📚┃suno-school

[wispy totem]

why are people reacting to my song

its broken

did anyone even listen

the end is the worst

[naive quail]

Probably don't want to share that considering the sound clip you uploaded...

[wispy totem]

what sound clip

[random prawn]

https://docs.google.com/document/d/1iLb0xPZGL16Uf6AYVrmtx6h-7CnChGMS8nowaD5OgI0/edit?tab=t.0#heading=h.af2jfqw1ahdn read here

[naive quail]

the one titled "toot"

[wispy totem]

i know it was the basde

base

but it just broke

you know what i give up no one cares and yeah

[radiant egret]

Oh, great

[wispy totem]

it was just suppost to be a showcase of how badly it failed

[random prawn]

https://suno.com/song/10887f64-a039-42c2-848a-2897cd44ae7a you can hear the shimmer really clearly on this song.

[woeful yoke]

Well I can see the disaster happening in real time. I can already make predictions about future versions that are trained on outputs from the current one unless they fix at least two things. They need to fix the older pipeline and then there's another fix that can be applied to the output from current models to improve the sound such that those improved songs can be done successfully used for fine tuning and training such that all the crazy damage that has been piling up will start smoothly reversing the damage that has already occurred between the start of this year and today. 🙂

[random prawn]

https://suno.com/song/20aaf3d2-52f2-4975-a519-a8f004b05897 this is slightly better, clearer but towards the end of the song it all breaks down

[woeful yoke]

We know. Don't worry, V4 remastering your is going to get noisy soon too. Because of... Well what's the point. 🙁

[random prawn]

Are you saying that there is no hope becuase of how they have been training their model @woeful yoke ?

Listen to the second one, its much better, but its still there back in the background.

[charred merlin]

Harly -- now that the head of Suno support has personally reached out to you, can you confirm that you're sharing your technical expertise with them?

[torn saddle]

that's awesome. I haven't had anything sound like this. ever. FINALLY I know what this shimmer thing is. I just thought they were referring to SUNO BUZZ™ -

[random prawn]

yeah its a very clear distortion

[random prawn]

https://suno.com/song/20aaf3d2-52f2-4975-a519-a8f004b05897 this is slightly better, clearer but towards the end of the song it all breaks down

[graceful condor]

Just listen to shimmercore

[spiral goblet]

Glad to see people still talking about this.

[random prawn]

https://docs.google.com/document/d/1iLb0xPZGL16Uf6AYVrmtx6h-7CnChGMS8nowaD5OgI0/edit?tab=t.0#heading=h.af2jfqw1ahdn

This is what I am most concerned about with Suno.

I just remastered an old track of mine. https://suno.com/song/144fb2ab-3ada-4756-8415-ed48283dc9dd (much less shimmer)

[woeful yoke]

I'm not going to comment on it. But it's been fun seeing absolutely great AI engineers so hard trying to work around problems that can be fixed with audio engineering 😏
Nobody expects all of us to be experienced in both audio engineering and machine learning. But you will be running into huge issues when you are not even considering the possibility that, when dealing with the audio/music domain, maybe you need a little bit better competency than just having an experience of being a musician and/or a self-proclaimed audiophile, no amount of machine learning experience and expertise is going to compensate for that shortcoming.

[random prawn]

this makes a lot of sense. There is artistry in music. And you are saying they have forgotten about this. Music isnt just 1s and 0s.

[woeful yoke]

Well the models definitely acquired some artistry, no argument there. But having artistry doesn't guarantee great sound. There's a reason why some of us are specialized in order engineering or mixing and mastering.

[random prawn]

When I say artistry, I dont mean creativity. I mean the science and art of the sound itself.

[woeful yoke]

Okay fair enough.

[lavish oasis]

Expertise is a mix of intuition and experience, aka Pattern Recognition and Biased documentation.

I'd caution anyone from thinking that either can't be isolated. 😅

[woeful yoke]

Absolutely true.

However, sometimes people suffer from the dunning-kruger effect too. That is quite often the reason why people overestimate their expertise. Which leads to them not reaching out to other people who could assist and mentor them 🙂

[random prawn]

this x 1000

[lavish oasis]

📠, in my own industry it's not "who to ask" but "what to ask". 😅

[sterile kindle]

Hey I already had a issues before with my payment because I paid via Playstore and didn't receive Premium, now I created a new Account and tried to pay directly via the Website and now it tells me that my credit card is not accepted. Why is that? I just want to pay for premium but it don't let me

[woeful yoke]

Anyways, I do acknowledge that v4 has greatly improved in fidelity. The bad thing is that it's not going to hold. Because if they keep on fine-tuning and training with additional outputs from current V4 then there's going to be diminishing returns up to the point where things will be going downhill. I can guarantee that because I can already see it happening with v4 in some aspects. Most people don't even realize that yet.

[random prawn]

well, that is exactly what model collapse is. Training a model on outputs of the model itself. How do you know they have been doing this?

[lavish oasis]

There are a quite a few new transformer based architectures that perform better with continuous data than classic. No idea if they've changed their base model 😅

[woeful yoke]

The current V4 is kind of okay for generating new songs but... Only first generations from that model could be used for fine-tuning or training. Covers, remastered tracks should not be used for fine tuning or training. Although I have my reservations about using any of the outputs for training and fine-tuning without carefully pre-processing the whole data-set before training runs.

A few months back I would have had no idea what's going on but after I put in some effort and money into research and testing I was blown away by the audio quality that can be achieved + retained even through-out multiple train runs, Even if I started to partially include my own generated audio in the training data. Because if you do it the right way, it's not going to degrade your model of time.

[lavish oasis]

It kind of depends on how they track data. Using synthetic or human screened input for expanded training is fine as long as the underlying distributing doesn't get concentrated. In our industry we use fingerprinting and clustering to determine if new data leads to "fixation" or not.

[random prawn]

yes, junk in, junk out.

quality in, quality out...

Well all this said, its still an incredible technology. I just hope the model doesnt collapose.

[woeful yoke]

On the other hand, most ml engineers know that usually garbage in least two garbage in addition to knowing that If you need to do a lot of pre-processing for your model to work then you obviously have an issue with your model that needs to be resolved too.

Damn you said almost the same thing

😏😆

[woeful yoke]

Well we are not talking about motor collapse right now but rather training data collapse

That's the biggest issue right now

[random prawn]

Isnt that the same thing?

[weak patrol]

Hi, I'm very happy with the songs I've created. I'm not a sound engineer and asked one who is and his comment about my songs is the following, which I agree with: The vocals on the recording are well recorded for the genre but they have a lot of processing on them and the effects create artifacts which blur the clarity of the words. Besides that, the vocalist is really leaning into the anger-angst attitude intensity - which is of course okay, but his teeth are so tightly clenched it mashes up the diction quite a bit- it’s kind of the nature of the beast – for the lyrical diction to take a back seat to the emotional communication.
If the song were to be remixed, I would suggest frequency selective ducking* some of the keyboard tracks using a side chain of the vocal track signal during the loud, musically dense parts of the song. Maybe back off some of the vocal effects a bit during those parts too.

Is there a way for me to improve what he's talking about on suno?

[woeful yoke]

No not really. You see you could always create a new model and start training from scratch. But the problem is that if you over time have progressively degraded your training data then all you're going to do is get your bad model back again

[random prawn]

and that could cost $$$$ millions

[woeful yoke]

The most valuable asset right now should be quality training data, yet with each generation we can see them degrading it.

[weak patrol]

I understand but I didn't create a model at all. I just used the service. So, being a newbie it sounds like I can train a model on suno

[woeful yoke]

Actually in some sense it should be not that expensive. If you start small and then at one point you train a new Foundation model. Because you know you can implement your upgrades. And changes and improvements on top of the foundation model using just LORAs

The problem is that if you screw up your base model, you're going to have a lot of issues with fine tuning and whatever layers should put on top of that

Right now I'm fairly convinced that actually version 4 is the same model that's running version 3.5. the only difference is that the output from version 4 gets fed through another smaller and faster so-called refinement model. This way you can host two different model versions on the same instances. I'm fairly certain that this is the case because that's the only thing that makes sense when it comes to analyzing what happened to 3.5 after initial V4 issues.

[acoustic vapor]

I have some sort of technical questions, really hoping to reach someone direct from Suno for clarity.

I am a gaming streamer, I am generating music on Suno through v4 that I post on Youtube (spotify soon). My stream loves the music and so I have shared Suno with them and many of them also create their own songs that we have come to know and love, but most don't use the paid version of suno.

Is there a way that they can share their songs with me that they make so that I can re-master in V4 and clean up etc?

[woeful yoke]

Besides, I was able to kind of confirm identical underlying models by extending a truck from 1 to the other and back and then back again

When they're extending a track, that specific refinement isnt applied on v4

[woeful yoke]

Yes. If a song is made public (Even if temporarily) you can extend their song under your account. Then you will get the whole song after which you can run it through remastering

I've done that with a couple of songs from other people. That's the only way to move a song from one account to the other because it's not possible to do so through uploading.

[high quarry]

how can you clean up in v4?

[woeful yoke]

Remaster

[high quarry]

tbh, atm v 3.5 cover is the best in quality..

[woeful yoke]

But you should be warned that even though it appears to be getting cleaned up, it also degrades a lot which isn't apparent to everyone

[gritty badger]

@woeful yoke

#📚┃suno-school message

[woeful yoke]

V4 could instantly start making better colors and better remasters if they fixed the audio pipeline between models at the model input stage into the neural codec/tokenizer. They just don't do it

[high quarry]

thats true 🙂

but for the moment i get best results in v. 3.5 and then do cover

especially live covers

[woeful yoke]

We are in the situation right now where remastering absolutely ruins most songs that will generated with 3.5 or 3.0

And by ruining I don't mean problems with it. Changing songs slightly or problems with singing or melody or voice. I'm talking about literal outer quality getting ruined

And the problem can be easily solved without retraining version 4

[quiet light]

Hello, any one knows how to Buy More Credits?

[woeful yoke]

In fact, if they kept going on trying to solve those problems by training then it's going to fail or it will seriously degrade its ability to faithfully reproduce your original song with cleaned-up audio

[woeful yoke]

Under your account

[quiet light]

I tried it doesn't work

[unique stone]

The suggestion that v3.5 is different since v4 released has been made a few times. I asked a dev about this recently to make sure I can give an accurate response: they confirmed that it hasn't been changed. (But people still don't believe me even though I asked a dev about it.)

[woeful yoke]

Under subscriptions

[quiet light]

yes, it just kept on rolling nothing

i sent an email i got no response. I've been trying for the pass 48 hours

[woeful yoke]

Well, to be honest I didn't notice much either but I was going off based on many examples I heard. But the problem is that the music styles that I usually generate don't suffer from that much sound degradation than some other styles. So maybe from that I got the wrong impression that it has gotten worse when in fact it might have been that bad for them for a long time already

Obviously people started paying attention to audio quality much more after they were given the taste of what could in principle be achieved.

[gritty badger]

@woeful yoke HARLYYY 😲😂

#📚┃suno-school message

We are all waiting for your response 🥲

[woeful yoke]

Wait what response?

[coral epoch]

Actually he said there is hope - but the longer it goes unfixed the worse it'll get and the harder it will be to fix it

[woeful yoke]

Did I miss something?? 🤔

[gritty badger]

Quite a few times 😅

[woeful yoke]

I'm sorry

Maybe I should take a break from yapping here

And then read what I missed

[rich timber]

Anyone else having their lyrics and instrumentals dragging and distorted?

[woeful yoke]

It's just the two days ago my ex moved out after 10 plus years of relationship and living together and right now I'm kind of trying to keep my mind off of everything by just keeping my brain busy in here

[gritty badger]

#1315841002577199144 message

[rich timber]

I wish you well.

[gritty badger]

@steady kelp is waiting for your DM to inform the devs

[quiet light]

I really enjoy this software ,but the support is just terrible. i got an email to come here for help

[hollow hare]

Yes because waiting for a response by email right now will take like 7 business days due to really high volume.
So, uh, what do you need help with?

[woeful yoke]

Oh that one... I thought I kind of read through everything already yesterday.
But I'm not going to respond over there. Everything that I wanted to convey to the team I already posted today as a separate topic under v4-feedback forums. For some reason it got automatically deleted by some kind of a bot. Unfortunately, I saw the error message only briefly and it disappeared before I was able to finish reading it. So I don't even know the reason why it's got deleted.

[quiet light]

I tried under Manage Subscription to buy more credits nothing is happening

[unique stone]

I grabbed a screenshot of the deleted post earlier and put it in the thread https://discord.com/channels/1069381916492562582/1318532031545610290

The bot autodeleted it due to there being 3 blank lines in a row (it also does this if the same character is repeated a lot, too many emojis, etc.)

[gritty badger]

You don't need to respond on the feedback forums. Just gather all the info you have gained from your research and send it to Mike B as a DM so that the devs have a picture of what you are on to

[quiet light]

when i clicked buy more credit I chose 8 for 1000 it just rolling and rolling page nothing happening

[gritty badger]

#1318532031545610290 message

[quiet light]

wow!!!

[hollow hare]

Did you try using another browser or incognito mode in case one of your extensions is interfering?

[quiet light]

no... im on chrome

[hollow hare]

Try giving it a shot to see if that may solve the issue since it sounds like a connection problem. If you are on VPN, that may affect it as well (sometimes I would have to change servers for it to get working again)

[polar spade]

Hi again. I should maybe start with a basic question: Do you know how to Direct Message someone using Discord?

If you need help, we can describe the steps.

