📚┃suno-school

[foggy shale]

oh duh thanks

[spiral bane]

site borked 😮

[hollow hare]

Can you be more specific. Error 5xx? Credits disappearing? Etc.

[spiral bane]

A timeout occurred Error code 524
Visit cloudflare.com for more information.

server is down according to cloudflare

[hollow hare]

Ah, pretty self explanatory
Usually fixes itself in a few minutes or use VPN (since it is a regional issue that pops up). I think some people says using another browser works too

[spiral bane]

seems like a network acritechture issue 🙂

[hollow hare]

#📚┃suno-school message

[karmic epoch]

What happened? I just generated 2 songs for 10 credits but it used all of it! I refresh and it returns but then it goes down again.

[spiral bane]

bug in front end

[hollow hare]

Visual bug, counter seems to continuously deduct until your generations are done. (Likely a dev was fiddling with something related to the counter and exerbated the issue).
Your actual credits are fine

[karmic epoch]

😭

[spiral bane]

devs pushing to prod without qa testing 🙂

[hollow hare]

Suno devs: screw it, we're doing it live!

[spiral bane]

with a good CI/CD tool (like circleci) you can usually get away with it

but a bug like this would fall through the cracks as it technically isnt broken the code is doin what it should just too often

🙂

[unique stone]

It's possible they might've been trying to fix the issue where the credit counter sometimes displays the wrong number of credits, but made it worse in the process 😄

[charred merlin]

https://tenor.com/view/fix-bug-in-production-fix-bug-production-ti-it-gif-13502120

[foggy jasper]

has this been addressed yet?

@glacial scarab confirmed the issue as well

[spiral bane]

well, i think i broke suno, i used the replace section feature and now its been giving me the whole song for about 15 min, and for some reason the section is 1.5 min long instead of the selected 30 seconds

1 min 55 seconds long is the sectoin lngth according to my screen *

[stiff briar]

Hello,
I accidentally created two accounts.
The one linked to my Google Account is where all my songs are, (I am calling this my real account)
The OTHER account is tied to my phone number.
I can't remove my phone number as it is the only form of ID on the account.
I can't delete the account (not an option)
I can't add my phone number to my real Suno account because it is tied to the other account.
Any ideas on how to fix?

[safe hazel]

Contact [e-mail] for account specific questions

[stiff briar]

Thanks

[charred merlin]

☝️ They're also the people you need to contact if you want to delete you account

[stiff briar]

I swear I looked all over the site and couldnt find any contact info - Thanks again.

[charred merlin]

There's an email listed in fine print on the account page, but there's a new help knowledgebase 💪 https://help.suno.com/

[lime igloo]

Hello, I’m using the Suno API intensively to produce music, but my accounts get banned after a while. Will these limitations be lifted if I upgrade to a Premium account?"

[charred merlin]

Not sure where you're getting API access from, but you're near-certainly breaking ToS by using it.

[lament cove]

Suno API doesn't exist, they've not even released it

[unique stone]

Suno does not currently offer API access, regardless of the kind of subscription plan you are on. You are only meant to access Suno using the website or official iOS app. Accessing it any other way is very likely to break the Terms of Service (as theharber mentioned) and would likely be why your accounts are getting banned.

[frigid orbit]

Technical discussion eh? I guess a genre would be technical in terms of music. What would this best classify as? https://suno.com/song/f1c44a37-f1b8-4ae8-9f5e-4ab18f07fe5c And would also like the #1154790975240155247 as well. I personally listen to metal when on spotify, youtube etc. Hell half my leg is covered in a slipknot tattoo. So I need help pls

[chrome ore]

hello. Why i still can't create a song ? always error message : a system error occured. Can I fix this ?

[hollow hare]

Unfortunately that error is way too vague for us without further information.
Please email [e-mail] to investigate if the issue is on their side (which it likely is).

Note that business hours are in Eastern USA time so they cannot answer right away because it is Friday evening

[lusty osprey]

Hello. When I select the "change Section" option, the page return me to home page. Anyone know what's going on?

[void spade]

i'm using the free version, can i monetize in spotfy??

[unique stone]

No. Songs created on the free plan may not be monetized and you have to credit Suno if you use them in e.g. videos or upload them anywhere.

https://help.suno.com/en/articles/2425345

[distant carbon]

Try to do it phonetically as in as it sounds. Now of course I do not speak Ukrainian but there must be a way to do it phonetically so "for u" could be "four U" or "for U" or "4 U" etc in English. Additionally sometimes certain vocal styles used have a problem singing certain words whilst other vocal styles have no problem, it really is random & will be down to how the training was done on the LLM for Suno.

[polar spade]

CloudFare time Again.

[polar spade]

Be careful. If it's pronounce, then you may have "for yu" .... just "U" is chance to be pronounced "oo"

[distant carbon]

Refreshing the page usually solves that. If you have multiple tabs up they can show different credits anyway as they are not updated live. But if it's a bug then it's could be something specific to a browser - maybe an extension. Could really be anything, i have not had that problem.

[polar spade]

. . . . And, regarding your use of songs you create for the Timbaland contest, you are agreeing to an augmented contract adjunct to the TOS which limits your use of those songs covering Timbaland stems.

[rose jasper]

And it's only for U.S residents

[split escarp]

Can't access the main channel

Can't open Create. Library, nothing

It's not a connection issue as other sites work fine

[polar summit]

Hello, I have an issue with payment. Suno don't want to billing me with my card, twice.

It's odd. It doesn't have issues before

[hollow hare]

Have you tried another browser, like Firefox/Chrome? Does the issue only happen when you log in? You may need to contact [e-mail]

[hollow hare]

Contact [e-mail] (may also need to contact the company of your credit card or the system that Suno is using for their payment system, billing may advise you on that matter)

[polar summit]

Idk how to contact any of them

[ebon vessel]

Is there any particular reason Suno playlists don't have a 'shuffle' feature? It's the one thing keeping me from using them

[south rapids]

#🪲┃bugs-web message
Its back now

[hollow hare]

• I literally gave you the email to contact the appropriate Suno department. E-mail [e-mail]
• Ask [e-mail] how to contact the company of the payment processor if they believe that the issue lies with them.
• Do you have a credit card? Look at the back of the card, it has a phone number of contact that company.

[brazen sparrow]

I’m a free user and have a recent song that I think are good. I know I can't use it commercially, but is there any way to do that? Or is my only option to upgrade to a pro plan or premium plan and use Suno Cover to recreate the exact same song? Any tips for that?

[hollow hare]

After subscribing, you can ask [e-mail] if it is possible to have the license granted for the particular song.

