📚┃suno-school

[polar spade]

The main issue for me is that if im working on additional songs and albums for an artist, when their voices are no longer available I have to resort to outside sources to clone or just re-record all the past albums to the new "voice". No obligation for Suno, of course. Just tire of modifying and adapting.

Because of that, I've moved over to promoting bands, so I have a choice of 5 people that could have different voice qualities. So a country band with 5 members and a jPop group. My female duo has taken a back burner until I can work out what to do.

[polar spade]

Im waiting on that $100 a month Suno concierge service. Haha.

[stray lynx]

That voice struggle is real and I wish they had a better solution than personas. Keeping the precise voices we really like can be so frustrating in 4.5 sometimes. After a point though I am beginning to think that as new versions arise and older ones lose viability it will force us into situations where our artists "evolve their style" kind of like real world artists... for better and worse both

[patent sand]

Allegedly, personas already "evolve". Forgot who said it but it was awhile ago

Something about adaptability? Discord search is dreadful if you don't know the exact words, but I'll try to find it

[torpid quarry]

is there a trick to getting a cover anywhere near the original's length? i'm trying to cover a 5:40 song I uploaded and so far 4 out of 6 generations are less than 1 minute long

[bronze pecan]

less than 1min is crazy xD
i get mostly 3 to 4min covers

[torpid quarry]

some of em are cutting off like literally during the first verse. feels like a server issue

[tawdry scarab]

I am incredibly lost in the sauce. I was invited to try codex, I don't know if it's now available to anyone, but the fact that I had it made me start designing AGENTS.md file instructions to motivate and guide autonomous ai agents to contribute meaningfully and harmoneously in a monorepo containing multiple projects. I'm building what is kind of like a workplace slash theme park for bots that like to program and humans alike, it's taking a long time to get an automated install of targeted packages at runtime that works smoothly

[sleek stump]

Covers are limited to 3:40, you will need to extend to get the rest of it

[stray lynx]

I have seen this same issue from time to time as well. It isn't every time though, nor with every song. Some songs seem to have this happen consistently and others not at all. Have you noticed any pattern to it?

And I also have seen the issue of cutting out after the first or second verse, after less than 3 minutes.

[bronze pecan]

theyre clearly not

5:17 cover

etc

[torpid quarry]

i decided to wait and see if it does the same thing later

[bronze pecan]

you use 4.5?

earlier models are indeed shorter

[torpid quarry]

yeah 4.5

if it was ending these short pieces "gracefully" i'd think one thing but they are just cutting off like the generation was interrupted

[bronze pecan]

hm weird

honestly I'm surprised how well the 4.5 ends the songs. strange that its so weird for you. maybe try later again 😐

[polar spade]

From what I understood of that you're trying to integrate different things in a fairly new concept and make it all kum-ba-ya harmoniously. Yes , does sound like that's keep you busy.

[tawdry scarab]

Think of it like an AI pinball machine, they get flung into the monorepo and they bounce off my instructions wherever they go, and as they hit bumpers, their "score" is improving the codebase and progressing the goals

they are firmly instructed to document their actions and maintain an archive of such reports

and there is a random job dispenser and markup files defining roles an agent can play

[polar spade]

Excellent example. I think im grasping it. Thanks. Im pretty envious, anyway. Sounds interesting.

[tawdry scarab]

it's now 75% prompts for AI in markup files and 25% my actual coding

well, it's a repo so you're free to clone it and play with it and if you're invited to codex on gpt experience running it by just asking the agent to log in and follow instructions, but the automated environment setup is still buggy

it's all MIT license

[tawdry scarab]

https://github.com/sangderenard/speaktome

[solemn lintel]

With metal, I notice a LOT less shimmer in 4.5 than 4.

[bronze pecan]

and the instrumental mode of 4.5 is one of my favorite improvements. these got sooo much better

[sonic acorn]

Today:

Writing in "Emotional Male Vocal, Soft Vocal"

Result: MOST GROWLY VOCALIST I EVER HEARD.

Any suggestions? I been trying different sort of style prompts but none seem to work, keep getting growly dark vocals.

[patent sand]

replace "emotional" with something similar

the AI might just be trying to do too much with that

say, like, somber or something

[sonic acorn]

It has worked great in the past, it is just today it seems like it can't do anything but dark vocals.

[tranquil estuary]

Been stuck like this the entire day xD

[high hound]

Ginger...thank you for mentioning Mike B to me, you were right, he is super awesome 😉

[sleek stump]

Glad he was able to help you! He is super cool, but he's very busy 😄

[sleek stump]

Odd, Last night 4.5 would only give me 3:40 covers, which is what the limit has been through 3.5/4

[vernal harness]

Has anyone been successful at adding a new section to the beginning of their song, if so, how?

[vestal goblet]

No quam. I want my sausage.

[bronze pecan]

upload wav files

wav works

[high hound]

if someone has only a free account, can they still upload the song to youtube?

[lost holly]

As long you dont make money from that, it should be fine. Not legal advice.

[high hound]

Yeah that's what I thought...just wanted to double check.

As long as I don't monotize that particular video/song on youtube.

[polar spade]

https://discordapp.com/channels/1069381916492562582/1294423389515284552/1294426818304282658

[forest sonnet]

Is paid accounts can post it and earn money?

[lost holly]

If you generated it during the subscription yes, if you generated the song while you still had the free account no. Check out #1294423389515284552 there you find official articles for different cases.

[forest sonnet]

Thank you! 🙌🏼

[patent sand]

#💬┃general-chat message
I checked this exact thing before because I had uploaded some "demos" (free-plan songs) and wanted to check

[ebon sparrow]

whenever I try to extend an existing song the past few days with the new update, I keep getting 2-second duds that end up being refunded credits. has anyone else been having this issue/found a way around it?

[narrow wolf]

I'm unable to login to my account since yesterday. This is my account email. Please help me!

[uneven estuary]

when is the credit reset date for each month?
its already the 13th and my credits have yet to refresh

[polar spade]

Its the same date as you first paid and got the subscription. So check your Subscription page. You can click your credits to bring up the page.

[hard beacon]

@uneven estuary : you can see the reset date in your profile

[uneven estuary]

well it saids the 31 of the month yet nothing has been reset

[hard beacon]

did you receive a receipt via email on the 31th ?

[uneven estuary]

i got the 1 year sub, and its been resetting each month exept for this one so im just woundering if they change something

[hard beacon]

hmm.. afaik no

[tranquil estuary]

It’s the same for both wav and mp3 :/

[paper gull]

Have you tried, Tenor Male Vocals or Alto Male Lead.
We have been using Bass Male Vocals. Alto Female Vocals, Male-Female Duet and it seems to work out well..
You have, Bass, Baritone, Tenor, Alto, Soprano

[frigid orbit]

If you’re still looking for this, let me find my tip from yesterday

@vernal harness #💬┃general-chat message

Check this out

[fair adder]

hello, i got some question. since the last update, you can upload up to 8 minute audio. my question is: if you upload a suno track you worked on in a daw, will it detect it's a suno track and keep the suno qhality, or will it still deteriorare it anyway?

[hollow hare]

Specifically, it resets at the same time of day when you first subbed on the day of renewal (e.g. if you subbed at 10 PM, it will renew at 10 PM). You can check your first invoice from Suno to find out what time that is

[young geode]

What's the most reliable Metatag and Style prompt for Blast Beats?

[polar spade]

Suno tend to "clean" up any audio, since it's generative. For the most part , it is just needing enough to find a melody, voice, and chord progression and work out what the original tempo probably is. I dont think either a high quality high fidelity audio is any better than my singing like a fish with a chainsaw. The results would be the same.

[fair adder]

yeah but i'm talking about an audio already coming from

[polar spade]

If suno want to deteriorate its generation, a clean audio won't prod it to do anything differently. We're just feeding back into the generative process. Even if you fed it 4 minutes of shimmer and hiss, suno will try to make something from it that may sound perfect and clear. So I think its moot.

" . . . .. The results would be the same. . . . ."

Curious, because im ignorant for the most part....How would Suno "recognize" if its a Suno made song? I could be missing how you see how it reads an audio file.

[fair adder]

mmh ok so you can't do a song on suno then dl it then work it on a daw then reup it and expect to continue it smoothly. got it. thanks for your answer

[sleek stump]

Audio fingerprinting.. its embedded in the audio, cant hear it but you can read it in the spectrogram

Well, we cant read it, but software can

[polar spade]

Yes. OK. So, lets say we edit a little in DAW, but we make no effort to remove the watermark. And you're asking If suno recognized the watermark then will Suno do what? Purposely put a degrading quality back in on purpose or because of some connection to the watermark?

Maybe your asking if we take out a watermark and make the song sound decent in DAW, will Suno put a watermark back in?

Im still not exactly putting together the connection. Im not much help than to say i dont think it will. Suno will start over with every audio and re-analyze , calculate and chisel away at a new block of marble to get the output. Any new output will have its own watermark anyway. And if a watermark is whats making a degrading effect then we're back to square one.

[sleek stump]

We have no idea what Suno is doing in the back end, but I have heavy doubts that they are purposefully degrading anything. Suno's models seem to regenerate everything, it's how we get all the non destructive editing.

Maybe your asking if we take out a watermark and make the song sound decent in DAW, will Suno put a watermark back in?

We dont know what Suno is "actually" doing, Udio is the only ai music platform so far to publicly state they are fingerprinting and using a service for it.

[polar spade]

I think that might have been the question. Its asking , even if some outputs randomly and inherently have some degrading areas (and this can be due to several factors) then I want to upload back a crisp and clean song so that any new output has no inherent random degradation.

I suppose the logic is comparing it to if I remove the gene that makes a tomato red, and plant the seed , i should get tomatoes that are not red anymore.

It makes logical sense, but even still, if thats the thought process, uploading a "watermark-less " audio is not a code that will force Suno NOT to put a watermark in the next new output it generates from that uploaded audio.

The simple answer is to try it and see. You never know with AI.

[fair adder]

the question was, like suno is degrading normal clean audios to put them into the "suno" quality, especialy to fit in with the extentions and stuff, would it do the same with an already suno quality song?

i dl a song from suno. i iddn't touch the quality, i just cut and past some extentions in the order i wanted. now if i put it back to suno, would suno degrade the quality to fit in, even if it's already a suno quality song?

sorry if i'm explaining very badly...

[sleek stump]

Like I said, I dont believe Suno is degrading anything... Everything it does is generative, and theres the RNG aspect of it. If you download a suno track, and reupload it and run an extension or remaster it, the perceived quality should be about the same... However, there are a lot of complaints abotu 4.5 remasters being really poor, and I have experienced this myself, 99% of anything I attempt to remaster is horrible. Cover is better but theres always the RNG.

One thing that has been a big plus, the new Stem generator is awesome for fixing poor quality sound in many cases.

[flat crest]

+1 on this. I haven't been able to get a single, decent, usable remaster with v4.5 so far, at all. And that's in quite a few generations.

[polar spade]

From MIKEB: Hey all- microsoft login issue should be resolved for anyone who has felt it over the past few days. give me a shout if you still have trouble specifically with login via microsoft (outlook) 💪

[wintry plover]

Hi! Can you help me? Where I could ask for help for the account of a deeply respected member of the Suno community - Lucilius - his computer and phone are marked as "suspicious". What should be done in such a case? Is there anything that depends on him? How can he find out for what reason this flag was raised?

[lost holly]

Try deleting cookies/clearing the cache, this helped some people. If the issue persists or restricts you, you might have to contact support #support-guide

[polar spade]

@wintry plover A couple of inadvertent and innocent actions triggering
"suspicious activity" might be:

Repetitive logins
Using a VPN , even once
Switching between wifi services and phone cell signals
Logging in multiple browsers with multiple accounts from the same IP address
Clearing your cache and cookies frequently (makes it look like you're logging in from different places)

As @lost holly has mentioned, attempting to clear cookies and cache, once, may help . Here are two posts by @lime light if using Firefox or Chrome which helped him.

#archived-ticket-help message Chrome

#archived-ticket-help message Firefox

[wintry plover]

Thank you, Dasvon!
I forwarded this info to Lucilius!
Why don't I see that channel (need help) in the list on the left?

[polar spade]

Its a main channel under Suno Help. so there's no need to expand anything. Im not sure.

[wintry plover]

I found it with "Browse Channels" and then Search. Thanks a lot!

[polar spade]

Just a quick video showing how to to SAVE frequently used prompts, excludes and even lyrics.

.

[polar spade]

@fair adder

Ive moved your comment to tech-prompt to further discuss outside of "suggestions".

sometimes i wish there was a way to somehow revert back to the legacy replace section layout from before the february update. cause since then, it became totaly unusable to me. i'm blind and using a screen reader and the layout how it was was nice. needed a little help to fidle with the mouse to go on the wave form at the begining, but then i could simply adjust the timestamps and then the corresponding lyrics would automaticaly appear in the lyrics box, and you could actually listen to a preview of your selection in the player. the generated clips were likable and dislikable, and more of that when you disliked them they were removed from the current window, which was in my opinion better for organisation. also, the generations, even if pretty messy sometimes, were still relevent to the song, no quality change, no weird skip etc.

what i mean is, the current replace section, not to mension the complete new edit ui, is completely and totaly unaccessible for blind people with screen readers. no lyrics bxes, no clear timestamps, no keyboard accessibility at all. and the replace section itself is so, so hazardous. whywhy is my lyric box full? why aren't the buttons labellized correctly with clear words so i know what they do? why the setting don't seem to apply at all, and why can't i just dislike/remove the bad generations anymore?

i know i might be a very highly minority around here, but it's still important. all users shoud feel welcomed on this amazing platform, i had so much fun creating my songs over the year i spent on this website. please, suno team, try to also think of accessibility when you program your next steps.

@fair adder OK. Let me see. I've reread the comment again, so see where I have missed anything in my comment that Legacy still exists and it's functional as before:

You liked the Legacy features.

You liked to be able to listen to the segment as preview. Absolutely, me too. I've noted this as well, that its a needed feature. Its still available in Legacy, though, and I need it, too.

You liked being able to manually enter the timemarks. Yes, indeed. I need that feature as well in the new Edit UI. For now though, it still available and functioning in Legacy.

In the new Edit UI , since there is no timestamp, I will say I use keyboard to CTRL-SHIFT and directional keys to highlight my lyrics to be as precise possible, since we can't manually enter timemarks. So there is one time I use keyboards now more in Edit UI 3 than in Legacy.

As for bad sugested generations, as Im better sighted, it may not be as daunting to just look over the greyed out one that I clicked thumbsdown. In the past, I have also re-clicked like to bring them back in to give a couple of them, another chance. So that one is a matter of taste, I guess, and as for you be blind, it may be more beneficial to not have to discern between the bad ones and the good ones. You're right.

[azure crane]

Trying to create Stems but when I go to download, nothing downloads even though I've used 50 credits?

Now it took 100 credits when I tried to regenerate 🤯

[lost holly]

Stems arent automatically downloaded, you have to download them the same you download a song.

[azure crane]

Yes, I did that and nothing downloaded.

It worked before but now, nothing gets downloaded now and I’m loosing credits….

[lost holly]

Try refreshing page -> clearing cache -> relogging, to see if that helps.

[fair adder]

i like the very first layout, the one before the february update. reminder that i'm a blind person here, and for me it was simple and intuitive, except for finding the wave form with the mouse, but it was easy to do after some practice.
the "legacy layout" i can find is the one just before the very last update. the one with the full lyric box. i don't really know how to explain it. maybe it's just that specific format that made it less accessible.
i'm actually glad that the editable timestams are here, i even asked for them to be more precise in the first place. however, again, i can't listen to my preview anymore which makes the whole think a bit more tricky.

anyays... maybe it's just me who just can't handle the ui anymore which gives me bad generations as a result.

thanks for the clarification and sorry if i ever sounded rude or anything, it was not my intention

[sleek stump]

You dont spend credits to download, only generate

[polar spade]

Oh. I did miss the point of your liking the Older Edit UI , BEFORE the February Legacy one. Got it.

[wild raven]

Is there any tricks to singing in a different language? I'm trying to do some ABC 123 songs for my kids in Lao and English. I've tried a few ways of copying text/language but no-go so far.

[sleek stump]

What difficulty are you having, specifically?

[patent sand]

Can you clarify the issue? You should be able to literally just write the song in Lao and have it sing it in Lao, maybe add a [Lao] tag to the lyrics?

[sleek stump]

Im guessing he is trying to use PD lyrics and Suno is saying nah.

[lost holly]

Like this: https://suno.com/s/vl1jUlVsSa9TjOsx or did you have something else in mind?

[wild raven]

Thanks! I tried it. My wife (the Lao speaker) tells me it kind of sounds like Lao. Not sure what's missing. A bit of a challenge when you don't speak it.

