#📚┃suno-school

1 messages · Page 181 of 1

abstract sable
#

Oh I have. I guess it's inconsistent. It's still AI I guess

pallid inlet
#

have a link by any chance?

keen iris
#

Guys it's just unacceotable 2/3 of all covers by v5 is buggy af.
The bug is like when vhs was doing this disturbing repeating thing, many rap guys and DJ's like this shii.
Sound quality is good, melody and voices as well, but this thing crushing my whole expirience.
It eats my credits like candies and I don't get needed result only because of this kind laggs.
I begging you fix it pls!
I'm gonna throw money in you as some junkie, I'm in love with v5, it makes me forget about everything, I'm literaly like a kid playing with lego, the real world is not exist anymore for me.
Please fix it or drop missile on my head, I don't wanna comeback to reality 😭 pls help!

hexed oriole
fluid estuary
#

So did anyone solve the intro without wailing or is it still a must to generate an instrumental first?

loud forum
#

hello goodevenig

#

hi friends have a issue with recording and uplaoding in suno pro

#

not able to upload and recording music is limted to 15 seconds

polar spade
graceful condor
#

If I wanna straight cover a song in V5 without changing much except the Language of the song what are the best settings now? Sliders how? And are any style infos even necessary if there is the original Cover Audio I wanna cover or should I still add the Style prompts or keep it empty?

rocky wigeon
#

Error generating replacement
Unknown error anyone getting this while replacing

#

how can i solve that

zenith stratus
#

Hey everyone,
how can I best add sound effects like a phone ringing? I've tried a bunch of different ways, and it worked once, but then the lyrics didn't sound right, and I haven't been able to get it to work since.

white dirge
#

đŸ”„ UPDATED: Suno v5 Studio + Phonic Edge GPT | BoomBoomRaps đŸ”„Search “BoomBoomRaps” in the ChatGPT GPTs section

buoyant cairn
#

how do you make the singer sigh?
ha ha haa does a laugh
but i cant get a sigh.... he just says it lol

# 
I ordered an army of penguins, ten thousand strong, sleek and precise!
My plan was to rule the cold heavens, with wings made of malice and ice!
I stood on my tower of sorrow, I launched them with villainous glee!
I screamed to the heavens, "FLY, MY ASSASSINS! BRING THIS WRETCHED WORLD TO ME!"
[Sudden stop]
thud, thud, splat, thud, splat,
*sigh*
[Deep male baritone - muttering in frustration]
...Rats. Turns out they're flightless. Now I have a letter from the R S P C A.
And a weekly course on "Avian Wellbeing." It's on Tuesdays.

it refuses to sigh, i understand the thud and splat not being possible but a sigh? thats pretty common? tried sigh, *sigh*, [sigh] what am i doing wrong?

wispy laurel
sleek stump
buoyant cairn
#

Looked up DAW, my music is not on that level, I make songs to make me and friends laugh or just for fun to entertain.

sleek stump
#

The number of songs released throughout modern history that include a "sigh" are microscopic in number. And since Suno focuses on what is in training Music, and not FX, it's pretty much almost impossible to prompt for it. You may get lucky at some point and find a specific genre subtype that had a sigh in it, but thats like finding a grain of sand in an ocean full of them.

tardy coral
#

Was hoping the new midi feature would be more on the nose and accurate..but it’s similar accuracy to converting audio to midi in Ableton.. which can be a bit sloppy

buoyant cairn
# sleek stump The number of songs released throughout modern history that include a "sigh" are...

must admit as suno models advance, the audio quality gets better, the voice gets clearer.... but its lossing the quirks that allow for comedy songs. I am even having trouble keeping accents solid in v5 (irish accent was my main test)
its depressing.... but may be the same issue... singers with a specific accent are going to be rarer than others.... almost like they are overtraining with v5 now

sleek stump
#

Irish accents have always been somewhat weak with Suno (I do speak Ulster Gaelic), so thats not really saying much. I dont see a big difference for v5 currently.

neat field
sleek stump
#

Ah, you do know Bark was depricated like over 2 years ago..

#

Suno devs have said that Suno does not get trained on FX.

buoyant cairn
neat field
#

I think it has to be in the style/genre prompt not the lyrics

sleek stump
#

Not to mention Bark targeting text to speach as its goal

buoyant cairn
neat field
#

"Geriatric" style/genre LOL

sleek stump
buoyant cairn
sleek stump
#

In other words, they are not likely to cater to a snow flake group.

buoyant cairn
#

I was going to say that I am not the sort to be easierly offended, then realised that snowflake may mean something else in other countries lol

sleek stump
#

snowflake group.. meaning a group that is very tiny and unique

buoyant cairn
sleek stump
#

Businesses dont generally put much effort into supporting them as they are a cost and not a profit.

buoyant cairn
#

still, spilt milk etc.... just got to learn how to bend v5 to do my bidding as best i can lol

sleek stump
#

I have found that Suno is really good at fusion of genre, as long as they are sympathetic or complimentary, otherwise it tends to butcher the track in its attempt.

buoyant cairn
# sleek stump I have found that Suno is really good at fusion of genre, as long as they are sy...

indeed, i have a song where part of the instructions are `
A fusion of a haunted music box waltz and a deceptively cheerful Victorian carnival melody and an eerie children's lullaby with an extremely high fidelity recording
worked well... I do like playing with different style of lyrics attached to different tune styles... A bubly AI singing that the satation is going to explode, they are all doomed, to bubble pop for example creates an interesting contrast.

last bay
#

I like very deep aggressive female, sometimes even with strong persona influence it doesn't work so good but usually is nice with

#

Hm notice non of my public songs use that...

polar spade
# sleek stump Suno devs have said that Suno does not get trained on FX.

I wonder how it determined what was music. If Suno was trained on Spike Jones, Nature Sounds or any of the Sound Effects recordings of Spotify, then that covers many sounds already. đŸ€ž

But once it even thinks its knows what Im asking, it may calculate the odds that it actually should be there. If I want a cuica in a flamenco song is 58% likely, but a paper-rip sound in a novelty Halloween song maybe only be. .002 % chance, so it doesnt produce it once unless I do 300 generations.

lapis snow
#

I’m really struggling with V5.
In simple mode, for example, if I ask for deep house music, whatever it makes sounds almost the same — like Midnight Echoes, then Echoes Mirage, and it keeps repeating that pattern.
And if I want to use the custom section, the Auto Lyrics option is gone, so I have to take an extra step and write the lyrics myself.
My workaround was using the iOS app, where in simple mode I don’t face the same issue, but the problem there is that most of the songs end up being under 2 minutes.
So the only fast option is using version 4.5+ on the phone app

viral pond
naive ibex
#

I don't need "upgrade", I'm on the maximum (premium) one. And I don't see something like "early renew" or "pay upfront" for renew. For example Bitwig allow to buy "year subscription" not just for a year, buy as many years you want.
**But the main question **is if I have let say 20,000 tracks created and I cancel my subscription and not renew (on the current account) - will I lost all my songs, or it will be forever kept on my "free" (not renewed) account?

abstract sable
#

Is there a trick to properly adding sections in-between generated sections in a song without it sounding completely out of place? (in the plain song editor not studio)

shell jungle
#

Is there a way to change the picture when I make a playlist? The Ai generator thing doesn't seem to be working and I just want to put an actual image on the playlist

lapis snow
velvet granite
#

I hope someone can help. Lately soono has been providing a very annoying female voice to almost every single song I'm trying to create. Ironically the music is spot on I could not be happier! But every single track is ruined by this awful female voice that is going up and down the scale with every single word no matter what prompt I put in. I can get it to veer away from that if I include something like indie influenced or blues influenced but then it completely changes the music I'm starting to feel more and more that it would be so helpful if we could get Suno to treat the vocals and the music like two totally separate entities.

I am at Wit's end. I'm generating over a hundred versions of these songs and I love the music but I just cannot get rid of that annoying voice. Does anyone have any insight for me here

quiet solar
#

It's not an easy exercise.

Your best bet is to define that there are a male and female singer in the style prompt and then, much as you have been, prompt in the lyrics as to who is supposed to sing what. I'd suggest using [male vocalist] or [female singer] rather than 'her part' but it will still get it wrong on a regular basis - all you can do is tilt the odds in your favour, and once it gets a portion of the song correct - extend before the next inevitable mistake and try again until you slowly build up the song as you want it.

Expect it to be a credit-eating exercise, though.

scenic sequoia
scenic sequoia
half dragon
torpid haven
#

Wow v5 is bad or I need to change my prompting completely from 4.5+

wispy laurel
# torpid haven Wow v5 is bad or I need to change my prompting completely from 4.5+

Kind of both?
V5 is very much an unfinished product which clearly wasn't tested properly before release
But also, every model responds differently to the same prompt so yes, if you want to get the most out of v5 (and, it can give amazing results sometimes), you do need to effectively learn how to prompt all over again because a lot of old tricks no longer work, or work differently now

torpid haven
#

There needs to be an official prompt guide depending on how it was trained on. They are the only one having the training data.

neat field
#

finally compelled to use the downvote button after using Suno since v3. some v5 generations in my genre's of choice do some very strange generations... but hey, the ones that work are insane (for my preferred genres) so .... i'm happy

vivid wasp
#

Why is it sometimes my song link shows up as just a link and sometimes it shows up as playable within Discord?

wispy laurel
vivid wasp
wispy laurel
polar spade
# lapis snow I’m really struggling with V5. In simple mode, for example, if I ask for deep ho...

Yep. If we're not having fun, then we're doing it wrong.

If 4.5+ is your strong suit, do it in 4.5+, cover in 5.0, then extend in 5.0 to complete the song. Thats another option.

Remember 5.0 is beta and tweaked frequently. So something that works today may be different tomorrow , for awhile. Find its strengths and just use all tools available (which seems ytour strategy already) instead of relying on just one.

loud forum
#

hello

gloomy needle
#

is there any actual support in here? because this is now driving me insane. why cant I BUY credits

loud forum
#

restarted the browser still not able to upload files its not supporting sound files images are shown

crystal obsidian
#

Any prompt to make a villain's song? Something like broadway and charismatic?

polar spade
crystal obsidian
polar spade
velvet granite
sour dock
#

This chat is called tech prompt talk but there's barely any of it lol

tender eagle
tender eagle
sour dock
tender eagle
#

Well feel free to offer some. Not sure what cool is. In this regard.

sour dock
#

I follow 2k people and only 2 or 3 of them even have anything close to complex

tender eagle
#

Oh you mean prompt meta from on site users

#

Yeah well I mean average is average. Which is why it’s the average. I wager much of the ‘cool prompting’ is probably going to be with the more experimental users. Not the average lot who are focusing on making their lyrics or idea sound just good to them. Without diving into the ‘what crazy sound can I push it to perform’ as more experimental users would probably do. Just have to suss them out of the other lot.