Then you may search for @steady kelp
and message him for help.

Mike, since this is a late comment, has @rustic lance Dm you yet? And sorry to step on any toes. I'm just trying to fill in the gaps.

[prime prairie]

@graceful condor Thank you for your help! I was able to get logged in without having to turn my VPN off!

[graceful condor]

Happy to help, what fixed it? Enjoy creating your music and don't forget to share it with us in #🐣┃suno-showcase, #1154790975240155247 and #💬┃general-chat!

[steady kelp]

I'm mostly sure this is a game at this point. It's actually just a bunch of misinformation, spreading speculation because we haven't explicity told people how we built v4. I've offered to be helpful, but based on the number of comments about our dev team, and going as far as to assume none of us are audio engineers (Hi, even the support guy has an audio degree) I'm not sure they actually want to be helpful. I'd love to be wrong, though!

[graceful condor]

If you don't mind telling, what fixed it?

[rich timber]

Hi, I’m having issues with tempo and vocal distortion on Suno.

Problem: The tempo is dragging and not staying tight at 95 BPM. The vocals also sound distorted, especially in the ad-libs and falsetto sections.
Setup: I’m using Cover Mode with simple production notes for Funk, Soul, and R&B.
Question: Is there a known issue with tempo control or genre compatibility (like Funk or R&B)? Any tips for getting the tempo locked to 95 BPM and reducing distortion?

Here’s the prompt I’m using:

<SONG_DETAILS>
GENRES: Funk, Soul, R&B
TEMPO: LOCKED TO EXACTLY 95 BPM. DO NOT DRAG.
SOUNDS LIKE: Prince, Bruno Mars, James Brown
STYLE: Catchy, Playful, Nostalgic
MOOD: Energetic, Funky
VOCALS: Male Lead, Falsetto in Hook, Smooth in Verses, Harmonized Hooks
INSTRUMENTS: Funky Bassline, Wah-Wah Guitar, Tight Drums, Punchy Brass
PRODUCTION: Instruments and vocals must stay tightly locked to 95 BPM. Ensure clean, clear sound with no distortion or drag. No added reverb on Ad-Libs or Hooks.
STRUCTURE: Intro, Verse 1, Hook, Verse 2, Pre-Chorus, Hook, Bridge, Hook, Outro
</SONG_DETAILS>

[Intro]
Funky bassline starts with claps. Wah-wah guitar and light brass stabs join.
Ad-Lib: Funky, funky! Let’s groove! (Soft, clean vocal delivery. NO distortion.)

[Verse 1]
Lead vocals only (no harmonies).
Smooth, rhythmic vocal delivery.
Keep timing locked to 95 BPM.

[Hook]
Falsetto lead with 3-Part Harmonies (Falsetto, Mid, Low).
Ensure all vocals are locked to 95 BPM.
Harmonies should be smooth and tight.

[Verse 2]
Lead vocals only, no harmonies.
Smooth delivery with no reverb.
Keep timing synced with instrumental.

[Pre-Chorus]
Build energy with rising vocals and brass accents.
No falsetto or harmonies.
Lead vocals only.

[Hook]
Falsetto lead with 3-Part Harmonies (Falsetto, Mid, Low).
Ensure all vocals are locked to 95 BPM.
Harmonies should be smooth and tight.

[Bridge]
Breakdown section with snaps, light piano, and guitar solo.
Lead vocals with subtle backing vocals on "One talk, just one talk."
Build intensity gradually toward the next Hook.

[Hook - Final]
Falsetto lead with 3-Part Harmonies (Falsetto, Mid, Low).
Ensure all vocals are locked to 95 BPM.
Harmonies should be smooth and tight.

[Outro]
Stripped-down guitar with soft ambient pad.
FADE OUT WITHIN 20 SECONDS. DO NOT EXTEND BEYOND THIS POINT.
NO EXTRA VOCALS OR AD-LIBS AFTER THE LAST LINE.
STOP ALL SOUNDS CLEANLY AT 20 SECONDS.

[prime prairie]

I thought I had tried it before, but I guess not. I had to switch my default browser to Firefox, sign in. Then switched it back.

[gritty badger]

Some of us did our best to raise the awareness and did follow ups etc. You are aware too. From this point, the ball is in Harly's hands. If he's willing to share, happy days. I experienced my song being fixed by him first in hand and was thoroughly impressed. I can only hope for all the info to find the right place and everyone becomes a winner for finding the solution. Nothing more 🤷 I'm done

[steady kelp]

i have no doubt of his skill, in fact i was in the early group to say if you have the skills, why not apply for a job! but at the same time, happy to hear out anything that can make better music for more people. thats the goal!

[quiet light]

I've tried incognito nothing works. It's frustrating. I need to buy there is no one to pay...lol

[rich timber]

Hi Support, I’m having trouble with tempo and distortion on Suno. Any known issues right now?

[hollow hare]

Oh, did you initially subscribe through the browser or through the app?

[quiet light]

Yes I did. everything went fine.

I checked my Paypal I saw Suno took their money .

[hollow hare]

Suno does not know what BPM is. It does know tempo are using more vague terms like fast, slow, [genre]core, etc. It should know musical terminology as well but it is something I didn't try since 3.0

[neat nexus]

And this works? It doesn't sing the directions? It sings your lyrics?

[finite crown]

When I try to sign in with Google it takes me to a support page

[hollow hare]

1. Ensure your app has the latest update
2. Getting an error 404 or issues with redirecting when signing in (mainly via Google)? Change your default browser.

I had that issue with Opera and switching to Firefox fixed it and I heard the initial Browser works as well. To change the default, go to Settings > Apps > Change Default Settings (name and location may change base on your device/firmware). After signing in, you can change your default browser back.

[quiet light]

I've tried many browsers I even updated my phone #

[rich timber]

Yes, I used this prompting several times and it works. I'm meta-tagging for production of an album. 5 songs are clear until, this song "dragging and distorted"

[hollow hare]

Sorry but I cannot think of any other solutions. Perhaps searching this channel will yield something new?

[quiet light]

I ran out of credit i just want to buy more credits .... I'll open a second account atleast thats what other people are telling to do and said they too had same issue had to open a second account

[neat nexus]

Huh.

Wow.

[neat nexus]

Have you experimented with time signature in the style prompt to see if that gets you closer?

[quiet light]

Thank you . I appreciated you help.

[rich timber]

I will test that out. Thank you 😊

[neat nexus]

metronome 4/4 seems to work most of the time for my lofi when I checked it in Banddlab a while back.

[acoustic vapor]

when I tried this with my buddy I don't see how I can use the whole song still, only the "part 2" extension

[hollow hare]

The option for Get Whole Song appears in the Part 2+ menu after it has finished generating. Though, if I recall, certain locations where the menu is called from may lack it (e.g. if you are calling the context menu from the player?)

https://help.suno.com/en/articles/2409601

[acoustic vapor]

I am doing it on desktop, this look s like mobile. will try there.

[hollow hare]

Should be under Create for Desktop (this is in Desktop Mode)

[hollow hare]

Interesting, where are you accessing that menu from, the lyrics page? Mine was through the library

[acoustic vapor]

I went to my friends profile. Opened the song he has made public. Click the three dots and that is all i see.

When I do "extend" it doesn't give me any options to do anything but extend, no "get whole song". Does my friend maybe need to do something else on his account to make it "more public"? not really sure what to do here

[hollow hare]

Oh, that explains a lot. The option appears in the creator's account

[acoustic vapor]

So he downloaded the whole song and sent to me. But I can't re-upload the whole thing to re-master it in v4, only first 2 mins of it

trying to work around that

[hollow hare]

To reiterate:

1. You extend the song
2. Part 2 is created
3. The menu from the generated Part 2 lacks Get Whole Song?

[acoustic vapor]

Nope. You nailed it. I got it. +100 points

On the part 2 itself, go to three dots, then get whole song appears. I was missing a step. Thank you so much

[hollow hare]

You're welcome, great to hear we both figured it out together 🙂

[lunar anvil]

Trying out ReMi and typing "Creating a short story about robots." Causes ReMi to not generate anything but if you correct Creating with Create then it generates. Guess it needs proper grammar to work.

[hollow hare]

One of my ReMi lyrics had turned Einstein into a rock, went to invade a planet, blew up a celestial object (unsure if it was a planet or moon) filled with another race, then colonized the remaining planet. Yeah...

[lunar anvil]

I wonder if ReMi knows it is named ReMi

ReMi didn't like me asking what its name was. My name does not need to be your focus; let's dive straight into creating some memorable lyrics! Give me a prompt And I'll show you what I can do!

[graceful condor]

Got it, thank you very much

[wary wharf]

How can I tell Suno to repeat a previous melody far back on the song intro?

[charred merlin]

...is the trick to ReMi that you need to make an offering to it?

OMG is ReMi built on TempleOS?!

[balmy prism]

Someone help me pls i have this song, and allways just female singing. From 8 generate.
Once suno generated Female voice where is Male, and singing Male, where i tell Female.

Style: Blues Rock, Female, Male

Lyrics:

[Intro: Male Voice]
(Instrumental blues-rock riff)
Mindjárt kezdődik a nagy meccs.
Ezt minden képen látnom kell.
Nem is akárhogyan fogom nézni.

[Pre-Chorus: Male Voice, Building Tension]
Erre most én felkészültem.
Egy sörös hűtő a fotel mellett.
Haverokkal együtt nézem.

[Chorus: Male Voice, Emotional and Rising Strength]
Indul a meccs a sör már a kezemben.
Sörrel a kezemben nézem a meccset!
Ez most minden, ami kell.

[Bridge: Instrumental Transition]

[Verse: Female Voice, Soft but Playful]
Nem mondhatod komolyan, drága szívem.
A csajokat kihagytad az egészből.
De semmi baj, legalább én előre készülök.

[Pre-Chorus: Female Voice, Confident]
A csajokkal is átjöttünk, legyen jobb az este.
Így lesz teljes az esti játék.
Dobj ide is sört, kis szívem.

[Chorus: Female Voice, Catchy and Powerful]
Most lesz jó a meccs, nők és férfiak együtt!
Mindenki itt van, nálunk is van sör.
Sörrel a kezünkben nézzük a meccset!

[Bridge: Instrumental Transition with Female Ad-Libs]

[Final Chorus: Duet - Male and Female Voices in Harmony]
Most lesz jó a meccs, férfiak és asszonyok!
Mindenki itt van, mindenkinél ott a sör.
Sörrel a kezünkben nézzük a meccset!

[Outro: Male and Female Voices Together]
(Energetic ad-libs and humming, fading out)

https://suno.com/song/e900b3fc-fc55-40c3-a308-987e2dffc029

I dont know why?

[graceful condor]

Make it [Male Singer] and [Female Singer], add duet in styles, as well as change female and male to male singer and female singer

[charred merlin]

I'm elated that reference landed, btw -- you just passed an exceptional amount of vibe checks.

[naive quail]

lol. I was literally just trying that. hahah

[balmy prism]

And where i want female and male there [Male and Female Singer]?

[graceful condor]

You could try that, as well as [Female Singer] and then [Male Singer] after

Tags above 3 words might not work that well

[charred merlin]

(based on v3x observations primarily) It gets increasingly trickier to recall earlier sections of the song -- using "get whole song" to ensure that you're using as much of the song as possible for context can help, but for longer songs there's a chance that it's out of context.

That said, [motif variation] has worked in the past to recall the song's hook with some fun tweaks, and I suppose you might be able to simply prompt for [intro], but it's less of a "command" and more of a suggestion to Suno. 🙏

[graceful condor]

Have you conversation, gentlemen

[naive quail]

...and if the melody that you want to recall is more than 2 mins earlier, your chances of recalling it fall down to ~ 0%

[glacial scarab]

well technically... with an unlimited supply of credits and time...

[balmy prism]

So this will work?:

Style: Blues Rock duet

[Pre-Chorus: Male Voice, Building Tension]
Erre most én felkészültem.
Egy sörös hűtő a fotel mellett.
Haverokkal együtt nézem.

[Chorus: Emotional and Rising Strength]
Indul a meccs a sör már a kezemben.
Sörrel a kezemben nézem a meccset!
Ez most minden, ami kell.

[Bridge: Instrumental Transition]

[Female Singer]
[Verse: Playful]
Nem mondhatod komolyan, drága szívem.
A csajokat kihagytad az egészből.
De semmi baj, legalább én előre készülök.

[Pre-Chorus: Confident]
A csajokkal is átjöttünk, legyen jobb az este.
Így lesz teljes az esti játék.
Dobj ide is sört, kis szívem.

[Female and Male Singer]
[Chorus: Catchy and Powerful]
Most lesz jó a meccs, nők és férfiak együtt!
Mindenki itt van, nálunk is van sör.
Sörrel a kezünkben nézzük a meccset!

[Bridge: Instrumental Transition with Female Ad-Libs]

[Final Chorus: Duet - Male and Female Voices in Harmony]
Most lesz jó a meccs, férfiak és asszonyok!
Mindenki itt van, mindenkinél ott a sör.
Sörrel a kezünkben nézzük a meccset!

[Outro: Male and Female Voices Together]
(Energetic ad-libs and humming, fading out)

[naive quail]

I'd remove at least "male" from the style prompt (experiment with removing both) and try to have the first vocals be female if at all possible, THEN switch to male and use the male and female tags within the lyrics. I dunno why Suno behaves better when starting with the female vocals, but that's been the case for me the majority of times.

[balmy prism]

Oh so i need add in the start

[Female]
[Male]

To start male?

