#📚┃suno-school
1 messages · Page 88 of 1
Oh I don't know I have never seen it.
This'll burn thru credits fast, but I'd cut out all of the lyrics except for the first few lines. If that doesn't render, then the problem is somewhere in those few lines. If it DOES render, paste in a few more lines and do another render. Keep doing that until it hangs again. Once it hangs, you'll have a better idea of where the problem is. That's how I found that "it's me again" is a protected phrase.
Just says "inappropriate" I never know what that entails. Lol
lol obviously
i was asking @charred merlin though
I was asking for myself
and did you get an answer lol
I want to help train the model
i mean, we all do haha
I don't see it on the app, just in my desktop browser.
I just asked the question. Sorry I shouldn't have fed off yours
yeah i asked it here. Maybe you missed?
Yup. Mobile and quick, didn't scroll up sorry. I saw the gist and jumped in
most people are im sure on the web, it would make sense to put it on the web version to help train
I have used the desktop site a couple times and it is so much better...lol. which is understandable considering how many layers of stuff goes on the interface
in fact it used to be on the web, and if im not mistaken you used to get credits back if the quality of what you asked for was bad
Hmm...if that is true then I feel very very dumb
Take your music to the next level with Suno Music Mixing and Mastering! Get professional sound without breaking the bank. To ensure your tracks sound amazing on any platform. From podcasts to albums, we'll help you achieve clarity, depth, and punch. Learn how to elevate your audio game and make your music stand out from the crowd.
Subscribe for...
The dude that brought back the item because he didn't read the instructions
ok
lol
you were talking about bad sound quality
I remember there were options to report specific issues but I don't remember it giving credit refunds for it
i am almost positive when suno first came out..unless it was the other company that come out at the same time
they should charge credits only when we finalize or accept the result 🤣
I wasn't around until v3 so I am useless in this part
but realistically, if you wrote a description and it gave you something completely different, why should you get charged for it?
you should be able to report that and get 1 credit refunded
it only helps the machine learning to do that
I agree
lets voice this matter
apes together strong
obviously the pro plan offers plenty of credits but thats beside the point
everyone lets demand refund
but thats why im asking about the "report bad song quality" option on the APP and not the web
Hmm well I think the issue there is that the compute power involved can't be refunded to Suno, but I do see your point.
I can not due to the nature of AI. It does what it does and we pump out songs in seconds. No way I can demand a refund from having to cherry pick
Yes I missed the reply...lol...didn't mean you..,sorry again
lol smh
I meant for you or at least your topic
yeah, i dunno what that person is talkin about lol
I'm just getting back to the computer, so this might have been covered already but if you are including the name "korn" or the name of any of their songs in your prompt, that could be why it's refusing to generate.
Missed that Toker caught that one 🤘
Today I Learned...
You can report them the same on Desktop through ... > Report > Bug > Poor Audio Quality / Doesn't Follow Style / Doesn't Follow Lyrics
You MIGHT be able to put your phone into landscape mode & request the desktop version, but the UX is absolutely designed for the desktop experience.
Yep i see now.
thanks
I just tried deleting an upload from the app & it disappeared on desktop, but I also didn't have Suno up on the desktop, so maybe if you still had the page up, it would keep a version cached for a bit? 🤷♂️
So, if you delete on the web, it doesnt delete on the app
didnt try otherway around though
may need to refresh on the other client
Isn't that the truth -- the fact that so many metal vocalists provided solo work on it blows my mind -- and THEY'RE SO GOOD
@worn flicker, credit where credit is due, Forsaken is a killer track.
That would be a SUPER user-forward thing to do, except the computers are still working just as hard whether we like the output or not :/
Instead of it being always on like a lightbulb, whenever we generate a song with Suno it causes a surge of processing power -- THAT'S what you're paying Suno for, the server processing power.
If we could generate songs for free until we felt satisfied, competitors could generate millions of songs without ever "liking" them to drain resources from legitimate users -- the first thing to go in an attempt to fix this would be allowing free access to the site. 😦
If you're not satisfied with the output you're receiving from Suno, try asking for help here in tech-chat 🎯 and we can try to figure out how you can be just as satisfied with your songs as we are with ours 💪
Hello!
why covers aint working for me?
In anticipation of V4 I am considering buying some mid range headphones, for not more than $100, what do you all recommend?
I want my covers to sound gooooooood
sennheiser
But I'm also not independently wealthy enough for a $700 set of Bose
Thanks I will look into those. Edit: The ones I saw were like $250 but maybe I can get a Black Friday coupon code. I would love AirPods max but toooo expensive.
or get beyerdynamic's
Now you're talking! A quick google found me a pair for $150
Okay -- first you need a couple subwoofers -- then you find some duct tape...
I'd rather not get tinnitus but thanks for the thought
For that price range, I'd go with the Sony MDR7506 - there's a reason that so many studios still use those cans
those were my first "higher-end" headphones. Shelved now, but not forgotten 😄
When you say "not working", do you mean it's either returning the same song perhaps with a change in pitch or tempo, or a song that sounds somewhat related but nothing similar to the source?
its just loading and then i get my credits back
Ahh -- there might be a chance the song didn't pass the automoderation on the second time through -- 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...
well, its me singing so idk if its that
also its in finnish, and i dont think that suno's moderate for finnish is the best
I don't think the source being a vocal track should cause any issues (it doesn't) -- worst case you might need to email support@suno.com with the song link for your upload so they can look into what might be causing an issue.
I have done a lot of stupid songs in finnish and no moderation there
is there something going on with Suno because I get a lot of "system errors" now
I think that the vocal detection MIGHT have gotten better, and it might be picking up on "stupid songs" 😅
haha! Maybe
There's still areas that area oversensitive, for sure, and I'd imagine that as long as the songs don't violate ToS / Moderation page I linked above, it should be okay to rephrase some sections phonetically (if I ate my own arm was flagged for being too violent, but it was about what I would do if I woke up as a gingerbread man, and is arguably "innocent", saying I eight my own arm to bypass the automod flagging it seems like an acceptable grey area 💪 )
Style of Music has gone up to 200 instead of 120?
i still have 120 😦
A, An error has occured hahahha
LEt me try it out... it still says 200 
Nope! Doesn't work hahha, that's funny, what a funny bug
well, i think mine struggles even with 120
maybe you got a peek under the dress of V4
You can add more genres as meta tags
In the Lyrics section
wat
I just use Style of Music as an initial template, and then I add more meta tags into Lyrics to enrich it
Just add them at the start of the lyrics like this, just as an example:
[Clear Audio] [Clear Vocals] [EQ Mixed] [Lo-Fi Drums] [Panning Audio Left & Right] [Reverberation] [Angelic Flow] [Calming Flow] [Soft spoken lyrics] [Melodic Lead] [Smooth Vocals] [Intro with calm tones] [Indie] [Whimsical] [Soft Voice] [Cute Voice]
🤯
I'm shocked this is not common knowledge
oh it probably is
when i find something that works, i tend not to stray lol
but now I'll definitely be giving that a whirl
Are you also using Interlude?
anyone else getting an issue where Suno won't make a song over 2min? this is the 10th time in a row noqw
trying to dictate male and female duets now and it's a little resistant
hah. We can't have "Dog" as a style 
you sure there's nothing in the extend thingie at the bottom?
that's the only one I've seen hardcapped at 2
v3 only does 2 min renders, too
also check the cover thingie at the top
main songs go 1:45 tops...extends are giving me anywhere from 10 to 30 seconds tops.
i've been getting 3-5 min songs for weeks now
this is just today for some reason, last night all 4 min songs
What's the max clip lenght for V4?
v3.5 does up to 4 mins. Are these 2min renders showing v3.5?
Thanks for pointing this out ❤️
Sounds like a bug if you're not extending or doing covers. May want to report it to the bugs channel
Just using Persona, and all of a sudden, it popped up
Maybe I got lucky on over 4min then?
It's on my end as well, same as v3 mistakenly showing 3000 instead of 1250. I'm tugging on shirt sleeves about it
waaaaaat
yah, its happened at least a dozen times
we are on the verge of a new day my frands
and its a whole song not just noise and stuff.
HAHAHAHAH what the
legit though, if you got a 7m44s generation without "Get Whole Song", that's wild 👀
This is common only to us who have spent time using Suno. Since there is no instruction booklet, any time a new user comes about it is going to be new for them. So keep throwing out ideas.
[Panning Audio Left & Right]. Has that actually worked for you? That is a good test of control.
yes it's 100% legit and it was a normal render. not something I asked it to extend or make a full song. I was shocked and happy
when it's happened in the past, its been like 4:40 or 5:01 but that's the first one I have over 5 min
and last
that's why i just sit and watch this chat lol
Anything over 4 minutes is definitely breaking the norm, and I love how it's just commonplace to you at this point 😂 Not even phased.
the chosen one
It's like when I found out that you can get hybrid segments of oranges & potentially grow a whole new species of orange -- like, you could be the person who cultivates and popularizes the "cotton candy orange" or something, but most people just shrug it off and eat the segment without a thought.