#

I can’t think offhand it’s been a while since I’ve actually dug through a bunch of other users profiles. But a good bit ago there was quite a few people who tried a lot of interesting things. Who was it that wrote the prompting guide, I don’t remember he had a lot of interesting stuff back with 3.5 and such I forget who that was though. But people like him you can find some nifty stuff in their profile

primal juniper
#

Every song I've made with the v5 recently, the volume goes up drastically towards the end of the song.

ornate spindle
#

I cant find the difference between v 4.5 and 5

wispy laurel
rapid bluff
#

I'm finding that V5 is not as context aware as V4.5 and in general just takes less risks with the creative nuances, which in my opinion is where Suno has always shined and stood apart from everyone else. 4.5+ absolutely loves an exclamation point (sometimes too much), whereas v5 will disregard it more often than not. But these things might be more or less genre specific

cedar owl
# ornate spindle I cant find the difference between v 4.5 and 5

I find with v5:

Style 50%: a very safe approximation of what you asked for
Style 100%: the same thing but louder

It also seems to generate stuff very samesy even when weirdness is to the moon. Where with 4.5 it would give you more variety with the typical stability risks, with v5 I feel like I just get the same thing with the instability issues.

But it does seem to respond better if you provide audio for it to cover. I'm probably going to do things in v4.5 like I was and try covers into v5 for that reason

polar spade
# sour dock Yeah but nobody have even a smidge of cool prompting really

After being here close to 2 years, ive read plenty of posts of what people use as their prompts. Having said that.... Its typically best to leave it for someone to ask for specific help with a specific goal.

Otherwise, the prompts, styles and tags that work with Suno for the multiple outcomes is nearly uncountable. Hard to make a post to fit in just under a small paragraph about that.

So, what are you looking for here? Often we just refer users to the community Workflow channel for anything more extensive than "How do I make an extension?"

"Cool or complex prompting" is relative to the outcome. Again, when someone gets something they like, it's fine to infrequently post exactly what they did/prompted as well as a link to their song.

Here , there seems more the need for "practical" prompting, as in how to get a spoken voice instead of a singing one.

polar spade
# rapid bluff I'm finding that V5 is not as context aware as V4.5 and in general just takes le...

I must say, ive cover many older songs into 5.0 and find it "follows" prompts in a different way at least.

For instance, where I use "gospel" in 4.5 to get a minor key and vamping vibe, but set down the line in the prompt list to just give nuance, v. 5.0 finds it and give a gospel choir in all my choruses, to the point I am needing to delete many of my prompts to keep 5.0 from making them so prominent.

rapid bluff
#

quality is definitely very good. Personally I just want the creativity and contextual awareness of 4.5+ into the quality of V5

quasi dagger
polar spade
rapid bluff
humble verge
#

It's really gd annoying how v5 crams everything into 2-3 minutes. I'm also continuously surprised at how many instruments SUNO apparently doesn't know. Venture just a bit outside of the mainstream and SUNO goes

frigid willow
#

Is there a way to see the prompt you used when generating a simple song?

rigid umbra
#

not that I've seen

livid trout
#

seems that in studio, trying to change lyrics in the voice channel just adds all kinds of artefact. Shouldn't the algorythm know that it's only supposed to change voice ?

slate agate
#

If I cancel my annual subscription, let's say after 28 days, will I get the refund for the remaining 11 months or what?

for eg, Annual sub is currently $202, which equals to about 16.8 monthly. If I cancel my sub after a few days, will I get $185?

amber ivy
#

Is there an option to turn off the 'snap to grid'? I have a clip I need to move slightly but it's only letting me move it in full beat increments.

#

in studio that is

#

ah I found it by trial and error. ctrl drag

visual stream
#

Hi,
With Bark, can I currently generate a voice (for example by prompting “male, 32 years old”), keep that same voice—maybe by saving the seed ID—and then use it to generate audio of that voice speaking different sentences(not singing) with specific emotions like crying, anger, happiness, or sadness?

shadow burrow
#

I notice a good portion of songs now have 'parts.' Not like different verses, but where songs deliberately change tune and style mid-way or towards the end (e.g. Joji's Gimme Love has two distinct parts, one fast-paced that covers 2/3 of the song and the final 1/3 is much more hushed and slow). Has anyone tried this with Suno?

amber ivy
#

You can add tags in the verse markers where you put your lyrics. IE
[Verse 1 - Tags]
Lyrics
Lyrics
Lyrics
Lyrics

[Verse 2 - different tags]
lyrics
...

shadow burrow
#

Oh interesting, cool.

amber ivy
#

I don't typically make music with lyrics, but I do the same thing (In the lyrics portion) to gett different sounds for different segments.

#

haven't tried that with v5 (though I can accomplish much of what I wanted to do there in Studio)

amber ivy
#

This is especially useful when you want the chorus to have a different sound - or a particular bridge section

#

same idea
[Chorus - tags]
or
[Bridge - tags]

quiet solar
# crystal obsidian Any prompt to make a villain's song? Something like broadway and charismatic?

I had success with something similar using the style prompts "male vocals with a theatrical and nasal voice, fast-talking delivery, cartoon villain flair, humorous tone, speech-like verses with exaggerated enunciation, character performance"

This was aimed at creating a song for more of a cartoon villain, and did very well in the round, except for struggling a bit with a section that required over the top villainous laughter.

If you were looking for more of a Broadway musical style villain, then you'd definitely want to tweak the style prompt accordingly.

lapis snow
# polar spade Yep. If we're not having fun, then we're doing it wrong. If 4.5+ is your stro...

Yes, you’re completely right. Just to mention, the problem for me started when they changed the interface before V5 came out. Even version 4.5 gives me the same result. So I have to take an extra step: use V4.5 on my phone, which works well, then go to the computer to listen to it and, as you mentioned, sometimes make the cover in V5.
I really rely on producing albums and time is essential. Before the interface update, I had no issues. I hope they fix this soon.

bitter trail
#

i am using suno studio im trying to replace some stems to make them a higher quality but it is generate a whole different song anyone know why

round bobcat
#

Ok, [instrumental intro] seemed to get rid of the 'huhuhu' but now I have it in the new song - lol

wispy laurel
#

Yeah it loves doing that, I get it all the time too now

autumn valley
#

Anyone else have noticed that v5 cannot generate songs that are more than 4 minutes long? This is not an upgrade, this is back to the bad old days of v4...

livid crescent
#

is there a prompt that will successfully stop all this ad-lib yodelling in the intro of a song? all of a sudden I can't get a single song to generate without it, no matter what I try.

smoky crow
#

so i did more digging into the sound quality of suno v3.5 and v5 mp3 and i found that : V3.5 -> 320kbps @ 44.1khz VS V5 -> 201 variable bitrate @ 48khz

round bobcat
#

I have done several songs which are longer than 4 minutes

#

To my observations song lenght depends on the server load when you use Simple Create.

ripe galleon
#

if i can ask, could you by chance hand me a .wav file of one your (assumably generated) V5 songs?

#

i wanna see if it covers more of the spectrum than previously and stuff

livid summit
wheat turtle
ripe galleon
#

but this is coming from a free tier user

#

so idk

wheat turtle
ripe galleon
#

places where File limits are in place, yeah that too

#

like you said

heavy pelican
#

hello, I've noticed that SUNO (v3.5) generates a song as it should, but when song is over there is usually a small brake and then its plays a small part of the song in completely different style. First I thought that is because end the end of the song I had ... but now I'm generating without and its the same. Any ideas?

smoky crow
# livid summit Why would you export mo3 if it can export WAV.

I was testing the difference in quality with mp3 and wave. When i download the mp3 and wave and a/b them in my daw i noticed that the mp3 had more stereo mid widths and a slight of eq with a some extra transients added. I then tried to see what the sound quality was between them and i heard almost no difference in quality between the wav and the mp3. especially with house orientated music the transients on the kick and the hi hats sound really digitized. as a sample on 128kbps. So i tried another way and i edited the track put it on steady 128 bpm and then saved it as a new file. Then i downloaded that file again as a wave and also the original file as a wave. compared both waves in my daw and now i noticed that the edited version (i just loaded it in and saved it. did not do any editing) sounds better then the normal wav download option.

wheat turtle
smoky crow
#

I used moises to extract the different stems from the wave file. I get 13 different individual stems instead of 6 with suno. It even splits the drums into seperate parts.

livid summit
livid summit
livid summit
smoky crow
smoky crow
# livid summit Did you extract from mp3?

both as a test. You can hear in the audio spectrum that they 'fixed' the mp3 a bit. but the created hats, the transients going out of sync on the kick, claps which sounds harsh, keeps the same with the mp3 file and the wav. so you won't get suddenly better sound drums etc with a wav file. The ai has a hard time still to create consistent quality drum sounds.

nocturne tiger
#

Is the download button broken for anyone else ?

smoky crow
#

but the issue i found now is that when you edit a track and save it, the extra effects in the eq and stereo mids are now present in the wav as well. So if you download the wav directly without editing the wave file sound less edited. but if you edit your track and even don't edit it but save it away after setting it on a fixed bpm, the exported wave file sounds exactly the same as the mp3

#

remastering a track in suno is a nono. i don't what the ai is doing but it is messing up the whole track as in bad audio quality

twilit orbit
#

Hi, is possible to use only one colour of voice for all songs?

untold briar
smoky crow
#

Ok did some eq testing on the wav file and as i suspected the information above 16 KHZ is zero. A reel wave file should have information above this frequency, so it is cut off.
Did an analyse on this :

Key Findings:

  1. Suno v3 (2024): 320 kbps @ 44.1 kHz
  2. Suno v5 (2025): ~200 kbps @ 48 kHz (sounds better despite lower bitrate)

Verification:

  • File size test confirmed the 201 kbps is real, not fake metadata
  • Your 6:24 track = 9.2 MB, which matches perfectly: (201 kbps × 384 sec) / 8 = ~9.4 MB

Critical Discovery:

  • Spectrum analysis (FabFilter Pro-Q) shows Suno's AI generates content only up to ~16 kHz
  • Above 16-17 kHz there's a steep roll-off with virtually no content
  • This explains why WAV and MP3 sound identical

Conclusion:

  • Suno v5 uses efficient encoding: lower bitrate but better AI model = better sound
  • WAV downloads don't provide better listening quality - they only contain the same ~16 kHz that the MP3 has
  • WAV is only useful for professional editing/processing without generation loss, not for audio quality improvement
  • The 200 kbps MP3 @ 48 kHz captures everything Suno actually generates
#

so the whole WAVE is better in suno is just nonsense

#

it's a wrapped mp3

untold briar
#

There's a difference between "the Suno v3 audio problems are not a result of compressed mp3 encoding" and "Suno v3 wavs are just mp3 files wrapped" which would be a bug.

smoky crow
neat field
#

will you please also test the Remaster function with EQ analysis? Curious if remaster does anything to influence outputs

#

(I suspect "Remaster" is an audio super sampling and upscale model)

smoky crow
#

What you need to add in your prompt to get a 'first time right' is to use the following prompt : clear audio, no compression, best audio quality. <- V5 AI is better in understanding sound design till a point.

neat field
#

so cymbals might be the most impacted audio events, might be perceptible to the human ear too

smoky crow
#

if you do want to use suno in your daw i highly suggest to use the application : MOISES dot ai . There you can split in high audio quality up until 13 seperate stems

neat field
#

I like Suno built in splitter, but sometimes I will also play with Intels openvino one in audacity (local)

smoky crow
neat field
#

I'm still personally ecstatic with v5 outputs so far, while I appreciate the nuance of technically correct EQ, the model is "good enough" (being SOTA) for me

Worth making a bug report in one of the other bug report areas of the discord though with your full analysis, I suspect the devs would be more likely to read there

#

"technically correct EQ is a failed task for V5" or something like that

torn spoke
graceful condor
#

Man I really hope they remove the limitation on signs we can use entirely or at least raise them to something higher than 5000 and 1000. It makes no sense why there is even a limitation the begin with. I mean in most cases we need to Extend anyway so it makes no sense why limiting our creativity. I hate it if I have to painfully remove stuff that would have been great in the song only because it is longer than 5000 signs.

stray ruin
#

Does anyone know how you can see the prompt used to generate a track that has already been generated? I thought this would be stored with the track but I haven't been able to find it in track info.

drifting rover
#

I have an question on suno studio, are separate files with artifacts or is the quality good or is it similar to downloading separate files in pro

half dragon
random tide
#

Forgive me if this is the wrong spot. I was suggested to come here for this question but the upload button on the mobile app is not there. I do have pro.