If they do not grant it to you, you will need to recreate it (extending will not grant you the license as it would be considered a derivative of the original, which made during the free plan, you do not have the necessary rights).

If you need to recreate it, chances are you will make a better version using what you know now. But usually support should be good about this I think?

Just be aware it is currently the weekend thus they will not be able to get back to you for a few days

[lost scaffold]

For the Timbaland contest - I've noticed that when you recreate a section of the song, the icon that usually shows below the cover art indicating it's a cover for Love Again is missing. I have a few songs I've submitted but because they don't display the icon, I'm nervous they won't be included in the contest. Please advise. Is this a visual issue?

[jade nebula]

I have about 11 remixes and the same situation. The only ones that have it are the ones I haven't edited much.

[glad cobalt]

question ??? Or you can make it so that in the track title line, instead of Untitled, the nickname of the users and the date of the track, or the date and the track that he created today, are automatically entered. example 1: rokrest 20204 10 26-1
example 2: 20204 10 26-1 last.

The number at the end of the name indicates an attempt to create a track, and if the author wishes, he can rename it first.

[hollow hare]

I love the idea of ISO-8601 being used as the default title placement. Unfortunately, this is not the place to make suggestions, you definitely should forward it to #💡┃suggestions-web (or maybe even #🚧┃suggestions-layout ?)

[glad cobalt]

Sorry, for some reason I thought that this could be done here.

[hollow hare]

It's all good 🙂 Oh, you may want to fix the typos in the 2024

[unique stone]

Heh I was just breaking out mspaint to make a similar illustration 🙂

[glad cobalt]

https://tenor.com/view/mr-bean-panic-got-to-get-to-work-in-a-hurry-gif-11642828

https://tenor.com/view/三度c-smack-slipper-3dgree-gif-23958998

sent a message there if it clogs up the discussion wall, then I’m ready to understand and get a slipper for flooding in this thread

[charred merlin]

This chat has April 2024 vibes atm, I love you guys.

[unique stone]

I'm slowly renaming my tracks manually to have this kind of timestamp in the title, after a short label so I know what song they relate to (I switch between 2 or 3 different ideas.) It'd be nice to have a choice of automatic title formats to suit user preferences.

[charred merlin]

No chancla for you -- this is a solid idea.

[polar summit]

I found the issue. My card has reached his limit.

I forgot to tell my helper to empty it.

?

[charred merlin]

If your card is declining, but you believe that you still have enough money available, you need to contact your bank, or whoever gave you the card.

For issues related specifically to your account billing, you can contact [e-mail].

Is there something else you're wanting help with?

[polar summit]

Don't worry, I have plenty and I get the billing today. Thank you for your help.

Actually, like I say, I reached like the limit of my credit card and forgot to tell my helper to empty it during that month. So... that's why Suno didn't wanted to billing it yesterday

[charred merlin]

I see -- you followed up a while later with just a question mark, so it looked like you had something else to state or ask about.

Fingers crossed all gets worked out regarding your payment 🤞

[polar summit]

Yes, it did it. I get the billing.

So, thank you for the help.

[faint wraith]

Please help i cannot download any of the songs in wav format

it just says download failed and i have tried multiple browsers

[charred merlin]

If it's relevant -- we've noticed that .wav download doesn't seem available for songs stemming from an audio upload.

Beyond that, if you're on a chromium-based browser, try using Firefox, as that seems to resolve the matter in most instances I've tried.

Hrm, if you've already tested with Firefox, it might be worth reaching out to [e-mail].

[south rapids]

And there it goes again, Suno down for anyone else? 😎
A timeout occurred Error code 524
Visit cloudflare.com for more information.
2024-10-26 15:15:01 UTC (Helsinki, both times today)

[rain flint]

Suno has poor sound quality, why not generate vocals and music separately for subsequent connection or changing each separately, This would be the breakthrough of the year!

[south rapids]

ITS ALIVE! (again) 😅

[restive valley]

Nah, it's still down

Alright. Took a while. It's up.

[hollow sapphire]

Suno is unable to load! 🚦
On all browsers...
Same for you?

[calm estuary]

Ah okay so its not just me.

[fossil ore]

I've noticed that there is a significant amount of background noise to my creations; some worse than others. Wondering a)if there is a prompt I can use to prevent this and/or b) will Suno improve on this so there is no noise. Thanks

[south rapids]

You can try to add

Soundtrack Master

to your Style of Music instructors. Other than that I think most times when there's a lot of noise (other than the devs most likely is working on fixing it) it would have to do with transitions between sections in the lyrics. Like, in a chorus, there are often, I dont know the correct terms, but I go with "air" like singing in church things like that.

[wheat prism]

Good morning.
The Style of Music description does not distinguish between male and female vocals. Despite entering "male voice," "male," "male vocal," "baritone," "male baritone," and other similar terms in various combinations, the result was a female voice. This is unacceptable and a waste of time. The credits spent today are all wasted!

Furthermore, the vocal shaping today is severely lacking. The voices not only confuse male and female but also sound like amateur singers who have accidentally stumbled into karaoke and are simply pulling one note. This is unacceptable.

[south rapids]

Share how it looks I can give you some tips

[south rapids]

Another one 😎
A timeout occurred Error code 524
Visit cloudflare.com for more information.
2024-10-26 16:07:18 UTC

I guess there are many that's making does Timbaland Remixes 🥰

[hollow hare]

From what I had seen here, this tends to be localized. Some solutions you can try when you get this error:

• Wait a minute or two then refresh. Worst case for someone was like 10–15 min
• I heard from some people that switching to another browser helped (e.g. Firefox, Chrome)
• If you have VPN, that will definitely help

[wheat prism]

I am unable to provide you with a link to the song at this time. It is still in development, and I am not able to share a work that is not yet complete. Additionally, the copyright has not yet been registered.

I believe the issue may be a Suno bug. I would like to request a refund for the unused credits. Please advise if this is possible.
Could you kindly let me know when this bug is going to be fixed?

[south rapids]

Thanks, then it has happen I have not been alone, so dont think its on our end. Seem to effect Europe the most due.

[south rapids]

Not sure what you mean by most of that.
The Style of Music has the instructions, simply take a screenshot of that, or show the text.

[safe hazel]

You not liking the output is not a bug

[hollow hare]

Oh, definitely something finicky between the server Suno uses, for sure. I think when a major 524 happened before, people in Germany and California were both affected, somehow

[south rapids]

It says the host is the problem, meaning 99% sure its a resource issue at Suno. Also understandable considering the Remix challange.