You ended up with Cotton Candy Suno growing in your backyard and don't understand why us cityfolk are making such a big deal about it --"oh yeah, we've been generating 4m30 songs since I was knee high to a grasshoppper out 'round these parts"
HAHAHAH probably a bug. I mean today I'm getting a whopping 2 min on all my songs. or less time. so maybe its like "hey you owe us"
I've had many stop at 4, none have ever broken 4 without extending. That is crazy
Like taxes...lol we gonna get ours...
I assumed it was normal if i'm honest.
My colloquialism stands -- "5m songs catching you off guard? My pappy swears he used to generate entire albums at once, him and his band, god bless his heart -- losing his marbles in his old age"
Please invite me to live on your Suno farm.
god I remember those days, we'd sit in our jam room smoke a bowl and write a song in like 20 min. I did that for all my Pimp X stuff back in the day, i'd sit my home studio (office) smoke and make a full album in one sitting.
now that i'm sober, takes hours...days. LOL
Incredibly rarely, and mostly only on the first few bars
what are "scenes"? I just noticed this on the web app
~$sudo Suno
I see what is happening. Your like an undercover boss...
Lol
wait..do I wish? how much they get paid?
With how much time we spend on it, does it matter? To get paid to do what we pay to do would be amazing...lol
If you have an example I'd love to hear it. Honestly! That is major control.
hmmm maybe. i got 5 kids to feed (Total recall quote, i actually have no kids and love it)
It's funny you mention kids because of mine, I'd have to make sure everything was all good and go hide to put both ear pieces in. Maybe it has been panning and I'm just in a mono world...
Wait...I'm late to this party. what are you using these for specifically? can it be used in any type of music?
Some of those I have used and they work just fine, I just want to hear control over stereo sound.
I always add more meta tags on my prompts
Oh I use tags, tons, too many...haha
are there any prompts to unbury the sung lyrics from the music? the balances are touch and go
Example for a song I'm working on now:
[Isolate Vocals] [Isolate Instruments]
i thought Suno went nutso if you put too much in the song lyrics area?
It's kind of my trademark at this time, I tend to focus on getting more clarity
I've experimented since June on which one works well, on the genres I'm working on.
Basically, alt-rock, emo, indie <- These are my genres that I work with
i wanna hear this when you're done please
looks like story time lol
Actually, it will be added to my XMAS playlist 🙂
brava
That many tags would just confuse me...lol. I wouldn't even know what was working and what wasn't.
Yeah but... story time with prompts only lol
wonder what an instrumental of that would sound like
Oh hahhaha
You can make Instrumentals by using Lyrics method, you have a much greater control as well, and more characters for prompting
Now that is when I go nuts with tags and instructions
it's like i've never done this before... i've learned more here in the last hour than i have the past few months 👶

Even adding structure to your Instrumental will help. You can put nonsense but with structure. It will make Suno go longer than the norm of the 1-2 minute Instrumental.
[Intro]
&&###, &&###,
##&&&, ###&&
[Verse]
.....
what's this do
It makes Suno infer that you want something there
because i asked earlier this morning if hashtags did anything and someone said they were for older versions
Since it isn't lyrics then it infers you want instruments
then got into a discussion how hashtags don't work in discord
Most of the time
right on
Also, in can be useful putting in [intro] even without lyrics, keeps the ai from just jumping straight into the verse with no build
yeah, now THAT is how I've been doing it
other than just generating a bunch of blank instrumentals to extend from
I did get either by luck or it actually worked, I got it to actually follow a key pattern. I think it was like a one time thing.
Chord progression
cool
I think it was luck
i tried, but not a clue how to make the one that goes above
never worked for me lol
so i blushed, giggled, and scampered off to do it without anything
Lol. It was actually yesterday. I was trying to get the rocking motion of the opera house theme from Battlestar Galactica. So amazing.
lol
[Intro]
Slow, pulsing B minor riff to establish a haunting, deliberate pace.
[Verse 1] [Steady Rhythm]
Bm /// G // D / F# [Violin & Cello Syncopation]
so it does recognize actual written tunes and such?
very cool
Ehhhh..as I said luck
did it do what you were looking for?
It worked but I also did like 20 iterations
I do Kpop, Techincal Death Metal and Drum and Bass, Darkcore
That is why I say go for it. You aren't making something that is going to make or break you, so go nuts try everything
i've been on a weird goth industrial kick lately with an 80's coldwave flavor
Suno is broken for covers. Normal songs work fine just not covers
defintitely scratching some itches
Covers are in beta, so broken may be strong. What is happening?
I've been churning covers for a few hours now... i haven't refreshed my page or anything tho
Still workin' here, even with a refresh
It says that the clip is stuck
Every cover i try right now just gives the exact same results as the base, so, holding off on covers for a bit
Ah, OK scroll up
To where
Wait, Suno is working for you all? Like right now? Im getting errors anytime I try to generate anything.
Long conversation already happened about new safety implementation stuff
Even on songs that went through before
It wasn't my conversation, I just saw it happen
which is weird, because I'm pushing the edges of a BDSM song right now
Lol
it works but its popping out songs 2min or less for me, up to the 20th try now
It doesn't make any sense, as I only used my voice and sang about sunshine lol
It's probably because you're over the 120 character limit of the genre(mine says 200). They pushed some things through that don't agree with the backend.
AND...all the songs sound the same and are lame.
I do the occassional K-Pop, most of my most liked songs are not English, which is fascinating
Now that could be why. Vocals
Yeah, it says limit 200... you are saying it fails if over 120?
I don't know enough, this is new to me and I think everyone, that's why I refer to the conversation above between non normal users
Everytime. I tested it at 121.
'can confirm!
Male Vocals, Dark, deep, Heavy, Hard, creepy, demonic song. Listen and make your own with Suno.
gdi, I dont want to remove anything from my prompt, but im at 137/200chars
Someone had said something about v4 prep leaking through. No idea on the reality of it
That's it then. The 3.5 limits are 3k for lyrics, 120 for style.
i dont think that is possible unless
- they merged the code by mistake into the same code line
- they are doing live testing which can sometimes make that happen
that was just me speculating
this is awesome dude
but if they did option 1, we'd have all of V4
Lol
Is Suno down right now? It always tells me "An error occured" when I try to create. I did delete my chache and tried another browser, but still error.
For a friend of mine, shes way into bdsm, has some uh.. fantasies..
Yea, but at the same time go and change to v3 and see it stays at 3000 characters
got to verse 2 lol... wow
i need better friends
Tbh I wish the devs of this mess called suno would just try their bs on a development server and not on a production
hmm that's a bug. they probably were not even aware of it. someone said they pinged them about it.
I added "Dog" to push it over 120... and it error'd out.
Good friend, no more questions...lol
OMG I just got a song over 2 min...it MUST be xmas
i knew you could do it
i got this
Rip. I reduced the prompt but now its not making dissonant/industrial stuff anymore. Its just making like industrial rock music now.
On v3?
3.5
Oh, thanks...now I know why my account gave me an error, too. 😄
I'm gonna sneak away now. Everyone with errors, there is a long conversation above that may help explain things
@worn flicker's post is where I first learned about the bug
do caps or asterisks around specific lyrics help for emphasis?
or nah
parentheses definitely affect them
allcaps CAN have an effect, but it seems to really depend on the lyrics/flow of the song. They can also be completely ignored: https://suno.com/song/b6e6d14b-1d01-4556-9ed2-4e54b89c75c1
Acoustic Dobro Guitar, Slide, Blues, Resonator, Funk, dobro resonator guitar, acoustic song. Listen and make your own with Suno.
okay, yeah that's been my experience as well
depends heavily on the flow of the song
I have a couple playlists with experiments like that:
https://suno.com/playlist/48ec1c9e-836e-465b-9dd8-2daf43be3e42
https://suno.com/playlist/c9e59d46-82bd-4732-9a02-c7010ca291be
"Enraged Kittens" 😂 🤣
fear them!!
that song is amazing all in itself
i like the evil voices test too... how dependent was that on the persona?
wait... am i using the right channel for this kinda thing?
i don't want to be spamming stuff in a tech channel if it's meant for like... tech stuff and not a ctach all
That didn't use a persona - I did make one FROM it, tho.
I think it's fine when there's not much else happening. Just don't wanna derail any existing conversations
right on
I also have an SFX testing playlist:
https://suno.com/playlist/e5cf28f1-3415-4a7e-8b31-f9299512984a
Hi Sunos!! Question: Is there a way to show Suno a style of music? Like can I give it an example and then prompt what I want based on that style / exmaple?
