that can work actually
Do you need to use [] in the style of music section? also, does style of music recognize tags like "video game lobby" ?
yeah [Drop] is pretty solid. However something like [Progressive buildup toward hard drop] sometimes gets sung and it's hilarious
Hey, are personas working? I'm trying to create one but the menu pops up for a second then closes
only noobs don't add the 'greg rutkowski' tag
beksinski
I'm using mine. I need to save more of them. I have multiple AI personality disorder. LOL. I have 3 going on 4 projects with seprate feels.
its still funny though when it sings the instructions 😄 😄 😄 https://suno.com/song/46bd05c3-099f-4c85-800b-0b7704cc1c8a
dubstep, instrumental breaks, crescendo, club music, female vocals, song. Listen and make your own with Suno.
its because you have "the bunny" v4 thinks this is lyrics, its almost the same as () but not really really xD
You can [surround instructions with brackets] however this is more for stuff like [guitar solo] than complicated instructions. but positioning things like [synth lead] (repeat the tags if longer or not working) and later [drop] tags will sort of kind of start with a lead and then do a drop, at least, with a little luck
You just try trying using your literal instructions, they probably do something, and "morph" is a descriptive term