[wheat prism]

Here's a screenshot. But I've already written above that I've tried different options. So what's wrong here?

[cunning siren]

Not for me 😦

[wheat prism]

Are you talking to me? But I didn't write that, whether I like the song or not! I gave a specific description of the error!

[safe hazel]

Yes I am. It isn't an "error"

You are not guaranteed to get exactly what you put in the style box

[south rapids]

Have you tried
male voice
male vocals
male emotional vocals

Also, what's in your lyrics has a higher priority than Style of Music. This means if you write about something sad Suno will prioritize that even if your Style of Music say "Happy" In terms of it being a "bug" I dont think so, its just hard for the devs to figure out how to make it more reliable.

[south rapids]

Not me neither but it was a while sense that crash, this is new "crash" 😎 Are you also in Europe?

[cunning siren]

Yes I am

I am being routed via Frankfurt

[south rapids]

Me to, more or less. However it says on the error the fault is at the "host" meaning Suno.

[hollow hare]

As it stands, good luck getting copyright status. The law move slowly and US Copyright Office had already deemed the output from the use of AI itself are not eligible for copyright. (Though with some company bribing donations, that can change in the future.)

This case (involving Zarya of the Dawn) was quite interesting. In the end, the artist could copyright the title of the comic, the story, the name of the characters, the arrangement of the panels/comic book, etc. but she could not have the appearance of her character registered for copyright as it was generated from an art AI and thus had "insufficient human effort" (or something like that). If you want to know more, see:
#💬┃general-chat message
#🗣️┃chill-chat message

That being said, you should be able to register the lyrics, the mastered audio (with "sufficient" human involvement), etc. But likely the Output from Suno itself would be ineligible for copyright protection. Something you can inquire about from your local intelligence properties office (e.g. US Copyright Office)

[cunning siren]

True, this kind of things happen lately a lot. They should consider to scale up a bit. On which cloud service provider does suno operate? Is that known?

[wheat prism]

Yeah, I've tried all that. As for the Lyrica, there's no indication of a female voice. But I also tried to specify a male voice in the Lyrics, but the result was also deplorable: a female voice sang the main text and sometimes lines were inserted by a male voice.

[south rapids]

Think they are working on that too. They also hires a lot more support ppl that are under training (if I got that right) the reason for the crashes now is most likely to the remix contest with Timbaland

[wheat prism]

Thanks for the useful comment.

[cunning siren]

Ah yes that could be but it was forseeable and they could have prepared themselves to it ... not really professional 😦

[south rapids]

I see, well I would have to see them to help, but even so, there are times when it does matter what you do. I've spent 1000 upon credits at times before giving up.

[south rapids]

Yeah, but then again, probably cost lots of money doing it? image nobody would be interested in doing the contest. Its also the weekend today so ppl dont work, I dont think

[wheat prism]

Okay, I'll have to write to support again.

[safe hazel]

Support is unlikely to refund you because you wanted a male voice but got a female one

[cunning siren]

Ok, looks like you have waaaay more understanding for them than me 🙂

[south rapids]

She is probably just looking for a solution 😎 and I get it, when you know what you want and Suno dont get it, its "meh" ...

[safe hazel]

Support probably wont teach them how to prompt better either

[south rapids]

Maybe because the competition out there is not even close to this super Suno 😉 plus, Mike the boss at support is a cool dude.

Suno is back!? at lest for me

[cunning siren]

Yeah, looks like. Maybe now a different region is experiencing outages 😄

[south rapids]

Yeah, fair is fair 😎😅

[cunning siren]

This is for us to shut up and not complain that much haha

[south rapids]

BTW, check this out https://www.suno.wiki/

[wheat prism]

I read all the info a long time ago) But thanks - it might be worth browsing to see if there's anything new that I haven't had a chance to get to know yet.

[hollow hare]

Have you tried having additional male keywords to provide more weighting? E.g. Add "male vocals, male, male, men, man", etc. In the style prompt (more terms = more weight)
(For info about weights, see #💬┃general-chat message and Harber's demonstration with images at #🗣️┃chill-chat message)

[Male] metatag variants in your lyrics may help as well (though, in my experience when working with duets, the AI swaps the gender like half the time when a duet does work...). That was how I was able to turn a female-only style (Girl group) to use male vocals (albeit with a 25% chance, came up in the second gen in just one track):
https://suno.com/song/4578b480-f46c-47ac-a414-53b5cada6dc7

In that example style prompt, I would try something like:
"Clear male vocals, male, men, emotional [genre]"
Having the male keywords first for more weight

[winter thistle]

please🙃

[cunning siren]

Welcome to the relay race 🙂

[brittle furnace]

yeah it's down too here, sadly

[winter thistle]

my gosh is three day or more like that

[cunning siren]

According to @south rapids we can blame Timbaland for it 🙂

[brittle furnace]

oh really, for me i just experienced now

[winter thistle]

oh sure is three days that at least 3/4 per day like that

[south rapids]

Best thing would be for us in Europe to get things "done" in the morning.

[calm estuary]

I was able to get on it in one tab, but now when i try to get on it in another tab its failing.

[cunning siren]

Right now it's working like a charm for me. I even get instruments that I never had dreamed of to get like Timpani

[south rapids]

Check out my remixes posted in the #🐣┃suno-showcase message
Well, that would be when Suno gets back again, its down for the 4th time or more today for me 😥

[brittle furnace]

i see, seems sporadic, it went well for me for last couple days, guess lucky here

[sage spruce]

I am from Indian and I'm not able to buy premium of suno due to card failure. Is there anyone who can help

[south rapids]

Maybe, but need more information than that. What error do you get?

[cunning siren]

Will certainly have "an ear" 🙂

[sage spruce]

I enter my card details and after entering otp it just shows card decline in red. I have enough in my bank account. It may possible that my card cant pay I dollar, if there is any way I could pay successfully then let me know.

[hollow hare]

Your first point of contact should be talking to the bank/company of your credit card. Are you sure you have sufficient credit (after conversion + fees)?

For issues related specifically to your account billing, you can contact [e-mail]. They work on eastern US hours so you will not get a response for a few days

[south rapids]

Does this help? 🤔
https://help.suno.com/en/articles/2421185

[sage spruce]

I use a debit card for the payments and I had contacted them but didnt get the solution.

I am using mastercard

[south rapids]

Here is is your best option. Go to www.perplexity.ai and start asking it to help. It will find all there is to find on the web for you (might take some time) so be patient, its free.

[wheat prism]

Thanks for the detailed reply) Yes, I tried it that way too. But for me the percentage was equal to 5-7% of successful attempts to realize the male voice.😢

[charred merlin]

Unfortunately there isn't an official support presence on Discord, so no one will be able to provide immediate assistance.