Any assistance would be greatly appreciated.

Screenshot_20250929-092941.png
half dragon
random tide
#

I did try mobile web browser. It immediately opened up the app with no other apparent option. However, I can just use my PC at home. I was just hoping it had an upload feature on mobile as suggested by Google.

novel pollen
half dragon
random tide
novel pollen
#

Hopefully we get a built in options on modern phones to choose if u wanna open the app if it’s linked to a site sometimes my Samsung just opens the app when tryna access the site 😭

raven turtle
novel pollen
half dragon
keen crane
#

how do you guys force studio to create new stems (eg drums) to sound the same as the song working on? for me generates totally different sounds and i just want to change an intro or a break

round bobcat
#

So shimmer was above 16 kHz? haha

random tide
half dragon
# random tide No joy.

That's still mobile interface - click 3 dots there on top right and switch on Desktop mode.

random tide
random tide
#

They didn't stealthily remove that from pro subscriptions and give it only to premiere did they?

half dragon
#

actually...

#

Here's what it looks like even without Desktop mode on my phone browser... see that +Audio? click that and you'll see Upload option

AISelect_20250929_090417_Chrome.png
random tide
#

There it is!! That's going to make life so much easier. I appreciate you sir! Thank you very much

raven turtle
novel pollen
raven turtle
novel pollen
#

Rn v5 has really nice quality so max mode kinda got me excited

raven turtle
novel pollen
#

V5 is so good I’ve had a song I did on loop for a bit now 😭

graceful condor
#

As of now is there even a way to include Sound Effects? Because it seems not to work I tried a lot. Like phone ringing, TV static, bike ring, stepping on a cat, machinegun shots, people screaming etc. None of them work.

half dragon
graceful condor
graceful condor
half dragon
novel pollen
neat nexus
half dragon
graceful condor
novel pollen
raven turtle
neat field
#

getting some interesting outputs with various seeds by adding "[Track 9]" {or other track numbers} at the start of lyrics when doing an extension of another song.

novel pollen
polar spade
novel pollen
polar spade
# novel pollen Oooo cool I may try when I upgrade my sub this month

Ive been having fun just grabbing segments or choruses from songs I have in a Playlist, cascading each one into its own added track. Overlap and Set the fade between them. And I have a DJ mix or fun medley of all the hooky parts in my Playlist songs. Takea about 30 minutes.

Initially, Suno tends to adjust them all to the same BPM so the beats line up, it appears. So im working out whether thats a good thing all the time.

novel pollen
#

It sounds fun I come from producer AI so when I heard suno V5 was really good I decide to try it and so far I really like it

polar spade
#

Use both and you'll be a Power User! 😁

novel pollen
raven turtle
polar spade
novel pollen
#

For GPT it’s the most limited on usage without a sub

#

But it’s ok

polar spade
# novel pollen But it’s ok

All are really sappy to compliment your songs though. Haha. Apparently not easy to train to be "critical". ( keeping the subject on Songwriting)

orchid obsidian
#

Is there a way to prompt that the song is not to be sung fast or is it the lyrics that causes suno to sing it fast

sweet flax
#

Are the stems having issues right now? I keep getting "A System Error Occurred" on every attempt. Reloaded browser, too.

#

Same issue in Studio.

novel pollen
novel pollen
polar spade
novel pollen
polar spade
# novel pollen I need to learn how to use a DAW seems really fun

If you just want fun speed rap, you can use all or some of these prompts.

polyphonic speedcore, chopper rap, hyper-glitch, frenchcore, japanese metal, hip-hop, machine-gun rap, スレクトロ, hardtek, ポップ, j-core. [Frenchcore just in there for flavour]

wispy laurel
#

Now I have to go and look up what "Frenchcore" is đŸ€Ł đŸ€Ł

polar spade
fluid estuary
sweet flax
#

Ok, if it's not just me, all good. thumbsnail

wispy laurel
novel pollen
novel pollen
polar spade
novel pollen
polar spade
novel pollen
#

Ts is pretty cool

polar spade
neat field
smoky nebula
#

Has anyone had trouble using the "Extend" option in v5, everytime i try and use it , it comes up with an error and dosnt generate anything?,Thanks

supple moon
#

Hey guys.
Is there a way to trim a song?
Reason I’m asking is because Suno generated exactly what I want but at the 30 second mark, the song restarts.

novel pollen
magic shale
#

What I noticed after burning 9000 credits with V5 in many melodic genres (I did not listen to all the songs yet):

  1. Many times the songs have the same voice or very similar voice (no diversity)
  2. Many times the song start with the same key, chords or sound (no diversity). With 4.5+ it was the same problem, I even got the same songs 7 times in the row... is this really a random seed ????
  3. Many times way too much reverb (missing the diversity in the vocal fx and way of singing)
  4. Some time a whispering way of singing and too "pussycat" ...like the vocalist would want to impress someone

But still amazing compositions and arrangements.. the V5 is WOW, but it still need some tweaks.

What I noticed also is that a new model makes amazing songs 1-2 months and than average again.
This could be due to the further training and learning (with bad music).,..
I noticed that after release lf 4.0 and 4.5 and 4.5+...

novel pollen
neat field
novel pollen
#

I love V5

neat field
polar spade
# novel pollen Can u give me a example I’m a little confused 😅

Example of just a flowing train of thought instead of in verse form...

[Spoken vocal] [rap]
Electric storm, pixels bleed raw, data spike sharp as a lion’s claw., Words hit faster than machine-gun law, Chaos conductor, I’m the last outlaw, Riffs collide in a feedback war,, Bassline booming, it shakes the floor.skyscrapers crumble, they beg for more, language of fire, I’m the metalcore lore.

But my prompts worked with a few outputs tweaking until I got it right. Adding or subtracting prompts or changing lyrics will always get a new result. But its a good starting point

novel pollen
hot surge
#

is anyone else getting a system error when trying to generate stems?

polar spade
safe hearth
hot surge
#

any fix for that? or is it on the system side?

hard beacon
#

i think it's a system side error

novel pollen
hot surge
#

cool thanks

#

!

safe hearth
#

tbh i don't think it's very cool 🙂

hard beacon
#

technical errors can happen 🙂

hot surge
#

it's workign again!

hard beacon
#

Problem solved 😉

#

v5 likes to start with "ohhho ohhh oh ohoooo! :S

novel pollen
hard beacon
#

my last 20 generations start with ah "ohhh" 🙂

#

and no.. it wasn't in the lyrics..

#

update: last 24 generations

#

^^

hollow hare
#

Hm maybe try adding some sort of instrumental at the start? I haven't tried that yet though
Like [Short catchy guitar riff], etc

hard beacon
#

in this case it's electronic music 🙂

hollow hare
#

Ah, replace guitar riff with just instrumental

hard beacon
#

26 generations 🙂 i think it's a part of the elctronic music game... but i also have some nu-metal songs.. suno likes to start with ah "ohhhh ohh ohohhh" 🙂

#

every model has it's own style.. 4.5+ likes to scream.. 5 likes "oh" in the beginning 🙂

gleaming solstice
#

Any classical composers here interested in following Suno because of the possibility of generating AI mockups of MIDI?

raven turtle
#

I’m getting that damn ooh ahh every generation too! It’s v5 shimmer!

heavy pelican
# torn spoke Why use 3.5?

I don't understand the question - because is the newes (free) one? does version has something to do with problem I have?

torn spoke
wild egret
silver dirge
heavy pelican
daring cloak
humble verge
#

I'm increasingly getting vocals added to instrumental tracks, so just leaving the lyrics box empty clearly doesn't work.

wheat drift
#

Its only me or after V5 release you can not give a "theme" on the custom lyrics tab? I was very used of just adding a theme for the vocals, and it would create a song using that theme, put for exemple, if i put on lyrics " lyrics about night drive " it keeps that as lyrics, repeting the theme. Am i the only one who noticed that? Now i need to generate a full lyric and then create on top of it

proper berry
ebon ingot
#

Not sure if anyone needs to know this, but I --finally-- found a repeatable way to create a prompt, that allows for chorus melodies, where the lead is male and the female vocals are just harmonizing.
The key here is: JSON prompting.
here is an example prompt I used:
{
"style": "Gothic Darkwave with Irish Folk",
"tempo": "Allegro, driving",
"vocals": {
"lead": "Deep bass-baritone male, tense, bitter, always in front",
"support": "Ethereal female siren, Irish folk style, light and haunting"
},
"instruments": {
"bass": "Urgent, melodic, shaped with Celtic modal scales",
"guitars": "Flanged post-punk, sharp edges, echoing Irish tonal patterns",
"synths": "Layered analog pads, darkwave foundation",
"fiddle": "Prominent Irish fiddle, traditional Celtic ornamentation, modal phrasing"
},
"mood": "Angry, atmospheric, folkloric edge"
}
Song can be found here:
https://suno.com/s/hnw1gfidTLBkncry

Suno
Guard Your Heart

Listen and make your own on Suno.

▶ Play video
proper berry
#

impressive

half dragon
green pike
#

dear lord on this uptade I swear they do something to the "Extend" feature. I can't extend the song further than 5 minutes on a 4;20 minutes song. All the extensions are just so short It's barely called an extension. Do you guys have a secret formula to fix that?

ebon ingot
magic shale
half dragon
ebon ingot
half dragon
ebon ingot
quiet solar
# ebon ingot I pasted the lyrics section as I used it here... sorry I had cleaned up the lyri...

Ah, I was going to say - it must have been a mild miracle that Suno correctly assigned lead and support vocals without prompting.

Even with a prompted version, I still found it as hit and miss as narrative style prompting that specifies two seperate singers. For example, switching it to a female lead and male support with prompting still led to a male lead two times out of four, and in two out of four unprompted cases the 'support' alternated verses with the lead and duetted on the chorus without any direct prompting. Tests were done in both v4.5 and v5.