Well, their is a guy on here working on an app that you give title and artist too and it will give you prompts for it, its not public yet though.
@robust dome
There is a mood, beat, plus, vide of an ambient track, but I want to of course change the key, strings, etc
Not really. You CAN upload an audio file (assuming that you have full rights to it) and extend from that or do a cover of it (as long as it doesn't have vocals or anything that Suno detects as vocals).
Thats what I mean. Its a track I did years ago - ambient strings and dron style. I just can't seem to recreate it with a prompt.
No vocals - just ambient music
Amazing, thanks for doing all that work!
try uploading it and doing a cover
OK not familiar with that. Cover? and thank you for your help
via the track's 3-dot menu:
looking
OH so do I have to make a track first? Or can I upload the track? I see the three dots on tracks I have made. Looks like I have to have a SUB. Its not avail for free users.
If you upload the track, you can then cover it
not entirely sure if that's a sub-only thing though... i'm subbed
OK thanks. I think it is.
If you're receiving this when making a song, check that you're using 120 characters or fewer for your style of music prompt, and 1250 or fewer in the lyrics box if you're using v3 -- devs have been made aware and a change should be pushed shortly 🙏
I wish I could pin this
Did we make a post about this on bugs?
A lot of people won't read or understand, and even new users will come in and see the new limit at 200
And get the error
Those channels haven't been touched in over 3 months 😅
Damn, oh well, let's hope our mods can help then
Ali is quite active, maybe he'll see it
Caps can amazingly influence a song especially when ended with a !. Some of my best songs are due to all caps at the end of the Pre-Chorus and in the chorus. They can also be ignored as SixStringToker said.
yeah i've been trying out a few of the examples in his test samples and they work well
but they work best if they match the flow of the song
I'm gonna have to abandon this song until V4 comes in methinks
it's too complicated and it's chewing up my credits something fierce lol
cover/extend/cover/extend...
all good though, i wanna play around with some of the stuff i learned today
i ahve a song called Whip Crack that I wanna see if i can actually get a... whip crack in lol
Caps is great for metal!
I'm gonna clean up so many songs when V4 drops lol
I cannot wait honestly, I'm saving up my points
i'll throw a paycheck at it, I'm so excited lol
I was wondering if we can? like will it change the songs or keep the way it is and just improve things like the voice. cause I have some bangers here and they are fine but the voice could use a lil more "natural" sound to it
i've found that covering songs and not changing a thing tend to help clean them up
but there's also another feature coming i belive that lets us remaster existing songs
PLUS, i think the lengths of time are going to be increased for soem features, so that should help with the samples it pulls from
all speculation on my part, jsut what I've inferred from pieces i've seen in discord
Welcome to Suno
See Cover as v3.6
lol
So I'm excited too see how Cover will work on v4
yeah, it was a godsend
4:00 - V3.5
3:20 - Persona
2:00 - Extend
Keep these timeframes in mind
That's the maximum it will be using
in V4?
ah, yeah
I don't expect a time increase in v4, 4 minutes is plenty, maybe at some point 5 minutes, but time requires hardware not training of the ai
I'm mostly wanting increased samples for the others
extensions tend to change the voice too much for me by the time i reach the end
start off with Johnny Cash and end up with Zak Wylde
What do you mean?
mmmm.. wendy's nuggies 😉
if i have a cover of 4-5 minutes that cleaned up the song, it should strengthen the extensions from it right? since it has more of a sample to draw consistency from?
then again, perhaps the remastering part will take care of that
regardless, the melodies are amazing from the samples they posted
so even if i completely remake the song, I'll be fat dumb and happy
Extend does not keep the same style by 100%
This is a pain for sure
yeah 😦
That's why, I only make extend not longer than 1 minute
So to prevent the "Hallucination" from occuring
makes sense
yeahhhh
they also need a blend, take 2 tracks, blend the end of one with the start of another, or blend a segment of one into the middle of another
i'd also LOVE to be able to strengthen the direction for male versus female voices
hoping it recognizes that better
and more sound effects
sound effects control would be awesome, right now i try to figure out how to spell a sound effect.
do less than/greater than signs do anything?
I would love to take 10 songs, and then use them into one track, to work as an advanced inpaint function
We know how to do it already
(asterixwordasterix) & (asterixwordasterix-asterixwordasterix)
Just make sure to add Sound FX in Style of Music
ill try that next time, been trying to get rain in a song for so long.. this wasn't ment to be tagged to drew, sorry
oooo
Let me show you
maybe that's what I'm missing then
dark goth, breakbeat, german electronica, glitchhop, punchy kick, bass-house, experimental, sound fx, industrial male song. Listen and make your own with Suno.
its saying buzz
Yes, it's for you to keep track of where it is
@rose jasper how you getting that nice fatty distored drum kick
From my experience, beep, static, distortion (sometimes) work quite well
that is pretty epic though
anyone know how i can make like a repeating part?
like in songs where the singer has a backing track for the chorus, the singer would usually repeat the final part of a sentence while the backing track keeps going. so like
all - im lost in a daydream
singer - daydream
all - caught in a mystery
singer - mystery
Drew "The FX Lab" Everlong
Mix with dark goth, breakbeat, glitch, but you can also try to get it, by mentioning how the drum kick would sound
Thank @wet minnow honestly, he showed me how to do it
perf thanks
oh god
i put Sound FX in the lyrics instead of the Style
like a nightmare
like at the beginning, I put [Sound FX] in and it's going nuts
I've found that they work both ways, but I got better results in Style of Music
Just remember to write it like I did
Otherwise it won't do it
I use this a LOT
possibly too much
apparently there's no such thing... you see Drew's?
anyone else get an email from suno about v4 Loading?
Anyway to get rid off the hallucination honestly
i hear "its loading" and my developer brain says its in stage and ready for rollout.
lol
Hopefully soon then
They cannot tease me like this Brigg...they cannot.
well now i gotta check my email
--->> Developer<<---- I take it literal LOL
i got nada
hmmm
maybe. I hope this is why i'm getting songs that are all 2min or less and extends that won't pass 43 seconds. i'm burning thru credits here.
Save them
id assume a soft roll out, few people here and their over a few week time frame, making sure nothing crashes.
tried making a whip crack test like you did and i just made the coolest sounding gothic industrial trash lol
no whip crack, mind you
Yeah, I've noticed this as well, some Alpha Users have it
anyone know of a way to get Suno to do a break / bass drop in a song?
Maybe it's based around amount of likes, listens, followers
i've had good success with [Break]
guys, suno has https API?
[Break] or [Drop]
I've made close to 3k songs so...maybe./
and i've only been a user since late august I think
I don't even know how many I've made, how can I see this?
you can also try things like [Bass Hit], those are more hit and miss though
Suno has not, and will not release their API, so if you have found anything, it's not from the company.
I went to libray, counted how many on a page and then went to the last full page added it up. 3k
my first page has 4 songs on it but I think the norm is 20 a page
this is why we need a folder structure in suno
yeahhh my songs are ALL over the place
yours and everybody elses
7,280 songs
been begging for it, nested songs man. nested songs

if I didnt' have to work, i'd probably be doing this all day.
i mean. fridays are short days for us, we shut down at noon. Its 2pm here, this is all I'm doing until my girl gets home from work
this would help too, especially when you revisit a song after 2 months
Even better...
Statistics tab
Filtering (the current one is only on one place)
we have filtering, but it needs to be fixed, if I tape in the name of a song I don't want it searching lyrics
i'd love a stats tab.
- Nested songs (main, extends, covers, all go under the first iteration of the song)
- controls to tear out ghosting and bleeding
- update persona so it captures the voice and singing style and not try to emulate the song it came from
- individual music stems
the "controls" thing can be done with AI. it IS a bit complex but there's a plug in that already does it called Clarity. They can try to follow the same idea and do it in the web based. Personally
'I'd prefer a full blown APP of this opn my desktop.
Hmm support has replied to me about the watermark issue which was nice
but its still happening 😭
I find it strange, that Persona does not really keep track of songs you've made with it
Smaller Rips (Song Learning & iCloud)
Does anyone know what this new option is for exactly when importing?
Is there a list of changes anywhere for the new version? 🙂
anyone else getting told from the songs they generate 'this is a fake app go to suno.com/ios'
Persona as an idea is good but it is very flawed in my humble opinion.
no but I am on the desktop
I only use it for Desktop, so I cannot sadly comment
man...to have a true Suno DAW where you put in what you want and you can see it spit out the stems for everything on the fly...omfg.
even as a plug in for Ableton would be cool
well the android app isn't out yet...
im on desktop too using suno.com
I've never gotten that alert, what browser you on?
well I am not getting that issue at all, sounds like your browser may have been compromised
it happens no matter the browser or device
like i tried on my phone, tried on my desktop, tried on chrome tried on firefox idk
ok
omg...SAFE
Sounds like you've got something on the backend which does something honestly
Maybe you have some sort of script blocker
Or maybe cookiee blocker, adblocker latently on your desktop
Firewall, etc
did you ever try to use a fake app that got blocked and now your account is borked?