[e-mail] is the best way to move forward for official guidance.

[wheat prism]

I saw it too...

[sage spruce]

@south rapids ok let me try that. But in case it didnt work I can try to pay you for my subscription so that you can bye it for me

*Buy

[south rapids]

I get where you coming from but it would be a huge security issue on many aspect. Get yourself a paypal account, this will work.

[sage spruce]

But paypal is not accepted for payment in suno

[south rapids]

Your right my bad, Google Pay, Apple Pay, and Cash App Pay
use google pay then

[sage spruce]

Thats also not accepted

[south rapids]

Yes it is? says so on the URL I gave you

[charred merlin]

Correct -- also, I'm not sure that an AI would be much help, like perplexity or chatgpt, seeing as how they hallucinate answers or might even tell you to "call the hotline" despite there not being a phone number.

In the #1154772025940922438 section there's different regional areas that might be able to give advice on how they paid for their country.

[south rapids]

you cant compare chatgpt to perplexity...

[sage spruce]

@south rapids its not available currently in indian version

[south rapids]

How about the other mentioned?

[charred merlin]

In terms of asking an LLM for direct advice on how to solve a billing issue with a specific company, asking any LLM is equally as ineffective when there's a way to directly contact support agents -- this is coming from a dude who loves AI, but not all use cases are created equally.

[sage spruce]

The only option I get while paying to suno is credit or debit card with "Link". I guess I have to wait untill they introduce paypal

[safe hazel]

@sage spruce Just to go back for a moment to when you suggested that you give someone your info and they buy it for you. You should NEVER give your card info (or any info) to strangers on the internet

[south rapids]

Those aren't the terms. Perplexity will find anything avaliable on the web, ChatGPT most def will not. In terms of support being the best option, obviously. Are they working today? dont think so, even if, they would not respond today. Try to be helpful instead of questioning things you dont understand, thnx.

[charred merlin]

try to be helpful
👍

[sage spruce]

I am not talking about card info, I can pay through upi or bank transfer

[south rapids]

Yeah, but would still be the one entering my credit card details on your account. Meaning you could use it to subscribe when ever...

[sage spruce]

Yeh! I didnt thought that. I guess I have to wait.

Or try something else

[south rapids]

Like mentoring try Perplexity, at the very least it will give you "ides" on what to do next

[sage spruce]

Ok I will try it, and Thanks every one for the support. This was my 1st time seeking help on discord

[polar spade]

@charred merlin I've got a new prompt (new to me) ... "ensemble" .. More specifically " ensemble rock" .

[hollow hare]

Interesting, I assume it sounds like some sort of school musical band composition?

[polar spade]

Using for musicals. And it's great. I wanted a rock opera and thought "ensemble rock" along with "musical" and "broadway"

It's not a musical ensemble. But vocals. So "ensemble vocals" works as a choral sound as well.

I've got a contest example. It's long though, but just skim through it.

[polar spade]

Oh. Here's a shorter less than 3 min.... with applause and a little babble banter at the end.

https://suno.com/song/b94b6f74-079c-45fb-ae6e-bb4e186e3bf4

But on another song... it used the idea of ensemble, I think, and gave a new voice of three to each verse.

I'll have to test more.

Technically, I'm not sure if this qualifies for the contest. I'll have to review again. I'm "clearing" my songs now that I've gotten some sleep. 😆

[charred merlin]

I'll try it in isolation to see what reliable effects it can produce!

[polar spade]

Try ensemble duet, also just to see if that has any effect on flmaking more duets or even trios

[charred merlin]

ensemble by itself seems to evoke something similar to a background film score, but after 32 separate tracks, I'm not sure that I discern any difference between "rock" and "ensemble, rock"

[sage spruce]

Bro, perplexity didnt did the work. What to do now, I am still not able to purchase the premium from India

[south rapids]

You never answered if you tired the other available payment options

[sweet wigeon]

contact your bank, they can help you

[sage spruce]

There are no other payment options other than credit/debit cards

[sage spruce]

Cards only

[south rapids]

Why have you scratches over one of them? also, Billing adress would not be India but perhaps that just autofill your dident correct?
Also, you get way more value with the $30 subscription YOU CAN cancel anytime, so its all good, if your unsure how long you want to subscribe.

[sage spruce]

I scretch my last digit of card and adress is India cause I Live in India

Nah.. Im good with 10$

[south rapids]

Your digit would not show there? and no, thats not your billing adress, if you only put that it will fail. It should be the actual street and so on like
Streetway 123
City, Region
Postcode

etc.

[sage spruce]

I know what a billing adress is, I had entered the full adress with other details but still now working

Not working

[charred merlin]

@sage spruce -- I implore you to start the process and contact [e-mail].

It won't be an immediate response, but that way you will have taken a pro-active step to contact the only official support channel.

[south rapids]

Then I don't know. Unless you want to trust me in DM with your details not sure how I can help you. You'll have to wait for Suno support.

[sage spruce]

As you earlier said, I thought on that and find that if you do not save the billing details on web than it vanishes completely. So if you can buy that for me will bhe very helpfull

[south rapids]

No, I won't do that. What I suggested is that I remote control your PC and make sure you doing it right. (With your details)

[charred merlin]

😬

[south rapids]

Whats your problem, did you even think before you expressed that? what type of idiot would I be to first tell everyone here what am about to do then scam the poor soul

[sage spruce]

I did that correctly with multiple checks, I don't think there is something wrong. I can breef what I did while Doing the payment.

First I entered the email (18+) with my card number and expiry with CVC, than the name which is written on card with capital letters. Than I entered the adress (same adress which is use for my card) . Enter the captcha and then the otp. It loads for some time and shows "card declined"

[flint haven]

Please contact [e-mail] @sage spruce and do not ask people to purchase the plan for you with their personal cards

[charred merlin]

The advice that I've seen so far has been to 1) use paypal, 2) ask perplexity.ai how to solve the issue, 3) let you remotely connect to their computer to ensure they're handling the matter correctly.

It's not a problem that effects me, but none of those steps seem to reflect the official advice provided on https://help.suno.com. 🤷‍♂️

[sage spruce]

I did that but their reply was not enought@flint haven

[flint haven]

Ask billing for some alternative solution but please do not ask anyone here to purchase for you, even if you are willing to reimburse. It is getting to a scam-like territory

[sage spruce]

I didnt get the help from the bank. They say that there is issue with suno

[sage spruce]

As mentioned, no other payment options are available other that card in India for suno. I understands your Point , it is not Moral to ask someone the buy it altho I am paying back.