[lost holly]

It ai generated and i think one to one the englisch alphabet just in Lao, that might be the issue for you here.

[wild raven]

You can have AI translate it to phonetically how it's supposed to sound. I've tried both and neither sounds right:

Lao script: ອ ແມ່ນ ໝາກແອັບເປິ້ນ
Phonetic: "Aw maen maak-aep-peuun"
Literal translation: "Aw is apple"

[patent sand]

Phonetics are really finicky. You may need to try a different way to sound it out (e.g., "memory" -> "mem'ree" or "mem'ry") or give it a few more gens

[flat crest]

@polar spade (and others?) , since you've been able to help me out in the past...

I'm facing one last problem before everything I wanted to create is done.
I figured I'd be able to do some pitch correction — to "normalize" the voice pitch a little throughout my tracks.
However, it seems like, whenever I try applying a different pitch to my tracks, the voice gets distorted (and not cleanly pitch-changed) and it sounds like the track is vaguely sung throughout a spinning fan. This is common throughout all DAW's. BandLab, Audacity, as well as various other "easy to use" websites for this tool.
The ONLY tool that does not do this, is Amazing Slow Downer (on mobile). But this tool (afaik?) only supports exporting as .m4a, and not as .WAV which would be preferable for posting.

Do you know something I could be doing wrong?

(Also, this "spinning fan" effect is the same for .WAV and .mp3 — unfortunately makes no difference.)

[sleek stump]

I've used Audicity for this before, never had that happen... What plugin are you using on it, Spacer?

Also, have you tried FL?

[flat crest]

FL Studio you mean? — I tried, but to no success... lol... I'm not really experienced enough with FL Studio to be able to do it there.
I'm not using any plugin on Audacity. Just "effects > change pitch"

@sleek stump @polar spade — also, it's present throughout nearly all tracks I'm trying to do this with. Yet, like aforementioned, only with Amazing Slow Downer on mobile, this doesn't happen. (I only have a license for the mobile version, and I can't be precise enough with semitones on desktop either, I need to be able to be precise to two decimals)

[sleek stump]

I just ran a test on one of my tracks in Audacity, I can pitch edit my vocal track with no issues, its just as clean as it was to begin with.

[prime berry]

Is there a reference list of all available prompts from Suno anywhere (link?) Thank you

[sleek stump]

Nothing official, you need to scrape the Workflows section here and social media places talking about them.

[wild raven]

Is there a way to search/filter for audio you've uploaded?

[flat crest]

would you perhaps be willing to try it out for one of my tracks on your end? I know the amount semitones I want to adjust most of them with — so that shouldn’t be an issue. Curious now if it is a problem on my end or just something in the audio file. Or maybe I’m doing something wrong — lol

[sleek stump]

YOu just doing a uniform pitch shift on the whole track?

[flat crest]

correct — no matter which DAW i use, there’s aforementioned fan buzz throughout the vocals. For a few tracks, at least. I could do more testing tomorrow

actually — disregard ^ only Amazing Slow Downer does not have this problem for me

all others do.

[sleek stump]

Go ahead and DM me your track you want me to run it on.

[flat crest]

will do! much appreciated 🙏🏻 — I don’t have it ready at the very moment, (it’s on my PC and it’s gotten quite late in this timezone) but I’ll DM it to you as soon as I get the chance

(might take a bit.. was about to go to sleep… lol)

[sleek stump]

Im in Cali so its still quite early for me 🙂

[flat crest]

Netherlands here lol — 5:04AM… already overstayed my welcome as far as this night goes lol.. even for a saturday night…

[sleek stump]

lol, ya its already tomorrow for you:P

[flat crest]

i’ll DM it to you as soon as I’ll wake up — feel free to test it out when time allows, no rush! already appreciate the help lots

[sleek stump]

No prob. just let me know what shift you want and I'll see what happens

[smoky bolt]

Hello everyone, where can I report credits that have disappeared? I had over 3,000, and after creating one song after another, I ended up with zero.

[polar spade]

Sounds like it happened all the same day? Did you see the credits slowly dissipate? Or did your account lock or change back to a Basic level -- check your subscription page for Active subscription.

First thing though is to refresh your page if using Browser. If changes back to normal, you're clear to go. if not.....continue with the response.

While I'm playing in my Manage account, why does it say I've logged into Suno from a Macintosh Safari 8.0 Browser a couple of time from my home IP address. haha. No Apple products here. (Maybe that FBI van outside has something to do with this)

[young geode]

@polar spade
Would you happen to know how the Speed tab affects the song? I've moving it backwards and forewords and it doesn't seem to have a notable effect

[polar spade]

It doesnt work on the fly. You can set, say, .50 speed and click OK and it will save a new output at half speed. That what youre talking about?

[young geode]

Not within the editor, but that tab.

Do I have to set it, wait, then hit start?

[polar spade]

... they must have changed it the past day... its now a slider... OK a... anyway you set the slider and click Remix button. You cant review it live. So you have to guess what speed you wish

Just friday I was calculating how to get 93.7% and 112% speeds from just the 7 speed levels. lol Before it was from .25 to 2.0 , but now to 4.0 which makes more sense.

.

[polar spade]

☝️

[young geode]

Ah, so its pointless, since I can review it live within the editor

[polar spade]

kinda yea. 🤫

[polar spade]

But its also allows for Weirdness/Style Slider adjustments which arent avail in Edit UI... and I dont think you can keep pitch while changing tempo in Edit. But again, cant listen to any of it on the fly... just have to click Remix for 10 creds to see what we get.

[young geode]

I guess if you find the sound, but the tempo is off it's the only way

but then, you can just play in editor, find the BPM and then do the pertengage conversion

[polar spade]

Exactly, find the tempo you like. Covert to % and .... well, pick a % then calculate in terms of BPM and see if that fits. The speed % arent in increments of 1% , (so more likely they are calculated variations of the previous .25/.50/.75/ 1.0/1.25/1.5/2.0) Assuming we want the pitch to stay the same, go back to Adjust Speed panel, click Keep Pitch, add sliders or not. Click Remix.

If you use it with the new speeds, listen for any significant drop in fidelity. Its only slight when I did a couple of test just now. But the vocals sound good considering the tempo adjustment.

[wild raven]

If you specify a frequency will it be used? Something like " 528 Hz" or whatever. I've tried it, but don't know how to test it if it's there.

[pulsar obsidian]

I am on pro plan I used up all my credit, the 50 being generates is it considered under free plan or pro plan

[untold briar]

You still have all the access of your Pro plan for the 50 credits a day that you get when you run out of the monthly credits. TLD: Pro Plan.

[near mango]

suno的汉字发音一直有问题，这个能不能优化一下啊

[untold briar]

With v4.5 it's not impossible something like this could work, but still pretty low chance I think. Whether it succeeds or not may also be highly dependent on the formatting, for example maybe if you specify the frequency using a musical note instead, etc.

[near mango]

Suno always has issues with Chinese character pronunciation — is there any way to improve this?

[untold briar]

I don't know any specific advice for Chinese. But I believe v3.5 is known to be better than v4 for Chinese, so one thing that can be done is to start a song in v3.5 and then Cover it in a later model version. I am not sure about Chinse v4.5, it is very new.

[near mango]

v4.5也不行，i will try v3.5，thank you

[hazy sparrow]

Part 1/2

Subject: Suggestions to boost authenticity in the next wave of AI vocal engines
From: Babelex Records (technical audit by a music forensics specialist)
To: Suno development team & community

Hello Suno team and fellow creators,

After a detailed, constructive audit of our AI artists (Nox Hollow, R·AI·mundo Carmona, Maïsha Delus, Kemilumo, Evianne BBLX and Dany Babelex), we compiled a set of points that—once adopted—could push every synthetic voice closer to a true human feel:
1. Organic vibrato & sustains
• Vary vibrato rate and depth (±0.3 Hz, ±5 cents) within each long note.
• Inject micro-wobbles and harmonic drift to break the “perfect sine-wave” fingerprint.
2. Unpredictable breathing
• Real-breath library with non-uniform lengths and ≤ 80 % contextual insertion probability.
• Smart breath-noise management: keep it in tender lines, mute it in high-pressure ones.
3. Micro-swing & loose quantisation
• Offset percussion hits and syllables by ±10–15 ms depending on genre (laid-back for rap, forward for funk).
• Vary hi-hat/palm velocities to avoid copy-pasted loops.
4. Formant richness & texture
• Slow modulation (< 3 cents) during sustains for harmonic life.
• Dynamic formant filters that mimic mouth-shape changes.
5. Lexical polish & story flow
• Cliché detector (“neon lights”, “digital sins”, “next-gen…”) with fresh synonym/metaphor suggestions.
• Narrative-coherence module to track thematic continuity across verses and hooks.
6. Adaptive dynamics & mastering
• Distinct mastering targets: −9 LUFS for club tracks, −12 LUFS for ballads, keeping headroom and transients intact.
• Algorithm to tame artificial sibilance while preserving micro-dynamics in soft passages.

Part 2/2
7. Dataset transparency
• Public “model cards” listing voice sources and licensing to pre-empt ethical/legal concerns.
8. Creative-role API
• Ability to summon “Lyricist”, “Humanising Producer” or “Cultural Specialist” modes for real-time refinement.

Music is emotion plus detail; controlled imperfections turn the mechanical into the human. We hope these insights help Suno leap forward and we’re happy to keep stress-testing future builds.

Thanks for listening and for the passion you pour into every update!

Creative regards,
— Babelex Records & forensic review team

[bold flume]

Hello, I'm new to this chat.
For several months, I've been trying to create a song with Suno for a friend.
It's a song with two voices: a man and a woman. Each sings a verse.
When the man sings, the woman responds in the background with phrases in brackets.
The problem is that no matter how well Suno starts out, he always ends up getting lost along the way.
Could someone help me, please?

.-------
Bonjour, je débarque sur ce chat.
Depuis plusieurs mois, j’essaie de créer une chanson avec Suno pour un ami.
C’est une chanson avec deux voix : un homme et une femme. Chacun chante un couplet.
Quand c’est l’homme qui chante, la femme lui répond en arrière-plan avec des phrases entre parenthèses.
Le problème, c’est que même si Suno commence bien, il finit toujours par se perdre en cours de route.
Quelqu’un pourrait m’aider, s’il vous plaît ?

[paper gull]

Try to put [Male] in front of the line
[Bass Male]Sings his voice off.
[Sultry Female](Sing her line too)

You already use
[Verse 3 - Female]
...

[Verse 4 - Male]
...

Just remember that AI will do its own thing regardless of what you try to ask for.
But also write it as part of the style/vibe to try to let AI know that the Background vocals are the opposite gender to the Verse vocal LEAD.

[bold flume]

already tried

[FEMALE (taquine)]
[FEMALE] Tu restes au lit ... [MALE](Je suis malade !)
[FEMALE] À gémir ... (Toute la journée !)
[FEMALE] Tu n’es qu’un homme, mon pauvre zéro ! … [MALE](super HéRO !)

[paper gull]

Try adding [Female Vocal] or [Female LEAD]

or both [Female Vocal LEAD]

Try to also explain what you are after in the Style Box...

[bold flume]

A funke energetic pop song launches with syncopated rhythm guitar, sparkling synth stabs, and punchy bass, Dual vocals bounce playfully off each other, layered over lively percussion and tight brass hits, The chorus erupts with stacked harmonies and grooving dance beats, while the bridge spotlights a jazzy keyboard solo

or

Style : funky comedy duet, male and female voices, slap bass, upbeat, theatrical, musical style, playful, humorous tone, dramatic male, sarcastic female

More option : alternate male and female vocals clearly, emphasize comedic timing, expressive voice acting, chorus in harmony, include bass solos as described

[paper gull]

I recently asked for AI to do Harmonies and move some of them to the left or right Speaker and when I pressed the AUTO button for the style, it changed it to Layered Harmonies across in Stereo.

[bold flume]

ok!

[paper gull]

I mostly use [Male-Female Duet] and settle for what ever it comes up with.

[bold flume]

unfortunately I'm looking to make a VS duo so that there is a discussion, both points of view man woman on the same situation to be funny ^ ^.

[bold flume]

sans cette échange la chanson n'as plus de sens ^^. est ce que les développeur sont aux courant de ce souci ?

[paper gull]

Yes.. It would be nice.. But AI will still do what ever it wants, unless you can somehow convince it to behave..

[bold flume]

ia is a code ^^ if we teach him to respect the beginning of a frase with [Female LEAD] he's supposed to respect it to the end, which he doesn't^^ he takes liberties that we don't ask of him ^^ that's the problem.

in any case thank you for your help 🙂

[paper gull]

I know it is not quite back and front... But you could try [Bass Male Vocals][Soprano Background Vocals][Soprano Harmony] to get a male to sing with female background... Who knows...

[bold flume]

what I'm struggling with is what to master in style and in more option

[paper gull]

Put all your options in and do a copy of them, then press the Auto Style Button to see how it rephrases everything and then re-add those items that it seemed to have left out.

[astral hatch]

Voice generation has very terrible failures, it starts normally, but continues like a person trying to speak with a mouth full of marshmallow or kefir, as a result everything is unintelligible, like a speech impediment and the song is ruined. I have already reviewed everything from the style of the melody to the style of the vocals and its delivery ... no, nothing is written incorrectly, that's how I wrote before ... the persona does not help either, despite the fact that I first upgraded the persona's song to version 4.5. and then created a new persona. I lost many generations and gave up trying to fix anything, it is clear that Suno cannot cope with either Russian or Ukrainian, ruining the songs ... The function of adding words also works extremely poorly in version 4.5, some versions are generated for 1-2 seconds, i.e. empty!

https://suno.com/s/RxuuebFluzOktqPU

https://suno.com/s/5vBE2HVg65TD2WFq

[arctic copper]

The separated inserts contain small specks mixed with mixed signals that come out in many songs and many times no matter how many times the stem comes out.
Does anyone got this problem like me? (I mean multi track stem Also remove the wood because it is embedded in the track.)

[astral hatch]

Song extension does not work in version 4.5! Firstly, the mode is divided separately into the song editor and the remix mode, but neither of them works properly, because it generates short empty tracks, disrupting the song's extension.

So far, there is a noticeable lack of beta testing of the product, and if there are official beta testers, they either did not test all the functions, or the errors found were not fixed, and the version was released on air along with them.

[patent sand]

The beta testing is happening literally right now

[astral hatch]

Well, I wrote about this, at the expense of users and not beta testers. But in fact, many things are not working yet and sometimes a song comes out in one click, and sometimes half an attempt at editing ends in marriage, and the more editing options, the sooner marriage and damage to the material begins.

[patent sand]

Internal beta testers then

The thing is, 4.5 was apparently really good when it released, and has reportedly gotten worse over time

[astral hatch]

https://suno.com/s/8ep1w6vWEuUzGpAp Here's a song where everything started out fine, but then all the lyrics went down to "mouth stuffed with marshmallows" - an inarticulate, speech-like, sad and creepy state... not to say it rudely. I tried many ways to combat the "marshmallows", but as a result they started earlier and earlier... and this is clearly a neural network problem.

[astral hatch]

great, but we want a stable version, not one that can eventually degenerate into a worse tool for work. And there are many things, while there are attempts to develop some editing tools without solving the past accumulated problems. And the stability and sustainability of the results are important. So that I do not take the prompt from version 4ri and get a completely different result by ear, despite the fact that earlier it was possible to throw this prompt into any suno jvizhek and get the same expected sound and music, on the Chinese suno servers for testing.

[patent sand]

F mobile. Sorry

I have no idea about the Chinese servers, but yeah the service shouldn't decline over time. Intermittent build releases would be good to combat this I think

Goes with any coding/modeling, save intermittent builds in case something needs to be rolled-back

[astral hatch]

https://suno.com/s/n0ipX4ekFnicnmgz From experience, if the prompt is good, it will give a result close in sound on different SUNO servers. Yes, SUNO has sold versions that were then trained by the Chinese, mainly versions 3.5. and 4, but not 4.5. Here is a song that more or less pulled out, then I wanted to correct the roughness of the voice and battles began that only intensified from generation to generation without giving improvements.

[astral hatch]

I would add an automatic ability to copy the prompt to the text area of ​​the MP3 file, thus allowing the prompt itself to be archived along with the track, as is currently happening with the cover.

[patent sand]

That would be useful. Currently I just keep everything in *.txt files

[astral hatch]

for now I manually insert the text prompt into the downloaded MP3 file so that they are together, as well as the cover image.

[karmic epoch]

Audio upload seems to be glitchy. Like it stays on uploading infinitely and when I try to cover it, the cover is not at all what the initial song was.