Did you try to make a new account?
oh dammit
spent the last hour doing [whip crack]
but it's not brackets
you have parentheses
never,
but hold on it persists across devices tho
Oh yea, sorry!!! @glad badger MY mistake!
[ ]
It should be within this

I'm tired
that last one you posted uses parentheses tho?
well, i'll give both a shot
yyooo this watermark issueeee is crazyyyy...
Make only two verses, and use one with () and the other with []
I always use [] otherwise I get them singing things ...they should not...the sing
IP ban? Malicious software usage?
Try Cloudflare WARP and do it
well if i read support correctly its that my account is one of many that got flagged mistakenly but idk yo support said ive been 'helped' but the issue is still there
Yeah, try cloudflare just to be sure
damn now my account is stuck in create mode...its not moving.
first step, i would make a free account and try it, if the free acount generates fine, your account is borked
Suno ...she is the broken
What you mean exactly by that?
check thisss
aw man but if it is borked i was told it would be resolved 😭
It doesn't have enough information for me to understand it, it's just a short message
I do not see any watermarks on my creations currently. It was appearing after the maintenance, but not anymore.
The watermarks that cat is referring to are a bit of a different thing. There are a number of sleezebag sites out there pretending to generate AI music while actually accessing Suno's API without authorization... and charging people money for it.
Suno's response to this is to replace the lyrics on the generations done by those services with the ad that you hear in the track above. Unfortunately, their response seems to be a bit buggy and is also affecting some customers who actually ARE using the legit site.
Ah got it 👍
omg, thanks for explaining!
I didn't understand at all
Well, that's on them then, and great choice from Suno

Personas can be longer than 3:20 -- it depends on the length of the clip used for the persona
yes exactly
but :////// im just waiting for my account to get fixed
Well, I've told you what to try out
Well, it's certainly an entertaining solution (cept when it affects legit customers) but it begs the question... why was/is there no security on their api in the first place?
- Make a free account
- Use CloudFlare Warp
AI using AI to hack into AI, you need to hire Cybersecurity Specialists, and they want at least 120.000$ a year so yea...
Ok I'll try that rn
doesnt happen on the free account :/// 😭
Problem solved, you need a new account for now
Not problem solved, I want my account fixed
it was premium
was told it would be resolvedddd
Cancel that one, use it on the new account
...has the support team responded yet?
Or more accurately, did you follow up when they asked for information about your connection?
Follow up with support a few times per week, provide evidence, rinse and repeat
Yes and yes
They don't check this channel
You're welcome, at least you know that your main account is still not fixed.
maybe it will be soon
It might be the fact that the support lead was out this week, but usually they're able to investigate these matters relatively easily -- if they need to investigate the matter further, I'd take it as a sign that they're trusting that you haven't done anything fishy & want to help get to the bottom of why the algorithm is flagging your account -- if it was as simple as flipping a switch to "off", it likely would have happened, but they want to make sure it doesn't continue to happen.
Genuinely sucks to feel like you'e stuck in limbo, but it's a "help them help you" kind of situation.
And making these changes, might take sometime as well to:
- Re-produce
- Re-authenticate
- Administrate
- Setup preventions
etc, these things take time @noble sequoia
Ohhhhhh Ok I completely understand, then I know what to do i'll wait patiently fr
It might be the fact that the support lead was out this week
Ah sht well makes sense they can take their time 👍
Re-produce
Re-authenticate
Administrate
Setup preventions
etc, these things take time @cat
Oh yea absolutely, well they can do what they got to do i feel a lot more understanding now thanks to u two
if it was as simple as flipping a switch to "off", it likely would have happened
True
always good for them to be thorough ok hell yea thanks yall
Apologies for coming across like a jerk with the ellipses 🙏 -- I know you've been in contact with support, and it would be frustrating to not have an idea of WHEN you'd hear back from them -- if the roles were reversed, I'd be a lot less chill than you are 😅
ESPECIALLY if I know I've done nothing wrong, then being made to feel like the onus is being put on me to prove that's the case -- then feeling like the community support is brushing off the concerns.
When he's back from vacay, I'll see if the support lead's shirt sleeve can be tugged to take a look at this asap, if it hasn't been resolved by then.
Thank you and do not worry i didnt think u were a jerk at all 😯
if the roles were reversed, I'd be a lot less chill than you are 😅
Loll
When he's back from vacay, I'll see if the support lead's shirt sleeve can be tugged to take a look at this asap, if it hasn't been resolved by then.
But thank you! Also I hope they have a good vacation 👍
Sound effects can get weird. Sometimes * works, sometimes [...] works, I have actually never gotten a sound effect to work except for [Bleep] one time on a song I didn't care about...lol. have you heard the whip crack sound effect before on Suno?
only in like industrial songs
but i can't get it to prompt for nothin
it either just says it or it puts like an alien raygun in there
and for the love of god, do NOT put leather in there as a prompt
unless you want an old man singing
trying stuff now like, cattle whip crack, snap, pop
nada
Ok, so here is how I see it. After x amount of times it is time to either leave it be or find another way. I can tell you that I have not been able to get the Sitar to work in 3.5 but it does work in v3. I would like to see the Sitar on v3.5 I find it strange for it just not to be there, I'm not saying it isn't there but I am saying I have not been able to make it happen.
And that is an entire instrument
DO keep in mind..... soundfx is also defined by the genre you use.
yeah I'm prolly gonna give up after this next gen finishes
Have you tried v3?
yeahhh... trust me, this song should know what a whip crack is
If the AI has not had any data to train on/with that has any soundfx in the genre, it won't know what to do
not yet... i'll give that a go
yikes, system error for V3... is there a lyric limit?
Yea, but it is the fact v3 had something that 3.5 did not. I have never trained a model before, but I don't think it should prune a instrument
1250
You can't mix them anyway. I was just wondering if it worked as a test
hahahahaha
I really want to be proven wrong...lol. I want someone to have a sitar generated with v3.5. It is such a famous instrument to just not be there.
When Duduk is there and Sitar is not...haha. it is strange.
Be back
ohhh, that's what that is called... sitar
i was looking for that sound too
think the closest I ever found was a guiro
Yea, it was to go with my Battlestar Galactica stuff
Bear uses the duduk and the sitar amazingly.
Bendy strings...lol
i am going to punch suno...right in the dick today
I am glad I haven't tried to make anything today.
if I break 3min for more than 3 songs I may cry in joy
I've just been reading all this insanity
something is going on in the background for sure. I wouldn't be surprised it v4 rolled out this weekend.
Well something is definitely happening on the backend @worn flicker , which is hopeful
not sure if you can tell, i'm drinking whiskey while working on songs.
And also considering the e-mail that was recently sent out
you got one too Drew?>
I did
I responded saying I would give oral pleasures if I got it early.
think I will get banned?
when did they expand the Styles and exclude styles section?
Damn I don't ever read emails that aren't from college
Visual bug
for me it was 3 weeks ago
oh wait exclude...sorry
oh really? 200 Characters aint real?
Definetly not appropriate
@rose jasper noted
No, it's a bug
Could you pin that @final hazel ?
You done called down the Daisy....oh snap
It's not something to pin, this will be fixed soon. 😉
Zy
It's ok, I'm good reputation with them ❤️
i recently did a cover..ish...of Strutter from KISS but as a Soundgarden song. I can't stop laughing.
Ok good, glad at least they're working on it
Heya, anyone have any luck getting rid of shimmering/unwanted noise from vocals/instruments in Suno? I’ve tried with recreate and cover with no luck so far.
It's been times, where the admins have logged into Woodhouse as well and spoken to me, on multiple occassions @worn flicker , it's funny af
That is called drift.
It can be helped by breaking up the song. That is the easiest way
Does a man making mouth sounds to imitate a sitar count? https://suno.com/song/384e3a14-31ea-4227-b159-976f5b4c57fe
Sitar, Sehtar, Hindustani Classical, Carnatic, Haunting song. Listen and make your own with Suno.
Sorry didn't mean that for you...the nope
You mean stems and going into a daw, was hoping for something in house. Oh well ty.
Covers sometimes help. but...If its not awful I download the stems and go into my DAW and use Clarity or Ozone to clean it up
Sitar, Sehtar, Hindustani Classical, Carnatic, Haunting song. Listen and make your own with Suno.
Well, at least the whip crack sound effect seems to have an effect on the pronunciation and emphasis of the lyrics, so it's not a complete fail
Did you just v4 me?
Got any Mongolian Throat singing songs @naive quail ?