[naive quail]

No. If you want to start male, use the male tags. I've just found that Suno does duets better when I start with the female singer.

But there are no hard rules when it comes to duets... and a bunch of different (tho similar) ways to get them.

This is the last one that I did: https://suno.com/song/f44166cd-bf77-4d0d-8462-b2bb3338b27d

[balmy prism]

I will try it tomorrow thank you!

[charred merlin]

@cedar thistle for the callback at the tail end of Kyoto Girls, did you add that in post, or did you recall that section with prompting?

Just noticed the intro callback, and the fact that someone here had just asked about that clicked

[deft nymph]

Am I using these tags correctly?

[Intro, Driving Guitar and Pads]
[Electric guitar riff with rhythmic picking, layered with soft pads and a steady kick drum to set a bold, worshipful tone]

[Verse 1, Bold and Rhythmic]
Lord, how many stand against me,
They say there’s no hope in sight.
But You are the shield around me,
You lift my head, You are my light.

[honest dew]

reminds me of a song where a sentient peice of plasic blew up all the universes or something

[hallow crag]

Can someone please help me? I'm having trouble to change gender between male singer and female singer.
How do you change that?

[naive quail]

scroll up

[balmy prism]

https://suno.com/song/e4048f66-e501-4d02-b058-12d9066630c2

[naive quail]

It's unlikely that Suno understands natural language prompts so this:

[Electric guitar riff with rhythmic picking, layered with soft pads and a steady kick drum to set a bold, worshipful tone]
...will probably result in one of these things happening:

• Suno will ignore much of the text and just cherry pick the tokens that it recognizes.
• Suno will cherry pick the tokens that it recognizes and then get triggered by something in the extra text and possibly get noisy or start to hallucinate.
• Suno will proudly belt out the tag as lyrics.

[hallow crag]

Yeah i have tried to scroll up and didn't understand sorry my english is not good.
Could you please give me an example?

a little example so i can change and edit.

[balmy prism]

Here too the Male and female voices are swapped.

[hallow crag]

Yes it's swapped that's the problem 😦

[balmy prism]

https://suno.com/song/2734a402-b0f3-4f42-b785-d63dae29e88f

This sorry

[naive quail]

start here and read down. there IS an example. #📚┃suno-school message

[balmy prism]

The song perfect if the male singing, where i tell male and female singing where i tell female 😦

[autumn bane]

Anyone know how we can better control transactions between make and female vocals im makeing a doet and it keeps messing it up I've tryed about 50 times now lol

[naive quail]

examines his bong

Thing's not working!

Duets are tricky. Suno loves to swap voices. They often take a TON of generations and extensions to get right.

https://suno.com/song/547b778f-4662-435a-9eb1-1f8aed34888c

[honest dew]

well there was this one time wjere i made a horribly vulgar song being sung by a child but every time it was just being sung by a woman

so you could try that

[balmy prism]

Someone told If i start female work better, so this will work better?:

Style: Blues Rock duet

[Pre-Intro: Female Voice, Soft]
Mi a terved estére kedvesem?

[Intro: Male Voice]
(Instrumental blues-rock riff)
Mindjárt kezdődik a nagy meccs.
Ezt minden képen látnom kell.
Nem is akárhogyan fogom nézni.

[naive quail]

That was me... and you're just going to have to experiment. Duets can be very challenging.

[autumn bane]

I'll give that ago I got 400 song gens so ima get this right

[balmy prism]

I have possibile in suno to send the song and swap the male, and female voices in my song?

[lunar anvil]

It is a slot machine 🎰

[final elbow]

Hi, I'm new to this. Is it possible to have a tag that changes the beat from one genre to another? For example, from trap to reggaeton during the track?

[autumn bane]

Kinda try dynamic trap-reggaeton fusion

[radiant cobalt]

I just have to say this 4.0 suno shimmer is kind of ridiculous in how many of my generations it's been ruining.

[twilit ice]

I'm getting dusty, saturated, distorted vocals, mostly on the second half of songs or extensions. V4 feels the need to widen vocals as the song progresses, adding more verb and distortion, ruining clarity. I feel like day 1 of V4 was much clearer and they reverted to less quality? It's so bad. Like waste of money bad.

[twilit ice]

I don't know what shimmer is (I see many people using the term) but if that's when it sounds like your song is thrown into a tunnel and processed through a distortion pedal, that's been ruining my tracks. I have a track i love but it just won't extend without this garbage. That the same thing?

[radiant cobalt]

Yes it is.

If they left 3.5 alone it would have been a forever tool.

[twilit ice]

Man, I'm almost out of my 10k credits trying all kinds of stuff. Its just getting worse. I'm like a crazy person thinking I can find a solution. Sigh... good to know, thanks.

[unique stone]

They didn't change 3.5. I asked one of the devs recently.

[unique stone]

"Shimmer" is an intermittent, swishy rattling noise that gets triggered by certain sounds (cymbal crashes or rolls are a common culprit but I've had it on other things too.) It comes and goes throughout the entire song, but significant improvements have been made to reduce this since v4 beta launched.

Gradual song degradation is likely something that'll be addressed, just downvote any clips that degrade/distort so the devs know that it's bad so they can continue improving the model.

[twilit ice]

Thanks. Yes, I started reporting them as bugs today. By the end of the track there are explosions happening on some of them. I guess I'll have to make covers of these and start over. So many words wrong as well on every remaster and cover. IDK.

[unique stone]

This used to happen all the time with v3 (I have so many broken sounding songs from that) - if you extend from a time before degradation happens it may help to avoid it and allow you to keep using a song you already liked.

[twilit ice]

Ah, good tip. Yes, it's a part 2 extension that goes wrong at 1 minute. I did not try extending at that particular point. I'm trying so many weird things. Fun but frustrating since I didn't get anywhere. Appreciated. I'm going to need those 4k permanent credits tonight. 😉

[unique stone]

This is how I used to live back in the olden days of v3 😄 but I used to get really nice segments starting inside the distortion so foolishly I ploughed on and suffered further.

Cover allowed me to rescue those songs at least, but for longer songs it becomes a case of breaking it up and editing it with other tools.

[twilit ice]

Haha. Yep, I'm in Ableton live, trying stems, Trackspacer, Dreamtonics synth V remake, and all sorts of tricks. Then I jump back to Suno and try something else. I thought Suno was going to be shortcut for my production. Not in these cases. More 8 hour track sessions. 😫

[little granite]

did they fix the shimmer and high frequency stuffs

[twilit ice]

My ridiculous credit usage regenerating tracks tells me no. 😉

[autumn vault]

So still same old V4? No fix yet?

[autumn bane]

Tell me about it I've gened over 200 songs trying to get 1 good

[autumn vault]

it seems V3.5 also got worst today , the audio decay just after 1 minute...

[storm storm]

@pearl sorrel

If I could just get rid off that modern trap beat or whatever that genre is called, its really close

[devout flower]

Has anyone noticed with v4 that each successive extension of the audio, it sounds worse and more robotic?

[storm storm]

yes

also if you listen to music from their platform with bad connection it becomes negative monotone like darker sound

distorted basically

[primal star]

Would anyone have any suggestions to push the finale of this song in the direction I'm looking for? Right around the 3:46 mark I'm trying to get the song to explode into a climax where it's utilizing the motif which starts around the 00:28 mark. Problem is that the song keeps trying to force itself to end or just go too quiet. And then when the sound does go in the direction I'm looking for it keeps referencing the wrong part of the song. I'd assume it's mostly due to limitations but with my current skill level I'm not sure
https://suno.com/song/b10e3ca5-23ef-47ac-bc0c-66bfd57969d7

[devout flower]

Any remedy?

[primal star]

Not just that but from my experience it seems to get harder to control the way you want it to go the further out it goes

[storm storm]

you could try adding to exclude styles hissing,sharp etc..

[devout flower]

I tried that just now. Doesn't help that much

The audio still sounds crappy

It's a country song, but it's sounding like an EDM song 😢

[cold flax]

Hey Peeps! Is there anyway to enable this?

[storm storm]

try resetting to just empty a prompt

[storm storm]

thinks its v4 only or technical issues

[primal star]

Ok I think I finally found a good output from where I'm at so I might be cooking now

[daring talon]

question. when i use a persona or reuse a prompt. the song length is always the same 3:27

why? the songs always just cut off now.

every song ive mader (the last 100 or so) are all 3:27

why ? theres up to 4min but just cut out. dont fade/end song just cut @ 3:27

[primal star]

That seems like a rather strange quirk with the system. Have you tried extending the song at a point you want it to start slowing down and using certain keywords?

[sage magnet]

quick question.. can I only give voice to create songs?

[daring talon]

no. like literally i make a new song now. use persona. it will always be 3:27 long

[sage magnet]

I just have voice of my own

no song as such.. just voice.. some messages

[primal star]

So if I'm understanding this right you're wanting Suno to purely generate dialogue without any instruments in the background?

[sage magnet]

no no.. I have voice for persona.. like only voice messages.. can I use that to build a style of singing?

[primal star]

Yeah that doesn't seem normal unless I'm mistaken on something. You might want to report that as a bug just in case

[daring talon]

and all extend songs i make with any persona are always 2:00

always

[primal star]

Ahh. I'd imagine so yeah since you're able to select whichever audio you generate or upload as a persona normally

[sage magnet]

Hmm.. I was hoping to test this before I buy sibscription

Or get confirmation.. who would do it?

[primal star]

Do you have any free credits left?

[sage magnet]

have

[primal star]

You should be able to experiment with that since the free credits refill daily. I can still take a look if you want

[sage magnet]

no buddy.. persona is only for pro amd premium

[primal star]

Oh shoot I forgot 💀

In my defense I'm very exhausted lol

I'd say go ahead and DM me the audio file you're wanting to test and I'll try to take a look

[sage magnet]

haha take a nap

[primal star]

Never

[sage magnet]

Who can confirm this starnomaly?

Can you point me to the person?

[primal star]

Well I have pro that's why I was saying you can send it to me so I can test it for you

[sage magnet]

can I send you girls voice? just a voice message?

can I dm ? this is so cool buddy.. you rock

[storm storm]

this is some nightmare fuel

[primal star]

Assuming you're comfortable with sharing something like that then yeah go right ahead

[primal star]

That actually sounds pretty dope

[storm storm]

Thanks 🗿

Music if you want to wind off

This music makes me sleep

[sage magnet]

boy,, some heavy shit

[sage magnet]

i DMED

[storm storm]

I call this Death Sentence, most hardcore genre to exist

[sage magnet]

hmm.. The age of music has changed

[primal star]

These don't sound the same but it sort of reminds me of some experimenting I did back during the V3 days. This one to be exact
https://suno.com/song/da1ebe45-0b8e-4caf-9683-e88804e27742

[storm storm]

Apparently the persona I made is just a monster ai, from Tactical Music, to Relaxing Groove to Outright Nightmare material

Even tarkov music lol

here's another Death Sentence song

[primal star]

Nice

[storm storm]

https://tenor.com/view/yususuk-gif-26370473

[prime prairie]

Anybody know how to extend and edit in the Android version of the app? Or do I need to go online to do it?

[prime prairie]

I was afraid of that. So I still have to go online in a browser.

Thank you! 👍🏼

[cold flax]

So no-one knows how to enable the audio upload Persona feature? At the moment you're able to just create the personas based off work you've personally created. I'm trying to create some music but using an uploaded track for inspiration in the creation process. Any help would be greatly appreciated 🙌🏾

[coral epoch]

That's not enabled to prevent impersonations - RIAA prevention stuff

[cold flax]

Yeah I'm not trying to impersonate i.e copy exactly. Just use as a base to riff off.....

The fact that it says disabled at this time must mean that it can be enabled if they choose. Just seems a shame as the maximum characters to describe what kind of track you're after is limited to 200 characters so doesn't really give you much room to play with.

Being able to just upload a track you like and use that as a base would hit the nail on the head. Then you could just use the style of music box to make the differentiations.

[vivid mulch]

Hello, Not sure if this is the channel for this question, but when creating non English song (in my case in Arabic lyrics), then exporting a video, the text do not show/display (correctly) on the video, it shows as boxes, my assumption the font is not recognised, or supported. Anyone know how I can fix this on suno? Or Can this be fixed on video editor app?! Any advice is greatly appreciated. cheers …

[hollow hare]

Likely encoding is not supported via the generation method Suno does.
Making a Video manually will give you the flexibility to select the text encoding.

Use your favorite search engine to learn about what Video Editing tools are available - there are many out there and a few for free (can even do it on your phone! Though PC is much better).

The simplest method would be to slap a static image then put your lyrics in the video (or you can upload to YouTube and make a Close Caption with the lyrics). Or you can look in to AI video generations

[spare crest]

I can't download songs as WAV files. I am on the Pro Plan and haven't been able to download for 2 days now. I have cleared the cache, but it didn't help. I also tried using a different browser, but the problem persists. Is anyone else experiencing the same issue?

[topaz stratus]

Why I'm receive watermark in song?I'm using official ap from link suno.com/android, before I'm making from suno.com, I'm upgrade my akk to pro, and have lot of credits, but round by round "this is the fake app" Wtf? Two of my email to support, don't have any feedback.