But the thing is.. "I AM HELPLESS"

[south rapids]

So? I am not part of Suno support. Im just providing ideas.

[sage spruce]

Yeh ,I got your Point.

I thought it might work

[flint haven]

Unfortunately, we cannot due much to help you than what was already given to you as advice. This is a primarily community-driven Discord server.

Perhaps get in touch with any family members in other countries who may be more willing to help you out.

The only thing I can suggest is that if both Suno and your bank is saying it is not their issue, then it may be the payment processor.

Please contact billing on how to get in touch with their payment processor and try contacting them.

We apologize for not being able to help you here but this is out of the control and above the support level you would expect of almost everybody here as there is no official support here.

[sage spruce]

Thanks for you Advice and help

[open trench]

Hey! I would love some feedback from the community here. Where can I find a list of freelances experienced in Suno? I have a project for a non-profit that I'm needing assistance with that's just a bunch of suno work.

[hollow hare]

I mean, you can try creating music and asking for feedback or ask for tips with your generations at #💬┃general-chat

Just in case: mind Rule 4 as listed in #📜┃rules

[open trench]

I read the rules before posting. 🙂 I’m pretty familiar with Suno myself. What I’m looking for is the broader communities knowledge on what places have been successful to find freelancers who also get it.

[charred merlin]

Can you elaborate on what you mean by freelancers? Technically all of us are freelancers, since none of us have been hired by Suno to make music 😅

[vocal belfry]

Hi all, huge fan of Suno. As an infrastructure engineer with little knowledge of AI web applications, I am beginning to learn about AI models, etc. Do you guys have any technical blogs, diagrams, of how Suno is built and how all the web components work? I would love to learn!

[hollow hare]

Nope, the company did not release any of that information. Microsoft and Meta have github entries available for their works if you are interested

[tawny canyon]

Does anyone know what's up with the song image generation? Yesterday the resolution took a huge jump up and they looked much better than they ever have, and today it's back to normal. Super disappointing. Here's the comparison: https://imgur.com/a/zcDjmKQ

[vocal belfry]

That's cool thanks for that!

[charred merlin]

While Suno is closed-source software, but back when it was chirp, it was developed alongside their text-to-speech model bark, which they've open-sourced 👀

https://github.com/suno-ai/bark

While it's a different implementation from Suno.com as it exists today, it may serve as a good place to start 🙂

[buoyant gull]

I got bark to work on my machine but it's pretty slow on a 6GB RTX2060 and requires some settings to be tweaked. It's good for learning how the model works but not how it was trained. For that I think you may want: https://github.com/lucidrains/audiolm-pytorch

But for a full-noob, you may also want to start with simple Neuro Network concepts also. I know I had to at first to get the fundamentals of larger things like LLMs, Diffusion/Transformer types. Is all super dense material but pretty eye opening once you get it.

[vocal belfry]

very helpful! I work with AWS a lot, and I am now learning about LLMs, etc. I have a 4090 so, I might as well try to use both my technical background and my GPU to learn about music gen!!

[misty dust]

Can covers be used if I want to keep the song's instrumental but have the vocals sound like they're actually singing instead of just talking?

[lost scaffold]

PING - anyone know?

[charred merlin]

The last section of the contest rules say s to contact [e-mail] for questions about the contest 🙏 I'm not sure anyone here knows

[rich timber]

Now, Suno is singing my prompting for [vocal style] and [Instrumental].

I used these prompting to communicate with Suno and Suno is NOT listening 😐

[Vocal Style]
[Super Suno Prompt:
"Soft, airy R&B vocals]

[Instrumental Style]
[Super Suno Instrumental Prompt:
"Punchy"]

[Lyrics Begin]
[Verse]
I don’t love you no more,
I don’t care about you crying on the floor,
I don’t need you no more,
I don’t want you to love me no more
[Lyrics End]

What can I do to get Suno to STOP singing my prompts?

[safe hazel]

Use fewer words in your prompts

[lapis jewel]

try [{....},{....}]

[safe hazel]

you dont need [Lyrics Begin] and [Lyrics End], this doesnt do anything. There are just a bunch of unnecessary words in your prompts

[rich timber]

Thank you

[charred merlin]

Zothmire raises a good point 🙏
https://www.suno.wiki/faq/metatags/

[night hemlock]

why does suno sometimes completely ignore the first verse and IMIDIATELLY jumps onto the chorus? Is there a way to prevent that from happening??

Like sometimes the ai would just ignore the first part of the lyrics

[lost scaffold]

If you use structure like [Intro], [Verse], etc. it will help tell Suno how and where to begin. If you use too many or too few lyrics, it can have an adverse effect.

[night hemlock]

Thanks

[violet mesa]

Suno down... again

[final hazel]

works for me

Give it a few minutes

[void spade]

Is the paid version monthly or lifetime?

[unique stone]

Anything created during a paid subscription can be monetized/used commercially forever, even after your subscription ends.

[void spade]

And the musics i created before, can be monetized?

[unique stone]

No, only the ones created whilst subscribed.
But you can write to [e-mail] after subscribing to ask for permission to use the older songs for commercial purposes. They might let you.

[void spade]

The musics i create ARE really awesome.

[void spade]

I converted the price in dollar to my country's currency and its Impractical 😭 😭 to me R$540 annually. Suno needs to create other prices/plans for different countries.
Maybe 2 dollars monthly but with the same 50 points daily.

[safe hazel]

you dont need to buy it annually

[void spade]

It says if i buy annually i save money.

[safe hazel]

you do, but if you can't afford it annually, then there is another option

[charred merlin]

Buying a years supply of anything will typically get you a better rate, but sometimes it makes more financial sense to buy only what you need at the moment.

[surreal tapir]

question: would it be possible to add the sounds of a steam engine whistle to a song using it as, like, a specification in the lyrics such as [steam engine sound fx outro] or using "steam engine sounds" in the style?

[rancid herald]

Hey guys, Can I ask a couple Questions?

Im going to ask...

To get the music the way I want it I use the extender to create 10-15 similar versions of the songs. Then I go into premier and chop it up and lay out my chorus and verse. My problem is parts of the song end up different. Is there a way to copy my song using AI and have it play my finished song but all in the same style?

[quaint holly]

Hello, I have a question why when I upload a sample from a song via YT (without copyright from the NCS channel), why does the suno application detect that it has copyright and does not have any? Someone will explain it to me here. I found a nice song and a nice moment and I wanted to mix it somehow. using suno but it seems that the algorithm is not working properly or it's something I don't understand. Please explain this situation to me.