I used a silence persona to tap into the enhanced v3.5
Why do you hate me so much Suno? all I do is show you love
https://suno.com/song/6316e33e-de4e-4e99-b519-77e6d5b6d5df
EDM, Dark Core, Intense Fast beat, Drops. High Fidelity, HIgh Production, Deep Bass song. Listen and make your own with Suno.
Oh yea, I couldn't get it doing the silence cover either. I only tried that a couple times
veeery mixed results with 3.5. It'd be interesting to see if using a silence persona will get us there
https://suno.com/song/f9409ddb-c354-4ee2-b9a6-644300b2a046
I tried to get the guturall noise, but it's difficult...
mongolian throat singing, throat singing, deep bass voice, acapella, deep male voice song. Listen and make your own with Suno.
oh why...why do you hate me so much
https://suno.com/song/61078345-6786-4e37-b847-0d2bc680da34
Technical Death metal, Extreme Metal, Low tuned guitars, Gutteral Vocals song. Listen and make your own with Suno.
Regardless, my whole thing is for a version to be pruned of a famous instrument. I am not even saying it has been pruned, but I am saying I have not been able to find it. In 3.5 I am very intrigued
One thing you should try out @verbal mortar : dog, dog, dog, dog, dog, dog
In Style of Music
No tricks just 3.5
what does this do..this "dog"?
Amazing is what it does
hahahahaa where do I put? at the top
Style of Music
i put.. i try now
So as of now we are all in agreement that in strait up v3.5 the sitar is just not there. I am very curious, since there is no open source or model to see how it was trained or anything it just makes me wonder how it can happen like that
thanks for the convos, i learned a LOT today!
@rose jasper
for you..using our dog dog and nothing else but that
: dog, dog, dog, dog, dog, dog song. Listen and make your own with Suno.
sitar is there... try adding another instrument near it, or something like indian-instruments in the style.
To be honest, I don't think that I did. Even though I hand out that same advice daily...lol. aight. I'll try
Haha
It's definitely hard to get though 😄
Hmm, I guess V4 won't drop for paid users this week. Hoping at some point next week then.
See...get so caught up in details at times...lol
Wow, the f bomb gets dropped and nothing, but dumb a s doesn't work out
Ahahah...
Good job woody
ugh...I wanted a cover of a song Suno just did to clean it up and it kept the 'dog dog dog" prompt I did for Drew....this is gonna be a nightmare. @rose jasper you owe me a quarter
Sitar pretty consistently appears when I say "Indian" somewhere in the style. https://suno.com/song/31cdffa2-d561-4757-8914-b2cf6f086ba6
Indian sitar song. Listen and make your own with Suno.
Hmm. I'm pretty sure I managed to conjure up a sitar with v3.5 recently
@weary wharf already schooled me...lol thanks for digging it in...haha
Ah sorry I did my classic, jump into a channel, see one thing, comment on it, don't see that others already responded etc
Thanks, yes I asked for it..,anyone else want to rain down on my stupidity...haha. I successfully ignored my own advice. Enough to complain about it. So Sys gets his licks...if that didn't sound profane I'd make it into a song
I love this community, I'll be honest. Ya'll are just as crazy as me and have a different perspective. That makes a very potent combination for coming up with new things. Just saying.
I thank you, you jumped in to help me, how can I possibly say anything other than thanks. Sorry I should have thanked Ya'll
Thank you everyone that tried to help me out
Give thanks with a Sitar song... 
I will. I will add it to my current Instrumental. It's why I was so shocked
Is cover generation not working for everyone?
was many tries, different source tracks, but it doesn't work out in any way. At the same time, the usual generation works fine
not only cover i cant generate any song i think they are fixing it @final hazel
is drew everlong still around
We've been seeing that error message in place of "your content was flagged for moderation", so it might indicate something in your lyrics or style prompt that's triggering 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...
why non of this working, yes I have sound fx and sound effects in style. (wind)
[Intro]
(thunder)
(leaves)
(thunder)
(*Walking *)
(thunder)
oh yeah, it removed the astrisks
asterisks at start and end of each word
Sound effects aren't readily promptable, it's trained primarily for music -- it seems like there IS some nature sounds or something present in the training data, since SixStringToker was able to prompt a rainstorm in the style prompt section, but that was basically all they were trying to prompt for at the time
its why I was hoping @rose jasper was around, he showed this earlier and he seems to have had great success with sound effects https://suno.com/song/6daf3ded-b3e7-4125-9d68-dad8a068c4c3
dark goth, breakbeat, german electronica, glitchhop, punchy kick, bass-house, experimental, sound fx, industrial male song. Listen and make your own with Suno.
try using it as
[Intro: thunder ecoing in the silent background]
etc
Multiple ways too approach this, and I've gotten most luck using edm, rave, trance etc genres to trigger sound fx's.
🤷♂️ It seems like beep is respected, but the rest are debatable. A lot of them are read outright as lyrics, and many others are interpretive/subjective.
Style of Music: SoundFX
(asterixsoundasterix-asterixsoundasterix)
[asterixsoundasterix-asterixsoundasterix]
Keep in mind that most sounds won't work, since they are bound to the genre that the AI is trained on.
so trying to do soundfx with grunge probably not going to work? but maybe if I go more symphonic, cinematic,
SoundFX's working:
Beep
Break
Siren
Clap
Birds
Traffic
I should really make a document noting down all the available soundfx's
After 5 hours have passed, still uploading songs, damn
https://suno.com/playlist/0f32a272-785d-472c-9e64-4e1918ab6116 Just tried it 20 times and it didn't work once :/
Same thing if added to the style prompt section instead of lyrics.
F, maybe it just works by setting tempos and stuff then?
hardstyle rawstyle poetry song. Listen and make your own with Suno.
Approaching 10k songs generated, I'm not convinced it does anything beyond "prompt noise", but it's my belief that you don't "code / command" Suno, you talk to it & it "vibes" with your prompt.
waait.. did this happen last night?? lol
I have playlists where I've tested bpm stuff in #1256301881966198805, but none of it seems to work with any reliability. Usually when someone has a super complex prompt that they suggest to someone, it doesn't work when others try it & the response it "I dunno, it worked in my song".
I like to spend my own credits so people can listen to examples of popular ideas to see whether or not they work before they spend their own credits & get upset if Suno "doesn't listen" to a method that's been proven not to work.
from 120 to 200?
Only 120 works -- anything over should give a system error -- the devs are aware and are fixing it. Just a display issue, like v3 saying 3000 instead of 1250
It's a bug, the support team are aware of it
hmm does scanning multiple sites for prompt and feed it to an ai count as evidence?
hmm. generated at 125, maybe it just took "orche--"
For thunder & rain (and other "naturey" things), you'll need to use v3 and keep the prompting to JUST that stuff. You can then extend from it using whatever version you want, with whatever styles you want.
This is a test I did earlier today: https://suno.com/song/b436644a-edb7-4439-8510-e5f5b6ba6d2a
Ambient Sound Effects, SFX, Rain Storm, Thunder, Lightning song. Listen and make your own with Suno.
and this is one that I did back in june and extended from to add music: https://suno.com/song/31a91d5a-1222-4a13-a1fb-6c3df8ca7ccd
Rain Storm, Thunder. Wind, Doom Metal, Horror, Ambient Noise song. Listen and make your own with Suno.
yeah in v3.5, just seems to make vocals more staticy honestly.
oh 3.5 does sooo much more than vocals, but it sucks at nature sounds.
this was done in 3.5 https://suno.com/song/75f07f11-cb9b-4b1a-8e8f-1d10a944ac25
Dark Horror Poetry, Spoken Word, Glitch Audiobook, Creepy Male Vocals, Eerie song. Listen and make your own with Suno.
what i meen is trying to have ambients and sound fx in 3.5 its like the vocals get worse with those tags in it
@frigid orbit has done tracks with TONS of cool effects
"field recording"
I'm high what's up
lol. I know that song. Just talking about sfx in suno
He means, for your songs, do you use soundfx @frigid orbit ? OR are you uploading sound bytes etc?
Man this is a mess, 3 days suno now hasn't been working for me. first day was infinite waiting in line, last 2 days are log in, and create button spins indefinitely..............bummer
Worked 2 out of 2 gens. @worn flicker https://suno.com/song/04979164-25c8-400f-95e8-ac7983791dce But I will add, if I try "sitar, indian-instruments" I constantly get piano.
dreamy sitar-infused psychedelic song. Listen and make your own with Suno.