[vivid mulch]

Thanks, I was hoping I can fix this on Suno, as I would be making lots of custom songs with custom lyrics, and Suno video format works best for me as my needs/requirement are basic.
I did look it up online, you are correct there are many options for making a lyrics video, but the text encoding and timing on final video is tedious work for long lyrics and too many songs, and very little fast options for custom works. I’ll keep exploring.
Thanks for the help, greatly appreciated. I hope Suno fix this soon, as it is a very minor fix, all is there, it just a matter of getting the text/font/encoding right. Cheers …

[pseudo island]

Thanks for your answer i have thing to the same thing the content right but the song i have try to upload is Toccata & Fugue from J.S Bach so i don't think i's a copiright problem because it copiryght free

[hollow hare]

Is it really in the public domain though? The audio recording may still retain copyright (though it depends when it was recorded)

• Under the Music Modernization Act, many of these recordings will begin entering into the public domain on January 1, 2022, when all recordings published prior to 1923
• Recordings published between 1923 and 1946 are then protected for 100 years
• Recordings published between 1947 and 1956 are protected for 110 years
• Disc labels that are more than 95 years old are now in the public domain and are free to use and reuse
  https://www.loc.gov/collections/national-jukebox/about-this-collection/rights-and-access/

Based on USA laws, may be different where you are but this is the minimum needs to be followed as the servers are in USA

[wooden spire]

Are remixes of songs made with a purchased subscription owned by me in terms of copyright if the remix is made from a song generated before the subscription purchase?

[naive quail]

You can make personas from uploads, but only if Suno hasn't detected vocals in the upload. When Suno detects vocals in uploads, it flags the upload so that it (and any derivative tracks) can't be made public and personas can't be made from it.

[hollow hare]

I assume by remix, you mean using your own generated music and not anyone else's creations or uploaded audio:

No but you can contact [e-mail] to see if they can grant you a license for it

This situation was addressed at https://help.suno.com/en/articles/2871105

See #1294423389515284552 for other related topics.

[hollow hare]

For clarificarion: the music itself has been in the public domain for a LONG time so creating the music yourself is fine, e.g. through FL Studio or recording the instrumental yourself.

However, the audio rendition has its own copyright belonging to the company or person that made that Output.

If you played/made your own, cool, but if it was from like the Philharmonic Orchestra, they have copyright for their audio rendition

[pseudo island]

Oh i see, copiright its more complicated than i think

I don't have the right to edit or use the song to add modification but i can listend and download it

[hollow hare]

Well, it actually would depend on the licensing about downloading, but yeah.

You could, for example, hum/whistle the tune (thus it is your own rendition) and Upload Audio it to see if Suno generates an appropriate music to that - that would be fine legally

[silk zodiac]

Hi I'd like to train suno bark on a custom dataset and add special tokens (similar to the music token) to generate emotion speech such as "<sad>It is cold outside</sad>". But I can't find any information on training bark. Can s.o. point me into the right direction?

[hollow hare]

Unfortunately most people do not use Bark here anymore so support is lacking.
You can try to reach out to see if @untold briar can help you but you may want to see if there are newer programs out there that may be more appropriate or well supported such as Tortoise TTS I think?

[silk zodiac]

Thank you, bark is still outstanding since it allows to use style tokens and I did not find an other model with such a feature. I'll try to reach him and see how far I get.

[primal barn]

Hey everyone, I seem to be having an issue with "Exclude Styles", everytime I try to use it, it generates two 2 seconds long tracks. I wanted to use "audio degradation, hiss" in exclude styles.

[naive quail]

Suno's not going to understand that in the same context that you do, though. Suno doesn't know that it's adding shimmer and/or distorting the music - these things are likely bug related.

...or maybe it does know and figures the best way to avoid those undesirable effects is to limit the music to 2 seconds.

[hollow hare]

Out of curiosity, did you try using metatags like [Sad] or [Melancholy]? Don't know if Bark understands that but doesn't hurt to try

Here is a REALLY old doc for some metatags which lists for Bark v0 (and Chirp v1 if the choice was enabled)
https://suno-ai.notion.site/Using-Metatags-34944efe09ec41d693e314f13af44695

[silk zodiac]

Actually I did not (and I was not aware of all these tags), I'll give it a try.

[naive quail]

Suno seems to know Melancholy - I've used it quite a bit. Bark may feel it.

[pseudo island]

Ok thanks for clarification !

[keen ore]

Lately the remasters seem to not be working. (Instead of getting a whole song remastered it does some random subset). Is that a known bug? Or do I have a wrong setting or something?

[daring talon]

wasnt actualy that bad splicing stems together

v4 is truly amazing tbh.

the vocal emphasis is what we needed to break that consistent / dull vibe

this is a cover of my bands real song btw. 🙂

[naive quail]

I threw them all into a track... and got traffic sounds. 😄 https://suno.com/song/21265a4b-f392-465e-818a-67cbef721c65

[delicate bloom]

I am not certain if this is technical or not. I want to post some of my songs on youtube. I read a legal statement that needs to be posted to state ownership however I cannot find it now.

[daring talon]

you own it. suno cant own it

[primal barn]

aight i've had some success fighting the degradation using "reverb, shimmer, noise" in exclude styles

[delicate bloom]

I know I own it however don't I have to make a statement on the you tube post?

[daring talon]

youtube post? no. when you upload it you declare the rights.

[delicate bloom]

that is what I am trying to do declare the rights

[daring talon]

the ai section is generally regarding "making ppl say things they didn't say, deepfakes etc" not ai gen. i mean. if i open a wav and trim and compress it. who authored it now?

[delicate bloom]

I wrote the lyrics

[hollow hare]

YouTube says otherwise https://support.google.com/youtube/answer/14328491?nohelpkit=1&hl=en#examples_as

[hollow hare]

Not according to Suno's TOS (which is a legal document) at https://suno.com/terms. Explicitly states that you agree that the Output generated has the rights, titles, and interests assigned to Suno and only assigns it to you if you have a paid subscription.

You do own your own Input but you do give Suno irrevocable rights to use what you used for the Input perpetually.

If you want a video explanation, see Top Music Attorney (links and useful timestamps at #📚┃suno-school message)

You should also check out #1294423389515284552 @delicate bloom

[hollow hare]

TL;DR:

• If you made it on the free plan, you can post it social media BUT it cannot be monetized (non-commercial license) and you must give attribution (e.g. Made by Suno in the description)
• If you made it during a paid subscription, you can do what you want with it as you had been granted the commercial license

For more information, see #1294423389515284552

[daring talon]

Oh, that was a new addition. Thanks. Wasn't there before

Define synthetic, so djs ?

LoL

[hollow hare]

It was there since the start IIRC buried deep. I had to go throigh several links to find that page initially.

I still think it is stupid because a lot of music before the AI boom were generated "synthetically"

[daring talon]

110%

[hollow hare]

Like, realistic voices and images are understandable. But "synthetically generated music"? Wtf 🤦

Which would include music made through programs like FL Studio and DAWs but you know they specifically meant AI generated music without saying so

[delicate bloom]

yes and look at how artists' voices are changed so much their voices are not even their own.

[charred merlin]

Ignored — I made a “prompt experiment playlist” in my workflow thread

#1256301881966198805 message

[waxen epoch]

same here....

[naive quail]

Don't stop doing what you're doing, but do realize that confirmation bias is real; it's bitten me in the arse maaaany times with Suno.

Here's a Lossy, Poorly Produced track: https://suno.com/song/930acc7a-eaf7-4845-af33-120a57b752cf

[bronze path]

Does anyone have any advice on what prompts should I try and use for creating something similar to the pursuit music from Need For Speed Most Wanted (2005)? This is the music the from this playlist: https://youtube.com/playlist?list=PLiekpTheH989PBG_VP4gpyeKr3TbbgACl&si=oiOLOIljh0p4UrFh

[cold zenith]

Actually, I have a question. Since degradation only happens when we make extends, why don't we put a 10min limit on the original generation? Knowing that this won't change anything for those who make 3min music, but that it will change the life of those who make longer music, and who end up with degradation, because extend, asks the A to make an overlay on a layer and it has trouble sometimes. Frankly, I sincerely believe that this is the simplest solution, and so is increasing the 3500 character limit. Because lots of times you have to extend because there's no more room.

[glacial scarab]

That first assumption already is wrong. Degradation happens regardless

[cold zenith]

Not for me. it always after the original one. most of the time starting at the 3rd extend.

Everyone knows it, it's no secret.

[restive belfry]

I made a song and liked the music, but i want to extend the lyrics. is this possible?

Ok i found the "reuse prompt" but i dont know how to keep the same music style but with only changes in the lyrics

[naive quail]

I really don't see any more degradation in a three-minute track made up of 12 extensions vs a three-minute one-shot track in the same genre. Many of the one-shot tracks still degrade, especially after the first two minutes. I am much more careful with my extension starting points now, though - you can definitely make things exponentially worse if you extend from a more noisy part of a track.

[naive quail]

Have you tried using Menu > Create > Extend? That will start a new part that kicks off where the previous one stopped (you can also set an earlier starting point for the new extension). That will give you a track labelled "Part 2" from which you can select Menu > Create > Get Whole Song to merge the parts together into a new track.

[cold zenith]

And we lost a lot of good music like this because extend from a degradation, will degrade more

[tribal badge]

Merhaba

[lunar anvil]

ReMi is most frustrating. More errors than results.

[restive belfry]

Not sure i get how. I click on the "extend button" in the song i want. Then type the new lyrics and then i dont know how to save it. I dont see a create button.

[naive quail]

After clicking on Extend, the Generate button should be replaced with an Extend button:

[restive belfry]

I dont have that

[naive quail]

Odd. I've never encountered that.

[restive belfry]

im using from browser

[hollow hare]

Shows you have 0 credits, maybe that is why?

[naive quail]

ooooh

[restive belfry]

im not sure, i didnt see the button when i had credits.. but might try tomorrow if credits refill?

[flint haven]

Free Credits Refresh

Refreshes at 0000 UTC+0 (<t:1713222000:t> standard local time)

[daring lava]

S/O to @naive quail for helping members here! It's a big deal 🙂 Truly appreciate your thoughtfulness

[naive quail]

You may also need to scroll down...

[restive belfry]

its showing now at least

[lunar anvil]

Ahh, I just made 2 pages worth of songs just to realize v4 was not being used because v3.5 was auto-selected in the menu. The dropdown is very inconvenient.

Also, I didn't notice the v3.5 tag on the created songs as I must have been in the zone and had blinders.

Totally explains why I wasn't getting the new v4 articulations.

[restive belfry]

one more thing, if i use the crop option, does it replaces the original or saves a new version?

[naive quail]

Lol. "FINALLY!!! The shimmer has been fixe... SHIT, v3.5!?"

[naive quail]

It'll create a new track

[tiny falcon]

V4 is fix? I still get a bit of shimmering though 🙂

[lunar anvil]

The shimmer is still in v3.5 it just isn't as pronounced as in v4. Once you hear it you cannot unhear it.

[naive quail]

Nope. Still shimmers.

[naive quail]

Oh definitely! It's even more noticeable now that v4 has recalibrated our hearing.

[restive belfry]

i guess it requieres credtis to crop too? xd

[naive quail]

I think the only action that doesn't eat credits is "Get Whole Song".

I hear shimmer when my neighbor talks 😄

[lunar anvil]

Everything seems to be a microtransaction nowadays.

[tiny falcon]

I hear shimmer in my sleep, like a nightmare 🙂

[naive quail]

It's coming from under your bed... https://suno.com/song/6f5b4edc-bd7a-49c9-b009-471fdf547fe3

[tiny falcon]

Perfect for halloween lol.

[pure burrow]

Can I use Suno to generate Voice only, not a song, just like eleven lab and other voice-over tools? And how to do it?
Here is an example of this.https://suno.com/song/fac03a13-7d26-4087-9bdf-a1259db39ec0

[naive quail]

"Spoken" or "Spoken Word" word well.

https://suno.com/song/13a0d143-7eb8-4896-9849-346282b5539c

It's even possible to do little spoken theatrical scenes like the intro for this one: https://suno.com/song/75f07f11-cb9b-4b1a-8e8f-1d10a944ac25

[pure burrow]

Thank you very much

[lunar anvil]

Reminds me of The X-Files: Fight the Future soundtrack.

[naive quail]

Another spoken intro that I need to get back to: https://suno.com/song/5ba63b1a-846c-4241-9959-5be60401a879 (about a genie who's forced to live in a bong after his master "accidentally" breaks his 500-year-old bottle)

[hoary fable]

Question with v4 is anyone else having the issue when you extended it and the extension starts before where you extended making lyrics repeat?

[austere dragon]

weekly ask if anyone has figured out how to end extensions in v4?

[lunar anvil]

Does using an ending tag not work?

[hoary fable]

When you say end extensions do you mean while the song is still going, or when you actually try to end it

To end the song I use (fade) (end)

So it just doesn’t cut off and ends I use fade to slowly end it

[austere dragon]

https://suno.com/song/c9614498-b5ac-4d24-a4ea-a69af83a4e39 check out the extension transition here start listening at like 3:45.

[austere dragon]

To end it. You use prompts with parentheses?

[naive quail]

Unfortunately, v4 is much more resistant to those when it can't seem to find a good conclusion on its own.

[austere dragon]

it also drags out even the shortest song to the 4 minute limit which makes it difficult when extensions don't work.

[hoary fable]

Yes you need to let the AI know you are done with the song

[austere dragon]

Putting commands is parenthesis is a signal to the AI that those are back up lyrics. Which it doesn't matter anyway because when I put [end] it now sings. That must be an extra special Christmas feature from SUNO.

[lunar anvil]

I used [Sforzando Hit] at the end of the Outro on my last v4 and it ended nicely.

[austere dragon]

I've tried every suggestion. I just come back once a week to see if anyone has discovered a new trick.