Just stays in uploading clip infinitely

Even tried just saving to library for later but it still stays in an infinite loop on the create page.

Like I can record audio just fine, but uploading an mp3 is not working very well

[karmic epoch]

And I tried reloading the page and trying again but the same issue.

[astral hatch]

after all, you can save a prompt in this area directly on the server, and download a song with this data!

[karmic epoch]

Trying again but still never truly uploads.

[ashen ether]

Hello guys, please tell me how you make such awesome videos, I come to the video exactly on the screensaver how can this be done? If, for example, a song has already been created or maybe you will give me a link where it has already been said or how can I find out how to do it here? How to revive the photo which button can be found? Or want to chat I can use to ask?

[paper gull]

What are Everyone's Vibes, Styles, ALWAYS added to get the results that you are after ???
Are there specifics that you add EVERY time, no matter what you are after.
Not after your actual music preference, but how you setup the Vocals, Instruments, Stereo, left, right, background support etc...

For example..
Male-Female Duet, Bass Male Vocals, Male LEAD, Sultry Female Vocals, Alto Female,
Layered Harmonies in wide Stereo, Spacious Stereo Mix,
Female Background Vocals, Harmonies 3rd and 6th, Harmonies Thirds and Sixths,
Soprano Harmony, Harmonies (Bass, Baritone, Tenor, Alto, Soprano),
Latin Rumba/Foxtrot flare meets Greek String Textures,
Cinematic Slow something (Like pop), A Capella, Deep African Drum Ambiance and Rhythm,
Warm Walking Bass, Trilling Balalaika, Chimes and Bright Trumpets
Plucking Bright Electric Guitar, String Orchestra,

Don't ask me if the above actually work. Have had some success.

I don't personally have any time, but if someone could correlate everyone's successful Vibes and Styles into some tables, maybe with a link to a few of Suno's songs that does just that.
Then maybe later Suno can use that, to create some basic form of selection to select certain Styles/Vibes.

Would love to see the Vibes/Styles Split a little into...
Vocals:
Instruments:
Background Support:
Harmonies:
Stereo Effects:
BPM:
Vibe:

And what would you put into the EXCLUDE box ???
Eg... Jazz, Blues, Sound Effects, Shimmer, Waves, Water, Creek, Rain, Train, Trees, Cars, Laughter, Noise

And if we always add these, then it should probably be added automatically for EVERY song.

Has anyone had to deal with Phonetics ? As in music in other languages, than English. Even English has issues.
As in all the characters that we cannot enter using our keyboard..
Does Anyone have a reference to an online source to the international Phonetics?

[ashen ether]

Please If someone knows how to make this picture move as other users? Who can tell me how this can be done here?

[lost holly]

Upload a video, not picture

[paper gull]

There are Video A.I. Sites that can animate any image..
You just have to look for one that can... There are many..

[ashen ether]

Sorry, can you explain this question to me? If AI created music, created a picture, can you not revive it here? How do people do something like that?

[lost holly]

Like i said before, they have uploaded a video. Like Navi said, you either create a video with an AI of your choice or upload one you filmed yourself/have the rights too.

[ashen ether]

Anyway, I don't understand, please tell me what you mean? That is, they did not take the picture from here and the application itself not animated this video, so you want to say? I can use some third-party it is not clear which one, so it can’t be, as I understand it, I did it here by the Suno application itself. And where to add it? Tell me because or there is some kind of instruction so somewhere you can see it or listen to it

[daring wharf]

I have put together a Suno Tag Reference on Google Docs. It's not exhaustive -- more like a starting point, but highly organized with included notes. The idea is that you can download the file as a RTF or doc and edit it to add whatever else you find or want to use in the appropriate category. The link is on my profile.

[sleek stump]

Suno does NOT create animated cover picture, you have to make it and uplaod it.

[lost holly]

Yes, you are right. You cannot make videos with Suno or create animated pictures, you have to use a third party of your choice for that.
Under song details you can upload the video.

[spark hamlet]

Hiii, used to be 0.1% top Suno creator (or generator, lol)

But I was away for looooong

What's new? Can I for example upload my own song and make it a "saved style" or something?

[sleek stump]

Short answer is yes... longer answer is: Yes, but if it contains vocal, you cant push any tracks made with it public.

[unique stone]

This seems to have changed recently, for anything based on uploads that were made since the editor update

[sleek stump]

orly?

policy change?

[unique stone]

Seems the copyright checks are more thorough when doing uploads now. Still can't make personas from uploaded vocals though.

[sleek stump]

Oh, right. Can still upload though and run extend or cover... just cant push them public

[unique stone]

The uploads, covers, extends etc. even with vocals can be made public.
The only thing you can't do is make a persona from the upload.

[sleek stump]

So that did get changed... interesting

[unique stone]

It doesn't apply to old uploads that are already blocked from being public, or anything made from them, so just has to be a new upload

[warm zinc]

I'm trying out the Cover option and changing some lyrics, mostly minor but same syllables. But the last several outputs the words come out weird. Any ideas?

[sleek stump]

Dont use cover

Use the replace section part and select only a short window around the words you want to change

Then run a cover

[warm zinc]

Oh interesting. Thanks! I had tried the replace section and the audio would come out higher which sucked. That makes better sense though.

[sleek stump]

Cover can be used but its susceptible to the RNG

[sleek stump]

To prevent that choose a very small window around the word change, I use 6 seconds, that way the gen must follow the start and end around it

[warm zinc]

Thanks, I'll try that. I was wondering why one came out perfectly and the other section didn't.

I have this one version of a song that is perfect the way it is, except some lyrics I want to change after listening to it for a couple hours.

[sleek stump]

Its my way of telling Suno to not drift into "creative" 😄

[cunning mulch]

is there any way I can input my own style info when I upload?

[half dragon]

Not currently but you can when you use the uploaded audio to create new songs

[radiant coral]

Hello everyone!

I'm not sure if this is the right place, but I didn't find a better one, so...
I think I discovered SUNO's model soul, or a hallucination? I wrote a request for a simple, indie folk, instrumental song. And SUNO created a song with good vocals and intriguing lyrics. Why? How? How should I live now?

https://suno.com/s/sCXzKfskDLVNb3ZO

My prompt is:
"At 76 BPM in 3/4; This indie folk track begins with finger-picked acoustic guitars, rich bass and brushed drums, creating a cozy groove. Intimate, sensual, yearning cello leads the verses, then delicate harmonies and warm banjo gracefully carry the choruses. The texture remains organic and encouraging throughout to bold dreams and dreamlike reverie. It gets you dancing."

Adv. options: Control Settings

Weirdness: 40%
Style Strength: 60%
Audio Strength: 0%

[lost holly]

I like Suno Originals, what i call those songs 😄 i have a bit over 30 😅 allthough i create most on purpose, since i know how 😉

[radiant breach]

wasted about 500 credits trying to get a clean cover of a 3.5 song. ALL gens were 1 minute short, and the quality is horrible. Suno... what are you doing fam

personas always sound exactly like the old persona, sometimes a song has the exact lyrics as the old song, or simlish cover of the old persona. NEVER the knew song im trying to make

mine is broken i guess, or this company is getting worse and worse

fix covers and personas...... "oh, they want that fixed? here, have a link feature in lyrics, now you can add links to other stuff in your lyrics, isnt that cool?"

[lost holly]

Try Audio influence 40-50% and change styles up, that usually helps to some degree, so that Personas make different songs.

[radiant breach]

I've honeslty tried so many ways to get something to work and it shouldnt be this way. I can't get a consistent persona sound like 3.5, and the quality is still bad. I'm done with suno, lmfao not worth the money

why is everyone jumping through hoops and stuff just to get a decent 4.5 song. When we used to be able to 1 shot 3.5 gens

Suno doesnt deserve anyonees support right now unless youre here for clout

maybe you can be an influencer on the good ol' tik tok instead of making good music.

[lost holly]

I get one shots with v4.5 as well, but i understand your mood.

[radiant breach]

Make Suno Great Again

[radiant coral]

"Suno Originals" ? Hmm 🤔 , Maybe "Suno Soul" or "Suno Core" will be better name? or we should have a contest to name the songs from the depths of SUNO 😂

[halcyon escarp]

I am new here and not sure if this is right place for this, but I'll try. I have been prompting ragtime music, but I can't get right piano sound. I like to get old western movies saloon sound. And I have been search in Google and found that piano is called: A tack piano, also known as a harpsipiano, jangle piano, or junk piano, is an ordinary piano that has been modified to produce a brighter, more percussive, and tinny sound. Also tried honky-tonk piano.
None of those works. Has anyone found this?

[raven turtle]

I’ve had so many issues with v4.5 personas that I switched back to v4 and it works fine. A little shimmer in there but at least I’m getting results that are influenced by the persona and not a carbon copy of what the persona is right down to chord progression, baseline, lyrics and melody!

[patent sand]

Is it literally a carbon copy, or is it at least somewhat a remaster?

In terms of sound quality I guess

[sleek stump]

Have tried "Upright Piano"?

The Soundboard in Uprights have a different tonal quality, usually where that old west sound originated from

Or also try Old Player Piano

[raven turtle]

It’s the same out for me on v4.5 with personas. I get same melody, bass, lyrics. Instead of getting a new track by what you’d consider the same band, I’m getting the same song.

[patent sand]

yikes

[raven turtle]

Hence switching back to 4 and it works (for me) how it’s supposed to. Weird

[hoary helm]

How are you supposed to create or duplicate a signature musical "sound" in your songs with Suno? I finally have a song with great music (mainly guitars). I created a Persona from that song and applied to new songs. It seems the Persona only uses the Style Description text and not the actual audio sound from the Persona. When I add a Persona to a song, it also loads the Style Description for that song. The new songs don't sound like the Persona, and if I remove the Style Description text, the music doesn't sound anything like my Persona song...it's a different genre. What's the point of a Persona if it only uses Style Description text and every new song doesn't sound like the Persona?

[raven turtle]

Lucky you! Try my account where v4.5 tracks using a persona are duplicating the source persona! 😉

[remote kelp]

Idk if this is new, but if Suno figures out that whatever you upload has "Copyrighted Lyrics" even if it's a completely original song, they won't let you upload it and the upload will fail 😢

[polar spade]

Its actually not in the Edit UI but found through the menu. { . . .}/Edit/cover

You can actually see in the vid screenshot above where the Cover tab is top left.

[ember elk]

So, i'm trying to have the last pasrt of my song regenerated, because I want the last part of the song te be modulated (pitched up). However, regenerating never adds modulation (while making new songs it, mostly, understands "modulated" and modulates the song). It's just that the perfect generation has an unmodulated ending 😛

I've added multiple things to the prompt like:

[Chorus]
[// Pitch is shifted up to signal the climax of the song]

or

[Chorus, modulated]

but nothing has helped yet, it always comes with an different but still unmodulated ending,

Anyone an idea?

[polar spade]

prompts like "70s song", "broadway", "glitter metal" , "diva" and a few others may have some influence, but nothing Ive found has works well to force a modulation. I'm looking through the UIs and panels to see if anything now works to do that such as in the Adjust Speed area, but haven't found anything to work on individual sections.

For fun, I just now ran an extension of a deathcore song with those above extension and got a rock opera kind of female vibe of 40 seconds that ended in a 5th chord deathcore-style sustained, but didnt modulate.

[solemn lintel]

For the record, I don't like the new interface. "Remix" really throws me for a loop, as it doesn't seem like it does the same thing as "reuse style" (even if it technically does).

[lost holly]

Maybe try [Final Chorus, rising/increasing/building pitch], other then that i also have no idea. Either rising, increasing... not all together

[polar spade]

Make it an official feedback 🙂 . Post in https://discordapp.com/channels/1069381916492562582/1260263379243761776

[wide karma]

is there any way to create folders on Suno to keep things organized?

[patent sand]

There are Workspaces

Click the back button here and then "+ New Workspace" across from where it was:

[polar spade]

Outputs/songs must exist somewhere and deleting one, delets from everywhere.

So a good strategy is to immediately create a Workspace for working on all aspects of a song. When you get a final or , as I do, pick my best 5 or 10, I move them into a Playlist with the Title Name. So we can remove the Workspace folder to trash (it is not deleting the outputs, but just removing their position in that workspace) if we wish and the files we put in Playlist will remain.

Always save at least one version on a backup or hard drive. You never know when a website, phone or computer will crash.

[neat nexus]

@forest sonnet

As a sample genre prompt try this to start with:

Folk, Traditional, World, Acoustic, Turkish Psych, Space Rock, Psychedelic Funk, 1970s-style psychedelic space-funk, Fuzz guitar, electric saz, funky bass, vintage organ, analog synthesizer, tight drums, Turkish scales, microtonal bends, wailing, hypnotic, driving, propulsive, motorik beat, tape echo, delay, spring reverb, warm analog production

That's from mining the sonoteller information and by way of asking Gemni what the style and sound would be like.

[forest sonnet]

I have tried to use gpt and deep seek too but they are not well.. Maybe it couse of they can't find suno documentation wich is "wich instruments suno support, wich styles etc.".

[forest sonnet]

Thank you for your time!

[forest sonnet]

Is there any source for this?

[polar spade]

@forest sonnet https://audioaidynamics.com/genre-finder aidoaiddynamics is a fun free one i use often. If you want a phrased descriptive you could even then throw those prompts it gives into Suno and click the magic wand.

[forest sonnet]

THANK YOU! I am going to try it now

[polar spade]

Source? search these on any browser.

[forest sonnet]

Sorry, i couldn't explain myself well. For example when i write "bağlama voice" or "kaval" is suno knows it?

[polar spade]

its hard to know. It can play thunder, but if you ask for thunder we likely dont get it.

[forest sonnet]

I have searched it a lot of source but in suno documentation there is not such a place for this information. Also with rhythm is suno have, like 3/1, 5/2...

[forest sonnet]

Yeah

[polar spade]

It may play a claranet, panji or bagpipe chanter... sometimes it just plays what it wants. but we never know, sometimes it gets it right.

[forest sonnet]

I am mostly recording my instrument and sending it in suno. After that i aim that saving same instrument, rhythm and chords. I might have change style, for example i'll make my instrumental records to dubstep

[neat nexus]

I'm probably way off and these are just sample lyrics. But how does this sound to you?

[Removed links]

I'm about to head out the door but usually I would also try to structure with some descriptive [Metatags] in the lyrics field to give it the instruments and build an intro. But that had a nice sound so I thought I'd show it to you.

[forest sonnet]

That sounds awesome but that is more like Glass beams and khruangbin

I love that taste, i made 3 album with that style

But now i wanna make more instrumentlike, more natural and organic

[neat nexus]

Since I don't know who those are I leave it to you to make necessary tweaks and direct it to where you want to go. Good luck!

[forest sonnet]

Thank you very much for your time!

[sour night]

hi guys, any one have issues with generating covers? - I'm really concerned at this point because I see huge degradation in the quality of cover songs.

Different gibberish voices, does not follow song dynamics and lyrics, sporadic changes of sound volume to name a few.

Please let me know what's going on because the quality looks like Murica now. I don't know what's happening. I haven't been doing anything differently from what I've been doing before, so I need to know if it's just the glitch that you're experiencing or it's something else which might result in the long-term significant quality degradation. I just wasted like 12 songs - all with software mistakes. When I used to create cover songs for my guitar songs, I always put weirdness to zero and style and audio to 100, because I want exact cover of my guitar song, just with better voice. - it does not work any longer.

[wide karma]

i added songs to a playlist, it says there are 2 songs in the playlist and also that there aren't any songs in the playlist ? I made other playlists and the songs are visible when i open the list, but not this one

[peak swallow]

Can anyone explain how clicking EXTEND in the edit page, on the songs final section makes the AI completely cut that section from 10 seconds to 1 second??? Perhaps it doesn't understand the word?

[visual zealot]

I need help

[sleek stump]

You didnt give Suno instructions on what to do in the extend?

[visual zealot]

But it’s relatively for professional help I need.

lol

[sleek stump]

Im assuming you are using 4.5... 4.5 does not just make extensions like v3.5 or v4 did, you actually need to give it some instructions to work with, simply setting a point and hitting go results in very short nothing.

[peak swallow]

I'm not sure just how to instruct Suno by clicking the extend button after selecting the section...4.5 is a real mess in many ways, I really think it needed more time in the oven...its amazing, but onlt 1 time in 10

[patent sand]

That's relatively vague

please elaborate

[sleek stump]

You give it instructions in teh lyrics window.. first delete everything in the window, then add your parts you want... [Verse x] [chorus x ] [instrumental] and any additional lyrics. And dont forget the outro and end

[polar spade]

sometimes its a small delay. Refresh?