Have I upscaled this for you yet?
don't think so. Or part 1 for that matter
👀
OKAY -- turns out the fix the devs were talking about was ACTUALLY upgrading the style prompt limit to 200
@frosty cypress -- my apologies, when they said they were fixing it, I pessimistically assumed they were changing the displayed limit back to 120, when they meant they needed to fix the fact that it was still limited to 120 instead of getting boosted
THIS ROCKS
so its 200 now?
sweet
not sure if thats a good thing or a bad thing though, not sure I need more styles to confuse the music ai with
its bad enough I can start in country and end in hardcore metal in the same song, do i really need more?
v3 is still limited at 1250 characters, despite saying 3000 in the lyrics section -- but a 66.7% increase in style prompt size is wild 🤘
I still want to be able to cut/splice. 🤨 if crop was only made to remove instead of keep, that would have resolved two problems at once.
.. I know... I know... facepalm suggestion box.
goes back to my blend request, being able to blend 2 things together, creating a fill between 2 different sections or songs
Anyone familiar w/ post editing AI music using BandLab? I am mastering songs via SoundCloud and notice afterward the dB is maxing out consistently even though the volume is still relatively low. Tried using Splitter feature to find what part of the song is causing the excessive sound but was not successful.
It already blends, sort of, when using replace, so if we could just "crop" to remove instead of keep and let it "blend". But I feel since it seems to have to interpolate to blend, there is no set point that it really cuts to, thus we have emerge , "replace", instead of trim or blend. But set margins to replace and it's varied lengths in samples are difficult to work with. I.e. designate 10 seconds and replace 2 words, and the result is a blended 23 seconds and an added chorus and different vocal. 😂
I know it's a lot of hours just to get it to do what it's doing. All i do is play with it. So, I'm just trying to find the sweet spot and noting issues, not complaining.
Ok, so I can see how this new limit, will really help v4
You can narrow down in such a greater power now
I've needed the style prompt space to add styles for nuances than for genre. So, as example, if I need a more upfront voice, less/no reverb or nudge to get female vocals , I could use opera, Broadway, or kpop, though the styles are prominently, for example, Americana, folk, neugrass, banjo-infused.
stable diffusion prompting doesn't work right? couldn't do like (Clean Female Vocals:2)
just wasn't sure if the :X controls worked or not, never tried. even though my first style is MALE VOCALS, it still likes to ignore it
You are working with a duet?
not normally no, only one song I am working on is a duet, mostly i just want my vocals male, but apparently I write like a chick
it will ignore that 3 times in 10 generations
While duets are possible, it's pretty random, but I found if you include "Musical" it's more commonly used
"male voice"
"female voice"
"duet"
[duet]
[male voice]
[female voice]
oh i've kind of nailed the duet thing, even figured out how to switch male and female in the same line
I've found, if you add the meta tag, it's easier to get it too work
in playing around for some one elses experiment in layerd singing, discovered ((doing this(will give))) 2 different singers and not be background or a whisper.
with duet and opera in the style
If you add inflections within ( ) they will linger throughout an entire section as well
we never really did figure out polyphony in suno, not sure its possible do to it on purpose or consistently
@naive quail how did you structure your lyrics and style to get that output?? That’s amazing!
could of sworn he was typing an answere. lol
woodhouse is probably gonna spank me over all the links, but here we go:
lol
there's a bit of secret sauce in there, but most of it can be found in other tracks I've done. For styles like that, I always start the lyrics with something similar to:
[Ambient Sound Effects: Dark, Horror] followed by some sfx tags or prompt noise (usually in the form of slashes //\//\\/\\\//\ ). The sfx tags don't always trigger the desired result, but they frequently do trigger SOME kind of effect that often still works.
This one still has all of the tags:
https://suno.com/song/135504dd-0dc6-4ec0-9d75-dbbdf0118622
and the tracks in these testing/experimental playlists should still have all of their tags:
Sound Effect Testing: https://suno.com/playlist/e5cf28f1-3415-4a7e-8b31-f9299512984a
Testing [Tag] Styles: https://suno.com/playlist/48ec1c9e-836e-465b-9dd8-2daf43be3e42
Experiments: https://suno.com/playlist/c9e59d46-82bd-4732-9a02-c7010ca291be
Dark Horror Poetry, Spoken Word, Glitch Audiobook, Creepy Male Vocals, Eerie male vocals song. Listen and make your own with Suno.
so whats with the slashes? i've seen those in a few others noise prompt?
it's just text that Suno won't sing... It can keep instrumentals going for a longer period and can trigger cool weirdness.
oh sweet. ill try it out
lol added it after a thunder tag, it went . poo poo, poo pooo.
https://suno.com/song/275b4792-6507-4683-ac66-bc9daaa71fac
192 style of music, and it turned out pretty good, much better than the original
emo, emotional, piano, alternative rock, male voice, melancholic, slow, ballad, flute, male voice, Christmas, choir, carol, positive, orchestra, ensemble, flute, sleigh, violin, blues, lo-fi song. Listen and make your own with Suno.
And gives it a bit more of a better focus, just like what you mentioned @polar spade
Explora las profundidades de [NUCLEON] un track dark techno ambient.
Un mundo de sonidos oscuros y atmósferas subatómicas, acompañados de visuales intensos y la letra en pantalla para que vivas cada verso. Esta combinación de electrónica, techno y dark ambient crea una experiencia inmersiva para los amantes de los sonidos oscuros y misteriosos....
The first part is actually a poem. I have the intro mostly done for the 3rd part, The Ghoul, which is actually Harber 'cause... he didn't run fast enough and got bit. lol https://suno.com/song/173ac910-bc00-4c67-8834-ca70eac7de79
Dark Horror Poetry, Spoken Word, Glitch Audiobook song. Listen and make your own with Suno.
... and broadway
I love Dark Horror Poetry, they are so much fun too make
The explorer in the Barrow's Whisper is actually the creature in Run, Harber Run.
我觉得只是轻微相似,这是日本风格的虚拟歌姬声音,并不是京剧中的嗓音,我觉得这并不完美,因为流行歌曲中也有类似的唱法,但并没有学习到,不然它会是一个比较有趣的玩法。👍
Always good to hear samples to compare. Thanks for the response.听到样品进行比较总是件好事。谢谢您的回复。
Since it may be, as mentioned, a difficult attemp to get a Peking / Beijing sound, it's helps to find how close we can get by trying other prompts. So I threw in Kabuki, vietnamese folk, even anime prompts and got closer to a sound than with just prompting "Peking Opera" . I'm sure things will improve with more AI modeling in the future. 😊
Ok I don't feel so bad. Trying to add it to my Instrumental is not working out. The only time I am able to get it going is when I have just it in the prompt and it won't even bother with extend.
I'll have to start a song with it and then wrap around it with the other instruments. I'm tired. I'll have to do it tomorrow.
Is anyone else having trouble making songs with custom lyrics? I need getting the "clip got stuck while generating" error
negative..
aw man
lol
tired.
This happened to me earlier today. It alleviated itself eventually. I guess it just comes with the territory.
Nope I've got some good ones and can't decide which one to pick. Without any hick-ups.
fair
had a lot of weird glitches last night after the server maintenance and chicken nugget fiasco, but its been fine for me most the day.
Yeah, i've been having this issue since like 7am
eat some chicken nuggets, maybe that might help. lol.. ehh. last night was.. strange..

Wrong thread sorry
PLEASE HELP
i amtrying to generate song with lyrics but its not working. i can generate instruments and stuff without lyrics
Can you provide us with a song link to see what is the issue?
(To double check: are you in Custom Mode and is the instrumental switch turned off?)
Same here
Been waiting for it to work almost all day
Hmm, what is the problem, system error appears?
It says “clip got stuck while generating”
Also please provide more details, like if you are using v3.0 or 3.5, persona being used, etc.
A common error being seen today:
v3 is still limited at 1250 characters, despite saying 3000 in the lyrics section
– TheHarber
Ah, I see, unfortunately I am unsure how to deal with that issue since I never experienced it. Can you DM me the lyrics and styles you are trying to use so I can try to duplicate the scenario?
Oh is that a new thing? the lyric limit right now?
For 3.0 since the maintenance, yes. Thinks broke a bit including the max limit display
One of my very old v3 got hit with a bug that makes it appear as v4 too (but definitely does not sound like v4 lol)
oh that may be my issue because my lyrics exceed that, thank you lol
I'll just need to wait for V4 then

Another similar issue is an old 3.5 bug with that character limit exceeding 1248 character limits. I got hit with that once and took me a while to remember of that bug
https://discord.com/channels/1069381916492562582/1259902708765167709
So this lyric limit is new for maintenance for pre- V4?
I cropped this from the 2nd gen: https://suno.com/song/dff0c129-b6eb-4100-84fc-9f9bc2f892a9
I want that as my ringtone
Delay Sooooon
the website is telling me to wait until my generations are finished even though i have none being made
I typed "test
this is a test
thank you" and it generated this as lyrics
I hoped suno trained their own lyrics model 😭
Refresh your browser it may solve it
is there a way i can edit the lyrics, and maintain the same song?