If necessary, I will provide the title of the song that is considered copyright but is not copyright

[flint haven]

https://cdn.discordapp.com/attachments/1069381916492562585/1267223169979256913/sunoaudiouploadtos.png?ex=671ea80b&is=671d568b&hm=8a9c9a25acf4910c14863ccfc673623c63c7c643193d61cf21f078dd54c62af4& @quaint holly

[quaint holly]

I even checked YouTube via Alogrytm and no copyright was found

But I use a work that is copyright-free, so what is written here is pointless

[surreal tapir]

but you don't own the actual rights to the song you're trying to upload, correct? there's a difference between copyright and owning the full rights to something.

[quaint holly]

If it's No Copyright, it means I can use it

[safe hazel]

No, it doesnt

[surreal tapir]

no copyright just means you can't get DMCA'd for playing the audio. Doesn't mean you own it.

[quaint holly]

You have a super cool Alogrytm XD

[surreal tapir]

For instance, Steamboat Willy is public domain for copyright reasons, but Disney still technically owns the rights to him as he is still Mickey Mouse.

[flint haven]

Note it does not say anything about copyrights, it is specifically referring to the rights. Do you have a mechanical license for the use of the sample? For more info, see https://dittomusic.com/en/blog/what-license-do-i-need-for-my-music/
Even if you did, as you can tell by the wording, you would be violating Suno TOS.
TL;DR: if you did not create it yourself, chances are you are not allowed to use it. @quaint holly
P.S.: Back in March, a 20-year-old rapper on TikTok was successfully sued by Sony for $800,000 USD for sampling music from a copyrightten 1986 Japanese instrumental song https://www.billboard.com/business/legal/tiktok-rapper-pay-sony-800k-illegal-music-sample-judge-1235642526/

[safe hazel]

I know who is puppeteering Woodhouse right now 🤭

[flint haven]

🙊

[surreal tapir]

no you don't. hides

[quaint holly]

Okay, but what's most striking is that I have used many songs from NCS and I have never encountered anything like this 🙂

I don't know if NCS is blocked by Suno applications? because I really get lost in it

[quaint holly]

I know this as well as every company that thinks that everything can be done 🙂 But sometimes it's not like that 🙂

I will obtain licenses from the Creator

[vapid nymph]

The second verse of doom is super aggravating tonight.

😠

[hollow hare]

I am certain Suno uses YouTube Content ID system for the copyright checks in Upload Audio so that is strange. A possibility is that someone made a video with that specific NCS music and copyrighted the video, which added a fingerprint of the music.

Regardless, according to NCS' TOS (section 5.ii), you are not allowed to modify the music outside of trimming for timing purposes, so you are not allowed to use their music in Suno (because the music would be modified) which would violate their terms of service

[polar spade]

If parts of a multi-worded prompt is not understood by Suno, it's my assumption it won't acknowledge any of the prompt? Are commas recognized as deliniators in metatags? I know they are needed in styles to discern the difference between a unique style or a modified one. But I'm not certain that also applies in bracketed metatags. So many questions. I'm not certain without testing.

[frigid orbit]

I am lazy, but I like to put all the styles used on my lyric page. Thought I'd share here.

GPT is useful, just not for lyrics

[hot trout]

I seriously have to wonder if someone is actively somehow poisoning Suno training data at this point. Example: Since when does a style prompt of "Narration, Male Voice, Gritty, West Country Accent, Emotive Bass Vocals" mean "use music, and sing like Celine Dion and Rhianna had a (EDIT: FEMALE) crack baby"?

[small gull]

The new one sux

[hollow sapphire]

I'm looking at your prompts:
[Vocal Style]
[Super Suno Prompt:
**"**Soft, airy R&B vocals]

[Instrumental Style]
[Super Suno Instrumental Prompt:
"Punchy"]
I don't think the AI ​​made a mistake, including "..." in a prompt structured with brackets [...], tells Suno, that it should be spoken or sung...
Try removing the " " .

[hot trout]

Does anyone know why audio uploads to Suno are not working?

[charred merlin]

I was just able to upload a clip, so it's not a system-wide issue.

[frigid orbit]

I have yet to share in these issues. Simplify your styles and prompt, shuffle them around or try a new angle.

[hot trout]

It only CLAIMED it wasn't working. Now it actually has.

[hot trout]

The whole reason I am confused is that this exact prompt has yielded me many many good results in the past, and now suddenly yields hot turd smears.

Like, in the past, I may not always have LIKED the results this prompt (and several others I use stylistically) gave me, but the output made sense in the context of the prompts given. I'd say 98% or so of the time. Now, so far, it's closer to 2%, and I am being generous there.

[frigid orbit]

It could just be that this specific purpose doesn’t really align with the prompt. You have bunch of extra bloat against vocal styles and quite varied.

It sounds like a similar issue to when I tried using South African Vocals with english lyrics. All O got were “Gggghghghhghghhghhggg” with proper lyrics.

But it fixed itself by consolidating the rest of the prompt, and sticking to one vocal style in styles.

It could just be that for this specific outcome, it just doesn’t work. Even reordering your styles can go miles.

[polar spade]

Our knowing Suno can throw in some hallucination gens, how many gens did you consistently get this result?

[hot trout]

If we count each individual generated track, more than 50 so far.

[hot trout]

Food for thought. Maybe I can use in-line tags to add the required nuance. See, each element of the voice is important to the whole, but I have found that if I try to combine it all, Suno loses it's poo HARD.

[charred merlin]

Let me guess -- v3.5?

[polar spade]

Wow. 50 all similar. You're right then. Even within one spectrum of a genre ( Celine and a Crack baby) suno would have a broad range. For it to do the same on the first run of the same prompts is odd indeed.

[hot trout]

Yeah.

[charred merlin]

Hopefully future models will "poo less"?

[hot trout]

LOL Well, in the case of one long vocal prompt, I get it. Asking the AI to interpret something like "Emotive vocals sung by a gritty bass voice in a West Country accent" would be a bit much.

[hot trout]

Actually, hell with it. I'm gonna try exactly that, just to see if it changes things up on me.

[hot trout]

There was still plenty of variation, but all of it was some twisted combination of pop and R&B style vocals sung by a female voice.

[polar spade]

Oh yea. Each word modifies the following word as each, i understand, are weighed as prompts. When it doesn't know what a "by" style or an "in" style is, it likely can get confusing.

[frigid orbit]

That’s where I use extend. If I want to mix genres, I do not do so on my initial gen. I for instance start leading with one genre, do a switch to another. Then go back to a full song, extend shortly after when the genre flip kicks in with a combination of both style prompts. Tags are meh if you can trick Suno into a mixed context. Why go the hard way when theses a sure fire way to combine things within Suno’s context.