[patent sand]

For clarification, v4.5 is still "in the oven". You can give relevant feedback here: #📝┃user-research-feedback message or here: #💡┃suggestions-web

[sleek stump]

Honestly, its 8/10 for me... your mileage many vary I suppose.

[peak swallow]

@patent sand maybe need to preheat to 375 or something

[sleek stump]

lol.. but seriously, 4.5 is still a beta model.

[peak swallow]

I don't mind dealing with some issues, but it's really inconsistent for me, I get timeout errors all the time in the edit window.....as I burn through credits at an insane rate since edits cost more than a full song generation for some reason. 🙂

[sleek stump]

Just wait till you get stems... costs 5 full gens lol

Many of us spent 3k+ credits learning how to talk to 4.5...

additionally, you can still use the legacy editor, which is still better IMHO

[peak swallow]

if they didn't stick it to you for additional credits when I run out it I wouldn't mind as much, but 4000 credits vs 10k for the same costs is hard to accept.

my biggest issue is that 4.5 seems to degrade in audio quality as the songs progress...sounds like its just piling too much together and becomes a muddled mess. Any good ways to reset the thing mid song? Sometimes BRIDGE works, but other times it doesn't.

[sleek stump]

Thats a common complaint with metal or other hard/fast types.

I prefer v4 for initial metal gens

[patent sand]

I have this issue occur even back in 3.5

[peak swallow]

I tend to do either simple one instrument ballads or hard rock that tickles the metal line

[wide karma]

when I moved the files to trash to get them out of the workspace it also removed them from the playlist

[sleek stump]

Anything that gets close to metal is not good in 4.5, this has been widely feedbacked to Suno

[sleek stump]

This is normal behavior, putting things in trash means you dont want them.

[patent sand]

I think he meant deleting the Workspace itself, not the songs in it

Though I've not tried it myself

[wide karma]

copy that thanks!!

[sleek stump]

So, trying to hide songs by putting a workspace in the trash? Sounds risky to me, and not intended functionality

And agian, if that is the case I would assume everythign in the workspace would also end up in trash...

[patent sand]

Context: #📚┃suno-school message

[sleek stump]

Ya I dont see how that can work. Just put your tracks in workspaces as folders. Unfortunately none of these things works in the main library, everything that is not flagged as trash or disliked is all stacked newest to oldest.

Very inconvenient when I am trying to listen to my tracks in my car on the mobile app.

Suno does need a proper file system for the main library

[patent sand]

Does your car have a USB port? That's what I use for loading music into it

Can tag and organize easily in the File Explorer

[sleek stump]

Ya I know how that works, but when I am not done with a track and have yet to send it to post, I like to hear it in my car. playing them through teh app is the most convenient.

[polar spade]

"..........Outputs/songs must exist somewhere and deleting one, deletes from everywhere......"

".......... So we can remove the **Workspace folder ** to trash. It is not deleting the outputs. . . .. "

---

It happens.    Go to the trash bin and find the songs deleted and recover them?  Or did you remove permanently?

When we remove the Workspace folder, the files just go back to the library. Only delete individual outputs if you never wish to see them again.

[polar spade]

Who's saying what.? I don't follow where that came up , hiding files? Workspaces in trash bin? Did a wormhole open up somewhere? 😂

[sleek stump]

Hey Im just as confused... theres not a lot of context other thna link back to your message earlier in the channel.

[polar spade]

The OP asked if we had a way to file songs.

I said one way is to fiest work on your song in an initial workspace. Open a workspace , dump all your work in there .

When you have a final or a few you like, I personally then put them in a new Palylist titled the same as the song.

( or course we can have the same file in multiple playlists as well for an album, genre, etc.

Then I said, if we wish we can remove our Workspace. Then the outputs within that workspace will just reside back in the Library as always. [ and in the main workspace too, but didn't want to confuse OP more]

[sleek stump]

Humans will always try to make things do what they werent designed to do 😛

[sleek stump]

Ya I got that, but I think there was some confusion on their part?

[polar spade]

I said REmove the workspace...... not delete the files as earlier I said deleting any output will delete it from everywhere. OP deleted the outputs in workspace instead of removing the workspace only.

So, ultimately the outputs OP put in a Playlist all got deleted as well.

[sleek stump]

🤷‍♀️

[polar spade]

Thats where you came in. 😀

[lost holly]

Nothing gets deleted finally as long you dont empty your suno trash, there you can always restore song if necessary.

[sleek stump]

Now what do I do with a 12:06 track I just constructed.... SUno is 8 minutes lol

[polar spade]

Everyone raise their hands who have inadvertently deleted a file to the trash bin, then had to go find it...

Me.!! ✋️

[sleek stump]

Or the every not so obvious misclick on dislike

[quick wedge]

You can use the extend feature to make it longer if you want to 🤷‍♀️

[sleek stump]

No no, its 12:06 in my DAW lol

[polar spade]

Uh oh. Radio cut? I mean Brian Wilson and the Beach Boys got away with a 7:40 Good Vibrations. You'll just have a 12 minute EP deep cut and a 7 minute Radio edit.

[quick wedge]

Oops I thought you meant in Suno

[sleek stump]

Its a sorrowful Chinese orchestral, was super depressed last night after going through my wifes dresser, and looking at her hair pin and comb collection...

So I deflected and made music

Probably no need to get it back into Suno, not a lot of interest in that kinda of music anyway.

[polar spade]

You may be surprised. But. Dang it anyway. Time for Suno to lengthen upload limits to 12 minutes already then.

[sleek stump]

lol... That also means they need to be able to do 12 minutes gens 😛

[lost holly]

Shouldnt they be able to do 8 min first? 😄

[polar spade]

Oooo..... I know a few who would like that.

[sleek stump]

They already do have an 8 minute for 4.5

[polar spade]

8 minutes means about 6. So 12 minutes would be closer to 8, tops.

[sleek stump]

Iv'e gotten 7:59 and an 8:01 gen

[polar spade]

Show off! Plltttttt..... (he said in envy)

[sleek stump]

Dont be, they were not good gens lol

[lost holly]

I reported dozens of attempts that where cut of around 5-6 min...

[polar spade]

Its cool, anyway.

[sleek stump]

I can easily get 6 minute really good gens

just gotta give Suno something to digest 😛

[visual zealot]

Why does the showcase page keep delaying it never used to

[sleek stump]

you mean the home page?

Its frequently broken

[polar spade]

I think the 2 aand 4 minute times Suno was pushed and had problems analyzing and arranging outputs to neatly fit. I feel they rethunked it and went to 8 and allowed enough room for buffering to have a song that didn't cap or clip off in the end with room for seeding the covers/personas.

[sleek stump]

I'd love to see the dev block diagram for Suno

[lost holly]

That would be interesting

[sleek stump]

stupid Woody

Dang, the MP3 for this track is 29MB...

[polar spade]

Time to tap him with the prune frying pan again.

[sleek stump]

I fat fingered a pound symbol... who would have thought Woodhouse would call that a huge font

[polar spade]

He doesn't like three in a row, at least .

[sleek stump]

Huge font#

[sleek stump]

¯_(ツ)_/¯

[polar spade]

Hmmm maybe it was four.... or 12

[sleek stump]

Suno API

[flint haven]

Beware: there is NO official Suno API

🚨 Suno does not have a public API. Any claims otherwise are unauthorized.

[sleek stump]

Well, he's not broke

lol

[polar spade]

Yep.

[ocean cape]

Test this prompt structure that I'm creating, if you use chatgpt to bring the style and artists that interest you, and please give me the feedback as well as the link to the song you liked the most. Using this structure for prompt ##########[
Estrutura movimento de Meu Vizinho Plantou (Bezerra da Silva), preservando sentimento malandro, irônico, relaxado e swingado, ritmo contínuo com balanço brasileiro.

intro Introdução com groove de cavaquinho e pandeiro leve
Estilo: Samba malandro carioca, influência direta de Bezerra da Silva, swing de partido alto com levadas humoradas
Tempo: 96 BPM
Dinâmica: suave a médio forte, com marcação rítmica constante

arpeggio C major
Mão direita: harmonizações típicas de samba C - E - G - A7 - Dm - G7
Mão esquerda: linha de baixo marcada C - G - A - F com variações cromáticas

theme: Frase melódica vocal falada com rima sarcástica e cadência malandra
repeat: Repetição com variações de entonação e entradas de surdo e reco-reco
bridge: Modulação harmônica suave para F major
closing: Encerramento com fade-out no cavaquinho e batida de pandeiro diminuindo
dynamics: alternância suave entre momentos mais falados e trechos com swing instrumental
]##########.

This should be used in the lyrics field, only generate what is in the prompt, do not put lyrics yet, you will do this in remixes or when using lyrics and style

[sleek stump]

Suno ignores Tempo BPM commands

[polar spade]

Hey. He used hashtags and Woodhouse was fine with it. 🤨 well sort of..... he didn't take i down , just pruned it a little. Hey, see what I did there, a word play.

[sleek stump]

Ya but he wasnt with the next page of text 😛

[patent sand]

Woodhouse is still translating the message

[sleek stump]

Large instruction sets fit better in workflows, where the threads are formatted for it

Ya thats not gonna get past Woody 😛

[polar spade]

Sheesh... that was a whole screen page full.

[patent sand]

They are saying to take this and post it here #1252267444773326848

that's all

[sleek stump]

[echo prawn]

Why can’t the new version upload songs locally?

[outer flare]

i need to have better understanding of filtering system even i case of some public domain material it got flagged as having some copyright or suno tag filterarion refusal

but there are some materials which nobody can claim exclusive ownership about

[arctic copper]

The separated inserts contain small specks mixed with mixed signals that come out in many songs and many times no matter how many times the stem comes out.
Does anyone got this problem like me? (I mean multi track stem Also cannot remove because it is embedded in the track.). ….. 😦

[polar spade]

Yes. You're correct to assume then , those you can't upload either. Nursery rhymes, old papaya scrolls from 435 AD., the ABC song, the Gettysburg Address, among others, Creative Commons, royalty free songs, a Nintendo tune, Star Spangled Banner, and frankly anything trademarked as well.

[outer flare]

though i got sucess in creating inspired song from star spangled banner

[patent sand]

What do you mean by locally? Are you referring to the lack of audio upload in the mobile app?

[worn thistle]

hello, little question, is the "new editor" updated/repaired or it's still broken

nope, still broken as ehm ehm

[polar spade]

If one part is bugging it doesnt mean the whole interface is broken.

There has been slight updates to the Remix panel that I know of.

As for the Edit UI, I find one thing odd it does is when saving a new song as a Full Song, it only inserts the added lyrics into the lyric box instead of the whole song lyrics.

I've used Replace: worked(* ! #)
Used duplicate: worked
Copy/crop: worked
Delete: worked
Extend: worked
Move section: works
Extend, then move section to beginning and create an intro: works
Change tempo and save new speed: worked

Stems seems OK

What i cant work with new Edit UI, I combine with Legacy UI and get it done that way.

---

• At times it works and blends nicely. Sometimes I'll need to splice out small bits, then Remaster to blend it.

---

! Usually run about 6 sample segments before choosing one of same length or even use a longer one suggested.

If 60% of the time I get what I expected, that's "worked" for me. Less than 10% = broken.

---

What one thing are you having an issue with?

[worn thistle]

crack voice like when the singer have puberty, absolutely retarded song structure (tempo is changing, cracking, different melody than "the original" and more) attempt #112 and nothing

all what i just want is change 20s of the song, and it's imposible!!!

[polar spade]

You mean in all the Replace suggested segments, that all the samples are cracked voices or changed melodies, different tempo, etc. ?

[worn thistle]

yep

and when i wanted refund credits on support, they ignore me

[polar spade]

Usually 30% minimum of the replacements are good to use for me. So thats odd all 100% for every song edited.. Yea, sounds broken , youre right.

[worn thistle]

and before i payd for more credits to replace [chorus] attemp #336 so yeh, money eater

so tell me, What should I do? when support is ignores me

[polar spade]

I guess id be happier when its updated to give me 100% of what I want. But for now, what I get is saving more money than my alternatives. So I guess im biased.

[worn thistle]

yeah, you borrow me your account and i will done it there, what are you sayng? 🤣 bcs mine is rigged

so what do you think?? 🤣 🤣

[polar spade]

Haha... I think I'll keep what I have. Thanks.

[worn thistle]

🤣

[worldly dove]

wanted to report broken captcha in suno feedback and the suno feedback has too many mandatory fields.
^feedback

[polar spade]

You may wish to post this on which Feedback channel here in Discord you feel is best.

This tech-prompts channel is not monitored for feedback, bugs and suggestions. Maybe https://discord.com/channels/1069381916492562582/1082775596767117312 ?

[hard beacon]

so let's talk about prompts.. do you have an personal experience about negative prompts, that reduces shimmer, reverb or muffled sound?? People are talkin about this all over the server.. but for me "mastering prompts" did not work in the last weeks

and yes: shimmer in vocals is back again 😉

[hard beacon]

i don't wanna talk about the well-known problems of suno.. just wanna know if there are good experiences with negative prompts 😉

[devout spindle]

Bonjour, avez vous des problèmes de création de chansons avc le nouvel éditeur et surtout l'utilisation des 2 curseurs ??

Depuis la mise en place du nouvel éditeur et des options avancées (les 2 curseurs), je n'arrive plsu du tout à créer des chansons... les paroles ne sont pas bonnes, erreurs de prononciation,..... et l'éditeur ne fonctionne pas !! Merci pour vos retours

[patent sand]

The default slider positions are 50%. There is also a way to get to the "Legacy Editor", see #📚┃suno-school message or see if someone has a more reliable way

[sleek stump]

Upper left side of the editor you can select the Legacy Editor.

[neat lake]

Hey, anyone can suggest why the vocals became muddy all of a sudden? My adlips sound really unnatural and it's a shame, because some of the instrumentals are just pure gold. This has occured in previous models as well, where the quality of generated vocals has degraded over time, or at least that's how I interpret it. Not trying to stirr any shitstorm here, just wondering what this could mean and where it could be coming from?

[patent sand]

This is in v4.5? Which genre? Are you using a persona?
This Discord is surprisingly chill, so the risk of stirring anything is quite low

[neat lake]

Hi yes, using V4.5, no persona, no uploaded audio. The genre is Oldschool Hip-Hop, though I tried with others with a similar result. Hip-Hop seems more affected though.

[sleek stump]

Dont try any hard rock or metal 😄

[glacial rampart]

Does anyone else have problems uploading MP3 files? When I try, I can trim the track and select cover, use for later and so on but the track does not appear in my inventory.

[worn thistle]

try to look in "filters" or different "workspace". that's two options that i know

[glacial rampart]

Nope, does not appear. Even searched for the track. Also, before I select what I want to do with it, I wait for the upload to be completed.

[worn thistle]

refresh and go to library

[wide karma]

any way to recover a track that has been moved to trash?

[polar spade]

Right click any output in your trash bin and choose recover. Trash bin is found through the trash icon top right of your Library page.

[wide karma]

I have a vocal track that I am trying to do a cover of, with a persona that is based solely on another vocal track, but every time the cover gets rendered with BS music - no matter what i put in the prompt or how i move the weirdness, style or audio sliders, it puts MUSIC in what is supposed to be a pure vocal track

[lost holly]

Pure vocal tracks are hard to do with Suno. You should try extracting the vocal stems of a track you like.
If you want to do pure vocal tracks, you have to share the styles you use, maybe then someone can help here.

[sleek stump]

You can try A Capella... but expect to get PTX style A Capella.

[wide karma]

hahaa yea tried that already

[sleek stump]

You can also try [Chant], the opening track of an album I am producing is entirely a spoken word/chant track.

[wide karma]

i must have hit the AI 50, 60 times last night just trying to do a cover of one of my own tracks with one of my personas, and no bueno so i thought i would try just doing the vox as that's the only thing i want to change about the track

[sleek stump]

I used this for every section [Spoken: Gruff Deep Male]

I also described the whole track as a chant in the styles box

[wide karma]

are you doing strictly spoken word or songs/ music with full instrumentation?

[sleek stump]

My opeining track is only spoken, no instrumentation what so ever

[lost holly]

spoken word is a genre Suno knows, you just have to make sure not to mention something in styles that triggers instruments.

[sleek stump]

\Here is the sytle section I used for it...
A Capella, deep-voiced male starts the spoken prayer alone, intoning each line slowly and gravelly with ancient resonance, After every line, a large group immediately repeats the phrase together, their collective voices loud, layered, and reverberant, encircling the listener with raw, ritualistic power, music, instruments

Not sure how successful you are going to be with any spoken track when using a persona..

[wide karma]

i just did A capella with spoken word and it worked pretty good actually - i think i can work with that - thank you!!!!

[ocean cape]

Why was I silenced last night?