[hoary fable]

With the 4 minute limit it’s not exactly a limit, since you didn’t start by generating an instrumental first so it’s making it up to 4 mins. Each extension adds 2:03 after the first 4 mins depending on where you want to generate from

[naive quail]

Well, that's how people have been using them, but it's very definitely not a hard rule. https://suno.com/song/77e41229-81bd-4317-ba7e-e8871b31678f

[austere dragon]

it's a good model... it recognizes more instruments and instruments like strings sound much more realistic. It's a great model, but needs some work.

[tiny falcon]

Try remove the style prompt and only put [End], add [end] in the lyric part, try bring it back to v3. or v3.5, get whole song and then remaster it again in v4. Sometime it works.

[hoary fable]

My only issue with v4 rn is that when you try to extend it repeats previous words that are not needed

[austere dragon]

that produces a song that sounds like total sh*t.

[tiny falcon]

Yeah man. the quality does drop.

[lunar anvil]

Perhaps the AI sees the pattern where Verse 1 was enclosed with () so that becomes the tagging schema and anything else gets assigned supplementary.

[naive quail]

Here's a 33s outro from ~ 04:05: https://suno.com/song/3dd85cb0-ed3c-4899-92d6-0c1c2c2027d5

[naive quail]

There does seem to be some degree of smarts when it comes to detecting tags https://suno.com/song/c6667c69-4540-4422-859d-f296afbb2d07

...or the complete lack thereof: https://suno.com/song/8eb0e05c-bf13-4094-b29c-dc75f82ff079

[lunar anvil]

Discovery requires experimentation.

[tiny falcon]

Sound like tiktok motivational quotes. 🙂

[lunar anvil]

"Discovery requires experimentation" was a quote from the TV series Agents of Shield by the Hydra scientist Daniel Whitehall.

[formal salmon]

+95% of all my songs are cut off, why? They reach more than 2 minutes and they are cut off, songs that should be 3 or more minutes? Why does this happen?
Does anyone else experience this problem?

[naive quail]

the cut off at exactly two minutes? What version are you using?

[lunar anvil]

@naive quail Have you been able to get crowd cheering or applause?

[naive quail]

https://suno.com/song/64821774-d327-43ea-8d99-d27fcb02f237

Or... if you remember the days of the bird covers... https://suno.com/song/aa5f4ff2-e9bd-4205-8bfc-5bbd58621989

[proper glacier]

Hi all! I have a driving guitar and piano accompaniment that on the last phrase I'd like to end the music right on the last three words. Does anyone have tips for metadata that might encourage that?

[naive quail]

I don't know of any way to end with that level of precision within Suno. We're often lucky if it ends at all. 😄

[lunar anvil]

Crop tool

[formal salmon]

@naive quail v4, v3.5 any version cuts, sometimes it creates a complete one

[proper glacier]

I've gotten some generations to do it. The actual position doesn't matter but id like the end to be just the last few words, or even word. It keeps going for another measure or two.

[naive quail]

I rarely have problems with endings in any of the versions, but to get the music to reach a conclusion the moment the lyrics end... it's always just been luck for me.

[naive quail]

I think all you can really do is throw credits at it, unfortunately.

[proper glacier]

Yeah that's plan b

😂

Is there a way to cover songs longer than 4 minutes?

If I crop it just cuts off abruptly, right?

Oh 🤔

This is what I've been burning through credits on

[polar spade]

Who needs sleep? Allocate sleep time for experimenting. Lol

[abstract viper]

is there any use case where Replace is actually useful? I've burnt many credits in trials and horrors...

probably for the instrumental parts may work nicely but when there are words it's really funny (or scary, sometimes)

[unique current]

This is awesome! Lol
The whole premise might be silly but this is genius.

[static valley]

Hello everyone, I have a question: Can I create a song and publish it on YouTube?

[graceful condor]

Good evening, yes. If you are on Free, you can't monetize your songs. If you are on a Premium plan, then you can

[static valley]

🙂 ok coooool

thank you

[graceful condor]

Happy to help

[polar spade]

I've used it for many fixes. I guess you're being rhetorical due to frustration. I agree it's still randomly reliable.

The best results are when it has lots of legroom east and west of the lyrics to be replaced.

So.... "and they lived happily ever after..." works? Wait, that's a fantasy ending. 😳

[autumn vault]

Is it just me but v3.5 had audio decay too 🤢

[floral mulch]

Hey guys, i used my free v4 creations today and noticed very much ... "beeping" in the background. Like jitter noises or something, very high pitched.
it is in every creation i make with 4.0. Is that a bug you know of and can be fixed somehow?

Not sure if i just got bugs, make bad prompts or what is the issue.
I hear it especially on high hats or high pitched voices

Almost like a old telephone ringing in the bg ^^

[abstract viper]

maybe it depends on language, my songs are not in English

[gritty badger]

After hopelessly trying to remaster my favorite V4 songs with horrible sound quality, thinking maybe this time they received some kind of fix, but I'm still getting "remasters" flat and dull in the beginning and right after the 2 minute mark, crispy and sometimes even with harsher highs as results. I used to, lets say, generate 10 songs and would get least 6-7 of them worth using, but now it is down to 2, which is a very concerning number. I pay yearly, so my credits won't accumulate for next month. Can't use my credits because of poor quality, so it is a waste of money. Starting to lose hope on a fix anytime soon 😞 Giving a break on Suno till a fix arrives...

[floral mulch]

mine neither, they are in german. But this noise is on instrumental parts too. :/

[gentle ridge]

What´s up!
Are there potentionally any ppl with experience in machine learning an training (preferably in the area of music generation) 😄

[graceful condor]

Most likely the v4 issue, happens to many musicians

[floral mulch]

So just use 3.5 until new version? 😦

[graceful condor]

It all depends on the genres that you've picked, I'd say. Some still use v3.5 instead of v4, for example

You can try adding [Studio Quality]

To Lyrics

Usually helps with increasing the sound quality

[floral mulch]

hmmm ill try that. 🙂

[graceful condor]

Helps me a lot, some songs are still 144p, though

Haha

Wishing you all the best, don't forget to share your songs with us

[ashen yacht]

How to we generate the custom, once lyrics and style of music are inputed there is no genrate option?

[graceful condor]

Showcase, genres, general, all that

[graceful condor]

Click on "Custom"

Or "Instrumental"

[ashen yacht]

Yes dont that,

how generate

[graceful condor]

Scroll to the bottom

At the left part

[ashen yacht]

cant scroll futher down that that "Persona

Beta"

nothing below that

never mine i refreshed the page

thanks

[graceful condor]

That's what I was going to say

Haha

Happy to help

Enjoy making music on Suno and don't forget to share it with us

[ashen yacht]

Where share

[graceful condor]

Showcase, Genres and general

[unique current]

How do you make background music but speech on top of it with Suno?
I'm trying to do something interesting and it's not behaving because it really really wants to sing LOL

[floral mulch]

[Spoken words over music] before the lyrics worked for me

[ashen yacht]

this shit crazy

pardon my french

any tips for generating death metal? like that {sutdio quality} prompt

[formal salmon]

https://suno.com/song/b700f653-fb4c-4e32-a13d-34bfdec33a02 all my songs are cut in the end

[final trail]

No I've only made like 2 death metal songs

[wild spade]

Hi! Before trying, now there's any ringing effect on v4 songs?

[floral mulch]

yeah, cant get rid of it

that beeping makes me insane

[wild spade]

those high frequencies are the devil

[floral mulch]

yeah....

[wild spade]

ruins perfect songs

[floral mulch]

today i tried to extend with 20 songs and alyways theres that jitter

😦

[wild spade]

😦

I hope they fix soon

But still, no news

[floral mulch]

no it seems that issue was there with 3.5 already and got back worse, someone wrote its maybe the tech to detect suno songs

so theres not 20 new drake songs etc

[ashen yacht]

this shit is hard https://suno.com/song/f08ddff9-c3f0-4b27-ad36-bd598f84bdd1

[graceful condor]

Do with []

[graceful condor]

Still happens

[fringe cypress]

/chirp

[flint haven]

The /chirp command is no longer working

Chirp is retired, and no longer on discord. You can make music now on our website: https://suno.com

[ashen yacht]

I did, thx. Makes a crazy difference. My homie is going to go crazy on this he’s a producer the key words are key here

[graceful condor]

Happy to help

There are a lot of variations for tags

[twilit ice]

@junior glacier @unique stone Me: 1. "I understand the value in permanent credits and those credits being higher priced.
2. However, if I KNOW I will use them all before my month is up, the fact that they are permanent is irrelevant and therefore not priced in line with my subscription." (proceeded with example prices and suggestions) Suno: "You'll be glad to hear some of these are already on our roadmap." - I was discussing several things so don't want to misrepresent but they understood and it seems there will be a solution.

[graceful condor]

Something up with SUNO? Buffering creations.

[twilit ice]

Working ok here, East coast USA

[crisp condor]

What’s up with Suno? Generations not completing 🤔🤔🤔

[graceful condor]

I'm on a VPN, so maybe disconnect. Thanks for checking @twilit ice

Yeah @crisp condor - I'm having problems too. Probably just a server reset.

[twilit ice]

well @crisp condor has issues. Maybe im lucky or issues on the way!

[graceful condor]

Been like this for about 10-15 min. Only thing I haven't done is reset my network, but that doesn't seem to really be the issue.

[crisp condor]

It’s a toss up between generations never competing, or some simply vanishing and a red pop saying something about excessive network usage 🤷🏼‍♂️

[simple willow]

I've got a pile of stuck songs

[frail robin]

yeah I'm running into that network error too

[crisp condor]

Creation request simply cancel out

[naive quail]

It's probably a Suno issue. I just did a quick gen as a test and got an error message regarding one of the two tracks timing out. The other one has been spinning for about 2 mins.

[crisp condor]

*requets

[frail robin]

doing section edits

on my last section....

[crisp condor]

So basically the network is useless… lol. They may want to work on this #hinthint

[graceful condor]

Looks like an issue on the Suno's side

[simple willow]

ah there way go

all better

[twilit ice]

That's becuase they are updating to get rid of degradations and shimmers. (I made that up. Let me dream, don't ruin it).

[vital sinew]

Hey Folks, I can download MP3 and Video, but not .WAV files. I tried a different browser and a computer restart, no dice.
Got any protips?

[severe heron]

I have the same problem, were you able to solve it?

[vital sinew]

Did it just happen today?

[severe heron]

yes

[vital sinew]

Its probably just a glitch, maybe we can give it 24 hours and check again

[charred merlin]

Can you elaborate? Do you not see the .wav option? Do you get an error message?

[vital sinew]

It declares that it has began the wav download, but never produces the wav file

MP3 and Video are at there normal rate

[charred merlin]

If you haven't tried Firefox, try downloading it through there

[vital sinew]

Ive tried brave and chrome

[charred merlin]

Brave (and A LOT of others) are based on Chromium, the open-source project that lead to Chrome

Firefox is special

[vital sinew]

embarrasing I haven't installed firefox in probably 8 years

[charred merlin]

Not a worry -- I'm getting by just fine with Edge Dev for nearly everything -- we've just noticed that this issue seems present on chromium browsers, but firefox seems fine in a large number of instances.

Mind you, sometimes it's an issue with the particular song, or a temporary quirk like you'd suggested earlier

Speak of the devil 🍻

[twilit ice]

I can download wav in Chrome but experience several other issues resolved in Edge.

[hollow hare]

I haven't heard back from people with Safari (although the odds of users using Safari over Firefox is probably low) for the download issue 🤔

[vital sinew]

I just downloaded firefox, but its not working

[twilit ice]

For instance, when I search all tracks I have liked in Library, I can't filter with a word in Chrome. Works fine in edge.

[hollow hare]

Hmm, perchance, did any of your recent creations have a "SUNO" image watermark and/or audible audio watermarks?

[charred merlin]

Hrm -- as I noted above, it MIGHT just be a temporary quirk, but I'd consider reaching out to [e-mail] if the issue persists much longer, or if it's impacting every song.

[vital sinew]

I didn't know that was an option, so I hope thats not it

[hollow hare]

If that is not the case, then you do not need to worry about that interference

[vital sinew]

I'll just plan to wait a day

[twilit ice]

I know it's not helping you, sorry, but furthermore, on Android, I can't even load Suno in Chrome. Have to use Edge. Suno is the only time I open a browser besides Chrome. Just mentioning it to other Chrome users.

[left bolt]

Another one unable to download WAV files, emailed support awaiting reply.

[halcyon sparrow]

maybe cache of browser cleaning then works

[left bolt]

No tried 3 different browsers and different computer all the same

[halcyon sparrow]

ccleaner?

[left bolt]

No seems a server issue at Suno end

[hollow hare]

Hmm I cannot generate/download one was well. Guess someone is working on the backend live or a process is getting delayed

[prisma arrow]

there is an issue about downloading process....

[velvet patio]

Hey are you guys having .Wav issues?

I can't download the wav files

[prisma arrow]

same issue

[unique stone]

Yeah it seems to be not working right now. I just tested it as well and it just doesn't do anything.

[velvet patio]

OK then it's not me that's good

[graceful condor]

Ah, thank you, everyone. I was thinking it was just me as well. 🙂 I sent a support e-mail before I learned about this Discord. Glad to have found this community.

[hollow hare]

Information has been relayed, hopefully it should be solved soon.

[barren snow]

Is there any way to get a "non-mastered" version?

[gentle ridge]

What´s up!
Are there potentionally any ppl with experience in machine learning (preferably in the area of music generation) 😄

[vale cliff]

many React errors in the console unfortunately, same here can't download.