However, what a flipping amazing sound; one of the v4.5 generations was absolutely gorgeous. JSON prompting clearly works, and it's useful for having a very character-efficient way to communicate a lot of information to Suno.

Not a silver bullet to male/female duets though, at least in my trial run.

sweet flax
novel pollen
jolly mantle
#

Hello, I tried to upload one of my songs to Suno and get a copyright error for the lyrics. However, the lyrics are absolutely mine. Has anyone had this issue?

quiet solar
velvet granite
#

I am trying to use an uploaded snippet of my voice as a vocal guide, but suno keeps returning 5 second songs since thats how long my snippet is. the snippet is for tone. I keep reading that it's a minimum of 3 seconds my clip is six. Help?

hidden yarrow
proper berry
green pike
#

I'm going to commit a war crime how is possible to THAT be COPYRIGHTED

image.png
#

........

image.png
#

I've literally generated the song fully song, wanted to change the final part because it's too generic, try to put ANYTHING on this section (even blank, as you can see) and it's says it's... copyrighted?

image.png
#

well, I should stick with the original one then.

#

what a shame.

royal kelp
#

Dear Suno, please train your nu-metal model on some Korn

polar spade
steady granite
#

anyone got issues with projects getting stuck after exporting a song..i.e. the new song gets stuck saving/processing for the lst hour

green pike
polar spade
#

.

polar spade
# polar spade Got it. Its also common for Suno to find groups of letters between words that c...

"...Suno is not allowing me to turn off "Allow Remixes" for my own songs if I used my own persona to generate it...."

@twilit comet I checked a song made with my persona, and was able to turn off Remix. You may check both Visibility in the menu and Song Detail / (More Options) to make sure both have turned off. Also, sometimes when turning one off and checking the other one, the first one might turn back on. Turn the second one off, wait a minute and see if the other then turns back off.

polar spade
ebon ingot
velvet granite
velvet granite
young geode
# wild egret will this work better for v5? do you know if this new version incorporates more ...

I've yet to use the new version, so I don't know.

Is assume that the new version reacts the same as 4.5, so all you need to do is focus on the main genre you want and layer in aspects of other genres, with their specific instruments/play patterns, using a combination of focus prompts, weightings and scales:

Main Genre: Ambient Low-Fi. Secondary Genre: Dreamwave (40%), Etherealwave (35%).

Weirdness: 70%
Style Prompt: 75%

And then whatever else you want.

quiet solar
hidden yarrow
rain spire
#

I’m on a Premier plan, but the Make Persona button is always grayed out when I try saving from a cover. Other Premier users seem to have this working. Do anyone know why?

reef quarry
velvet granite
heavy pelican
steady solstice
#

Quick question: is there a way to prompt SUNO to do a "BOY BAND"? You know, those choral parts where all the boys in the band sound like Nsync or Backstreet Boys?

rapid bluff
raven turtle
#

The problem I have with v5 is all song generations do not resolve. They clip the waveform. Meaning there's some jiggery pokery to get a proper 'ending'.

Screenshot_2025-09-30_at_07.47.04.png
hollow hare
wintry compass
#

I’m getting copyright issues with my lyrics. But it’s bullshit because I’ve literally made this song , it’s produced it before and now there’s an issue?

steady solstice
#

do you mean CO strike OUTSIDE Suno? (like in YT?

torpid haven
torpid haven
# reef quarry

I hope there is a page where I see what my credits are used for when. The same way it is with any other service I use.

candid moat
#

Help, Model was automatically changed to support your selected conditions.

raven turtle
sour stream
#

howdy everyone, question - if i want heaving breathing, IE britney spears used to feature heavy breathing like in slave 4 u, how would i prompt it? ive tried using () and [] but it either says it word for word "heavy breathing" or ignores it

young geode
torpid haven
#

Did you test with 0%?

#

Put one genre with 0 to the front and see if it sounds still like it

young geode
# torpid haven Did you test with 0%?

No, I would test it with a Genre like Rock, Retrowave, Pop against:

Rock (50%), Retrowave (25%), Pop (25%)

Rock (25%), Retrowave (50%), Pop (25%)

Etc.

torpid haven
#

Yeah but the 0 would show you if weighting works at all

torn spoke
#

The end result of trying to get a near identical verse to the first one using extend and a persona.... after like a 100 attempts

image.png
torn spoke
#

in the prompt

#

might influence it to care about the inline tags

sour stream
acoustic glen
sour stream
#

When I name artists it blocks it

torn spoke
torn spoke
#

what you can do is find out exactly what gengre the song is from, and the period which it is from.

#

in the same way that, instead of just saying "rap" you'd say "west side rap in the 1970's"

sour stream
#

Would I put the "heavy breathing" in () or in style prompt like [female vocals, heavy breathing]

torn spoke
#

() is for lyrics, [] is for effects or specifics

sour stream
torn spoke
#

LIke

[Chorus]
Cause I'll
ALWAYS get up (Always get up)
ALWAYS get up (Always get up)
I may fall but I'll
ALWAYS get up (Always get up)
ALWAYS get up (Always get up)
...
#

"chorus" is in brackets, and the backing vocals in parentheses

#

I could put [Screams]
In a line there somewhere and it'll likely try to include that somehow

sour stream
#

Brb gonna try

torn spoke
#

woops

flint havenBOT
glossy spoke
#

anyone know how to get it to follow exactly the original audios melodies and chords? I have found since v5 release the quality is a lot worse. especially in its ability to strictly adhere to your prompts even in older models

torn spoke
#

... there was no swears in that

#

what

#

oh well

flint havenBOT
#

đŸ’© is on the list

torn spoke
#

Ohhh

torn spoke
# sour stream Brb gonna try

You can specify what kind of chorus/verse/bridge you want too

[Crescendo Bridge]
Healthy? Lol fat chance
Tasty? In my hands
Cause I don't care what I eat
Don't play by life's rules, I cheat
You see, my life rules, all meat
So baby turn up the heat
Because from where I stand
I know you'll understand
That the risk of cancer, strokes and blood clots
aint got ****
on mother **** hotdogs stands!
#

Let's go! No bad words

#

:D

sour stream
#

letsssss go- it did it!
[Female heavy panting and breathing over bass drop]

torn spoke
#

if you like those kind of female vocals, like sensual, go to my profile and try the song "just like that"

sour stream
#

Imma take a listen in just a sec

torn spoke
#

It's not a nod to anything in particular, just an experiment to see how spicy I could make a song

sour stream
#

Oh - she says "like that" several times in "slave 4 u" so I thought maybe Britney fan as well

sour stream
torn spoke
#

There were a lot of lyric manipulation tricks that had to be done to get it to work

hidden yarrow
wary portal
#

Question: if I export my stems in MIDI format, can those me used in my notation software?

polar spade
# glossy spoke anyone know how to get it to follow exactly the original audios melodies and cho...

Just using the sliders and adjusting as needed seems to be the most influential.

As Suno would not "clone" your voice, it can do a good job at generatively emulating it, but ive never gotten anything consistent. Only a close proximate once sliders were introduced.

That said, your question was exact "melody and chords". Suno does well with that though by setting Audio slider close to 100% on a Cover. If it misses a note or chord, perhaps thats likely. A very clear, ambient-noise-free audio upload is preferred. But ive had great results with 4.0 in the past with a very incoherent and noisy recording. With version 5.0, I have not experimented as much with uploads yet.

And have no style prompts.

If there is singing, lyrics are preferred to Void Suno hallucinating over an instrumental.

What have you tried that seems to not work well?

glossy spoke
polar spade
glossy spoke
torn spoke
#

Ultimate consistency trick discovered:

Take the song, cut it off after the chorus, right before the verse starts. Copy paste that onto the track, essentially creating Verse1 (V1), Chorus1 (C1). (see the image of the track)

Then extend from the timestamp that the original, non-copied version stops.

ALSO, create a persona from JUST the verse you want. And then use the settings shown in the image when generating

image.png image.png
polar spade
subtle trench
#

Hey there, I'm just wondering if it is even possible to download songs that are in my library through the Suno AI Android app as the option isn't showing. Can someone please send a screen shot of the option?

#

I genuinely can't see it. There's the allow remix option instead of download.. Btw, I'm using vivo y20g running Android 13, the Suno AI app is also updated to the latest version.

Screenshot_20250930_191109.jpg
polar spade
#

@subtle trench Yea. You're missing the download option. You have the other options in order, same as mine.

#

@subtle trench You might clear the cache in your phone or restart the phone. Your latest model update should be 1.30.1

Last Android update was September 18th I think.

polar spade
manic saffron
#

I have a really good song but it skipped a few important lyrics at the intro part I keep trying to edit them in with the editor but it doesnt work what are ways to add them without ruining the rest of the music?

graceful condor
#

Why does Studio need a sound file at least 6 seconds long? This is so dumb. I mean if I need a sound that is only 2 seconds long why should I stretch it and later cut it again. I am not allowed to upload a 2,5 Second SFX file. This is beyond logic.

polar spade
# graceful condor Why does Studio need a sound file at least 6 seconds long? This is so dumb. I me...

I dont know. But banking off the idea that, as a generative AI, an audio upload is treated same as a Replace in the editor. It needs a relative amount of noise to compare to the left and right of any point in order to tokenize the audio later into something coherent enough to then be able to cover or extend it with a similar melody line or chord progression or even vocal quality.

If you could at least upload just two seconds, all it could ever do is give you a recording. Thats it. It could not generate anything from it. It would hallucinate and do its own thing, similar to extending from a 2-second extension.

Totally a guess as Im an idiot to these things, but in my mind, seems a comparatively reasonable assumption.

I often add enough silence in front and behind an upload of a 1-second audio and just clip off the edges, as you mentioned. A bit tedious, but simple as well. It tends to be a "recorded sound only" , but now, with Studio, we can splice it anywhere and remaster or cover to blend it in. .

glacial cobalt
upper shell
#

Is there maybe a good guide to follow?

hard beacon
#

remix with "new vocals" ?

#

just an idea 😉

#

@sharp vector : short question.. are you the guy with the BoomBoomRaps GPT ?

jagged narwhal
#

Suno changes my vocals to another one how can I keep?

quiet solar
sharp vector
#

No in either case.

hard beacon
#

ok 🙂

quiet solar
jagged narwhal
hard beacon
#

well.. you type in your vocals and suno is changing them in the output?

jagged narwhal
hard beacon
#

so we are talking about uploads with vocals and then the cover function?

timid gazelle
#

3.5 not working with persona? Why

half dragon
jagged narwhal
#

I also cant add later because it changes everything 😄 time etc

quick grove
# jagged narwhal

I saw a trick to kind of bypass this using either RVC or Kits.ai where you have Suno sing your song, then make a clone of your voice with RVC or Kits.ai and use the voice clone to convert the song to your voice. đŸŽ¶

hard beacon
#

@jagged narwhal i think suno is the wrong service with the idea you have 🙂

#

sounds like you have a "ready" song and you want to pimp it up with suno.. or something like that 🙂

quick grove
# jagged narwhal

You can always have Suno make the song and you just record your vocals separately and add them together in a program like Audacity.

jagged narwhal
hard beacon
#

you see.. we don't have enough informations about your intension 🙂

quick grove
# jagged narwhal huh? 😄 😄

He's asking about what you're trying to do. For example: Are you trying to make a song that sounds "like" your voice? Are you trying to make a song "using" your voice that you uploaded with the song? Are you trying to upload your vocals to use as a persona? That kind of thing.

hard beacon
#

thanks 😉

quiet solar
#

@jagged narwhal I'm still not 100% sure what you're trying to do, but if it's "I'm singing something as an upload and I want Suno to add music but leave my singing unchanged", then that's quite tricky.