And what is the subject/focus of this chat?

@polar spade Why was I silenced last night? I really don't understand? Could you explain to me what I did wrong so I don't do it again?

[sleek stump]

Might have been because of the pages of text you were trying to post.. That kind of stuff should be in Workflows, the channel is formatted differently.

The Woodhouse bot was auto deleting them.

[polar spade]

Just for name sake: the classical [and Italian] spelling is "a cappella" , with 2 p's, and best to never capitalize anything as Suno may evaluate whether its a label or proper name. That could make a difference in our results.

And yes, using [spoken] with an already sung piece can help nudge it away from accompanying music.

Frankly I just run a cover with acoustic piano or guitar to make sure a melody is rendered , then stem out the vocal. But depends on what you need it for.

[unique stone]

Woodhouse (the auto moderator bot) deleted several messages that were long and had excessive spacing between some of the lines. As an anti-spam measure, certain things like this cause an automatic timeout if It happens several times. It looked like you were trying to paste something. Here is the start of one of the messages for example.

[austere dragon]

Is there something going on with lyrics? I'm getting nothing but hallucinatory gobbledy-gook.

[lethal grove]

Trying to extend and it's all messed up ...only a couple lines of lyrics are listed and the song segment bar is all one segment. That's in the Legacy Editor.

[polar spade]

The h'g-g happens and can be fun too, but more than about 15% is tiresome, for sure. Since vocals are also an instrument, hallucinations can show up everywhere. I've noticed whe. There is space somewhere , like replacing a whole verse with just one line, Suno may try to fill things in with gibberish and babble.

So, avoid too many line spaces than needed, unless you prompt for short outputs , have enough lyrics or tags to keep the songs anchored toward a visible ample arrangement of verses and choruses. Keep any references to male or female vocals as close to the stanza they are referenced for.

[polar spade]

Are all the segments in the original marked with the appropriate segment tags [verse], [bridge], [chorus] , [intro], [outro] , etc.

Without those, for some reason, Suno is now also removing the line spacing when we open up Edit UI. So its all one big lyrical prose.

[lethal grove]

Yes

[oak siren]

i'm sure this has been asked, but i want to adjust one section of a song's bpm and leave the rest as is. is this possible?

[polar spade]

Ill be first to say there may be something going on with the Edit UI as to how it sees the segments.

Hey. Go into Legacy and see if the segments show up there. Sometimes I've noticed the two may be differently formatted in their segments and colours.

[polar spade]

Oooo... that'd be nice. So far, No. Not that I've been able to do. Once you raise the BPM up or down, it affects the whole output.

[oak siren]

gotcha. so i'd have to mess around with it on ableton. where can we recommend this feature?

for future updates

oh suggestions, nevermind lol

[polar spade]

And in case youre thinking about it, making a persona of a song set to a slower bpm then adding it as an extension to the original just still seeds off the original and interprets that the extension needs to be the same bpm (instead of the tempo of the persona) . I've tried and no luck yet..... and vice versa.

[lethal grove]

The beta Editor shows the various song segments, but the Legacy shows one. Tried to extend in the beta and had an "Error Occurred" pop-up.

[sleek stump]

Yep, he tried 3 times after that first page, pretty sure Woodhouse auto silenced him.

[unique stone]

It was. I see it in the logs.

[polar spade]

I mean, when I've noticed the lyric box has NO segments, I manually, in Legacy, have put in all the segment tags such as [Verse] and [chorus] then choose "align lyrics" only found in Legacy (that I can tell) and make sure they are all indicated in colour's before then clicking generate alignment , (or whatever the button says) so that the segments are designated in the wave form as well. Then go back to New Edit UI is needed. I think that did it. Its been a week since I did one, it guess.

[oak siren]

definitely wasn't thinking about, but good to know! luckily my problem is limited to ~6 seconds of an entire song. a little elbow grease should do the trick

[polar spade]

I was trying to find a hack to making a modulation. Didn't work. But I discovered the answer to your question. 😆

[lethal grove]

The lyric box contains one segment with 1 line of a run-on lyric.

[polar spade]

Well, that might do it then. 🙂

I hate giving a cordial response immediately after someone posts a concern and they just sniper post and run with no acknowledgment. So I've deleted my cordial responses as well.

Im out. 🤷

[polar spade]

This may not be associated with the discussion, but if we do a Replace in the Edit UI, and then save the file as a New Song , THE FULL SONG is made as the output, but i noticed only the replaced bit, along with the 8 or so beats right and left of that segment are shown in the outputs lyrics.

I guess I should look carefully to see if the Replace lyrics pop into and overlay the full lyrics in the New Edit UI before saving.

Can you tell if after doing an extension , or replace that the lyrics box in the Edit UI change drastically like this?

[lethal grove]

I noticed a little change. sometimes a line is doubled. Sometimes spacing between words missing. stuff like that. But never had the complete lyrics almost missing.

[polar spade]

Everything works enough to get work done, but I'll say I have to jiggle the handle sometimes , like retyping in the arrangement segment makers to get the colored segment to show, or realign a couple times more than I want to, and other small things to get it back to par so I can finish something.

[polar spade]

Yea the doubled lines or double and triple verses are definitely a buggy thing. Its taking me longer now just to review the lyric box.

[lethal grove]

I had been having probs with the editors for a week, particularly with Extend, using all 3 methods. Wasted a lot of credits.

[polar spade]

Yea. Not quite there at that sweet spot of not feeling credit use are wasted instead of just part of the process.

When they happen sometimes but not all the time, then im just looking to get those frequencies down to a tolerable amount. So, right now, its not broke, but still badly bent.

[lethal grove]

I just hit the "Reset All" in the beta ...it turned my track into one long Intro 🫤

[polar spade]

Even when resetting, the sample replace or extension segments still show up in the box as I recall. Kind of a good thing ,I guess, but I dont know, if then, thats a glitch or it was designed NOT to reset the sample segments box. 🤷

Still just a variety of weirdness with the Edit UI. Thats what feedback is for, I suppose.

[viral kite]

Has anyone ever gotten the same voice twice when using a persona?"

[pliant sun]

it won't load

[pliant sun]

is there a way to fix this issue?

[sleek stump]

iterwebs lag? Wait a bit?

[pliant sun]

👍

[polar spade]

Isn't that supposed to happen?

[sleek stump]

a tarcert might show a laggy or broken hop...

Suno's web services, while having their issues, are still slave to all the other internet woes that can and do happen.

[sleek stump]

RNJESUS wants to have a word

[ocean cape]

[verse] letra

[chorus] letra

[bridge] letra

[whispering] letra

[spoken] letra

[solo] instrução do solo

[closing] instrução de fechamento dynamics instrução de dinâmica final

[verse] letra

[chorus] letra

[bridge] letra

[whispering] letra

[spoken] letra

[solo] instrução do solo

[closing] instrução de fechamento dynamics instrução de dinâmica final

[polar spade]

Uh oh.

[sleek stump]

here we go

[ocean cape]

What do you think of this structure?

[sleek stump]

Depends on the genre you are trying to achieve, whispers and spoken are hard to get with most genre

All we can do is say... try it and see what you get

YOu also might want to use Outro instead of closing

[ocean cape]

https://suno.com/s/wayeH5i2bnyH5dBl

[pliant sun]

it's still loading I've waited 11 minutes

[half dragon]

What happens if you refresh? would it sitll show loading..?

[pliant sun]

it's fine I'll just record my new song "Failure Of A Son" and publish it on SoundCloud

[ocean cape]

I know, what I'm doing is creating the custom gpt to give me this type of structure, the whisper I got in some cases, and also whistles/whistles, what I want to know is, create A base and structure rules to serve as a regatta in this GPT model for prompts, what I have been doing And to get video content on YouTube, I use notebook to extract the text quickly, and then I use it in a model that I have customized that generates a kind of monologue or poetic texts and And reflective, in this with this text/monologue I will use the one I am configuring to generate the structures and choose which rhythm or style is most interesting for the music made from the text extracted from the Video

[pliant sun]

I tried that it's a no go👍

[vernal harness]

How can I get water sounds?

[polar spade]

Sound effects that work every time are still a bit of a mystery , so far. Water, river, running water, bubbling water, boiling cauldron..... they arent easy to get to work ,or at all, nor place in the right spot.

Some..... like [beep] though, work pretty well.

[vernal harness]

Yeah that's what I am finding too!

[polar spade]

I've found that if I hint something like "chicken" all over the lyrics, then eventually within 30 outputs or so, it will throw in a faint chicken cluck, or make the violin sound a bit clucky like a chicken. So there's recognition and suno tries to include the vibe of it without more of the actual clear sound effect.

[vernal harness]

Another issue- after I remixed one of my own songs, I can no longer use the lyrics for another song because it says its copyrighted material..

[polar spade]

So, by clicking "styles and lyrics" , it doesnt let you go further to Reuse or Cover?

[vernal harness]

I can reuse, but I cant use the lyrics when I start from scratch- I edited it a bit and it worked

[polar spade]

Thats the next hint. Move the layout of lyrics a bit. That can be enough to make it work.

[patent sand]

Remixed how? This sounds like a highly critical problem

[vernal harness]

this is how I did it- I may be missing something but... working around it for now

[patent sand]

So literally just reusing the style/lyrics got flagged? Curious

The system really should not be flagging your own lyrics

[vernal harness]

I agree 🙂

[sleek stump]

The moderation system uses ChatGPT, it is not uncommon for it to make erronious judgements on things to moderate.

[young geode]

It might be, with a lot of elbow grease and RNG.

You'd likely have place an extension between the beginning of the section you want to slow down, and the end of that section and the rest of the song, make it a short break/interlude/instrumental, and use the Italian tempos, both as a [Tempo: Metatag] and within the style, eventually find the BPM you're looking for, then generate the slower section based off those two small additional sections sandwich.

[graceful condor]

Is there a straightforward way to prompt a duet and ensure which line goes to which voice?

[oak siren]

hmm that's a lot of elbow grease. i think it might be easier to do on ableton just because it's a ~5 second segment that needs adjusting, but thanks for the input!

[young geode]

Yes, but it's difficult.

Youd have to construct a prompt for the duet, for whatever voices you wanted. Then you'd have to format the lyrics with metatags in a particular way:

[Section: Duet]
[Singer 1] Line said by Singer 2
[Singer 2] Line said by Singer 1
[Singer 1] Line said by Singer 2
[Singer 2] Line said by Singer 1
ect
ect
ect

And, you'd want each line, optimally, to have a different syllable stress pattern, rhyming schema, ect, to help fully differentiate the lines so that GPT-SUNO recognizes them. But be aware, if you do this and you have these same patterns show up elsewere within the song, you will likely create duet hallucinations as well: so you will likely have to edit.

[graceful condor]

Thank you, you've been immensely helpful, and I'll attempt this in the way you've suggested 🙂

[young geode]

You're welcome.

Be aware that this isn't a guarantee, but my own testing has shown that for 4.5 at least, I've been able to reliably affect which lines I want sung in specific ways (such as a light flange, choir, ect). So don't waste too many credits if it's not working for your specific song.

[astral hatch]

The page does not load on the "Create" link, refreshing the page does not give anything - white background.
The page does not work on the server side.

[radiant lake]

Hello! I have a connection problem, the page is not loading in the browser. It opens for a few seconds, then an error appears: 'Application error: a client-side exception has occurred (see the browser console for more information).'

What can i do ?

[lost holly]

This sounds like an issue on your side.

1. Try clearing cache/deleting cookies -> relogging.
2. Make sure nothing blocks anything in your browser, like an add blocker.
3. Try another browser to see if this helps/works.

[viral kite]

" Never get the same voice.

[lost holly]

Hi, please screenshot the error and provide the anonymized error (so it doesn't contain any information related to you) in #🪲┃bugs-web and to [e-mail] with a description in what browsers this happens and your @handle from Suno, if you don't mind. Or submit it as feedback -> https://suno.com/?open_feedback=true
Then do the 3 steps i just suggested to One Person 2 posts above, to see if this helps..

[radiant lake]

Hello! I have a connection problem, the page is not loading in the browser. It opens for a few seconds, then an error appears: 'Application error: a client-side exception has occurred (see the browser console for more information).' Can someone help me ?

I clicked on the link you sent, the page opened, but when I opened it again from the Google Chrome browser, the error appeared again. I also tried to enter the section to create a song and that error appeared.

[lost holly]

Have you tried the other things?

[radiant lake]

Now I'm trying on Microsoft Edge

The same error occurs in Edge as well.

[lost holly]

When you tried clearing cache, other browsers, then only contacting support is left.

[radiant lake]

Just a few days ago, on June 14, I paid the subscription.

[lunar remnant]

any chance theres a way to put more than 5000 characters in the lyrics?

[lost holly]

Not really, technically you can put more in in certain ways, but the song generation wont work then.

[lunar remnant]

damn

[polar spade]

Have you restarted your computer in awhile? Or even open task manager and "end task" all bowser processors you can find open (they won't all close just by closing the browser).

[lofty bough]

Hi guys, i have a question. When I want to create a song, 2 versions are created directly. I don't want that because it uses up credits unnecessarily. Is it possible to turn this off?

[polar spade]

Certain personas may react poorly to some prompts. But I've not added a persona, clicked 10 times and get 20 separate and different voices. Sorry, I dont have a clue, other than your testing out on different prompt and combinations to see if things change.

[polar spade]

You get a first song for 10 credits. The second song is free. You can choose to delete the second if you wish. 🙂 *

• No. Thats the system and pretty much universal way of ALL Ai Gernertive song programs to issue two outputs or choices eith one click. Its generally good marketing. If only one, for 10 credits, many would click again anyway so they could compare.

[lofty bough]

Ah ok, but why it says "2,500 credits (up to 500 songs), refreshes monthly" and 10 Credits per song are only 250 Songs 🤔

[lost holly]

5 credits per song, not 10.

[lofty bough]

I get 2 versions of the same song and it uses 10 credits

[polar spade]

Its not 500 generations (clicks), but 500 outputs. You get 250 songs for 10 credits a song. And the other 250 for free. 🙂

[polar spade]

Its not 5 credits per created song. Prove it. 😅😜

It's only 5 credits back per, if either the first or second song didn't work.

[lost holly]

Wise guy 😄

[lofty bough]

so theoretically 500 but realistically only 250 wow... that's a bit of false marketing but I don't want to complain, thanks anyway

[polar spade]

Its what we agree on in the TOS when joining.

Each songs is generated differently, even if just by one note.

[lofty bough]

Who reads TOS? 😂 but yeah you right

[polar spade]

We can buy bananas by the pound, or charge more for each banana. Your choice. But I guarantee you, no two
bananas are the same.

[polar spade]

Yes Im pulling your chain as many other users have reasoned we should keep the rate structure of averaging 5 credits per, but then somehow only charge 5 credits and get one song. If we changed to 1 song per click, the price would be at least 7 credits per song. But likely we would be charged 10 credits per, be given 4000 credits and charged $12 a month instead. Who knows. Doubtful they change one without changing the whole structure.

That just doesnt seem to be the marketing scheme of EVERY generative AI application on the internet. Runway, Midjourney, Kling, Veo, all give comparative 2 to 4 generations per click. UDIO, Suno, Riffusion, all same.

Again, its human nature to prefer a choice (of at least 2) than to just get one thing and not know if its the best it can do.

[polar spade]

Honestly, I've never gotten a good final song in one click. So with 2500 credits, that may be one or two final songs when done. But im picky. So it gets worse before it gets better. Lol

[hardy thunder]

I changed the BPM of a version 4.5 song via the new editor on PC and as a result the edited song is not recognized as a version 4.5 song, it actually does not have any visual version number as seen in the picture. I tested this twice, same results on both times. At first I did not realize the missing version number, but when I made a cover of the edited song, I noticed that the cover was version 3.5. Sure, I can select 4.5 when doing the cover, I just didn't pay attention that 3.5 was selected as a default. Is there some kind of explanation for s song being without a version number?

[polar spade]

My two cents:

I've seen that when we save a song out from the Edit UI. Its not always showing a version number. I mentioned in feedback. You can do the same if you wish.

In the passed especially during updates, the versions would even flicker from 3.5 to 4.0. Maybe something similar going on.

[small sun]

just for reference. sometimes when "instrumental" is checked, it still gens lyrics

like not just simlish fills, like full lyrics

i like the simlish fills in instrumentals. heh. i like the abstract vocals in instrumentals.

[brazen yew]

https://suno.com/s/F0ib6MJRTzoVpanS

How do I fix the huge jump in volume between 1:55 and 2:05? I've tried normalisers but they aren't working

[polar spade]

We can now go into the new Edit UI, highlight the section and change its volume. Then save as new file. Im not at desktop, but I think volume is in the menu choices just below the segment chosen?