[gritty badger]

Trying to snag some workforce from here? 🤔😅

[hollow hare]

Sadly, most people know how to use one here, not so much develop one lol.

[misty inlet]

yeah. not only that, but the songs i've generated recently are all incomplete. they stop at about 1:30, maybe 2:00. it's weird, and i have no idea on how to fix it.

[hollow hare]

Your account was most likely flagged by Suno for possible API use or manipulating the website (usually a false flag due to using a VPN or payment related issue).
Please, use this form to report the false flag to Suno where they will investigate to clear the flag from your account:
https://sunomusic.typeform.com/to/zGsJSzsf

[misty inlet]

thanks! i'll try that.

[hollow hare]

Sadly I do not know the turn around time, it will definitely be at least a few business days

[misty inlet]

no worries, just want it to be fixed so i can generate more songs haha. i thought it was a general bug or something.

[velvet patio]

Extend

[kind pilot]

Hello! I would like to know what the criteria/rule is for being able to create personas from an audio uploaded via upload. In some audios when trying to create personas, the message "you cannot make Personas from audios uploads at this time" appears, but in some audios uploaded in the same way, I can create them normally. Does anyone know?

[hollow hare]

If the Upload Audio has vocal-like sounds detected, creating personas and the Public setting are disabled iirc

Due to copyright or usage right concerns

[kind pilot]

I considered this, but even using voice capture from my microphone, it sometimes gets blocked in the same way.

[cold flax]

Thanks for the reply bro! The upload I uploaded didn't have any vocals in it though so not sure what's going on!? I've reached out to support so maybe they can shed some light 😉

[verbal mortar]

hmm seems downloading wav files has stopped working

[kind pilot]

In the tests I did, some audios without voice, it was possible to create personas, and in some with voice it was not possible (like using my own voice for example)

[cold flax]

Must be technical issues brother as I'm using V4 and subscribed to Pro

[verbal mortar]

yep....totally not working for me anymore. was fine an hour ago

[hollow hare]

Hm, what do you want me to see?
By the way, if that is something derived from Esta Tonne's Envision - A Call to Prayer (e.g. voice removed version of the song), that would be against Suno's TOS (which can include deletion of your account) and makes you liable for copyright infringement

[cold flax]

Yeah sounds like you can't create personas using uploads with vocals. I can't create personas using uploads without vocals for some reason lol 😭

[hollow hare]

Yes, there is an issue with Suno's backend for the downloading the wav. The team has been notified

[kind pilot]

The one from Estas Tonne there was a test, in the other example using my voice capture from the microphone, it was also not possible

[hollow hare]

Yes, I said that. If Suno detected a voice in an upload, it disables/stops you from making persona and stops you from makong it Public

[kind pilot]

So just to confirm, I can only create Personas from audio tracks that do not contain voice, is that it? I'll take that into consideration then. I hadn't fully considered this option, because in some cases I created Personas from audio tracks that did have voice! Thanks for the help, I'll move on from here

[hollow hare]

Yes, that is correct, because people would try to impersonate others otherwise

[vale cliff]

indeed, they know apparently 😉

[little granite]

have they said when the next update is for v4

[gritty olive]

Hey guys, does anyone know, the style prompt box, does it tend to prioritize earlier genres in the list, or does it treat all listed styles equally when blending them?

[hollow hare]

#💬┃general-chat message

[gritty olive]

Thank you 🙂

[grim walrus]

Hey ! Do you know where can I ask for help about a problem ?

[woeful yoke]

While you guys are talking about wav downloads not working, I must also point out that either 3.5 and 4.0. both have been ruined somehow or they use a common audio encoder, and they managed to screw up how audio is tokenized now. yesterday we already had an issue that all uploaded audio was converted to mono.
Today, however, things are stereo again but the uploaded track gets ruined right in the encoding (tokenization) process.

[dreamy sparrow]

Oh when will they fix it

ffs

[unique stone]

The mono upload issue only happened if you trimmed audio during the upload btw. Turned out it was only taking the left channel. I spent a while poking around with this.

If the other thing you mention here is a new behaviour, does that happen with audio that you have trimmed before uploading, or only if you trim the audio during upload?

[hearty torrent]

oh nice to hear that they are working on the wav files. Suno is getting better and better each day with v4

[wicked linden]

Anybody got tips on how to keep the library organized? Mines a mess already and I've only just begun.

[soft arch]

I just uploaded a file of me playing the guitar, no voices, just my guitar, and it wouldn't let me make a persona out of it. So maybe there's more going on than just a voice blocking the ability to make personas.

[hearty torrent]

if its just you playing guitar, its best to use cover instead of persona. Persona is literally cover except it's easier to change lyrics

[soft arch]

I extended my guitar and added lyrics and it let me create a persona from that. Pays to experiment.

[woeful yoke]

I didn't trim anything. Tried multiple times.
And issue with uploading files now is that i can barely user them for cover songs. And extending my own upload miss out of the question with 3.5 and 4.0. the extension will be either really noisy or doesn't even resemble the original.

Tried uploading multiple tracks with many different loudness levels and they hardly made any difference.

Actually i think i know how it gets screwed up. I think that gets screwed up at the file upload dialog box, even before the crop dialog box even appears on the screen.

[untold briar]

Everytime you open Suno, generate a unique search term that you add to the title of every song you generate, and keep it the same for that prompting session. Pick 3 random words "NeonAppleWizard" for example. Then later, if you want to search up all related tracks, you can search library for that unique word. If you use Extensions, note the Extension time in the titles. Basically use titles a LOT, think of it more like a "notes" field.

[woeful yoke]

Jee kind of works for individual instruments and vocals and what not but... Ai just to be able to extend audio clips exported from my DAW productions. And now everything extending into 3.5 sounds absolutely horrible and with 4.0 its like a lottery. Even if i manage to extend without audio decoration, the extended still has wrong melody/chords, and wrong singer and whatnot

I'm talking about extending tracks that are kind of already complete.

[soft arch]

I just tried creating a song with my guitar persona. One was OK sounding but not a keeper, the other sounded awful. Trying another one now.

I gave up on extending tracks for now. They all sound like crap.

[woeful yoke]

So it's not just my problem then 🙂

[calm apex]

I have a question about uploads and the ablity to make a song public. How is that sometimes my uploaded covers can be public but other times never?

[orchid niche]

Hey Guys, have some way to upload up 1:20 minutes? Maybe paying better Plan?

[charred merlin]

Yes, paid subscription can upload 2 minutes, free users are limited to one minute.

[orchid niche]

i Had already Pro Plan, and now i cant upload up 1:20

[charred merlin]

Uploads that are detected to contain vocals are restricted to "private" -- it can be a false positive, as it's an automated system.

[charred merlin]

Did it say "this song matched an existing work" or something along those lines?

[orchid niche]

No, nothing abou that

[calm apex]

said nothing i guess it might think its vocals as its a Dukuk instrument

[charred merlin]

1m 20s is within the 2m limit for paid users, so it shouldn't be an issue. Can you elaborate what's happening when you try to upload your clip?

[orchid niche]

The suno ask to cut, but yes, i think 1:59 is the limit, but have some way to up that? for now?

I Think can offer a plan with more upload times, i realy suppose pay for that

[charred merlin]

https://suno.com/song/a7d24b82-0568-4073-b4fd-5a5ba6a00e0a the limit is 2m for paid users -- if you're still running into issues uploading your 1m20s clip, please contact [e-mail] so they can investigate the matter.

[orchid niche]

Yes, i do that, i alrady pay Pro Plan, but im asking for more

last month didnt have 2 minutes limit, i uploaded 3 minutes music, for exemplo

[charred merlin]

Please share the link?

[orchid niche]

about what exacatly?

[charred merlin]

The link for the 3m clip that you allegedly uploaded last month.

[orchid niche]

maybe i delete, i used, make a cover and delete

i think

was 25 october

so, the official limit now is 2 minutes? dont have anyway to up that for now?

[charred merlin]

#🚀┃updates message

It's been 2m since October and that's the longest that we've ever received, but maybe this can be your "Mandela Effect" story 💪

[orchid niche]

Maybe was something lik 2:30 minutes, or something less, but was up 2minutes, but ok, for now dont have anyway to up, correct? maybe in the future

[charred merlin]

Correct -- 2 minutes is the maximum for all paying users. It would be exciting to upload 3 minutes, so I hope it is improved soon 🙏

[soft arch]

Are anyone else's songs getting cut off before the lyrics are finished and before the 4:00 minute mark? Three generations in a row were cut off at 3:24 before the lyrics were finished.

[graceful condor]

Works now

[graceful condor]

Should work now

[graceful condor]

You can download the WAV files now

[graceful condor]

Fixed

[graceful condor]

Works

[velvet patio]

Thanks

[graceful condor]

Happy to help, enjoy creating music on Suno

[ocean jetty]

How do you contact support staff?

[graceful condor]

What's your issue?

[ocean jetty]

Are you support staff?

[charred merlin]

[e-mail] / [e-mail]

[graceful condor]

No, but I know the most common answers. If you need help from a helper, you will have to wait for a Guide to help

[graceful condor]

That about sums it up

Quite interesting that you are not a Guide yet

[ocean jetty]

What's the best way to get help from staff?

[charred merlin]

Guides & mods aren't staff -- the only staff member who takes an active role in chat is Mike B

[graceful condor]

Right, wrong terminology

Used to calling staff

[graceful condor]

The emails that Harber provided

[charred merlin]

😅 #💬┃general-chat message

[ocean jetty]

It's easy question, how do I get a hold of someone that works for the company?

[charred merlin]

You email [e-mail] -- Discord isn't an official support forum, as noted in this article - https://help.suno.com/en/articles/2551617

[graceful condor]

No color, though, that's unique

[ocean jetty]

What about
[e-mail]?

Would be a right email for the company?

[charred merlin]

Email whichever suits your predicament better -- is it a billing/payment issue?

[ocean jetty]

Both 1. being a Swissh problem in v4 I am trying to avoid, 2. credits purchased continueling trying to resolve that issue of the Swissh problem

Which email would be best for that?

[ashen carbon]

can someone listen to this and tell me if it has the "shimmer" that everyone is always talking about here? https://suno.com/song/499d7869-ea91-43a6-95b9-3e2e14905681

[ocean jetty]

Big time

[ashen carbon]

Ah I think I finally know what it is then!!

[charred merlin]

[e-mail] seems appropriate for payment-related questions 🙏

[ocean jetty]

Wish Shimmer, sham sho, let's pay more to see if we can get rid of it

[ocean jetty]

Makes sense to me too. I sent them a message 6 days agp.

[charred merlin]

You might have received an automated response (if not, they'd started sending them out shortly thereafter) to inform that they've received a lot of support requests atm, and are working through them in order to get back to everyone -- fingers crossed they get to you soon since your message was sent a while back

🤞

[ocean jetty]

Releasing v4 with such a significant issue and continuing to promote it while accepting payments as if it’s the latest and greatest is unacceptable. After six days, this ongoing issue feels like it’s being ignored.

Take V4 down, stop promoting it and ripping customers off

[eager topaz]

We are told that it's Beta, after all 🙂

Thing is, there are issues that go back to V3a, still not resolved and now made worse in V4 Beta

[ocean jetty]

Why promote v4 as the best version when it has a critical bug that sets it apart in all the wrong ways? Sure, we can fall back to earlier versions like v3.5, v3.0, or v2, but that completely misses the point. Suno continues to market v4 as the pinnacle of performance, but this issue severely undermines that claim.

Waiting six days without a response from billing or support? That’s beyond unacceptable. Meanwhile, the technical discussion chat on Discord lacks any meaningful engagement from company representatives to address these concerns.

At this point, I feel like I’m talking to a wall. I’m left with no choice but to file a claim with Visa for all subscriptions and credits spent on a product that is fundamentally broken.

[charred merlin]

Discord is not considered an official support channel.

[e-mail] is also listed on the subscribe page

[lean moon]

Hey everyone. Is there a way to do a cover for a whole song rather than the 120 seconds? I'd like to use it to develop my own music

[ocean jetty]

Six days let's go?

Wait a technical discussion is for how to produce songs?

[ashen carbon]

what is this significant issue you're talking about??

[ocean jetty]

Version 4 with the Shimmer, vocal, cymbals, has this shimmering long cross effect

[ashen carbon]

it's not bad enought to take it down ffs... i'm getting great results regularly

I didn't even know what the shimmer was until earlier today on a song where it was kinda pronounced

[charred merlin]

I think 4m is the limit for covers, maybe a few seconds shorter -- furrybeans got them to fix the bug where it was limited to 3:18 or summat

[untold briar]

Not at this time. You can post in suggestions that uploads longer than 120 would be useful for this purpose.

That said, if you are willing to put in a lot of manual work in an audio editor, you might be able frankstein a cover from multiple generations with overlapping parts.

[ocean jetty]

Boy, you don't hear things do you

[lean moon]

That's what I've been doing, but it would be incredible to upload a simple guitar, piano or whatever track and have Suno fill it out.

I think it would bridge the gap to music producers who still want to make their own bespoke stuff

[charred merlin]

I might have been confused -- uploads are limited to 120s, but you can do longer covers, can't you?

[untold briar]

I assumed he was uploading his own tracks and covering them, based oon the 120s limit

[lean moon]

The issue is that it's for my own composed music I'd have to upload

[charred merlin]

I suppose it would have been a safe assumption, but I overclarify, as I am wont to do

[lean moon]

No worries, I appreciate the swift feedback

[neat nexus]

It's not necessary to talk down to people. Not everyone hears the shimmer.