You can edit your displayed lyrics if that is what you mean. This is done via Menu > Edit Details or going to the lyrics page (/song/) and scrolling to the bottom.
If you mean fixing your lyrics, there is the Replace tool but that is limited to subscribers and I think it may change the music in that region?
I haven't had a reason to test that feature out since I am predominantly a mobile user and that tool works in Desktop Mode
There is also using Extend but that very likely will change the music
i am on desktop tho, will try it after i subscribe hehe thanks
Just don't subscribe only for that tool. Do it for the credits
Just in case: be aware that the subscription is a renewal/refresh credit system, not carry over. That means when the month ends, it will refresh back to the monthly credit amount, not additive.
The exceptions are if you buy top-up credits - which are more pricey - and Suno gift credits. They can carry over but you will need an active subscription to use the top-up credits
thats good info, thanks alot
I'm hoping they take my suggestion at some point and make it possible to do early renewals (instead of having to purchase credits if you run out)
Few days ago i asked about getting back onto my account. you helped me find my @ and a bit of trouble shooting. Thank you, I got it back.
Quelqu’un parle t il francais ?
Just use sonnet 3.5 for writing lyrics, it produces much better quality songs than you can get with any other method if you use a good system prompt
I'd share an example but idk if my songs are permitted to be shared as they blatantly sound like Eminem, and I think it against TOS to share the ones like that.
Could staff correct me if I'm wrong should they see this?
En utilisant suno je n’arrive pas à choisir si je veux une voix d’homme ou de femme y a t il moyen de faire ca ?
Using suno I can't choose if I want a male or female voice. Is there a way to do that?
vous devez inviter via l'invite de style avec un chanteur masculin ou féminin
vous pouvez également inclure dans les paroles certaines balises de métadonnées
No clue if that's correct translations, it Google translate
ils ont un wiki en ligne
yes ici we speak french
Has anyone had any success with getting Suno to use proper Old English/Anglo-Saxon pronunciation? It gets most things right already by default, but struggles with c/ċ, ġ, ǣ, and vowel length. Replacing those with k/ch, j, and ae (as a digraph) messes up the rest of the pronunciation and turns it into this weird modern German/Icelandic hybrid.
for example, here it pronounces the c's in āþecgan and cymeð properly, but not in lāc or swylce: https://suno.com/song/eab2858b-a555-4da4-be59-23af965dd3d0
old english, organ, hurdy-gurdy, fast aggressive, dark, low notes, choir, thunderous percussion, battle, marching song. Listen and make your own with Suno.
I know it's kind of an edge-case/not really what it was designed for, and it's impressive that it manages what it does already, but still
You might find results by getting a small easy for it to sing sample, making a persona, then using that persona for those kinds of songs
That is in fact exactly what I did there already! :P
Ah lol
I hadn't looked at it as I'm on mobile
Then idk, probably just do a refinement loop of getting more and more accurate personas
That's what I did for getting the vocals I want
I trained it on Beowulf, since I imagine that has (relatively) plentiful material for the model in general to be trained on
Are you writing the lyrics yourself or are you using ai to assist
I'm just copying them from the original (it's over a millenium old so no copyright issues lol)
Oh I wasn't concerned with copyright, just was wondering if the specific cadence or something(im not musically inclided) was throwing the model off, I'm not familiar enough with old English but could you try writing it in the phonetics
This is what claude outputed, as an example of what I mean
"Lay-oh-dum iss mee-num swyl-che him mohn lahk yi-feh;
will-ath hee hi-neh ah-thec-gan yif hay ohn thray-aht koo-meth."
Ooh, good idea, might try that
It's mainly just a pain in the butt to do that manually, I didn't realize AI could get so close on its own
Did you use a specific claude bot or just their general model?
I use both their web app and the api for making my songs, I got a long system prompt that guides it to write the positive and negative style prompts, and the lyrics for me. But their api is less restricted if you want a specific sound to the lyrics.
Gotcha, thanks a bunch!
Your welcome, let me know if you want that prompt, though it would need to be when I get off work.
Either.
Or both, how ever you wanna say it. Sometimes you get some funk on a gen, sometimes you upload your own funk
can someone explain shortly how the personas work?
it should copy the voice and style of the song, right?
however, every time i generate something the output is exactly the same song that the persona was based on, as it does not follow the new lyrics input at all
have tried 10+ generations.. only thing ive noticed that it generates something random in the end, either hallucinate or repeat some lyrics
could you share your lyrics and promtps
its likely the source song that might be a problem
the feature bassically grabs the vocals but also the general vibe
well, currently i let AI generate random lyrics and now it seems to work
is there a general rule..? like what i was trying to do is to replace lyrics, some words etc over the whole song + added a verse
(so instead of extending and replacing sections, thought persona might do the trick)
i would advise if you want your songs to sound good and original to not use the built in lyric generator, and instead use claude to write the lyrics.
Could you explain that a bit more, im not following
im not using it mostly, just for a testing purpose
sec, will create a longer wall of text
personas are for new generations
send it in dm if wall of text
the cover feature would be what your looking for im pretty sure
idea: change out some lyrics, some names, add a verse etc
instead of changing sections or extending, might as well use persona
original song
https://suno.com/song/8f2cbb71-dfeb-43ab-9e09-08ce3f1029c0
persona created from it
https://suno.com/persona/0e2c3a6c-d4bd-448f-b4dd-409f1331c5cb
example of a "new" song by using the persona + new (similar) lyrics
https://suno.com/song/0085a847-9f19-44c5-8bf5-9d916fe1d431
That would be better served by using the cover feature
the 1st verse has a change already, so no need to analyse the whole song
ill try the cover, thanks
just edit the lyrics in the text box first and then those lyrics should be changed
also make sure to use the same title to get the same seed as the original song, pretty sure thats a thing(?)
seems like the title does not matter
cover either creates the same song or a totally new song (though some sections are familiar)
well.. must respect the beta tags though and ill try to do it in the old and longer approach
You could just run through generations with the feature until its similar enough, thats what i do when i really want a 'mastered' version. but that burns through them quick unless you itterate on the prompt/lyric metatags
true, might even get something which i would enjoy more
My advice would be to over engineer the lyric metatags, though ive only played in one genre doing that so idk how well itd go for yours.
Thank you SUNO, for putting 200 to style instead of 120. Great Move
Greetings here. Can anyone tell, is suno works better today? Or it still bugging and glitching like before?
I mean 2 days
Must specify what a "bug" and "glitch" would consist of. Yes. There's been some... but ... well... ok... I'm a premium user and haven't experience any bug or glitch that wasn't explained by my actions.
I've experience downtime for maintenance. So rule those out. And I've had hiccups which after 5 minutes don't repeat. Often due to user use fluctuations.
Now , some things don't work well, which seems glitchy and buggy and I report them. Most, seem to improve.
So, yes, there seems to be an improvement on a daily basis.
I guess suno is working better today. Yes!
I read through as many convos, throughout the day. A mass number have been either from newer users that are having trouble with generating off the lyrics they use..... or older users who are now trying to reuse older lyrics from weeks or months old songs (or make persona's from them)...only to find the older lyrics get re-evaluated and filtered. All seem to resolve the issue or just stop responding in chat. So that's hard to tell if a Suno error ...or user error....but occasionally I try to repeat their error (but of course, not use their lyrics) and I can't repeat anything to give me an error.
So what's been happening the past 2 days.... specifically that has affected everyone?
@polar spade I think it's just that instead of outright saying "your content has been flagged for moderation", now it makes it seem like a random system error, so they keep trying -- in this instance, it should be obvious why it's being flagged for moderation, but because it's not as explicit of an error message, people will ask what they're doing wrong.
It's a confusing error... message. . Quite.
Beyond the lyrics, using artist names & their specific songs in the style prompt is the first thing they cover in the moderation page 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...
Someone recently said "I used to make stupid songs without any issue, and now those same prompts are getting blocked" -- I don't know what they mean by "stupid songs", but just because something wasn't flagged by the automoderator previously doesn't mean that it's guaranteed to pass the moderation every time -- sounds like the automoderator is getting better at detecting content they don't want on their platform.
In regards to replicating the issue -- I think the automoderator is interpretive to a degree, and not just a hard & fast filter of blocked words. The founders of Suno came from Kensho, who are working to reduce toxicity in LLMs, so this is right up their alley https://blog.kensho.com/an-introduction-to-mitigating-toxicity-in-llms-c8d1c5600a9b
However, I've tried using someone elses lyrics, and it worked once only to get blocked the second time and every time after that, so it might not have caught something on the first pass, then learned afterward.
I have also been a premium user for over six months. Since the last technical maintenance, SUNO has been generating low-quality outputs. I noticed several messages in chats about this yesterday, and today the issue persists. I have tried changing the text, style, and settings, but the quality remains low on version 3.5. I just want to know when everything will return to normalю
I don't want to waste credits and get low quality sound🥺
Excellent point. That can get a direct response.