Your best bet would be to get a Suno cover you like, with the changed voice, then export the instrumental stem as an MP3 and record yourself singing back over it in something like Audacity. Then you can export the final version with Suno music and 100% your voice.

If what you want is Suno to clone your voice based on a small sample of your singing and make whatever new lyrics and so-on that you put in sound like it's you singing, that's beyond the capabilities of the model.

hard beacon
#

in other words: you can't YOUR VOICE on suno tracks 🙂

subtle trench
#

I have updated my phone today morning, currently I'm using the September 1st sec update from Google which is the latest. Also, the Suno AI app is running on the latest version which you specified. It is downloaded from Google Play store as well.

Screenshot_20250930_212752.jpg Screenshot_20250930_213441.jpg
#

Is there a fix sir? Let me know.

hard beacon
#

did i miss something? where is the problem??

neat field
hard beacon
#

sometimes i don't understand the questions.. is it only me?? đŸ€”

neat field
#

no only you, i see the confusing post above. i think they forgot a word/sentence

hard beacon
#

phew... good 😄

quick grove
#

I was lost too

queen linden
#

Good morning! any tips on refining a good teck sound? I would like to incorporate a HAL (2001) voice... I've achieved the silky, sinister calm but.. it doenst have the bot feel... just sounds very james bond villian.

quick grove
#

Are they vocals that you're writing into the prompt?

half dragon
subtle trench
graceful condor
#

Anyone else having this problem? EVERY cover song I generate lately has this male country-western voice. It is not either of my two saved personas. (posted this question in the general chat, also.)

pure river
#

But it's possible using high quality models of vocalists and singers to do that

#

I tried that since a year and half ago .. headache.

#

In a couple of years it'll be possible max two to three

upper shell
pure river
#

RVC's are still being developed and they aren't perfect either.

queen linden
#

I cannot get vocals even close to what I want... was very happy when screeching, soaring swooping female vocals came under some control... Thanx Dasvon..... but now... its been replaced by breathless babydoll vocals.... deep disticntion...

spice birch
#

(v5) Well I knew it was a matter of time before The Voice start sounding like a robot.. v5 sounded good at first but now at the beginning of the song today they sound robotish and like its losing its voice. JS...

obtuse night
#

hi, anybody understand why we can't extend the song more than around 3mins ?
i try to extend my song which is around 3m and extend from 2:55 (ish)
All of my extension only around 13secs each.. 🙁

half dragon
obtuse night
placid aspen
#

F**k V5. It's as close to unusable as it's possible to be without actually having a huge sticker on it which says "UNUSABLE". 90% of iterations are now routinely ruined by the track-jumping bug. Like a delicious looking cake which, on closer inspection, turns out to be a dog-turd.

half dragon
#

It got more of those in the last couple of days.

neat field
#

v5 can do some interesting things. made with Pro sub, releasing this one is Apache 2.0 licensed. not public yet though.

post-screamo, breakbeat, math rock

half dragon
#

Mostly in Style - in Lyric, [..] are meta tags which might be considered mini-prompts for song sections.
Some also define [..] before lyric for prompts as well.

crude tundra
candid moat
torn spoke
half dragon
#

Like you I never got much success with it in lyric tags... so I rarely use it, but others use them extensively and I admit I can't tell if it's actually following the instruction.
I tend to keep it at minimal and enjoy the randomness.

polar spade
# jagged narwhal

It may be just me, but I feel 99% helps better than 100%, in that its similar to a volume knob. No one ever turns it up to 100%. 100% is more like infinite. Or, if it gives an exact copy then its no improvement over the original.

You may want to try these steps.

  1. Upload audio.
    A. Check lyrics for correctness.
    B. Delete styles or keep Sunos automated styles. Your choice.
  2. Extend audio from as near the end as possible.
  3. Get Whole song (which then is basically your audio upload)
  4. Remaster the Full Song version.
  5. Cover the Remaster version. Audio at 99%. Lyrics only. Not styles.
    Stytle and weirdness sliders at 50% or 1%.

See what that does for you?

raven turtle
half dragon
#

I don't have that issue in Style box (yet! lol) but I did run into some track jumping issue lately (last couple of days) .. other than that v5 is nice.

polar spade
#

If this were a recording studio, we would be ecstatic that we only need to edit and splice a couple of spots. Of course , its not live so Suno can be fire take after take, but the vibe is still there. Take whats good in what you got and fix the small parts. Now with Studio -- even easier.

Show of hands for anyone whose been a part of AV or a studio that much work had to do with splicing film or tape. 👋

Show of hands anyone after year 2000 who started using a DAW and needed to intricately splice tracks by eyesight.

I try to relate to whats positive about these things. 😉 It keeps the momentum going and finishes the songs quicker.

It would drive me nuts (and burn through too many credits) to keep looking for the magic bullet to male a perfect take, so I dont have to fix the small things.

half dragon
#

Well said.

raven turtle
polar spade
#

I thought it was cool at 5 years old that I could splice my little reel to reel with cellophane.

raven turtle
#

Yep. If I want scatman bee-bopping at the beginning of each track I’d have prompted it! 😂

copper cedar
#

I used to be able to use my voice and extend a song get pretty good results on the vocals. Now it won’t even do extended at all. I try it with cover and it doesn’t even do the voice. Why does Suno keep making their new versions worse than the previous one. And why should I even continue to pay?

raven turtle
#

Getting pretty consistent vocals with changing minimal stylistic prompts but the warbling at the beginning of each generation and the “needle skipping” is a pain in the ass.
Fidelity increase is the big bonus for me. If Max Mode does what I hope it’ll be the holy grail for less post production heavy lifting outside Suno environment

#

Same here! Not sure if it’s genre specific but it’s certainly a pain. I put “vocal hums” in negative prompt which can help

#

I’m Electroclash, electropop, industrial dance
 so it’s not confined. Damn.

#

I’ve got a sort of workaround generating in 4.5+ and then cover into v5. Seems to get round it a bit

#

Not bad at all. Worth a shot. I tried stem splitting in studio and cutting the offending vox out but it’s hit and miss

#

I’ve had that happen. I end up splicing other generation vocals back in using external DAW (Logic is my preferred poison)

dusty sparrow
dusty sparrow
dusty sparrow
# polar spade It may be just me, but I feel 99% helps better than 100%, in that its similar t...

yeah there's a reason why a bit above 80% goes into the red, because you're now telling it to Hype prioritize either the audio or the style, if you do both then it tries to fight itself on which one is more important, usually with audio I'll do 50-70 and style (which is your idea and direction) I'll make 60-85, and I always put weirdness between 0-15 and 15 is still high for my liking.

hollow hare
#

That's quite...something. How far were your parameters cranked to? Try to never set it to 0% or 100%, they tend to break things a lot. Even 99% weird, for example, typically sounds better than 100%
(I am just assuming here that one of the scales caused the issue)

#

Huh, even more weird

#

When you exclude something, the pattern is actively searching for that aspect to exclude.

Ironically, if your original prompts very rarely or did not have that aspect, the exclusion will cause that aspect to more frequently.

This is because of the biased pattern search where the AI was seeking out the excluded aspect

E.g. the odds of a guitar in trance song is very low. When you exclude Guitar, the AI will be seeking patterns related to guitars to avoid. However, because the chance is very low, the pattern used for searching for the guitar influences the final patterns used, causing more guitar to appear.

TL;DR: if the you exclude a prompt that normally would not appear, the AI will bring more of that excluded style

dusty sparrow
#

that's a weird/difficult one, if you don't touch the scales wouldn't weirdness be set to 50 as default? I would always turn it down to almost 0 personally. it's going crazy with the gender halucinatins, mine did that to me when I did a rap song ONE time, I'm an EDM producer with ableton so I was messing around and was super confused by that behavior. it could also simply be the formatting of [Verse 1, male rapping] and then [Hook, female vocalist]...

#

in the editor you can just retype out the lyrics and it should try to fix those glitches

hollow hare
#

As for why it would hum, it depends on the genre. For examples, raps woukd very likely have humming at the start or a yo, yo because of the prevalence present in rap music.

Typically being more specific about the subgenre may help 2wth that since the pool pattern is smaller

dusty sparrow
#

@unreal sleet so one thing I really want to suggest is learn song structuring/producing, example: when the song is stripped down to just piano for the female vocals, that is specifically a production "choice" people need to better understand these types of things to build the flow of the song/story they are telling, because then you can prompt exactly what you want it to do.

"i got a song that was perfect but its got some words glitching" is what I was talking about, you can FIX that type of thing in the editor

#

it doesn't matter if it's typed right or not, it's an AI halucination and you have to accept that and learn to overcome the issue when it happens

#

yes you should be able to, I and many others have, sometimes it's literally as simple as just retyping out the lyrics within the editor itself and saving or generating again, I haven't done it since before v4.5+ just v4 I think?

hollow hare
#

Ah yes, forgot to mention that. In addition to the first two lines in the lyrics, I added :

  • "At the start, have a plucky synth." To the second sentence of the style prompt
  • Excluded "vocals at start"
#

Hm I think Persona would be the better way to go. Or used to be, unsure about it these days.

Hmm, I would crop the portions you want influenced and do what you mentioned.

Be aware that iirc, inspo only samples 4 songs (random) from the playlist so have a 4-song playlist uniquely to the aspects you want

#

Good ol' randomness.
The other option is to extend from the portion you want, hm

#

I haven't made anything new for quite a while (been too busy at work) so not sure what other advice to give as I didn't experiment much with the latest features. Hopefully other people will chime in

#

It should be influenced by the source, though how much I think how much is random?
The tradeoff is you will have a lower max time for new gens based on these because persona, cover, inspo, etc. reserve a portion of the resources allocated to the song. https://discord.com/channels/1069381916492562582/1422252931943501935
Typically a persona made from a short clip tends to take less resources but may reduce the chances to study what you want

#

Hmm try being more specific about the type of pluck you want? E.g. plucky synth from a Hawaiian House (adding geographical location can influence the outcome a bit)

#

Though I think it was more effective in 4.0

hollow hare
#

Ah

humble shale
#

having issues trying to get a bells sound similar to that used in Hell's Bells or For Whom The Bell Tolls. Instead, the generator keeps giving me small, high-pitched bells that are used frequently. Not what I want. I'm unsure how to get the loud big bells that I want.

#

fr?