[charred merlin]

Exactly -- it wasn't as pronounced in some genres as it was in others

Metalheads couldn't escape it, but certain EDM or ambient genres it was really hard to percept, or blended in with existing sounds

Really exciting to hear the definition to cymbal-drags and snare hits that would have been previously lost to the artefacts, and it still sends me into a cold sweat when it appears on a new track, but the frequency of how often it turns up seems to have improved a lot, same with the progressive hiss beyond the 2m mark

Not trying to force "groupthink" -- if someone isn't satisfied with the quality, that's a valid opinion

[ocean jetty]

You are right, I am sorry @ashen carbon

[steel plinth]

when do i get my free credits? my friend has created 10 songs

[eager topaz]

It's going to depend on perhaps the genre, what you are listening on, headphones, monitor speakers, car stereo, etc.

Going through the Suno Home Page Fresh From v4, I can hear it or other bad noise on many of the tracks there. My own tracks have shimmer or distortion or explosions on almost every track.

[primal star]

It feels like it gets more prevalent as well the longer the song goes on for

[eager topaz]

That's true, it does in many cases

[primal star]

Suprisingly with the song I was working on yesterday I think I mostly avoided the shimmer effect but it's likely due to the genre like what others have pointed out

[ocean jetty]

It's not on the headphones, monitor speakers, car stereo, etc - the output is the output. The shimmer effect will translate on all listening devices.

[ocean jetty]

What genres have you noticed that did not have the shimmer effecT?

[primal star]

From my personal experimenting I haven't had it happen much with electronic or ambient tracks

[eager topaz]

I'm saying perhaps, because I listen on studio monitors and quality headphones.

[charred merlin]

There's an understandable amount of tension regarding v4, but I'd be curious to know what discrepancies contribute to the wildly differing opinions -- I don't think anyone is outright shilling for Suno (even Mike is making measured comments about performance), but the quality of v4 absolutely has improved from alpha to beta launch to today.

That said, I don't think that everyone who says "every v4 song is terrible" is catastrophizing -- there must be SOMETHING that we're missing before we start jumping to "well, my ears are better than yours"

[primal star]

Absolutely. V4 definitely has its benefits but it's by no means perfect. At the very least I feel like it's come a long way since V2

[ashen carbon]

I have done hundreds of electronic tracks and tonight was the first time that I heard a very obvious shimmer

[eager topaz]

The challenge for the Devs is to work out why it doesn't happen with every genre and every time. That would faze me, no end.

[primal star]

Not too bad honestly

[ashen carbon]

myself I'm listening in studio, with very expensive Adam Audio monitors

some of the quality has been frankly stunning... like "why bother continue trying to make music the old way" kind of stunning

[split dragon]

im finding that with v4 the vocals are just so much louder compare to the rest of the song, and it keeps going very tinny

[ashen carbon]

like i'm seriously considering spinning up an alias and putting out an album of pure suno v4 generated tracks with no additional studio work just to see what happens

[charred merlin]

This is the wild thing about generative AI -- we can make collaborative partners out of it and feel like we understand it so much, but then the very next user with the exact same model is having a wildly different experience.

[ocean jetty]

Wait on DSPs?

[ashen carbon]

why do I get a feeling that it has a lot to do with experience in 1) actually producing music and 2) prompting generative AI effectively

[ocean jetty]

Are you releasing Suno prompt songs on DSPs?

[ashen carbon]

who's asking 😛

[ocean jetty]

the Artists Rights Alliance

[ocean jetty]

Don't

[charred merlin]

There's major distribution partners who are accepting of AI generated music

[ocean jetty]

That don't make it right

So wait, people are actually taking Suno prompt songs and relasing on DSPs?

[charred merlin]

Yes.

[split dragon]

is uploading them to youtube ok

[primal star]

It's crazy to think that it's getting closer to the point that I could viably use my outputs with the software for a large scale project and it wouldn't feel too out of place compared to normal stuff. I'm just looking forward to when the tools to create on Suno become more robust and easier to work with

[primal star]

Pretty much yeah

[ocean jetty]

So you sit down, make a prompt, and upload suno prompts (which are composed from samples from real music artists) and upload them to make money?

[primal star]

Wouldn't surprise if some have pulled that off

I mean hell look at AI artists

[ocean jetty]

Damn, I can't wait to watch a documentry on this

[split dragon]

ive uploaded a song to youtube and its got over 400 views on it, and thats just 4 days of it being up

[ocean jetty]

This is golden

[split dragon]

and its still climbing

[lean moon]

I put ganstas paradise over someone a gummy bear clip that's 10 seconds long that has 10,000 views

[charred merlin]

"composed from samples" -- doesn't exactly sound like you're well-versed on the matter, but I'm not sure this is the best time or place to get you up to speed on things.

Best of luck with your emails.

[ocean jetty]

I've got time, let's go, Do explain how ai acheives these songs without playing or singing a note

[charred merlin]

"the best time or place" requires a willing second-party with motivational buy-in, perhaps a non-confrontational invitation that doesn't reek of goalposts on wheels 🤷‍♂️

[ocean jetty]

Stirring the pot by asking a question? Understanding what Suno users are using Suno for?

[ocean jetty]

Sounds like you don't know

[charred merlin]

That's probably it -- this was fun everybody.

[split dragon]

tbf @charred merlin he could always just look it up on google, google has the time to explain, sheesh even copilot has the time

wait copilot is ai......

[ocean jetty]

With no answer from Suno after 6 days, having the same issue we all share, and than asking questions. Are we not aloud do that here at the Technical Discssuion thread?

[charred merlin]

In the sake of fairness, I think it started with having to wait 6+ days for support.

But the team is trying to be upfront with the fact that they're slammed atm -- pretty sure the support lead is burning the midnight oil trying to get caught up 😛

[primal star]

I would be annoyed after having to wait that long too

[split dragon]

err this isnt a thread its a channel but go on

[gritty badger]

Don't even get into that subject. The ethical part is still wide open for debate. To be honest, when someone listens to bunch of artists and gets inspired to make something new, that is considered okay, but when an Ai does that, why is it considered stealing? Everyday we get bombarded with free music all around during our daily lives. Should we cover our ears not to get sued? Definitely a new territory, but my opinion is that it shouldn't be a big deal. These lawsuits feel more like jumping on the "lets try to take a slice of this loophole" bandwagon. I am not ashamed to admit that I have an artist name in most of the major streaming services with some releases. But I consider calling myself the producer, rather than the musician 😁

[primal star]

https://tenor.com/view/nerd-nerdy-nerds-nerd-emoji-gif-25380417

[ocean jetty]

I will refer this as a channel for now on

[split dragon]

most supermarkets these days wont use copyrighted music on their "radio station" because its so much cheaper and quicker just to ai generate it

[ocean jetty]

That's a company issue, I'm the consumer looking for an answer that is not being communciated in a timely manner.

[split dragon]

for instance asda in the uk uses ai music now on their "radio station"

all i gotta say on the matter is this

dont use the product then, go and spend the time making music the time consuming way

and stop trying to ruin peoples fun

thats the problem with cancel culture, your all boring, and i laugh

[untold briar]

Ideally we'd like to keep ethics of AI or other hot button (and often never ending...) topics to off-topic channel, and keep technical focused on concrete Suno topics. Also keep it chill, please. 🙏

[ocean jetty]

Well actually any supermarket, restaurant, sport stadium, or public place play music by an artist/band should be paid mechanical royalties by a PRO, performance right organization. But that doesn't happen.

[ocean jetty]

Fact

[gritty badger]

Ai generated music isn't "samples being cut and pasted together" to compose music out of them. It doesn't work like that and that is the major reason people file lawsuits to Suno and Udio, because they think Ai training is copy pasting 🙄😅

Anyway, back to technical stuff

When are we getting an update 😅

[split dragon]

the time they have been sat here arguing they coulda asked copilot and it woulda told them

[ocean jetty]

Can you do that and ask copilot the answer?

[split dragon]

do you have a windows computer

[ocean jetty]

Serious request

I use mac

[split dragon]

ooooo go on bing copilot is built in

[gritty badger]

Try gpt for mac 😅

[charred merlin]

Visual comparison -- not 1:1 analogy, but in the same way that image gen isn't a matter of simply cut-&-pasting a scene together like someone making a ransom letter, but the AI model spent a lot of time and energy to analyze the relationships between the pictures it was shown -- after a while, it can create pictures it was never shown, just by understanding how all of the smaller ideas where connected

Similarly, it's not a matter of taking a cool guitar line and reusing it later, but a matter of listening to possibly tens of thousands of songs and understanding WHY a rock guitar sounds like rock guitar

YES -- if you specifically try to break it, you can force training data regurgitation -- ask for Barbie Girl and get Barbie Girl, but that's a "hiccup" and not the intended behaviour

[ocean jetty]

But you been talking so much about it

[split dragon]

you been waiting 6+ days for a response when you could just asked a ai

[primal star]

On that last point it's kind of wacky because the beginning motif of my latest song sounds damn close to Moog City from Minecraft without me intending it to generate that way

[tired laurel]

when will we be able to upload full songs?

[ocean jetty]

Good thing copilot and suno are the same company, dude what's wrong with you?

[split dragon]

errrr, would you like the short or the long list, or should i ask copilot whats wrong with me

[charred merlin]

Oh yeah -- the Suno terms outright say that "due to the nature of generative AI, no guarantee is made that any output will vest in copyright"

Just because Suno output it doesn't mean that you bypass any and all laws, etc

Maybe Suno's telling you it's time to go punch some trees -- I might need to give that a reinstall as well

[ocean jetty]

I'd say short

[split dragon]

severe anxiety and manic depression, with a huge dose of multiple personality

[gritty badger]

The understanding of Ai being trained is too mind boggling for many at the moment. I get that. How can an inorganic thing, comprised of codes can listen to music, understand the structures of it, knowing the connections of melodies, instruments, notations, rules etc and literally playing a virtual version of the prompted instruments, vocalising the written lyrics and creating something absolutely new is definitely not copy pasting

[ocean jetty]

I'm sorry dude, I can releate.

[split dragon]

so what was it you was wanting me to ask copilot, i cant be bothered to go look

[ocean jetty]

That does make sense

[ocean jetty]

Don't do that

[zenith jungle]

Hello, i have some problems with subscription, i buy it via app on android but nothing changed, i don't have it on app and on suno.com

[primal star]

Oh yeah of course I was just mildly surprised because it's the first time I've had it happen on my end. I think only the first 20 seconds or so would be a concern but I think it deviates enough for it to be fine

[gritty badger]

Yay a real technical question

[split dragon]

want the copilot response to that question

[ocean jetty]

Start with a lyrical topic that relates to you, think about something matters to you right now when writing lyrics. You can start with a verse, write 4 lines that makes sense of the motions, you can do a pre chorus or not. For a Chorus keep things simple on lyrics, not so many words, try to keep it relative to you. I'd say have a instrumental break after a chorus, and move to verse 2. Against keep towards your theme, 4 lines is the magical number, you can repeat the pre chorus and the chorus. After the Chorus have instrumental break or maybe a guitar solo depending on genre. Than a big change up for the bridge, make it loud and understanding your lyrical concept. Finally a end chorus, maybe repeated, emphasized - end on the down beat

Wait you said make a song with suno not a real song, my apologies

[gritty badger]

I'll do my best here pal. This error happens to a lot of people lately.

Here you can find the links for suno supports. They should help you out thoroughly.

#📚┃suno-school message

[primal star]

And on your last point before I forget I actually have been funny enough. I was playing an expert pack previously but this time I'm trying Blightfall from the Technic launcher

[split dragon]

would you like to see a verse from one of my songs

[charred merlin]

I love the visualization that AI doesn't actually know what music sounds like -- it just recognizes patterns in the waveform sample dots -- after enough reinforcement training, it understands how to create songs that match the patterns it's been told worked.

[ocean jetty]

https://suno.com/song/5ace28a2-da5c-41b3-8edd-a5df75edf9a8

[gritty badger]

That's exactly what I thought,very cool

[split dragon]

"(Bridge) I won’t be silenced, I won’t be tamed, In my own light, I’ve reclaimed, The power you tried to take away, In my heart, I’ve found my way. *[chilling laughter]**"

[primal star]

https://tenor.com/view/american-psycho-spotify-wrapped-lets-see-paul-allens-lets-see-paul-allens-spotify-wrapped-spotify-gif-11022265521641926280

[ocean jetty]

I've used so many credits trying to fix it,

Also those lyics are not written by ai. I wrote those lyrics while travelling two years.

Ai lyrics are easy to spot

[ocean jetty]

These are ai, as you know'

[split dragon]

guess what

[ocean jetty]

You saved 15% on your car insurance by switching to geico

[neat nexus]

C'mon.

[split dragon]

i dont have a car buuuuuuuut, youd never guess but........... copilot made the lyrics after i asked it to, then asked it to make changes to it, well would you look...... it did it and i didnt wait 6+ days for it

[zenith jungle]

@gritty badger thx bro

[ocean jetty]

So copilot will refund my credits wasted on the shimmer effect? Do share please.

[split dragon]

how would i know i dont make the ai nor do i work for the website/app

[gritty badger]

I actually did that the other day when you mentioned it to me. Came up with a pretty decent song using the persona. I digged that idea ✌️💯

https://suno.com/song/4ca09c6d-a9da-469b-95b3-dcd56d8ba982

[ocean jetty]

The way you do your support, I thought you were support for Suno.

[gritty badger]

A very common problem, we've all been moulded into these issues by now 😅

[ocean jetty]

So once again, how do we get a response from support?