Personally, I've been bouncing around so much testing in the tools and not completing songs that I cant say , myself. My results have mostly been just random things i didnt want, some may include the fidelity of the extension or "replace" samples but I brushed it off for then getting nicer sounds later.
That will be something I'll listen for. Glad you narrowed it down. I was taking a wild guess based on other convos. I missed the chats about sound fidelity. Not that I can do anything, but maybe get details that will help me later.
You might have to ask again sometime later to get a more "worldly" response after others wake up. ☺️
i wouldn't expect any movement on that front right away as they're about to launch model v4 and i'm sure all hands are making sure that goes smoothly
Something I wanna contribute to this before I head off for work, I get this error when my lyrics are too long
Not consistently, but it can't be filter in all cases as my lyrics will work eventually
specifically for v3, it says 3000 but the limit is 1250 😅
v3.5 can use the full 3000
@crystal pagoda I think you speak Chinese? Does each character equal a roman letter equivalent or would suno translate consider , such as Ni, to be 1 character letter, or 3, as it translate to "you? I've wonder how characters affect lyrics capacity.
The issues resolve when I take a verse off and just add it on by extending it, I could try to get a video of this if I remember
I do not speak any language other than English
Also as for the characters Chinese and other special characters would take up more tokens I would assume, as they do with language models I'm pretty sure
Oh sorry. My mistake. I'll hold that question for someone else, then. Have a great day! Sitting back on stool.
Lol
😅 I misread it as "I get an error" instead of "this error" -- My bad 🙏
yeah, the "stuck in generation" existed previously, and could happen if the servers were getting slow, but it seems like it's become more of a "catch-all" message as time has gone on.
why does it sound distorted in chorus .. its overclipping
Do you mean 1) where she sings "wise men tra-a-vel far lands" or 2) something else ( you're saying the audio sounds distorted as if it's clipping)?
Yes during chorus/drop not just vocal all the sounds
Lots of good songs are wasted with that bad distortion
I couldn't hear it using ear buds and turning up volume. So maybe my ears don't have what it takes. Did you run the stems to see if it exists in either one, as a stem?
For what its worth: Android Samsung Fold 5 as a Premium user.
could you help me, when i click create, it loads for a long while then says error, failed to create. im under word limit, i tried everything.
when i click creat again though, it says please wait for your other generations to finish.
You'll notice that songs take a bit before you can see the "Extend" button -- that means the songs are still processing.
Free users can have 4-5 songs processing before they're limited, and paid users can have 20-21 before they get the error. You'll need to wait for your last few songs to finish before you can make some more 🙏
well its been 4 days since i actually opened the website
so i dont think any other songs are being generated
If they've been stuck in the generation queue for 4 days, you'll need to contact support@suno.com so they can look into fixing the account -- that doesn't sound like anything the community would be able to shake loose
okay thanks for the help mate!
unfortunatly there are only articles
this is the exact error im getting:
A clip got stuck while generating.
No credits were used and you can retry.
I don't hear anything that I'd describe as clipping on my iems, but there's definitely some noise that rises and falls throughout the track - not sure if that's what you're referring to or not.
There's a ambient-kinda background effect with that synth in the beginning, not sure what to call it; sort of a staticy wind-like sound. It's fading as her vocals come in at 00:08 and completely gone by 00:10. That seems to rise and fall a bit throughout the track, but really becomes prominent enough around the 01:07 mark that it does make the vocals sound a bit distorted by the final chorus.
Clipping sound over limiting compression
Hihats also can cause that "sss" sound around 6k hz, turns into a bad noise after 2-3 min mark mostly.
That error has been coming up lately in regards to content being blocked for moderation -- https://help.suno.com/en/articles/3198209 -- there's been a lot of instances of prompts that used to get past the automod that are being blocked now, like using specific artist / song names in the lyrics or style prompts, etc
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...
Is the song generation down? Every time I go to make a song it refreshes the page and the song isn't there
The system asks me to verify that I am human every minute.
Before we provide any further assistance, how many Rs are there in the word Strawberry?
3
So ChatGPT 4o? 😄
Okay -- Suno must be making a mistake, because I can assume that you're either human, or running chatGPT o1-preview, which is basically close enough these days. 😛
(stole my joke, you brat)
Other's have been experiencing that issue. Report it to the bugs channel. I had it happen just once yesterday, but others get plagued with it like that.
Thank you
I think that kind of security stuff is triggered if Suno is having a hard time verifying the authentic connection -- whether you have a VPN, any browser extensions that impact how websites load, any aggressive antivirus software, etc
An easy way to check if it's localized to the device or if it's account-wide is to try to use Suno.com from another device, like your phone's mobile browser if you usually use desktop -- if it's acting the same, you might need to contact support@suno.com for them to look into the account & help investigate what might be causing the Captcha to trigger
THX
Anyone noticed that sometimes, songs using personas, stop at 3:17. I had a few creation like that, ending in the middle of a sentence at 3:17 every time.
Yes, you'll get that every time, around the 3:20 mark. It's park of the persona package.
What's the best unofficial API Library or Tool?
None. Suno has not released any API and those that do use it are done through illicit means.
Accounts using such methods have been identified and, in most cases, penalized
Oh okay...
So instead of making an API they punish even them who do that for them (where you still then buy credits from Suno) to use that API, okay.
Yes because it is against the TOS (see the datamining clause) as Suno has not released an API system to anyone at all
They are just funded by investors and don't care to make money over API?
I honestly would not be surprised if Suno's legal team are putting a case together against certain sites and repos, at least against those that are based in USA
It is up to Suno how they run their service.
If they were in it solely for the money, there would be no free plan users, or not the system as is.
I mean free plan makes sense to get more users, but API is easy to do (they already have it for Frontend) and makes a lot of money. But instead of doing that even going against tools that make API possible? Insane to me
It is a small team and they are focused on other priorities for now.
Understandable, but then why putting effort into suing tools that do it for them?
You still buy Suno Credits etc.
And most people would use an API to make an opposing service using their infrastructure. Not really seeing a reason why API should be allowed, at least to anyone not partnered
Regardless, there is currently no API service being offered by Suno.
For further questions about this topic, please direct it to support@suno.com
Yeah
You could TOS that out, but they could also just make higher pricing for API so even if other did that they would earn much and could improve app faster
Again, please address such suggestions or concerns to support@suno.com
(I am not a member of Suno anyways so nothing I can do nor is anything I am saying on their behalf.)
If you want to have an official reason as to why Suno is not allowing API usage, at least at the moment, support should be able to address that
uhg. Library maintenance is so tedious
It's not 3:20 - that's only for personas that use 40s worth of context from the source clip
Boss Tones only seems to use 18s, leaving 3m42s as the cap for that particular persona -- it wasn't primed specially, it was made from a full 3m song
@charred merlin nice find. (I've never seen anything but ~3:20 mark myself, thanks for clearing up my assumption)
I honestly don't care about api. I assume there are other music ai generators out there that are literally copying Suno and what it does. They have similar quality to V3.5 and a 4 minute max clip...
copying Suno
Selling unauthorized access to the API, pocketing the cash, & leaving users with an inferior service devoid of early-access features? Yeah, there's a couple & it's a shame, since the users might have trusted that random site to provide commercially licensable songs, which may result in their "unknowingly stolen" songs getting flagged when they try to monetize them 👀
Precisely why I have no problem adding Suno to my budget
hard to find developers nowadays that actually try to make a viable product and have passion behind it
Hi all, does anyone know if Matchering has a maintained successor?
I haven't followed up with the DIY side of things, but https://www.bandlab.com/mastering has 4 different settings & offers .wav download for free, no limits
Tried Bandlab before but it is somehow creating muddy stuff or at least sometimes adds some plaque
According to the matchering github, it's been integrated into the UVR5 desktop app -- that might suit your needs?
Will have a look, thanks. Let’s find out if it’s different
Was there any issues you were having with matchering, or just saw that the last release was oct 2022 and wondered if there were any projects that had built off it?
The latter, two years of inactivity for an Open Source Tool usually means „discontinued“ or replaced 😃
Or sold 😂
I was lucky to have stumbled into LANDR early on, I don't recommend it outright because I prefer to suggest free (preferably without limit), but they've been phenomenal to my liking so far
Lately I saw a vid where a guy was claiming all the online mastering Tools including those at distributors would just make use of matchering
Cannot compare of course
https://youtu.be/wZRV2H4PK0Q?t=458 He calls out how LANDR is different at 7:40 -- they're literally the pioneers of AI mastering, with over a decade's experience -- looks like he also shows Bandlab at the end of that spiel there 💪
I can use more than 120 characters in the Style Box? When did this happen? 😲