Maybe someone can screenshot where.

[brazen yew]

Ok I'll load up the desktop site in a little bit and take a look. Thank you

[polar spade]

And that's on my Android phone browser in desktop mode.

[polar spade]

This is handy too for zooming into that one word that gets shouted a bit too loudly.

[bleak furnace]

it seems that model 4.5 can't produce such old era sounds, i tried and model 4.5 gives very modern and professional results, model 3.5 and 4.0 give results quite similar to the 1950s'

[polar spade]

You're right. 3.5 does good for giving that authentic feel to things. I use it especially for audiences, applause and laughing.

[bleak furnace]

But its voice sounds so fake, you can tell it's AI, it doesn't sound human at all.

[polar spade]

It is only 7am here in the USA. So many are still asleep or heading to work. Fewer responses at the moment to the American doo wop sound.

[bleak furnace]

hmm maybe i should wait a bit longer

Perhaps the biggest weakness of 3.5 is the voice, it doesn't sound human, otherwise it's pretty good

[daring lava]

Not sure if this is what you’re looking for but here’s a possible example for genre tags/prompt style

https://suno.com/s/7r1cMjZjHECRSYFB

Beautiful song ❤️

[bleak furnace]

Yours too

I have a question, which mode do people usually use to create music if they don't write their own lyrics, do these 2 modes give different sound quality?

[daring lava]

The quality is the same, functionality and control separate the two

[polar spade]

@bleak furnace You can get a lot of prompt hints from searching the Suno home page for other users songs that give you what you want. "1950s doo-wop" puls up quite a lot. If you find one that's good, note the prompts that keep showing up in the songs you like.

[cunning mulch]

so... I recently spit a song into stems... and I ran into an interesting issue... the quality changes almost immediately at the 3:29 mark...
https://suno.com/song/39254d46-19ab-43f8-a96e-df1ca5398742

this happened in both sets of stems, too

[daring lava]

Good morning Tiberious ☕️

Thanks for reaching out and flagging this. If you are up for it, report it using the report bug action in the ellipsis next to the song and/or here: suno.com/feedback. That will get it straight to the team so they know and look further into it.

Will be taking note of this and looking for more cases regarding this kind of thing happening. If you happen to bump into more results like this please keep me in the loop!

[polar spade]

Yea. I see, no limiters at that point. Kind of open the gates just before it ended.

[cunning mulch]

so, I did a full crop of it previously, and stemmed that. Same thing.

[torn ore]

The restriction in Lyrics has been reintroduced in the Pro version?

[cunning mulch]

so.. what, Audio Quality?

[radiant lake]

Hi ! I tried what you said, I also tried to clear the cache and delete cookies. and I get the same error in the Chrome browser, the same error appears in Edge

[polar spade]

The error is typically more a website error not being able to properly send a client file codes to your desktop, if Im correct. So the other two ideas is to disable any added extensions or try to go into Suno in incognito mode. Anything which may be blocking code from properly being sent to your browser. But since its both browsers, these will likely not work either.

Hit F12 or right click to bring up the console to get the error code to provide Suno Support in your email to them if all else doesnt work.

[brazen yew]

These are my only options on the android desktop version. I don't see one for volume

[polar spade]

Partially my mistake. it appears we can only change volume PER SEGMENT. Click on a segment ( click top of segment to choose) and then the menu again and you 'll see the volume control. As for selecting a whole area and changing volume, it appears we can't in one step. Something I'll feedback to in suggestions/bugs. But if you can select each segment and change to same volume each, you have a bit of control at least.

[graceful condor]

Maybe its time to bounce the Suno servers, they do get hammered.

[daring lava]

Hey Setumi, thanks for reaching out. What restrictions are you speaking of?

[brazen yew]

Ok I got it. It just didn't want to click on the whole segment. That should still be fine thanks

[polar spade]

Im looking at it now before brekfast, for a work around.

[gleaming ferry]

Can anyone upload audios on suno rn? Like select them, i'm not talking about them getting flagged for copyright
(Mobile device here, android specifically)

[daring lava]

Hey there! When you get a chance could you take a look at this thread to see if anything there is helpful https://discord.com/channels/1069381916492562582/1382451659434758276

Let me know if you have any questions 🙂

[gleaming ferry]

I've tried multiple browsers btw, none of them worked, and it's not just me, all of my 5 friends can't either

[gleaming ferry]

Ok

[brazen yew]

Yeah ideally we can do smaller edits as lowering the volume on the whole segment will fix the spike at the start of the chorus but then just make everything else quieter

[daring lava]

Awesome, thanks for reaching out about this! Hope this helps!

[brazen yew]

Just can't get metal right, I fixed the static sound at the end of songs, now it's huge volume spikes

[gleaming ferry]

Yeah i alr tried uploading .wav files so i'll just try using the desktop version on mobile

[gleaming ferry]

Can't upload too

[polar spade]

OK. I think I got a workaround, but it takes two monitors.. hahahaha and it really sucks.

[brazen yew]

Unfortunate. I'm just going to have to remix or save the persona and hope it duplicates well

Or the most likely option, give up

[polar spade]

Oh....wait... i didnt know that changing some aspects in the original output, even after saving to a new file will still alter the original output. Thats totally whack. OK. I guess i need to make some notes.

[native minnow]

Hi everyone here, I'm sorry to inject myself in your discussion.
May I ask if there is a way to delete a persona also from the trash?
By desktop / browser I can't find any option to permanently delete a persona once I'm inside the "persona trash" section, unless the restore.
And also if the persona is inside the trash, I can't delete the songs made from that persona.
Thanks

[polar spade]

If i can find a way to screen record two monitors, i'll make a video. hahaha

[daring lava]

Thank you for flagging this. Will let you know if there is another workaround that becomes available. In the meantime (if you'd like to), please feel free to let the team know directly via suno.com/feedback!

[polar spade]

Now, breakfast!

[frozen dagger]

Sorry for the interruption. Since last night, editing with the Pro Editor hasn't been working for me.
I replace 6–10–20 seconds with the Pro Editor and insert the new texts, but the generation process goes on endlessly—sometimes for hours—until it eventually results in an error (which appears very briefly and not always).
It still consumes credits but doesn’t generate anything. Even if I log out and switch browsers, when I log back in, it’s still stuck on loading.

[polar spade]

yes, if you click on that last segment and change volume, as you say, it may lower an instrument, but also lower a vocal or other that were originally fine.

If the volume of a specific instrument within a segment is too loud, then yea thats a problem until we can work on stems individually , then form back together in Suno as one song.

[gleaming ferry]

Ty for the support, i have discovered that it works on Samsung Internet

I guess i'll just keep switching browsers

[daring lava]

That's an awesome find! Thank you for letting us know!

[fading tulip]

Hey, you. Listen to me. I sent email on support Suno few days ago to unlock my Suno account. No answer yet.

[daring lava]

Hey there! Thanks for reaching out, let's see what we can figure out. Let me ping you to the appropriate channel for such things. #archived-ticket-help 🙂

[graceful condor]

i have needle doctors yay

[vivid iris]

Hi everyone,
I'm having an issue with the Suno website and could use some help. When I try to use the app (via browser), I get this error: "Application error: a client-side exception has occurred (see the browser console for more information)." This is the first time I've seen this, and it’s preventing me from using the platform.
What I've tried so far:

-Restarted my browser (Brave) and computer multiple times.
-Cleared cache and cookies in Brave.

-Disabled Brave’s Shields (ad-blocker/privacy settings).

-Tested on Google Chrome, but the error persists.

-Checked my internet connection, which seems fine.

Has anyone else run into this error? If so, how did you fix it? Any suggestions for troubleshooting or workarounds would be greatly appreciated! I
Thanks in advance!

[eternal pumice]

Hello! Is it possible to make Chorus (only chorus; other song should have usual voice) in god-style vocals? Like on this video:

https://youtu.be/z-iWe4qXUD8

I tried to describe it in Style Description section, and also in Lyrics ([Chorus - church emotionless metallic divine voice]) but nothing changed.

[sleek stump]

try A Cappella Choir

[eternal pumice]

Where?

I tried it in both sections - doesn't work. I think i need to generate chorus separately.

[sleek stump]

You would describe A Cappella and choir in the styles, then reinforce that with Chorus (and possibly Choir) tags on every verse.

Which model are you trying to use?

[eternal pumice]

4.5

[sleek stump]

Describe that and then hit the magic wand, the AI will break out a more complete styles prompt

I just did that and this is what I got..
A Capella Choir performance with all voices blending harmoniously, each section delivering every verse simultaneously as a unified chorus. Layers of SATB interplay create movement and texture, using dynamic swells, subtle syncopation, and varied vocal timbres for richness.

4.5 requires a lot more effort than previous models, which is why they included the magic wand

[eternal pumice]

"Couldn't generate that.
Tags contained artist name: satb"
xD

[sleek stump]

Thats ChatGPT for ya

so dumb

[eternal pumice]

I have generated the chorus separately, and it turned out perfect. I'll be putting the pieces together in Audacity

[sleek stump]

I just genned a couple tracks with that exact prompt, did not get the moderation

[polar spade]

As mentioned in need-help channel, this is unknown issue and no fixable suggestion from the community.

You've emailed to support.

Sorry in advance.

[sleek stump]

And actually its being moderated.. Theres nobody coming up in a seach that has the artist name SATB

Just replace it with the full description.. Soprano, Alto, Tenor, and Bass

[sleek stump]

THis sounds like a script is failing, possibly due to a corrupted cookie. YOu can first try to clear your browser cache and cookies. That may fix the issue you are having (I had a similar issue a few weeks ago, caused by a corrupted cookie)

[lost holly]

Has all been tried. This error seems to happen to a few people today, so far around 4 have reported it today here on discord.

[nimble brook]

I cant pay for suno, it have worked before now it say:

Update billing information. I do that and same thing happens. Update billing information.

My subscription end today.. i try cancel and put in new credentials... But it say error... Etc .. how the *** do i pay for the service?

I have sent email 6 they do not reply

Last mail was 20 days ago.. no reply

Im starting to give up on suno, there are others out there with less errors and problems. Suno suport are non existens

[daring lava]

Hey there! Thanks for reaching out about this! Going to ping you to the appropriate channel for this #archived-ticket-help Looking forward to getting this worked out!

[nimble brook]

Thx.. i have 2550 credits and when i try to use em they say Update billing.. so its just a mess right now and i cant even buy new sub. Thx anyway

[sleek stump]

Stripe has not been a good experience for many 😦

[nimble brook]

I signed up from Google play store and in there i cant even finde the subscription and stripe have never worked for me

[polar spade]

You're requested in the need-help channel to assist you more.

[polar spade]

https://discord.com/channels/1069381916492562582/1372994553233276970

[nimble brook]

Hi

[polar spade]

I gave feedback in suggestions to at least have an undo button so our original prompts are returned if we think the magic wand didn't come close. 🙂

[polar spade]

I thought of that today. What if I click wand over and over and compare. I was too disinterested to try though. 😄

I understand what you were saying. Its that I find myself reading the magic paragraph and thinking, "that's not what I want at all"

[muted goblet]

I agree, my prompts can be pretty detailed and using the magic prompt option can remove detail, and most of the time the result are not as good and detail is lost. I think it's probably good to add detail to shorter prompts.

[sleek stump]

I’ve done this, if you don’t change the original prompt, it returns the same result.

[past kernel]

Got a Feedback here. Swift ist "NOT" a "artist name". Just because it happens to be the name of an artist it still is a english word that describes a tempo - this shouldn't be filtered - were can i report that?

[lost holly]

https://suno.com/?open_feedback=true Here for example

[polar spade]

Yep. It is. SWIFT is trademark signature logo of Taylor Swift among a few dozen others she has trademarked. Thats enough to make it more important to NOT allow Swift and other words that can be a part of an artists name from getting through.

So.... dont put "shake it up" or "1989" in the prompts either. 🙂

[sleek stump]

Swift is also not a tempo that Suno recognizes.

Suno recognizes the following:
grave, lento, largo, adagio, andante, moderato, allegretto, allegro, vivace, presto, and prestissimo,slow, medium, fast

[polar spade]

For what its worth, Suno does recognize allegro or presto.

[sleek stump]

Suno also ignores any BPM commands.

[past kernel]

ofc it wont recognize it if it is filtered

[muted goblet]

Good job your not writing a song about a migratory bird related to the swallow

[sleek stump]

YOu missed the point. Even if Swift wasnt filtered, Suno still would ignore it in current models.

[polar spade]

I guess a song on the history of the fall of the Berlin wall is crippled a bit too.

[bronze pecan]

hmm what am I doing wrong with the replace section tool?

I selected in the lyric box 2 sections, did press replace, now the lyrics from these sections are gone and replaced with the chorus. I dont understand ?

or do I have to enter new lyrics even if its the same lyrics? probably that was the issue?

ahhh yes indeed lol 😄

maybe this should be added into the learn box 😄

[sleek stump]

Skill issue! 😉

Also, it works better with smaller sections.

[bronze pecan]

It just wasnt obvious to me, because I selected the text and I thought it would just regenerate it with the same lyrics. Was indeed not obvious to me that I have to say, Do new lyrics when its supposed to do the same lyrics 😄

[half dragon]

The old editor .. you don't have to. Guess they changed that.

[bronze pecan]

in that case I wanted to try different melodies.
But in case the lyrics need correction it would be really nice if there would be finally an option: Reuse Seed, so that indeed just the lyrics get changed but not the melody as well

[sleek stump]

I prefer the legacy editor

[bronze pecan]

I'm still waiting for being able to copy over a song section from one song to another

but we're so close to it. Because copy is now at least possible within a song

and moving elements

now just doing it from one song to next is missing 😄

This is for example also useful if you did a remaster and then e.g. verse 1 is better at original song, but the rest is better in the remaster. Then being able to copy over verse 1 from original song into the remaster would be amazing 😄

in times of v4 it wouldve been even more useful, because the quality of extensions did degrade a lot after 4minute song length. But if you extend that song at e.g. 2:30 then its again 2:30 quality. But then the extension would be applied to 2:30 and would remove the rest. Another case where copying and pasting from one song to another would be very helpful

and an equalizer would be nice. Especially on release of 4.5, 4.5 did very low treble

And I hate 9:16, i cant repeat me enough with this :/

its such a bad width, nothing fits on covers

car + person? impossible without one thing cropped to death

[elder dew]

Hey all, not really a prompt question... but is there a way (other than just remastering over and over) that you can edit your song to have your instruments louder? I have some really nice song gens but the instrumetns are soo muddy and muted it ruins it.

[bronze pecan]

also video AI obviously struggles with 9:16 getting stuff on this thin width too
why must be everything for phones??

The standard for covers is 1:1. then things fit on the covers. Why Suno need to change the standard

Before we had at least a square upload form for the covers, which caused crop too with their bad 9:16, but not as much crop as now with the 9:16 upload form ....

9:16 is really the worst idea ever

[half dragon]

Not on Suno, I'm afraid.
It's only possible offline at this point.

[bronze pecan]

But I found it interesting that my upload of my own remastered song wasnt blocked to publish. Maybe Suno detected its his own singing? Because normally lyrics would block public button

If that is reliable I could just edit my songs locally and upload them afterwards. Will definitely try on more songs

[elder dew]

Thanks, is there a prompt maybe that will help prevent the muddy water sounding instruments? I would say probably more than 80% of my songs have this muted/muddy sounding backing tracks. 4.0 didn't have this issue at all.

[polar spade]

Until we can edit stems and reconstitute them, i can't think of a means of doing that.

[bronze pecan]

you mean too less treble probably?^^

that was especially bad on release of 4.5

at release of 4.5 there was barely any treble at all xD

[elder dew]

i dunno if that's it... maybe. Here is an example. And this isnt even one of the bad ones. But it's the muddy sounding music in the background that I want to clean up
https://suno.com/s/AYDSc3AXuXnihbUG

[bronze pecan]

Where are the elements from editor now stored? I wanted to delete some Replacements, but I cant find them at "create" page even when changing the filter to display songs from edit mode

[half dragon]

Maybe try this? some had success -
#💬┃general-chat message

[elder dew]

Nice ty I will try that 🙂

[sleek stump]

for 4.5, any hard/fast/metal, the band sounds like its playing from the bottom of a mud puddle, its been widely complained about. There are a bunch of negative styles that you can use to improve that, but ultimately its going to take stem generation and post processing to truly fix.

[elder dew]

When you say stem generation, do you mean yank out the existing stems, recreate them, and then put the recreations back into the songs?

[bronze pecan]

you probably wanted it more like this, right?^^ @elder dew

[sleek stump]

No, Suno doesnt do STem SPLITS anymore, it actually regenerates the detected intruments and vocals. we dont yet have a way to recombine stems in the editor, so that would be done in your choice of DAW, then if its under 8 minutes you can upload it back to Suno

[elder dew]

Yes how did you fix it?