#

oh, I didn't read

#

💀

hollow hare
#

It's moreso for looking for a specific subgenre when you have an idea of what you are looking for (assuming that the LLMs do not give you the correct answer)

I found out about Crime Jazz back in the past because of that

humble shale
#

apparently "2,000-Pound Bronze Bell" didn't work in the prompt

#

"Church Bells" also didn't work

polar spade
humble shale
tall fractal
#

has anyone figured out a quick solution to fix the vocal clicks and pops in the generations?

#

external software etc

polar spade
#

Whether it works, i don't know. But for just prompting, you might need to make some spooky Halloween songs using deathnell so Suno feels the vibe, then make a persona out of the one which you'd then add to whatever epic, metal or concept song you got going.

polar spade
halcyon sparrow
#

anyone know how to improve song duration i mean i have a 4.5+ song 5:39 i cover in v5 i get 2:22 or 2:50 max

polar spade
tall fractal
#

ive noticed the exclusions dont work very well haha. but u can extract stems and remove them if u want

glossy spoke
# glossy spoke yeah I have found that the issues are now present in 4.5 as well.. it could just...

it seems after I posted this everyone is reporting similar issues. I have a bridge that is a different style to verses and choruses and it completley changes it. when I do the bridge as its own generation no problem at all but together with the rest of the song it freaks out. to the @shy yew team it would be great to be able to have some type of implementation where you can set the vocal melody and rhythm with midi and same with the unlelying chord structure it would be awesome to have studio be able to lay a vocal guide and chord guide that is rock solid and the song follows the core with eblishing around these core objects. I'm assuming the model is using the training on specific genrese to determine its production which obviously is breaking down when you do something outside of that genre or style. If you can nail how to handle that I think you will unlock a new level of creativity.

lunar epoch
#

Idk what happened but Suno Android app ate my entire data transfer for whole September in just few days

neat field
#

that moment when you accidentally delete the Embedding URL link in showcase and cannot get your discord embed to come back... feelsbad

near flume
#

Don't know what they're tweaks with the female vocals, but generations from yesterday and today are just crap. All the noise bugs from v4 are back (female voices only). And the voices are being cast across multiple stems (vocal, fx, drums) making impossible to fix. I've emailed links to support.

steady solstice
#

////
Hello, I need to prompt Suno to do the usual "all female voices chant the choir" in a stadium.
Does anyone have experience prompting that?
The lead vocals are a male country voice.

neat nexus
steady solstice
#

This is my prompt:
[Bridge / Many Female voices concert chanting]

#

but Suno only produces 1 female voice.

glossy spoke
#

It's really starting to piss me off that it is using the same hooks and structures from previous song generations and its become unuseable. I'm gonna try using a completly fresh account to see if I get better results. I think this is a major issue if its the case.

steady solstice
#

yes, It's like "caching" it's own previous generation. It's like the "seed" it's stuck in the same number. SUNO DEVS, is there a way to have control over the seeds (random number in the system that identifies the variations/generations of a song)?

lyric kestrel
sour stream
#

this, this, this ^^^^^

untold briar
#

If you are getting similar results but not changing the prompts much at all, that may be a characterstic of v5. In some ways it is useful, it means you can tweak and slowly dial-in a prompt to get the sound you want. But it also means if you want a diverse range of outputs, you can't just hit Create over and over, have to change things up. (Easiest way to keep chaning Inspo songs).

solid crater
#

Almost makes one think that - like Topaz VideoAI - there could be genre-specific models... 4.5+ gives me great prompt adherence for some things, a struggle for others where v5 seems to 'get it'. So I swap between them when hitting a break wall on something in my cranium.

humble verge
#

I've gone back to 4.5+. V5 just gives the most boring results. The vocals trend toward monotony, and there's so much hiss now. I can do 40 versions of a song in v5 without getting anything worth listening to. With 4.5+, I got great compositions on the first try using the same prompt. Sadly I think SUNO did something to it, because sound quality is often worse now than it was before. They really messed up with this release.

tall fractal
humble verge
#

I just got a song that began with an applauding audience. That's a first.

graceful condor
#

Does anyone happen to know which style I should use to get as close as possible to Linkin Park's (Numb, in the End) style? I'd be grateful for any help.

humble verge
# graceful condor Does anyone happen to know which style I should use to get as close as possible ...

Who knows? Here's what GPT suggests:

Early 2000s nu-metal / alternative rock. Emotional and melodic choruses with soaring vocals, contrasted by rhythmic, rap-influenced verses. Heavy distorted guitars with tight palm-muted riffs, deep bass, and punchy hip-hop-style drum grooves. Atmospheric synths or piano motifs provide texture. Production is polished but aggressive, with layered vocal harmonies in the chorus and a dramatic, anthemic feel.

graceful condor
humble verge
graceful condor
torn spoke
torn spoke
humble verge
torn spoke
#

Ehh, I find that tuning the advanced options values, along with being very specific in the prompts leads to cool results

humble verge
torn spoke
#

do you generate the lyrics?

humble verge
torn spoke
#

lol then we rule that out

#

hehehehe

#

had to check

#

I imagine that would cause generic outputs as well

humble verge
torn spoke
#

same

subtle trench
torn spoke
#

I have a tendency to put "unique" or "quirky" scattered in my prompts, you could give that a shot

#

I'm very picky though, so regardless of model I only end up with using 10% anyway

subtle trench
torn spoke
#

and what device?

subtle trench
humble verge
torn spoke
subtle trench
subtle trench
torn spoke
#

use desktop mode

subtle trench
torn spoke
#

the app is not good for creation

#

use the browser

#

simple as that lol

#

:p

subtle trench
#

Ik that, I'm supposed to use the Android app according to my use cases.

#

That's why.

torn spoke
#

@humble verge I think you'll just need to experiment more with prompts, turn up the style influence a bit, and specify that you actually want emotion, when you want it, and what type of emotion

#

I only get results by being super specific

subtle trench
#

The main reason is, I'm gonna buy the pro plan through the app because then I can use my Google Play store credits instead of buying it from the official website that requires giving card or account information and has the feature enabled by default called auto renew.

torn spoke
#

Then buy the pro plan on the app, and then just login with google on the browser

#

I don't see the problem lol

subtle trench
#

Makes sense.

torn spoke
#

You're not limited to any given browser/device just cause you subscribed on that device <3

subtle trench
#

Makes sense too. Well, perhaps I should use it like that for now and continue finding a fix. Saving time.

torn spoke
#

@humble verge what kinda song are you tryin to make? c:

humble verge
# torn spoke <@949837627685740634> I think you'll just need to experiment more with prompts, ...

I'm not new at this 🙂 I can mix up my prompts from here to kingdom come, and v5 still won’t give me anything remotely close to what I’m telling it to do. It ignores prompts far more than 4.5 ever did, simply because its narrow scope means it doesn’t really comprehend them.

V5 is great if you’re into common genres like pop, R&B, or rap. But I work in blues–folk fusions, I incorporate Balkan and Jewish world music, I write ballads, and I explore southern gothic. I’ve even come close to inventing my own genre with certain fusions between blues and traditional Balkan music—sounds I’ve never heard anywhere else. V5 just can’t come close.

A big issue is vocals, but it’s also the general lack of off-genre comprehension. It takes everything and pushes it in a generic direction. It compresses long songs into three-minute formats, rushing through lyrics instead of letting them linger the way 4.5 does. It doesn’t matter how I instruct it in Lyrics or Styles—the AI overrides it.

Sound-quality wise, especially the vocals, it’s excellent. But musically, it’s ignorant and lacks imagination.

torn spoke
#

Despite personas?

humble verge
torn spoke
#

You've made the music you want in 4.5 right?

humble verge
torn spoke
#

So does a persona made from the style you want still not influence it in the direction you want?

humble verge
#

remastering 4.5 songs with v5 is quite good.

humble verge
subtle trench
torn spoke
subtle trench
#

Already did restarting, clearing up app data and cache, reinstalling. Nothing worked.

torn spoke
#

Hmm, are you using a vpn?

#

If yes, try without. If no, try with.

humble verge
subtle trench
humble verge
#

ah. And I assume you tried signing out of google and back in?

subtle trench
humble verge
#

do you have the direct email to SUNO support?

subtle trench
subtle trench
sour stream
subtle trench
#

Totally agree.

humble verge
tranquil breach
#

anyone want to challenge this? it help to improve the model i guess

image.png
#

its success

unborn star
tranquil breach
#

its also a combination of two suno generated song

#

chaotic song

#

"Super Warrior"

#

you need to step by step to exam what is too chaotic and tell the model right way to act and motivate it

unborn star
#

Well

#

It's impressive, that you got something with weirdness 100, but I never tried Audio Influence 100% might be the reason that it works, since it looks like Audio Influence > Weirdness

tranquil breach
#

but its sound new on the result, creative

unborn star
#

It might have an effect Weirdness, but not as strong as you think.

AFAIK if you use a persona, you can put weirdness on 100% and get similiar results

#

New obvious

#

Since weirdness is the thing to get out of the base structure song/loops and so on.

tranquil breach
#

i first test it with 100% audio its gave very messy result'

#

when prompt down to tell what should not and what should, it getting better and better

unborn star
#

I see what was the final prompt?

tranquil breach
#

that one in the picture

sour stream
#

My one account has been double billed for months and I'm still waiting on it to be resolved months later

tranquil breach
unborn star
#

I see

tranquil breach
#

are you interested to test and train it without audio 100%?

unborn star
#

it wont work

#

if you don't have any reference it gets wild

tranquil breach
#

never try never know, you got motivate it?

#

its a important skill in engineering

unborn star
#

I tried already, that's the reason I asked you.

tranquil breach
#

hmmm, gave me your lyrics and genre? i try

unborn star
#

I have a documentation about suno over 220 sites long, just collecting data, that's the reason I asked you.

tranquil breach
#

i see, can i have your genre and lyrics , i try for you

unborn star
#

wait

tranquil breach
#

ok

#

wait, i can't see your setting, is this the 100% weirdness?

unborn star
#

You said I should give you a song?

#

Don't you?

tranquil breach
#

ok i wil try now

#

give me sometime

unborn star
#

go ahead

#

tag me, if you're ready

#

I might not be online at that time, but if you ping me, I will see it

#

depends on how much time you need

tranquil breach
tranquil breach
#

originally your prompt

#

together, we and suno just created history!!!!

unborn star
tranquil breach
#

@unborn star

unborn star
#

Ye but what did you changed?

tranquil breach
unborn star
#

but you changed style box?

tranquil breach
#

yeah

#

just add a few negative prompt

unborn star
#

Can you just tell me, what you changed? xD

tranquil breach
#

i tell the AI its has a creator and be steady

#

and a few things that cannot be done when in chaotic mode

unborn star
#

Can you share these?

tranquil breach
#

how should i do?

#

what should i do?

unborn star
#

IÂŽm confused, I don't see what you typed in negative prompt (exluded) right?

tranquil breach
#

messy chaotic destructive chaotic messy at [chorus ***}

#

and be patiently train it

#

thank you Jesus Yeah~!