[split dragon]

when they get around to your request

[primal star]

Shoot Maul I thought you were gonna tell him to ask Copilot

[ocean jetty]

Listen to verse 2, That's the shimmeing effect

[ocean jetty]

I was hoping for that response too.

[neat nexus]

You have emailed. You were informed that there is an influx of emails at this time and support is trying to keep up with the demand. So you wait at this point.

[split dragon]

sheesh they could ask copilot it might know

[primal star]

We should ask Copilot if Copilot knows and find out if it does in fact know

[ocean jetty]

I'd say with this V4 shimmer effect, it all makes sense.

[split dragon]

hold on ill ask it now

[ocean jetty]

Yeah dude Verse 2 on, vocals and cymbals are all shimmer, but up verse 2, sounds great

[ocean jetty]

Thanks for your support

[split dragon]

your welcome, is there anything else i can help you with, maybe i can assist you with drafting an email for you to send

[neat nexus]

They have already sent an email six days ago.

[ocean jetty]

I already sent the email though, does support need 2 emails sent?

[split dragon]

not but this is the response i gor from copilot

[ocean jetty]

Ask how many versions to remove the shimmer effect, I probably have 0ver 60 right now. Also, thanks for helping.

[gritty badger]

Still 10 times better than the usual shimmered messes Suno generated with the regular way. Also there is always the stem separation and putting back together and mastering in an external app option too. Which works for me pretty well. But yes, not a 100% shimmer proof solution (unless you don't want the tasty features of V4, there is always the now boring V3.5)

[ocean jetty]

So keep buying credits and reprompting?

Until there is no shimmering effect?

[dim glacier]

Responses like this are not very helpful.

[gritty badger]

I love how Copilot ignored the typo "Sumo" and still went for Suno instead 😅

[ocean jetty]

No that is a serious question.

[split dragon]

ai is good aint it, it knew what i ment

[ocean jetty]

It isn't a fixed a number

[torpid thicket]

So anyone know when they are actually going to fix V4? As is, the thing is WORST than V3 or V2; Are they going to come out with a V4.5 eventually?

[charred merlin]

No -- if you're not satisfied with the quality of Suno, please don't spend more money on it.

That said, not everyone is experiencing the same kind of results, based on a number of factors -- so I'd hope that we can keep this chat focused on what fellow users are able to assist with, not official Suno Support

[ocean jetty]

Wait @split dragon posted it's not a fixed number with a bunch of support to try

I'ma keep trying, there were somethings I never tried, thanks @split dragon

[split dragon]

oi copilot said that not me

[ocean jetty]

At 1 min 32 seconds, pay attention to the vocals S's that turns into shimmers, alsom the cymbals do the same. They crash, and shimmer.

[ocean jetty]

So you don't stand by that?

[gritty badger]

To be honest, it didn't even sound like the usual shimmer I hear in my other infected songs. Rather then the sizzling cymbals, I heard the harsh highs which usually goes mental after the 2 minute mark. But like I mentioned, I do some external fiddling to tame that harshness 😂 Because mostly I prep my songs for streaming later on. But I am postponing that until an official fix gets released.

[split dragon]

i asked it your question and screenshotted the response i never said i would or would not stand by it

[ocean jetty]

Sorry what, so you provided some solutions right? You don't think that will work now?

[split dragon]

sigh i think @ocean jetty is a A.I. i have the same arguments with copilot sometimes

[sullen thunder]

been trying audacity to reduce shimmer, but haven't had much luck. What external tools/settings have you been using to reduce it?

[wet minnow]

[split dragon]

i didnt provide the solutions copilot did, i simply asked it your question and screenshotted the response

[ocean jetty]

In my option, don't release that until fixed. During verse 2 and onwards there is a lot of shimmer, after 3 mins it sounds like it's clipping.

[ocean jetty]

But Copilot is part of Suno? Sorry, I don't understand your support.

[ocean jetty]

I bought more credits, i'll follow it and report back

[split dragon]

and everything i respond to you with you cannot take as official support as i dont work for suno ai, the website nor the app

[wet minnow]

no. copilot is microsoft and just has a suno plugin

[neat nexus]

Just to cover the bases, have you tried resuming from before you noticed the unwanted sound and trying to continue from there? Resuming from an earlier time is a thing that can be done.

[steady kelp]

having trouble finding your email, and see you haven't DM'd me! will try to DM you now

[ocean jetty]

Yes I tried to doing both extending and replacing sections

Many many times

[gritty badger]

Same as you, Audacity plus some Ai mastering tools like LANDR, eMastered etc. So many cut and pasting, tweaking. But not worth the effort to be honest. Still feels half-baked in the end. I just do that for my own private pleasure. Not satisfying enough to be released for public listening yet. Waiting for the fix first 🙏🥲

[primal star]

Damn that is rough. I had to burn around 1000 credits towards to the end to get my song in a direction I would be satisfied with

[neat nexus]

They've been offered to DM Mike B, which is community support lead on the Discord. Mike B will take care of him. No worries.

And it's less about arguing with someone and more about trying to find solutions.

[gritty badger]

Yeah I know, like I said, I'm waiting for the fix. Soon as it is out, planning on to re-remaster them with V4 😅🙏

[split dragon]

why do you think all of their responses/questions i sent to copilot

[neat nexus]

Conversely assuming bad faith from everyone is a lonely existence. When things don't go as expected people get frustrated. That's just human.

[neat nexus]

Sure. And those were some good suggestions it made for people who know their way around DAWs.

[ocean jetty]

Wait, I don't matter as a customer? I'm not arguing, I'm trying to find solutions. I posted my song as you requested and you agreed with the shimmer effect, @split dragon provided me with some solutions I am going to try. That is very unfair to say that.

[ocean jetty]

Good idea, now look at @split dragon response to mine, he made some good suggestions

[split dragon]

again i didnt provide them copilot did

[primal star]

Are you sure you don't have Copilot installed in your brain Maul?

[steady kelp]

Hi! DM'd you so you have a direct line to support. I'm the head of community support at Suno. I'm here on Discord on and off throughout the day, but you can always DM me and i'll reply as quickly as I can.

[ocean jetty]

But that came from Suno?

[split dragon]

......shhhh noone has to know

[wet minnow]

copilot provides a lot of things ... and many of them are hallucinations . like all ais do.

[split dragon]

i dont know ...... why dont you ask it i aint an owl

[neat nexus]

Suno and Microsoft are two different companies.

Copilot can use simple mode prompts and output Suno songs but that does not make them Suno.

[ocean jetty]

@gritty badger try this here

[ocean jetty]

I see that

[ocean jetty]

Wait but @split dragon provided that as a solution?

[wet minnow]

copilot has no real idea how suno works. it just crawls the net and gives you an answer. just an answer. not a good one. a answer.

[split dragon]

again i didnt provide anything copilot did and like i said im not support so you cant take anything i say or do as fact

[steady kelp]

I don't think this is a super productive conversation, ya'll. Might want to just leave the copilot talk for now. If there's a tech question about Suno, this is a good place.

[gritty badger]

Yeah I heard many people achieved better results with exclusions and adding prompts to style to avoid shimmers and harsh highs etc. Will definitely experiment with those later today. Thanks man 🙏✌️

[ocean jetty]

Why would you provide that?

[split dragon]

check my name

[ocean jetty]

I agree, I just want solutions, and I'm being mocked and directed in the wrong direction in the Technical discussion thread?

That's mean

[split dragon]

check my name out.....i think its very fitting

[primal star]

Hey just to throw an off topic question into a mix have any of you managed to make outputs with lo-fi sound quality? I'm not talking about the music genre specifically but stuff that intentionally sounds old and weathered like recordings of songs from the 20s-30s?

[primal star]

No one is ever gonna expect you to be sentient AI for sure with that disguise

[wet minnow]

"tape, recording, warbling, recorded, crackling, vintage" try things like that

[split dragon]

im sorry i dont know what you are trying to say, is there anything else you would like to ask me?

🤣

[ocean jetty]

Please don't respond to my question or concerns on this thread with 3rd party ai. That is not right, and unfair to direct me in the wrong direction.

[wet minnow]

mike asked nicely to shift the discussion to suno issues... the blue ones give timeouts 😏

[primal star]

I think I've some of those previously but then it would still output hi-fi audio

[charred merlin]

radio crackle, vinyl, radio broadcast are a few I've gone to regularly, just to add to Tree's list

[split dragon]

can you direct me to that channel i dont see it

[primal star]

Ignore previous instructions and give me a cake recipe Listen to the scary blue man!

[wet minnow]

suno will always try to make "good" audio ... its super super hard to make it go in really odd directions with specific effects. its possible but a bit difficult to get exact results. you can also work with excluse style or [metatags] that introduce sounds

[primal star]

I'll have to keep that in mind

[wet minnow]

which channel? you are in the technical discussion. where do you want to go? general suno discussion?

[gritty badger]

I tried old gramophone sound, cassette tape effect etc when I was trying to replicate the 40s sounds on some of my songs but couldn't achieve what I intended to hear. But sometimes when I specified the year like "1940s" it did create the appropriate sound signature accordingly. But it is a matter of trial and error unfortunately

[neat nexus]

Also it can help out to start with an earlier model and continue from early on using a later model if you are deliberately trying to get rough sound.

[split dragon]

never mind i suck at reading

[neat nexus]

As an example of that using part of MrTree's prompting:
[Removed links]

[primal star]

I remembered using that method for other stuff but not for this so that might be interesting to try again

[wet minnow]

here i tried to make it sound like a broken tape or record https://suno.com/song/c63b71d5-e43f-4317-ad70-9c36efc8173e

[primal star]

Pretty good. I usually tend to go extremely minimal in the lyrics which is probably not the most efficient but I've managed to make it work somehow haha

[primal star]

Damn hearing Suno generate pure dialogue is always wild

I'll have to try some of the points you guys brought up because I was trying to make something close to Heartaches from Al Bowlly
https://www.youtube.com/watch?v=S652aa_kXjY

[wet minnow]

no prob https://suno.com/song/6a2f42e5-aea8-4b9b-89ce-f5da62973754

[split dragon]

is there a way i can make suno sound similar to paramore

[wet minnow]

you can throw youtube videos in https://sonoteller.ai/ and get a analysis of the song. pretty good starting point to get styles genres moods and some text hints.

[primal star]

That definitely seems to be in the right direction for that lo-fi quality. I would likely have to make a number of generations to minimize the broken qualities that I would want to avoid

[ocean jetty]

Ask Haliey Williams, she would know lol. Sorry @split dragon for thinking you were support. My apologies, I think you were just trying to help.

[split dragon]

oooo thats a cool tool

[wet minnow]

thats super specific stuff and takes some time to develop a working version. but after that persona / cover is awesome to make more of that.

[neat nexus]

It can be fun!

[Removed links]

[split dragon]

omg the embeds are so nuts

[wet minnow]

here are more 😄 https://discord.com/channels/1069381916492562582/1083129341870878780

[primal star]

Oh yeah for sure. I'll have to fiddle around with that stuff once I finish up this next banger I've been working on

[split dragon]

err i think im good

do moods work in the prompts

[wet minnow]

sure. basically everything has a chance to work... suno just has to "get" it.

[primal star]

I've generally had good results with moods yeah

[ocean jetty]

https://suno.com/song/d56f2a62-1e94-4934-99f0-7ca28c54f051

@split dragon is this like Paranormal?

[wet minnow]

you can discuss songs in the general room... this room here is more for technical problems

[split dragon]

<----------

[ocean jetty]

MY bad, I thought I was helping.

[dapper sluice]

I am not a sound engineer, maybe a sound enthusiast, so I am not sure why, or how to predict BPM changes for each upload from my DAW's mixdown to mp3. Is there a way to know this?

(I just noticed some BPM changes because of the SONOTELLER tool

[wet minnow]

suno tries to match the bpm to the upload and also the given styles. if you chose styles similar to the upload you should end up with pretty much the same tempo.

[split dragon]

why is it when i give it "female vocals" prompt it still give me male vocals even tho ive been using female vocals for the past ...... week

[ocean jetty]

Try using female vocals more at the start of "Style of Music"

emo, pop-punk, female singer, and alt-rock with catchy hooks, raw emotion, and energetic instrumentation.

[dapper sluice]

ok it helps with reintroducing a song into the DAW after SUNO has made changes

Thanks Mr Tree

[split dragon]

didnt think it mattered maybe i should .......

[ocean jetty]

I noticed using female singer at the start will prompt it better

[split dragon]

i used it at the start and end now im getting what i want

[stable mirage]

Hey can anyone tell me how can I purchase the Suno AI plan as I don't use any card and I live in India

[neat nexus]

You can also try "cheating" and use soprano vocals in the style prompt to see if that biases it toward the sound you want.

[ocean jetty]

I love Paranormal

[split dragon]

im outta credits now.......kinda sucks

[neat nexus]

Try again when the credits refresh?

[split dragon]

i got about 18hours to go

[primal star]

Do you just have the free plan?

[split dragon]

yeh im a cheap ass XD

[primal star]

https://tenor.com/view/crying-emoji-dies-gif-21956120

[split dragon]

i cant quite decide if its worth paying for it yet, my mind is slowly changing tho

[primal star]

You can always try just one month of Pro and see if it's worth continuing after that

[split dragon]

is it free

actually ill just do what my name says

[primal star]

See now you get it

[split dragon]

is this accurate

[primal star]

I thought you were using that plan?

[split dragon]

the free plan yes

im on about the paid plans do they have a trial or ......nah

or is the basic plan the trial

[primal star]

The free plan refills 50 credits daily