[polar spade]

Emotive gritty bass voice (if insisting on combining) is likely to get a better result. As Opti, references, splitting things up can wash a heavier weighted style over top as another option.

[hot trout]

It's not actually a genre mix... Those vocal qualities I used in my initial prompt collectively describe a single voice type, one very common to the genre I am working in (in this case, symphonic pirate metal).

[polar spade]

Yea. Trying to get a Tom Waite vocal is not easy. 😁

[hot trout]

I don't even know what I would try, to get a voice like his, LOL.

[charred merlin]

@hot trout I'm not too familiar with trying to evoke that particular vocal style from Suno, so I'm not going to pretend I know what style prompts would work best, but all that comes to mind is to give v3 a shake if you haven't already -- v3.5 excels at song structure (beginings of songs SOUNDING like the intros to songs, building verses with actual cohesive flow, and in general just 'tightening a song', but it drifts toward "generic" without strong use of prompting or an existing melody to "hook onto".

v3, as noted in the model drop down, is a lot more broad and versatile in terms of what it can create -- using both models is an absolute superpower in terms of getting what you want.

However, if you feel like the style prompts should be resulting in different outputs, you can use ... > report > bug > doesn't follow style to help retrain the prompt weighting in future models. (not immediate feedback, but if this prompt chain is something that you want to work as is, this is the most impactful way to make it happen)

[polar spade]

Get Suno to smoke 3 packs of cigarettes a day?

[hot trout]

HAH!!!

[open sleet]

I actually switched from claude to gpt, and I'm loving it all the more

[hot trout]

The reason I am suspicious is that, until VERY recently, the prompt I gave WAS the ideal one to get the desired results... I have a dozen tracks that are beautifully polished that used it, or very similar. Now, suddenly, it's acting like it doesn't know what the sweet Sammy Fuckall I mean when I use it. It's just baffling. I will work with some variants, and see what I can come up with.

[hot trout]

And absolutely, will keep reporting. Hopefully the future is more like the past than the present! Will also try mucking about with 3.0 as well.

[polar spade]

Not trying to be funny.. but try , at least once, using "stand up comic" along with enough other styles to NOT get a stand up comedy routine, and use 3.5. The comic voice can invoke a bass style, but sometimes warbling voice.

If not worried about being "public" certainly can record audio of singing in a emotive gritty voice.

[frigid orbit]

Voices are still instruments for suno. Once you start thinking of them as any other sound. It’ll go a lot smoother, at least for me.

[hot trout]

I'll have to consider that more. Worth a shot.

[hot trout]

Though now I have to wonder, are all of my many past, first-try successes flukes?

[frigid orbit]

I use LLMs for suno just for the annoying stuff like a summary, quick punctuation changes etc. I don’t need an LLM for knowledge

[frigid orbit]

Depends, if in a simple method, one that can be replicated, it produces consistent similar results from your pov. Then it isn’t a fluke, but it’s only good for that specific outcome. Remember we’re guiding Suno what to do, we’re not programming it.

I do a lot of genre meshes through extends, then from a whole it becomes a lot easier to mix stuff together. Whether it’s Gregorian chants and hardcore dubstep, slowly glitching/corrupting vocals etc. Small steps make the big steps easier.

[hot trout]

I just want the same consistent vocal style throughout that I got with all of my other tracks, LOL. Ah, well... Time to try mixing it up and seeing if anything else yields results like I used to get.

[frigid orbit]

Not every consistent song starts with a consistent method to get going. It’s about finding the method that works for you and your concept. Go wild man! Have fun!

[bitter spire]

Quick question, are you planning to add something like HQ sound recreation? I love the extend feature but all my tracks become super mono - and also I got a lot of artefacts especially in stereo field. Also having an influence slider will be so cool especially for cover 🙂
But overall great job!

[alpine venture]

https://www.patreon.com/posts/look-at-using-114805423 for those curious about how prompt oscillation works

(obvi, as always, free and public)

I like this tool as it makes it waaay easier to show people how something can be made, and yea. It's just suuuuper useful.

[cyan jolt]

What prompts should I use to create the maximum 4minutes every time. Its not consistent

[balmy musk]

Would it be possible to add MIDI download support?

[uncut ore]

Hi everyone! I already have a song generated with Suno, but it’s missing the instrumental intro. What is the most effective way to create an instrumental intro that follows the same melody as the chorus? Thanks a lot!!!

[elder sable]

what did they mean about this discussion?, to use Suno for beat and lyrics but re-record using the artist actual voice?

[charred merlin]

If you generate a song on a free plan, you don't own the rights to perform that song elsewhere -- if you made in on a subscription, you have the rights to perform it, like by re-recording it in a recording studio, etc

Just like if I hear a cool song on the radio and decide that I want to make a guitar version of it and perform it myself, I'd still need permission from the song's owner if I wanted to publish or monetize my new version.

Suno is saying that you have permission to re-record your paid-Suno songs 🙂

[elder sable]

no need to credit suno as I am the owner of the song right?

[charred merlin]

If the song was made on a paid plan, correct -- you do not need to give credit to Suno 🙂

If it WAS made on a free plan, you aren't allowed to do anything with the song that would make money -- as well as you'd need to give them credit

[elder sable]

thank you so much for answering my questions
that part really confuses me, cos i wanted to use my own voice for the songs

[charred merlin]

Nice! One of my songs resonated strongly enough with me that I recorded vocals for it -- hearing an AI song that you've taken that far of a step to personalize hits you differently in my case 🙂

Wishing you the best of luck putting your next song together 🤘

[polar spade]

Covers, will sometimes throw in an intro, but then obviously changing the song perhaps not to your preference.

Back painting in or attaching an intro, I haven't found a solution. Yes, other sites offer that as an option. So it's possible, programming wise.

As an adjunct: I'll be experimenting with leaving 2 seconds of space BEFORE a 117 sec song I've self-recorded or sing to audio, and see if that intro blank space is a catalyst for Suno wanting to fill in with an intro.

To think of it, do instrumentals ALWAYS create an intro, or do they also sometimes "start in" abruptly?

I haven't gotten a generation WITH blank space (silence) at the beginning, so I'll have to make my own.

[elder sable]

did you mean you only generated beat with no AI vocals ?
or you replaced that AI vocal with ur own?