#

i prayed through it

unborn star
#

xD

#

You just wrote "messy chaotic destructive chaotic messy at ?"

unborn star
tranquil breach
unborn star
#

thats all?

tranquil breach
#

yeah, cuz that's all the negative effect the AI done

unborn star
#

Can you give me the full prompt you typed in excluded?

tranquil breach
#

‑messy chaotic, ‑destructive chaotic, ‑messy lyrics, ‑messy at [Chorus vocal voice [P1→P2→P3]]]

unborn star
#

That's all?

tranquil breach
#

do the song sound new to you?

#

yeah that's all

#

i think its hypnotic though

unborn star
#

If this really works, it would be really nice, since it gives easily varity in any song

#

Could do a lot with it, if it's really that easy lol

tranquil breach
unborn star
#

number on 6? xD

tranquil breach
unborn star
#

When I try to recreate what you did, it gets a bit messy

tranquil breach
unborn star
#

so you mean I should recreate pieces until it works?

tranquil breach
#

yes, it seemed to work within 20 pieces

#

at least its not unreachable

unborn star
#

how much credits do you have? xD

tranquil breach
#

i am premier user

unborn star
#

to test something like this

#

It might help me, thanks for sharing
me too

tranquil breach
#

you might need to go into suno editor and roll the dice to amend , or you can ask people with DAWs to amend for you

#

ummm

#

delete the backing vocals and fx i guess will do

#

depends on DAWs you have , if your DAWs can convert the lyrics into midi form well yes

#

midi form with lyrics*

#

if you are unable to do it, you can DM me

#

i can try

#

but now i got tutor classs

#

you didn;t say that 1st hey haha

eager compass
#

hi guys i need some help understanding how to prompt for drumprograming i been trying to get suno to spit out drum like in this song https://www.youtube.com/watch?v=drsgNdfhV0Q but i only get straight 4 to the floor typ drums whatever i do. When i tried to use "Contemporary RnB" in the prompt like i saw in the prompt of the song used in the youtubevideo from Suno introducing Studio, but then suno ignores my "124 bpm" and start outputing 70-90 bpm rnb songs.... please help 🙂

YouTube
ZaraLarssonMusicVEVO
Zara Larsson - Hot & Sexy (Official Audio)

Zara Larsson - Hot & Sexy (Official Audio)

"Midnight Sun" available at: https://zaralarsson.lnk.to/MidnightSun

Follow Zara Larsson:
http://www.instagram.com/zaralarsson
https://www.facebook.com/ZaraLarssonOfficial
https://twitter.com/zaralarsson
http://zaralarsson.se
http://www.zaralarssonofficial.com/

(C) 2025 Sommer House, under exclusiv...

▶ Play video
tranquil breach
eager compass
#

yeah but what do i call this style of drums.... i been trying to analyse it in sonotelelr getting prompts from othe ais but still get standrad 4 to the floor type straight drums....

tranquil breach
#

Crash

#

Ride

#

Clap

#

Those are drums type you can choose

#

Try get openai to help you write the drums prompt, it doesn't as simple as it was thought

eager compass
#

yeah i know the names of individual drums .... but im asking for the name of this style of drum pattern... like if ask for uk- 2step get a drumpattern like that style if ask for pop i get a straight pattern, but his pattern have backbeats and stuff like Hiphop and for exampler reggaton but as soon as i try to combine pop something is just goes straight and booring 🙂 ( hard to explaing im Swedish).

tranquil breach
#

Its worked well

eager compass
#

ok ill try that thx... my first idea was to find a song and import so i could see the analyse prompt i got from it but that didint help.....hmm im just gonna try to make my own beat in logic and se if ican refine that and use......... thx guys

tranquil breach
jade mist
#

Any info if v5max is actually coming?

tepid loom
#

I am a Pro member and it seems that the new version is only producing 1 minute long clips. This is very strange. Is this correct? Anyone else experiencing this?

#

Every prompt I put it. Extending. New Song. I've been using it for over a year so I'm somewhat experienced.

#

Did not have this issue with the free version.

orchid sigil
#

Prompt wise .. I have a general question, I don't know if anyone can answer it .. How can a poem written by Edgar Allen Poe, which is now out of copywrite contain copywrited lyrics??
I've had this come up a couple of times, with different poets from the 19th and early 20th century.

unborn star
#

Can you show it to us?

unborn star
tepid loom
unborn star
#

Like I said, show us your prompt you using

#

creating these pieces

orchid sigil
red ridge
#

how to make it just make the vocals for your instrumentals and not alter the instrumentals

orchid sigil
#

I can understand if the material is current (70 years or less old) the poem in question was written in 1829 so is almost 200 years old.

unborn star
#

Maybe you take to much inspiration of songs of others

orchid sigil
unborn star
#

Yo, looks like there’s no way around it, sadly.

tepid loom
unborn star
#

How about Style Prompt?

tepid loom
# unborn star How about Style Prompt?

punk revival, post-hardcore, nu metal, screamo, art rock, post-punk, experimental indie, angry, sad, hardcore, metal, melodic metal, dramatic, dark epic, angst, string orchestra, emo rock, kicks, punk

tepid loom
unborn star
#

Hmm

#

then it's a bug

#

Was wondering when I saw that it should work, normally.

jade mist
#

What is this max mode everyone talks about in Reddit... Max mode is not a thing in Suno, at least for now. The issue with v5 (which I have reported via suno's feedback) is that sometimes, the backend chooses v4.5+ even if v5 is selected.

tepid loom
# unborn star then it's a bug

Yep. Total bug. I emailed them but I just wanted to try to solve it here as well. Thanks for your attempted assistance!

trail junco
tawdry scarab
#

the way lyrical rejection works is to avoid copyrighted recreations

#

not to provide maximum rights

orchid sigil
sour stream
tawdry scarab
#

maybe it's not a perfect filter, but yeah

orchid sigil
#

That's fine .. it's not clear when it rejects lyrics .. 😄 now sometimes it is, but the copywrite one isn't to be fair.

tawdry scarab
#

yeah

sour stream
#

do we not have a known issues section on here or am i overlooking it?

tawdry scarab
#

afaik that just gets passed around word of mouth in the server and then the team picks up on it from posts in the bugs and feedback sections

#

frequently people believe there are problems that stem from their prompting

#

it's not like you can trust people when chance is involved

sour stream
#

seems like the team would benefit so that they (support) wouldnt have so many tickets for known issues

tawdry scarab
#

you could put it in one of the suggestions channels, like the web one, ask for a known issues page

covert meadow
#

Noise and electrical sounds- How can I avoid this? I even re-produced them with tag of "no noise, no electrical sounds." Any suggestions?

jade mist
#

I can tell when it's 4.5+ by the exact same characteristics in the songs, plus the fact that when it's 4.5+ and not 5, it doesn't understand the lyrics, it sings them through, but when it's v5 it actually understands them. For example: [spoken] Sing with me! in a live song in v5 results in the audience actually starting to sing the choru.

tranquil breach
tawdry scarab
#

if it's no it belongs in the avoid prompt with no qualifier "no", if it's in the positive prompt "no" barely means anything

raven turtle
grim gyro
# orchid sigil Prompt wise .. I have a general question, I don't know if anyone can answer it ....

Because the backend filters for the text of any public work and it's not an interactive interface, you can't direct the platform to understand that the text is within the public domain. However, you can get around this by using phonetic variants of words, and breaking multisyllable words into fragments by using a dash between them. Obviously this raises the issue of copyright ownership if you've recreated a song under copyright and attempt to publish it for the purposes of earning income.

tawdry scarab
#

mooted is a fun word in this context is that jargon or something you thought to say

raven turtle
#

Here you go

grim gyro
# raven turtle

I'm assuming this is being rolled out slowly as that option is not in my interface.

eager compass
#

not in mine either

tawdry scarab
raven turtle
grim gyro
hearty grail
#

hi there, I was reading above and noticed someone is putting style information into the lyrics section. Does this help? I've been adding a bit of direction via: [Chorus - sung], [Verse - growled+screamed], but I've been putting genre: X, Style: X, Mood: X, and the like within the style guide. Does adding style information into the lyric section help at all?

#

Also, I don't see Max Mode either.

hearty grail
grim gyro
hearty grail
#

neat! Thanks!

tawdry scarab
tepid loom
tawdry scarab
#

v5 doesnt fill just to fill like previous models did

#

so if it feels like 2 minutes of content it's a 2 minute gen

tepid loom
#

yeah. that's not the issue. its downloading as a 4 minute clip and only appearing as a 1 minute clip in the web version.

sour stream
unique stone
#

Max Mode was not a feature of earlier versions.

raven turtle
#

I don’t recall seeing ver having this feature available before as if it was I certainly would have used it as it sounds exactly what I need regarding fidelity

sour stream
#

I was just copying and pasting what Google AI had listed. Sorry

#

I'll remove

unique stone
#

It probably picked up speculation from places like Reddit 😄

jade mist
#

Why max time for v5 generations is not shared?

last bay
#

should also be 8?

neat field
#

getting some very unorthodox outputs by adding a hidden Track number to the start of a song. may be nothing, may not even be in the dataset. and interesting

Song lyrics
Stuff and things```
remote flume
#

Cover option is broken!!!!!!

#

It's just generating new song and not covers

half dragon
paper steeple
#

I think I've done some wrong?

remote flume
paper steeple
#

I was trying to cover a song but I couldn't create anymore

remote flume
#

Yup, so it's just not me

#

Cover is broken atm.

paper steeple
#

create in general won't work for me anyore

radiant breach
#

dang, who broke Suno

paper steeple
#

I don't know what I did wrong

#

i need help

radiant breach
#

its either working now, or clearing cache helps

paper steeple
#

I'm trying to email them or something

#

This is going to sound patethic but this is getting me anxious

remote flume
#

Custom cover is bugged. It's just creating new songs instead of cover now. I'm quite pissed off tbh.

#

I guess a lot of user in general chat don't use cover option đŸ€·

hard beacon
#

i'll try the cover function.. let's see what is happening 😉

paper steeple
#

If it's not that, my keyboard doesn't seem to work on it

#

I tried going into other browsers but it's the same

hard beacon
#

well... cover function does covers for me 🙂

sour spade
#

Wait, my bad, realise that this falls more under the suggestions and feedback category

remote flume
hard beacon
#

not for me 🙂

#

well.. of course it's a little bit different.. but it's not a complete new song

remote flume
#

I do see there's a black box that's there

solid crater
remote flume
#

I can see it, but it's blocking something

polar spade
remote flume
#

This is all I see now

weak osprey
#

Is anyone else getting "an error occured while generating the clip" when trying to extend a song?

Screenshot_2025-10-01_at_12.29.06_PM.png
#

I've had this error for like a week straight on multiple browsers

remote flume
radiant wasp
#

it's not just creating... everything is all out of whack. Been trying to contact support since 1 this morning. All the subscriptions are messed up. Can't do anything

paper steeple
#

Okay, it gave me a song, after I reloaded my page!

#

That was terrible to go through

remote flume
#

Instead of covering, you see it makes a completely new song, even if I'm covering from the same song as you can see in the arrow

severe kindle
#

suno offlie?