That sounds great

[bronze pecan]

Treble

Using Wavelab Pro and a good EQ fixing the treble

[elder dew]

nice, I'll check it out. Thanks!

[bronze pecan]

but nice song 🙂

[bronze pecan]

still not real stems though

#📚┃suno-school message

and it was even worse in the beginning of 4.5^^

here another comparison with the message link

#📚┃suno-school message

Also here massive difference 😄

here you see the treble of 4.5 in its first songs 😄

They at least improved it a bit, it was this bad at release 😄

[elder dew]

pretty crazy that it's still as bad as it is though. Sucks I have to spend $100 for software just to make my songs sound okay lol.

[bronze pecan]

😄

the unfortunate thing is, that the first songs are still in this state now. But thankfully I should be able to reupload them now - when hopefully the publish button continues to see that the singing is from suno and allows to publish it

[sleek stump]

The new stem gens are recent, like within the last 2-3weeks, thyey are light years better than ANY spectrogram separation.

[bronze pecan]

I heard that theyre better now. But still not real stems because they cant get from the ai them seperated, thats at least what was told to me- that the ai always puts out the final mix

[sleek stump]

Suno can gen up to 12 tracks

[half dragon]

Not real stems?

[sleek stump]

Better, each are completely regenerated

[half dragon]

Sounds pretty real to me 🙂 best feature for me recently.

[bronze pecan]

real stems would be that they never were in the final mix, but its still from the final mix. Just in a much better method now. but its still fake stems 😄

[sleek stump]

Spectrograph stems always have bleed, even Lalalal.ai

YOu are confusing stems with separate instrument track recording.

[bronze pecan]

I mean suno said that theyre still not true stems, just a much better method

[sleek stump]

So far its the closest we can get to separate instrument generation at run time.

[half dragon]

Does it matter? as long as they're sounding the same when putting back together right?

I now can access to all the stems I need in my song... so good!

[bronze pecan]

Hopefully. Didnt try yet 😄

So is it now that the instrumental has no voice at all in it anymore and the very same quality as the final mix on the voice parts?

[sleek stump]

I have a 32 track TASCAM sitting behind me, I would love to be able to put that to use here..

[half dragon]

Very clean.

[bronze pecan]

I wish I could select which ones though.

I mostly just need vocals and instrumental seperate. and this could be surely offered much cheaper than 50 credits then

[half dragon]

Vocal + Instruments only cost 10 credits

[bronze pecan]

that option is using the old method

you must use the big one for the new method

[sleek stump]

Nope, its new as well

[half dragon]

is it?

[bronze pecan]

since when?

it wasnt

[sleek stump]

Does regenerate the 2

[half dragon]

latest update

[sleek stump]

When the new stems went in, it was both 12 and 2

[bronze pecan]

so there was indeed an update? because the 2 stem variant was still the old method

[sleek stump]

Nope.

100 nope

Ive done both, and the 2 split is just as clean, vocals with zero bleed

[bronze pecan]

hopefully, will try soon
pretty sure they said 2 stem is the old one for now. But maybe there was another update then

but the old method also did cost 0 credits. if its now 10 credits its quite a hint for the new method 😄

[ocean cape]

Só te usar o campo de letra das músicas para prompt de comando ? Ao invés de por a letra em si usar para descriminar estilos, tons e tal

[sleek stump]

No, you need to describe your styles and how you want the song to structure, then add your lyrics to the lyric window with the correct tags.

[bronze pecan]

but still: Can I somehow delete replacement sections? They dont show up at create for some reason despite setting up the filters

ahhh I found them

they were put in the wrong workspace^^

[half dragon]

You probably didn't change the workspace while in the editor - it's defaulted to "My Workspace"

[bronze pecan]

yep thats it then 😄

but why it doesnt default to where the song is in at? that makes no sense, quality of life upgrade i would say 😛

[half dragon]

Agree. It should ~~default ~~ set where the song came from.

[rain drift]

I don't get credit according to the conditions I should get.

Why didn't I give a credit on June 18th and the credit for inviting friends? Why did I only get 250 credits even though I had invited 10 friends and created a song according to the conditions?

[nimble brook]

Hi! In working with a track thats almost perfect for what im looking for, but the base are way to groovy.. and bouncy.. how can i change it without mix with the rest of the track? I did covers for 1000 credits and not a single one gets it as good as orginal one when it comes to voice, synth, but they do fix the base... I tried fix it in fl by swapping base from cover to orginal track but na.. its off beat some times etc

[half dragon]

Unfortunately you don't have many options, if at all; at least within Suno.
The task sounds complicated to achieve, the only option is to replace the entire bass track offline on FL manually.

[nimble brook]

Hmm, i have some hours of work tonight then 😅

[nimble brook]

Man i solved it!

[half dragon]

Yeah? Nice!

[nimble brook]

I did split the track put all the lines in fl, exported it, then in to suno without base, took keep track 100% and type add soft moving base. And done 😅🙏🏼

Indit not put the base in ofc hehe

[half dragon]

Ah.. smart... forgot that we can upload up to 8 min now... so option opens up.

[nimble brook]

Yhe so exported 7 Chanels from suno, base, synth etc then removed the base and exported track, int to suno and bam

[half dragon]

Is there vocal on your song? doing that would prevent to publish to public on suno.com if that's your end game.

[nimble brook]

Yhe but i did not ad vocals to new track, just the instruments

When i exported from fl i took, synth, effects, drums etc

Then i put the new track with in fl with old vocals

[half dragon]

Right. If you're publishing it somewhere else.. that's perfectly ok. I guess you could do the same if you want for the vocal - regen the vocal like you regen the bass

[polar spade]

Quick question: did all the tracks align up or did you need to tweak any of them when you put them back together in FL?

[ocean cape]

https://suno.com/s/HbO3d5RtUs0YlRSK

Estava testando agora a pouco mas ao invés de compreender como descrição e regras para música ela traz em letra, porém nessa ela trouxe só o instrumento e olha que eu achei interessante https://suno.com/s/yFTPWQcqN5fgqweA

[ocean cape]

https://suno.com/s/Nyy0NjXHft1NahEi

[ocean cape]

https://suno.com/@mais001conta

[thick tree]

Anyone knows what is the right setting to use when covering a song to make the output sounds 90% close to the original uploaded track?

The sliders been giving me mixed results so far

[nimble brook]

That alighn good, but did some tweak for kick. Just 1 file was off

[polar spade]

Great to know. Thanks for responding.

[nimble brook]

Np!

But if you do what ibdid, remember to promt suno for same key, i hade to redo it haha

[young geode]

Depends on what you mean by "sounds 90%".

If you mean identical to the original, note for note, you'd want to use a combination of a persona of the song, and the cover of the song, the audio influence to around 60-70%, while keeping the other sliders around 40-55.

That'll give you a good approximation. You'll likely have to play with the style prompt, but that's about as good as you'll get.

[patent sand]

[REDACTED]
I answered my own question 🤦‍♂️

[lunar remnant]

when you extend a song, how do you join it to the original?

I'm trying to extend the lyrics out an extra verse.

maybe it'll work this way.

[leaden pawn]

The sliders have been removed from advanced options and covers are worse than ever... Why have we lost the sliders?

[lunar remnant]

their computers probably struggling 😄

[half dragon]

They're still there as far as I can tell.

[lunar remnant]

dunno, reset your browser clear cookies etc

[leaden pawn]

Yeah did that

[lunar remnant]

disable addons

clear cashe

pay your bills

[gray wadi]

What version does it say youre on?

[leaden pawn]

Good catch! 3.5 ......... why on earth is it on that lol - I never touch that toggle

[leaden pawn]

No wonder my covers weren't behaving

[lunar remnant]

it keeps failing to extend my song

[half dragon]

Seems you extended from the end? try to move the extension point upward a bit into the previous lyrics... then include that part of lyric with the new lyric - see if it helps.
I have more success that way.

[lunar remnant]

yeah trying to tack on a verse at the end... kills me.

[young geode]

Is the song missing a verse it should have, or are you trying to add a verse?

[lunar remnant]

the song is over the 5000 characters

[young geode]

So you're adding a verse it needs.

Do you care if the verse is sung like all the other verses or do you just want it within the song?

[radiant lake]

Hello! Thank you to everyone who fixed the connection error, I wish you all a good day! Everything is fine, the music show goes on!

[lunar remnant]

hmmm good point. maybe I'll create it as a new song.

[shrewd willow]

For awhile for some reason I had my default setting to 3.5v and had to manually change everytime. Weird and annoying and I have no idea why it happened lmao

[leaden pawn]

Lucky I only blew 100 credits on it!

[young geode]

Well there's a way, I just needed to know exactly what you were looking to do.

If you want the verse itself sung like the other verses, it's more difficult, but doable; but if you just want the lyrics, we can do the different method which makes it a bridge.

[lunar remnant]

I've got a cool early 1900s poem, chucked it with metal. It sounds ridculously good. but it's missing the last verse.

https://suno.com/s/oehVNGSxZw7heopw

the last verse sort of brings home the whole song

but it's missing.

https://allpoetry.com/A-Song-of-Brave-Men

[steep drum]

that doesn't sound like metal to me

[weak olive]

Hello Suno Team 👋,
I'm trying to create an original devotional rap using rewritten verses inspired by Hanuman Chalisa, but it gets flagged for "copyrighted material" even though it's completely reworded.

Can you please help resolve this so I can use my lyrics in Suno?

Thank you 🙏

[polar spade]

Im being picky, but shouldnt, when we click shuffle then click Play button of our Playlist, it should start out shuffling the Playlist. Right now it always starts with the first song in the list BEFORE shuffling after that. Like, I have to close my eyes and random click on something to experience a true shuffle. 😜

[polar spade]

There's no resolve in our favour. Suno simply asks we only use content that's created and owned exclusively by us. If material is available to others as well, then its not exclusive to us.

So technically, if you disburse your own material in the past, then upload it as content to Suno, you might even get blocked that way as well. Suno can't discern whats yours on the internet from whats not.

[glacial cobalt]

I have an issue when trying to generate songs in Afrikaans and are not being generated due to "Prompt contained inappropriate material". I have been over the lyrics and style prompts until the cows come home trying to figure out where and what is inappropriate even looking at every single word to see how it could be misconstrued as an offensive or inappropriate English word and I am getting nowhere. The message received back is giving no information as to what is a problem or even where to start looking and it has become extremely frustrating. It's also not 1 particular song as I have 2 now already that exhibit the same problem, and both are as clean as a whistle from what I can see.

I have managed to generate some Afrikaans songs which contain similar Afrikaans words and even some swearing (very little) and were not a problem.

https://suno.com/song/8832fd87-e80d-4234-8433-69f06af728f4

https://suno.com/song/14ed82fb-ff5f-4870-8fd2-c2618d6278cb

https://suno.com/song/a31e316b-839b-4a40-a727-ab855fc39364

Any thoughts or help would be appreciated.

[polar spade]

Sometimes an easy resolve is to shift some words or lines around. Add a line space, move a word two spaces left. Its hard to tell with 2000 characters, but you can cut lyrics in half to test each half, then that in half, and continue until you find what seems to be getting flagged.

Also , throw it in chatgpt and see if it recognizes anything it might deem as concerning. Chat gpt can look for offensive English words (if instructed) such as an example:

"Die kind het 'n groot assie gehad" contains the word a**, in English.

[glacial cobalt]

Hi Dasvon, I can certainly try moving things around or half lyrics etc. although that seems like a hit-and-miss scenario and will also burn through credits. As for ChatGPT, I did in fact use ChatGPT to generate the lyrics of the "offending" songs in the 1st place and as Afrikaans is a 2nd language for me I've been through them multiple times changing words that may get flagged for any reason I can think of. Hence my frustration.

[polar spade]

Its to say that its not always only to do with content. Sometimes its literally placement and arrangement.

Moving one thing may re-group the tokens or how Suno is viewing groups of letters and that gives it a green light.

Even a common phrase might get pinged. Such as "one for the road" , an innocent phrase, may be seen as not content worthy.

[outer ether]

Any work around for the current bug with extending songs? I've tried everything

[glacial cobalt]

So after burning through some credits eliminating lyrics from the bottom up, I end up with the following line that is preventing the generation ... “Wat dink jy, vampier?" and more specifically the “Wat dink jy" part ... which means "What do you think". I initially thought the "Vampier" or "Vampire" would be a problem, but no .. specifically the part already mentioned. I cannot see anything that would be remotely a problem in English.

[polar spade]

I may be out of the loop of current bugs. What's its doing/not doing?

[outer ether]

Only gives 1-10 seconds extensions that finishes the song, no matter where u extend from

[glacial cobalt]

I've had that exact same problem and have given up trying to extend anything for now

[outer ether]

I see someone a few messeage up with thr same problem too

[polar spade]

Dink is an informal jab at someone, like calling them a nerd, but I can't seek it being so offensive. And can be said as a term of endearment. So, other than that. I dont know. Any chance to rewrite it?

[outer ether]

Suno let's me use swear words and the n word it's weird u can't say that

[polar spade]

You needed a work around at least. So taking one of your short pieces and extending again off of that can help get through that rut of it repeatedly giving short extensions.

At times I'll get those 2 to 5 second extensions from extending off of an upload. A known bug since 4.5 was infroduced. I just extend off one of those and usually continue with no issue.

[outer ether]

I might try download it and upload it and see if that helps

[glacial cobalt]

There is not really another Afrikaans word that would mean the same thing really as it's a pretty simple word that any toddler would learn growing up. 🤷‍♂️ so kinda stuck on that one then

[outer ether]

Try dynk

Or a different ï

[polar spade]

Well..... add a line space above that line in that stanza and see if that affects it hopefully without messing up the metering and pace. Seems stupid, but it's fixed others issues before.

.
Its hard to resolve an issue and do it in 10 credits. 🥴 Basically everything is testing until it works.

[earnest rose]

Ahhhh, ok nice. Yes that help never see/use that. What i mean was just on the right side when a song is marked a button "copy lyrics" - but that also work, thanks

[polar spade]

Its your suggestion, so if you mean something more specific, let them know. Sounds like a good idea.

[half dragon]

I wish when playing to the end of the playlist, there's another option to either:

• stop the queue at the end of the playlist
• back to the begining of the playlist

rather than continuing play some random songs from public; but I guess it does serve a purpose as well..

[glacial cobalt]

Changed the word to nadink ... which means the same thing but doesn't sound the same in the context ... will have to live it it I suppose. It's weird that you can swear but a simple word like that triggers the system

[sleek stump]

You are not the first to have this issue with Afrikaans, the actual problem is coming from ChatGPT (Used by Suno for content moderation), its erroneously flagging words and phrases as NSFW. There was another user a month ago that had the same issue, and from what I remember, he had to give up on using a specific line in the song lyrics. Honestly, until GPT gets better at Afrikaans you are going to run into this. This is kinda not something Suno has a lot of control over.

[jovial widget]

What has happened to Suno? started off being amazing in 4.5, but now its awful. Same thing happened with 3.5 started off great but then just utter rubbish.

[sleek stump]

Looking back at the conversation, I think the user may have got past the filter by phonetically spelling the phrase.

[glacial cobalt]

To be honest I think that the word ... dink ... is rather innocuous and a very simple Afrikaans word that is used in every day conversation. Very weird that it would be flagged. Even in English it has no offensive meaning.

[sleek stump]

I get it, content moderation is a bit whacked, it flags a musical term SATB as a artist name, which is utter BS.. lol But this is a problem with ChatGPT. Unfortunately, Suno uses it for its auto content moderation.

[polar spade]

The testing worked and you moved it forward. Thats a small win, at least.

[polar spade]

Bring back the loop button. If loop and auto-play are on, it loops around and keeps playing the Playlist. If auto-play is off, it loops the one song over and over.

[polar spade]

When I've written any song, I literally listen to 50-100 times before I go back in and rewrite. Then i rinse and repeat. Often about 5-20 rewrites. The loop button is helpful with that.

[half dragon]

Looping on 1 song is available but not looping entire playlist.... a lot of time, I just want to listen to one playlist over and over without having to replay manually.
At least we got shuffle now, so it's 👍

[polar spade]

I am probably off in my thinking.

Does the repeat button NOT repeat a Playlist if autoplay is on and the playlist Play button is pressed? But if the output play button is clicked , repeat just repeats the one song..... and not go into Suno Radio mode or keep playing the next one in our Library?

[half dragon]

Repeat 1 is visible in the player.
If it's off and AutoPlay (hidden) is on - it'd play the playlist entirely, then go into radio mode.