#

i click create and nothing happens

hard beacon
#

for me it works

remote flume
hard beacon
#

just clicked it 🙂

severe kindle
#

i clicked create on pc browser nothing happening

hard beacon
#

i am on Mircosoft Edge and all works fine (in germany)

remote flume
#

I only have 2 options for the advance options

hard beacon
#

what else should you have??

#

there are only two.. there are just three in cover mode

remote flume
#

I think audio resemblance

half dragon
remote flume
#

This is pertaining to cover

#

Not create

hard beacon
#

ohm.. no.. you are in the creation mode

half dragon
#

Your screen doesn't look like a Cover

hard beacon
#

🙂

remote flume
#

Watch

#

I went to the cover and it's not popping up

hard beacon
#

hmmm... what's that browser?

remote flume
#

Google chrome

hard beacon
#

for me it the layout looks like the app but in browser.. if i open up suno in my google chrome it looks very different

remote flume
#

Yea, it was working fine an hour ago

#

Then now I can't cover

hard beacon
#

but that doesn't explain the different design for me 🙂 ...hmm...

#

@half dragon : are you using Suno Studio??

severe kindle
#

literally offline

half dragon
hard beacon
#

if i have an electronic dance track... is it possible to add an typical mixable dj intro with studio via the "add instrument" section om the beginning?? do you know that?

half dragon
hard beacon
#

ahh.. interesting idea...

#

i don't wanna ruin the good sounding output with cover versions till i got an intro 🙂

lunar hinge
#

not wrong

hard beacon
#

well.. i can also create the intro with FL Studio but.. i will try your idea 🙂

hard beacon
#

ok... this option (with my workflow) doesn't work well... but hey.. FL Studio and VST Plugins and Drums Samples will fix that 😉

#

GOT IT with cover function... style was "Figet House with a 32 bar dj intro (claps and percussion only)" only.. worked well.. now it's time to cut it in a DAW 😉

late mica
#

Hi, I need some help. The platform isn’t generating any songs for me at all. I’ve already tried different browsers and devices, but nothing works. It feels like my account might be blocked. Who can I contact to get this resolved?

My username is: @miguelangelrosasocampo

Thanks a lot!

hard beacon
#

you are not the first one today.. maybe suno has some country specific issues...

acoustic estuary
#

V5 keeps starting with "LIVE" sounding concert vibes, I DO NOT LIKE IT! I've done negative prompting both in the song [Studio Recording] and also prompted under the 'Style' tab and placed it as an 'Exclude' style, yet it still does it. Does anyone know how to solve this or am I relegated to using V4.5?

graceful condor
#

V5 is such a mess. Literally 4.5+ does so much better extends most of the time. V5 regardless how you change the sliders is all over the place.

sour stream
#

still at a loss of words as to whom at Suno gave the approval to release v5 as is

sour stream
hard beacon
#

i think all the "problems" in here are depending on music styles... i heard this "LIVE" problem the first time... and i never got his problem by myself..

#

what kind of music are you talking about?

sour stream
#

EVERY song of min has stupid harmonizing intro

hard beacon
#

u are talking about "ohhh ohh ohh" ?? 😄

sour stream
hard beacon
#

i know that problem 😄

sour stream
#

and ive yet to have a successful remix cover

hard beacon
#

but i got these problem only with EDM and pop style till now

#

*this

sour stream
#

electro-pop, rap, electro hip hop, trap - same issue

acoustic estuary
hard beacon
#

yapp... let's call it the "synth prompt problem"... 4.5+ was only screaming vocals... every model has his own character

sour stream
#

AND

acoustic estuary
sour stream
#

every song sounds the same,... like the AI "seed" needs to reset/cached reset/etc

hard beacon
#

Soapy 😄 ... do you have a link for me ?? i want to understand the "live problem"

sour stream
hard beacon
#

What the....

tawdry scarab
#

you have crowd multiple times in the style

acoustic estuary
hard beacon
#

maybe it's the word "cheerful" in you prompt.. and of couse "crowd-chant-chorus"

#

i think "crowd" is the the problem

acoustic estuary
hard beacon
#

try it and give us a feedback... by the way.. nice one 🙂

acoustic estuary
sour stream
#

may i be nosey and give it a listen @acoustic estuary

acoustic estuary
hard beacon
#

same here.. 8.3k generations... but if it's good it will be on spotify, deezer and other plattforms 🙂

sour stream
hard beacon
#

8.3k and no public one.. that's what i wanna say 🙂

sour stream
#

yeah i keep my suno prfile private lol

hard beacon
#

😉

sour stream
#

ill put my spotify artist profile in my bio tho

hard beacon
#

i am curious if the "no crowd" prompt will work for soapy 😄

sour stream
#

fr fr

naive sphinx
#

Does anyone here know how i can do that? Is it even possible? My problem is that Suno ALWAYS change the bpm and add some weird shi on my beat even tho i put audio influence at 100, what can i do plss?

image.png
hard beacon
#

ok.. i am confused at this point

#

😄

#

btw: nice graphic 😉 ....i think you should tell us your "problem" with own words... maybe i don't get it and other people do...

naive sphinx
#

okay basically I want to put my OWN beat on suno so that the AI ​​sings, and finds interesting singing ideas on my beat, to then be able to go over it etc... basically just for it to find reference ideas, NOT for it to change the beat, because that means it's no longer in rhythm with the beat that I make otherwise...

sour stream
#

u can

naive sphinx
sour stream
#

u have to record the audio

#

u cant upload it

naive sphinx
#

i dont want to record audio at first, bc i suck at singing, i want a ai to do it for me lmao

#

i want suno to do a reference tracks

sour stream
#

one sec sorry driving

hard beacon
#

still don't get it... what do you want??? you have a selfmade beat you wanna use?? you have self siinging lyrics you wanna use??

naive sphinx
#

like, he do too much

honest forge
#

trying to get a single simple vocal but I keep getting harmonies lots of extra vocals. Any ideas how to force simple living living room unplugged style vocal - one singer...

hard beacon
#

@naive sphinx : Suno is a music generation AI.. uploads will not used 1 to 1 .... in other words: you can use an upload of your own voice with a harmonry... suno is creating an instrument out of it... the only thing you can try is to upload the audio and use the extend function

#

to all other people in here: correct me if i am wrong 🙂

naive sphinx
hard beacon
#

the upload is an interpretation of the sound...

#

there is a very interesting video about it on youtube.. but i can't find it

naive sphinx
#

i feel like it's not as hard for it to create lyrics over a already made beats, i mean thats what he do but create beats too lmao, so thats sad there is not a options for that, i thinks it will help producers and musicians a lot

half dragon
tawdry scarab
#

lol i just wanted to mention the image is gold

hard beacon
#

the image is lit 😄

tawdry scarab
#

i'm worried about the dc offset in that waveform

#

there's like a whole portion with the tops gone

#

unnatural frequencies

hard beacon
#

hahahaha

#

that no saw or square lead synth 😄

#

@acoustic estuary : and??? does it work???

acoustic estuary
# hard beacon hahahaha

I have company over and we're working on another project, I'll get back to you shortly though.

hard beacon
#

no prob mate

naive sphinx
hard beacon
#

ok.. it's 11pm.. time to go to bed 🙂 have a nice evening....

naive sphinx
naive sphinx
acoustic estuary
last bay
#

i get nothing when hitting create :I

shrewd grail
astral fossil
#

is the Suno promo the extra discount on yearly commitment still going on, please can anyone tell me how to achieve this from within the android app, it seems a necessaty , but can,t find any button or field that supports this methodology Please someone can help ?

last bay
#

should be till october 5

shrewd grail
astral fossil
# last bay should be till october 5

thank you, i assumed the 2nd of october , fact is i can't engage from desktop the subscription commitment. I am asked to do this from within the android app , that's because how i made my ongoing monthly subsciption wich was on an android device. Any way on the android app there is no possibilty to find to do so. Except for regular subscriptions thus the normal pricing let's say, but as i like to change from my ongoing monthly subsubscription to the yearly wich includes the promo i am stuck. Perhaps it does not apply to already subscribed users even if it's on a monthly base. Can't find any pointer saying this, i just assume this, it might be a reason of absence of a visible 'promo subscribe button' in the android app.

naive ibex
#

Could someone at Suno please add an option to disable autoplay?

Right now, after I play any of my own tracks or playlists, the player automatically queues up something else — and almost always the same track:
"The GAZE, pt. II | GameIsHere x LJ."

With over 2 million plays, it’s hard not to wonder how many of those are just autoplay padding. That doesn’t feel like genuine popularity — more like an algorithmic glitch.
And when the same track starts playing again and again after your song — like a broken record — it quickly becomes annoying. After hearing it a **hundred times, **it really starts to wear you down.

I’ve tried reaching out, but the “Feedback” section feels like a black hole. Please give users more control over playback — specifically, the option to turn off autoplay of third-party content. If I’m listening to my own playlist, it should end when the playlist ends — not force me to listen to the same track over and over.

This small change would make a huge improvement to the user experience.

@stuck junco @steady kelp

boreal harness
#

I'm so tired of reviewing generations in my own library or workspace, only to have the player start playing a track from some random user. Just STOP playing.

tawdry scarab
#

i started using repeat

#

at least starting again is neutral

velvet hornet
#

Suno doesnt have a timestamp feature anywhere that I might have missed? Like where vocals starts and so on.

wild egret
#

Can this kind of music be created with suno? https://www.youtube.com/watch?v=RZSfotB_qgQ

YouTube
Retrocausal Memories
Get in the Zone: Ultimate Focus Sci Fi Ambient Music 🎧 to Stop P...

Get in the Zone and Improve your Focus with this Relaxing Sci Fi Ambient Music track perfect for long hours of coding, studying, creative work to Stop Procrastination including Binaural Beats (15Hz Beta Waves).
Perfect ambience and background music for studying, background music for work, ambient music for coding, programming and gaming, sci-fi ...

▶ Play video
#

i've tried this

image.png
#

using V5 here

tawdry scarab
#

try legato, arhythmic, free meter and drums, percussion in the negative. your prompt positively defines what appears but does not preclude rhythm or drive

wild egret
wild egret
tawdry scarab
#

music tends to do that, instead of making more whole generations extend before it explodes

sour stream
#

v5 really doesnt want to let there be male intros does it lol 😭

wild egret
tawdry scarab
#

sure, yeah, though it does mean you can't export its entirity from studio, and at 16 minutes you lose the ability to remaster

wild egret
#

because if the maximun lenght is 8 minutes, it will always explode somewhere within those 8 minutes

tawdry scarab
#

but those are the only limitations

wild egret
tawdry scarab
#

so the negative prompt is in advanced features

#

you seem to have put drums and percussion in the normal prompt

#

this second one gets rhythmic but softly so

tawdry scarab
#

ghost jazz and silent jazz can quiet things down a lot too but with a little lo fi beat

untold briar
white dirge
#

When you finally decode Suno dialect and it’s just LUHHHV MEEEEE on repeat like a sacred scripture.

2c8ba6d4-0856-4c1c-9024-55cacd5bde46.png
steady solstice
#

I mistakenly SENT a song to RADIO, and the only option I see is "EDIT / PUBLISH". How do I "UNLIST" my song? (I don't wanna publish it)