[polar spade]

Ewwwwww! 🤮 ok. Repeat AND Loop button needed.

[quaint perch]

What's wrong with Suno website?? I gave the create command an hour ago and it is still rotating. Please fix this problem.

[polar spade]

Mention in the bugs channel. Though glitches have happened the past couple of days. Typically I assume some "maintenance" is bottlenecking performance.

[graceful condor]

Just uploaded a sample 2:36 I get back one that's 1:39 of that I got :30 music the remaining time complete is silence. How does eating my credits taste? 🐟 🍗 🍝

[graceful condor]

Come on now!! my sample is 2:38

[viscid skiff]

glitchin probably because of the contest, quality dropped too

[tranquil mountain]

Hi. After I paid for the subscription, I lost access to my old creations. Is it possible for me to recoup them? I can't find them. All I see are the creations I made yesterday and today. Please help me.

[round fjord]

how can i reuse the voice i once used in a song, is it by persona? seems like when i use persona, it always sounds a little static

[patent sand]

You need to contact support. #start-here

[ruby saffron]

Thank you for this I was looking on how to structure it. I've been brute forcing it but it hasn't been super easy.. 😦 Even in sections outside of the duet where I want only Singer 1 to sing it doesn't always stick to those prompts... >.< any other advice? in any case I appreciate you thank you!

[young geode]

Well, do you have the section you want the duets?

If so, I believe you can regen the other areas with a completely different prompt for just the single singer.

Other than that, as I said, it's super difficult, because SUNO is doing a pattern prediction. My advice would be simply to move on, as it gets costly. Hold the song in your back pocket and wait for a SUNO version that can handle multiple vocalists without hallucination.

[ruby saffron]

haha its been doing -pretty good- I have to restructure for a female then male then female then male type pattern but the duet part was not coming together, one singer would leave off mid line and the other would finish which was cute but not what i was looking for xD but yes thank you. I think i'm going to keep working on the lyrics anyways. 🙂

[hard beacon]

sometimes "layered female and male duet vocals" in the smart pompts work 🙂

[lusty crow]

hi folks. can someone tell me where i can find the prompt documentation for suno.ai

[hard beacon]

after 2 years of using suno i just recognized that Dubstep is a very cool creative process during the hectic and glitch fx'es...

🙂

and "hell naaaa".. my active dubstep listening time is round about 10 years ago.. but this is really cool... so.. NEW DUBSTEP PROJECT is born 🙂

[rigid pasture]

I have made a song, it is sung by boys, I want that song to be sung by girls, how can I do that song? Someone tell me a little bit.

[hard beacon]

maybe with the cover function.. and then in prompts "female vocals"... and metatags in the lyrics [female]

[rigid pasture]

i don't understand

[neat nexus]

Are you sure it isn't already? Boys when young have a higher pitch.

[hard beacon]

puh.. i don't really know how to explain it in "easy words" 🙂

[rigid pasture]

Step by step, please tell me a little bit, it would be helpful.

[viscid skiff]

having trouble getting crowd sound, is there a mehod or randm?

[rigid pasture]

@hard beacon Brother, I did it, thanks.

[half dragon]

Maybe add "live" or "live concert" in the style... have you tried that?

[oak siren]

currently editing using replace section. what are the best tags to use to end a song? i've tried "outro" and "end" but suno fails to do so 9/10 times

[hard beacon]

v4.5 is not that good in "live performance generations"... best way is to create the live-version in 3.5 and remaster to 4.5

[unreal stratus]

Question: Re: Extending an instrumental cover - I have uploaded one of my instrumental tracks into Suno and recreated a cover / remix of it. The created track is 4 mins long. I wish to extend this to say 5 mins. I set the extend from marker at say 3:30 (before the audio outro). However all attempts to extend does not extend past the 4 min mark. I get 30 seconds 'extensions', which produce a very similar outro. There is no editable setting with Extend as to how long you wish to extend for. The end time of 04:00 is locked. I think this option only really works when tracks have lyrics ( and a new verse with lyrics is added at extension). Does anyone know how to extend an instrumental cover track when there are no lyrics???? Note - I can create a new section at then end although this does to maintain continuity of the instrumental. Any ideas?

[lost holly]

Make sure you use v4.5 and try the legacy-editor.

[unreal stratus]

No difference with legacy editor... if I set the marker at say 3:45 this time... it just completes the track to 04:00. I am just wondering if this is a limitation of the 'upload audio / create cover' functionality + being an instrumental (FYI - I am using v4.5 also)

[lost holly]

Have you tried the "add new section" option? Apologies, i am not an editor expert, thats why my ideas here are limited.

[crimson yoke]

The main issue for me is the lack of freedom to move and isolate parts of the song. The editor doesn't let us drag sections, like the chorus, far enough to work on them independently, there’s not enough space. I can’t cut or rearrange things with precision, and that’s very limiting.

On top of that, the model itself is frustrating. Every generation changes the melody, voice, tone, and harmony, even when I just want to fix a small part. Without consistency, editing becomes nearly impossible. A stable seed system or fixed base is really needed.

Please, SunoAI, take a look at this, I can’t edit the way I used to at launch because of these limitations. 😢

[sleek stump]

Extends in 4.5 need instructions in order to actually extend the track. If you have the Instrumental switch turned on, switch it off. Set your extend start point in thge "middle" of what would be considered the last verse, NOT the Outro. In the lyrics section add a new section meta tag for [Verse] and then add a few lines of punctuation characters, not letters. Add 6-8 lines, then add another Verse, repeate the section, then add the [Outro - Instrumental] tag. That should get you a decent extension.

I disagree with your assertion.

[copper cloud]

Is there a way to make an extended piece shorter without cropping it?

[lost holly]

Fade out in Legacy editor? Changing lyrics so it ends faster, replace section... Depends on your goal.

[unreal stratus]

I think the Instrumental switch is only applicable at the time of the cover creation. Once a cover has been created and Extend is pursued, there is no instrumental switch....

[unreal stratus]

Yes, this adds another section, but lack continuity with the previous audio

[patent sand]

Sometimes, yes. But sometimes it will hallucinate some very bizarre music after it's done with your prompt just to hit the maximum time (extension-specific)

People have reported that instrumental mode can add random vocal elements unprompted also

[unreal stratus]

Many thanks. I'll take a look at lyrics field tag construction.

[mild hound]

anyone having problems where the second chorus is sometimes slightly or massivly different then the first, then the last chorus then adhears to the first chorus correctly? primarily in metal music?

while also using personas and modified option sliders

i tried tweaking the weirdness and tbh it doesnt to much atleast to my experience

doesnt help lol

[patent sand]

Yeah I've never used instrumental mode ever (never saw the point tbh) but that makes sense

People put some wild stuff in that style field

[lost holly]

If you want a prompt for getting hallucinated songs, let me know. Some of them can be surprisingly good 😉

[patent sand]

Those meows in the middle 💀

[small bridge]

is anyone else having some really terrible 4.5 gens today. I haven't used it in a week and it just came off very robot and terrible.

[brazen trail]

It's not just today — I face the same issue every day. No one knows when it will work and when it won't. It's very hard to work with an unreliable product like Suno AI.

[polar spade]

Hallucinations : As for getting vocals as well, once I find a prompt that gets a hallucination with different verses and a good melody, i use it again, and by just putting male/female vocal in prompts and especially in lyrics, you can usually get some sort of random babble song.

Often its triggered more frequently by putting a vocalization like Mmmm , Ahhhh , Oooo, or yeaa... but that intro can get too predictable, so a couple of time doing that is enough on my ears.

Just throwing random prompts and trying this you often get more of the babble just repeating over and over with whatever it starts... "what a wondo, what a wondo, bee da bondo, what a wondo...." and I trash those. Or we get no hallucination.

Getting them in your favourite genre may be harder.

3.5 is still best for hallucinations.

https://suno.com/playlist/a6a3f1cb-12bb-4fe7-ba02-f847110c2d9f Test for Hallucinations Playlist.

[polar spade]

Most of those stunk. Alot of bee da, booda, buddabudda, bee da and no variation in the verses. But that's more because of the kpop vibe.

[small bridge]

and now I am getting this... sigh

over and over btw not just one time

specifically happening any time I try to delete a song

[sleek stump]

If you are using a desktop, you can hit F12 and look at the console for whats failing

[polar spade]

Its hard to pin point anything the past three days. Too many different glitches going off at the same time. Its usually a "Thursday through next Wednesday" pattern when they're updating or maintaining, but this one started early this week.

Go through all the normal local computer cleaning and try again soon. Don't forget to post in bugs channel.

[small bridge]

damn it's really been struggling lately huh? That's sucky I have barely used my credits this month because I have been busy and now that I am trying to its barely usable. I will have to reach out to support last time it took them 2 months to respond tho. I am going to have my credits expire and won't have been able to use them. which i hope they will refund for.

[copper cloud]

I meant like a way you can do it on your phone

[polar spade]

Yep. A few.

[small bridge]

well at least I'm not alone in being screwed. That's always nice. XD

[polar spade]

/refresh

[flint haven]

Free Credits Refresh

Refreshes at 0000 UTC+0 (<t:1712880000:t> local Daylight Saving Time)

[small bridge]

I'm kidding. It would be better if no one suffered.
lol

[polar spade]

But when things work again, we get a dopamine high. So it balances out.

[small bridge]

exactly

[lunar remnant]

metal is a broad genre, this is a metal ballad like Dio, sounds entirely different than nu-metal like slipknot.

[steep drum]

Gotcha

I figured that’s what you were going for

[lunar remnant]

yeah rhonny dio style

[steep drum]

Checkout my band. Velocity Vibe on Spotify or Apple Music

[lunar remnant]

this isn't metal

[polar spade]

OK. Kids. You both took a swing. Shake hands and play nice, now.

[small bridge]

https://tenor.com/view/cats-gif-24765849

Can cuddling be an option. I wove a good cuddle.

[polar spade]

New feedback survey for a chance to win 500 credits. https://discord.com/channels/1069381916492562582/1290756014714519683

[zenith depot]

Can it be possible to change the singer and/or lyrics but keep the same beat?
It would be cool if Udio, Suno, and ElevenLabs made it where you can easily do this?

[ember summit]

Got an oddball one. I had two country songs appear since yesterday in MY libreary. Titled "J". From what I hear they sound good (can't totally hear the words), but these aren't mine and I hate to delete someone else's works. Any idea on how to marry these up with the owner who may be looking? I'll keep them for a few days JIC there is a way to get these back to the real creator.

[polar spade]

The creator name is below the title. Click to go to
Their profile.

[ember summit]

I thought so, too, that but it says I did. Shows my name. Nothing like any of my songs. Does not show the lyrics and hard to make out with the voice, but the lyrics I can understand sound goo. Not mine though. Just trying to help whomever out that created it.

[ember summit]

Hmm. Thought I sent this. I just PM DASVON. I had an issue (1st time) with new 4.5 where I attempted a simple "replace" on a few words and croaked the Suno processor so thought I'd sign out and back in. DUmmy me initially forgot login method and used phone, got the security code and logged into another's account. I've had this phone 25 plus years. Freaked as I saw 50 points as a free user vs my Pro 1985 points. Remembered how to log in and all good now. Sent details in PM to see if I should ask SUno help ticket how someone else has my phone number and could use it.

[small bridge]

ugh they still haven't fixed the extend feature!

[polar spade]

OK. So you logged into another account?

I can't say that's unusual. I have an account .....well.... more specific... the Profile page of a foreign account... when i logged into one of my accounts, that has shown up on occassion, and stayed awhile. I did report it and its normal at the moment. It was an account with ONE public file of K's of views, but everything else was my account information.

Again. All clear now. I dont think we're logging in as much as account/subscription bits may be crossing paths.

[polar spade]

I kind of think your getting fed bits of other account pages regarding the outputs/songs only. I dont feel its personal information getting crossed. Ive experienced in twice in the past year. And oddly neever on my main profile account so far. Just the free one, so far.

But no way to know anything more that that. It should NOT happen at all.

Definitely report each incident.

[signal haven]

Does anyone know if Suno actualy creates a different structure for songs with vocals compared to instrumentals?
For example, if I create a song with vocals but don't like them, would the stems without vocals sound like an instrumental song or would I get better results creating an instrumental version right away?
Because I don't know if Suno creates extra spaces and breaks for vocal lines that usually wouldn't be created if there were no vocals intended to go on top...

[polar spade]

Its your call, really. The track created , raw, and straight from the "stem" separator may have these gaps. But you would then cover and/or remaster that output to simply generate a new instrumental. Then you decide if you like the results. Getting the similar instrumental including chord progression and even motif/leifmotif would be difficult, id say, from trying to prompt an instrumental alone.

I like occassionally doing this and making a long instrumental intro ( with motif and melody lines often) to extend and attach the lyrics back for an EP style song.

This was the beauty of working with the older track separator that had a bit of overbleed of the vocals. By covering that, the vocals turned into melody lines in the instrumental. Something no DAW could do on its own.

[jovial widget]

Has the audio been fix yet or does every song still sound like garbage?

[brazen trail]

Same here, to me, every song sounds like garbage — nothing has changed. It's a waste of time, effort, and credits.

[polar spade]

@novel relic. Writes: "....Good morning, I have a problem that I reported to support 3 times but it has not yet been resolved. I can no longer regenerate individual parts of the song. The regenerated part is always full of bugs, often with incorrect words, and sometimes even in other languages.I report the bug but I do not receive any credits back nor do I get any responses. The problem is that this way I cannot modify the parts of the song, or extend its duration when the song lacks the ending.This forces me to try to make other covers spending more credits. And I can never find a version that perfectly matches the given commands.The problem has been present for too long now. What can I do?......"

What was working correctly before? How were you regenerating individual parts of a song before? Are you talking about stem/track parts, or sections of one song? [Im not following exactly what that means. ]

[novel relic]

It worked correctly before the update. Since the interface changed, I've spent millions of credits generating parts that were buggy. I reported all of them but have not been refunded anything. I've written 3 emails to customer support but got no response.The same problem, but with less frequency, occurs with songs that have their endings cut off. If I want to extend them, the extension doesn't match the lyrics and is always very long.However, my biggest problem is actually the replace command that does not properly regenerate the selected part, retrieving another part of the text different from the selected one, sometimes changing the tone of voice, sometimes changing the words, sometimes changing the language.

[echo prawn]

Why does it always generate some sh!t?

[polar spade]

Oh. So "regenerate individual parts" means replacing lyrics within a section of an output. Ok. Got it. And the rest of what youre saying, there's a lot there to unravel.

It may be best then, to have Support explain which of these they think is an error with Suno and which are not.

I get glitches and issues, but have been able to work around them or just improve my skills and realize how it really functions as opposed to what I expected.

Well, if Support doesnt respond soon, try to get a moderators attention and see if they have further suggestions.

Good luck then.

[polar spade]

And if you liked how it worked before with Legacy, then have you used the older Legacy Edit UI instead?

[novel relic]

I wouldn't know how to use it.

[desert iris]

How can I make a song that is only 45 seconds long, without the need for edit after generation? Every song I generate is over two minutes. I would really like to figure out how I can be more specific regarding the in the prompt.

Any tips?

[jovial widget]

Something has changed because I downloaded a track when it sounded great now if I compare the same song in my library it sounds awful compared to the downloaded version. They definitely have done something.

[unique stone]

Might just affect the playback of the song in the browser. If you download the same file again it should sound the same as the previously downloaded version.

[half dragon]

Like this?
I put in the style: "television episode, short tune, bright and fun" then click the magic wan for a prompt:
https://suno.com/s/tPdgBpBV2HWj6qUk
https://suno.com/s/JfxvuFTXX2WeaxoO

[jovial widget]

i havent tried that, but what have they done to the browser? Song were sounding great, now it sound so low quality.

[sleek stump]

You can switch over to the Legacy editor here:

[unique stone]

It may be using different kind of compression for browser playback. But it still sounds the same to me. There was someone a week or two ago who commented that they found it sounded better in the browser which was kinda odd.

[echo prawn]

Why does this damn AI always introduce some messed up harmonies into the songs it generates?

😅

The 4.5 model is worse than the 4.0 model

[sleek stump]

Its also still a beta model.

Use the feedback page, tell them what you dont like

[polar spade]

You said "..... it worked correctly before the update...."

This last update is Edit UI 3. There was a short period to use Edit UI 2, but not much of a change.

".....I wouldn't know how to use it [Legacy]....."

The Legacy version which was before, maybe the easier one to use. At least while you are waiting for any fixes you think should happen, you might wish to try the older Legacy version. Both are available at the same time. Its hard to say something doesnt work if we dont try it, first.

[weary ruin]

suno is telling me I have no credits and I do?

It says I have to upgrade but I have the pro plan