[polar spade]

It's been answered, but think of the instrumental part as a "karaoke" track, a track with everything (oooo ahhhs, background, instruments -- in a perfect world) BUT the vocals, where you can remix the whole karaoke track and yet sing over with new vocals. This was a technique many used in the past. The fidelity may not always be top shelf, but it, the finished song, has its uses. Use at weddings, for example.

[charred merlin]

Instrumental that I recorded original vocals for -- for the large part, I only use the AI vocals for 'joke songs' -- not intended to throw shade for using or enjoying the AI vocals, it's just how I enjoy using Suno as a traditional vocalist/guitarist

[elder sable]

thank you so much guys for answering my questions
very appreciated

[hot trout]

By the way, thank you to all of you who have helped me with learning the technical aspects of Suno. I have multiple albums in multiple genres about to be released, and it could never have happened without this amazing software and the community that helped me hone my craft and turn my words into music. 🙂

[thin stream]

broken.

[rich timber]

Thank you. I wished they have a separate box exclusively for prompts. This would prevent singing prompts and Suno getting confused.
The " ", it works once in a while. Example, for me using [punchy]. I only used that word to experiment. It didn't work until I used the " ". Ironically, later I only tried [sultry], it will only work without " "
Good figure 😆 🤣
I really think Suno gets confused because of how data was imprinted in production.

[safe hazel]

The Style of Music box is the separate box that is exclusively for prompts. However, Suno can even sing the things written in there (but much more rarely)

[fleet meteor]

Anything that I can type in the Style of Music section to keep the cover closer to the original?

We really need some settings on the intensity of the cover.

[bitter spire]

That will be so cool! Also Replace Section does not work for me, it creates a very muffy sound.

[summer cove]

anyone know if the dislikes we give will affect future song creations?

[charred merlin]

Not in the short term -- for the new features like Covers & whatnot, they encourage you to give feedback through thumbs down to help train future models, same with ... > report > bug > ..., but it won't impact your subsequent results immediately thereafter.

Pretty sure you're allowed to use the Love Again lyrics for the purpose of creating contest entries, but I don't think we're allowed to use the lyrics however we want indefinitely.

Might be worth reaching out to [e-mail] if this is a question about the contest eligibility, otherwise I'd imagine that the usual rules about using existing lyrics still hold up.

Timbaland is allowing you to use his lyrics specifically for contest entries.

*this refers to the Stems, and doesn't comment on the lyrics -- I'd need to look further, but you're likely right that the answer is "not allowed", but you're going to do it anyway.

[round bobcat]

Hi folks!

Does anybody know trick how to get Suno to pronounce words correctly in a specific language? I can not persuade Suno to pronounce the word 'Labor' correctly in German. Is there an advice for that? Like the once in [...] ?

[charred merlin]

The easiest would likely be to spell the word phonetically — earlier today Suno kept pronouncing “bungalow” as “boongalow”, so changing up how I presented the lyrics seemed to help. “Bung a low” for whatever reason seemed to work in that instance 🙂

[round bobcat]

It looks like some nice features don't work anymore anyway, after the last changes. When I typed my own lyrics and let the style input field empty the text was read (nice for underlaying videos with another voice than your own), now it generates music from a random style. 😦

[empty whale]

Anyone else has an issue in which Suno keeps repeating the same mistakes when I press generate?

I tried extending a song and it just refuses to follow the lyrics and scrambles everything

I report the bug every time but I feel like I'm just wasting credits

[untold briar]

As general advice, if Suno keeps making the same mistake, start changing things even a little bit each time you retry. If it's an Extension, then change the Extend From time by even 1 second, then 2 seconds, etc.

[round bobcat]

Yeah, I already tried more than 10 phonetic combinations with very poor results in this case - sigh

[empty whale]

I'm feeling like Suno struggles with songs in Hebrew, something always goes wrong with them

[round bobcat]

anybody here who can recommend a software to improve the audio quality of Suno songs? I tried Audacity with noice reduction and amplify, but got only small improvements. To be honest, I could not get any improvement by working on music and vocal stems seperatly. Too poor knowledy of mastering music. 😛

I need something for dummys 😄

Especially orchestral music tends to low quality. I think is because of the many different instruments which are in their sum to close to noise and stable diffusion works with (static) noise.

[untold briar]

Before trying anything fancy, try using the Suno Cover feature to "cover" your own song, but in the exact same style. With some patience and luck you can often get a cleaner version with less noise, that is very close to the same as original.

[nova ravine]

Hello and thank you for the amazing creative opportunities that Suno offers. I'm not very technical, so apologies if this is an obvious question. I'm trying to upload tracks to Spotify but it's declining them because the hz measurement is too low. Can I change this on the tracks?

[round bobcat]

@untold briar Funny - I have the impression it gets worse when covering. 😄

Interesting - a male voice could pronounce the word, a female not

[marsh adder]

Try https://www.bandlab.com/mastering, plenty of people here use this (including me 😛 )

[untold briar]

It's not a sure thing you will always be able to "remaster" a song better with Cover, but I am sure it works sometimes. And sometimes really well!

[charred merlin]

HTTPS://www.bandlab.com/mastering is really nice, 4 different style settings to choose from on their unlimited free plan

I personally use an iterative process;

• AI mastering from LANDR (pioneers in ai mastering, they literally started doing it in 2014) — this brings all of the levels up, makes it sound phenomenal by itself, but more importantly gives something more tangible for the following steps
• revision pass through LANDR to accentuate bass/mids/treble to taste
• (if needed) stem separation in Audacity or BandLab, then mixing and mastering each individual track as needed, atm getting by happily with just the LANDR instrument FX plugins and the default effects built into Audacity (distortion, limiter, high/low passes, etc)

[charred merlin]

😅 didn’t see you already suggested it — I get excited to introduce people to BandLab, because it’s legitimately mind boggling how artist-forward they are.

❤️

[harsh geode]

I change the picture on the song but when I download it the new picture is not there.

[round bobcat]

hm.... Looks like I have to get better headphones first.

[errant merlin]

Click "Regenerate Video"

[harsh geode]

I do not see that option

I see it thank you

[round bobcat]

I think I already tested one of those online tools, but was not satisfied. Looks like orchestral music is tricky. Pop, Rock has much better quality which does not need much impovement.

And it is bedtime - again... See you later

[harsh geode]

I am subscribed to the pro but it is telling me to subscribe to regenerate

[errant merlin]

Sorry, I didn't know that was a subscription only feature.

Weird

I'm on pro and I don't have that issue.

Logout and login again?

[harsh geode]

The good log off and on worked. Thank you

[void spade]

Some musics i created Disappeared, i can only find them searching in my Library or in the notifications tab. It's a bug??

[bitter spire]

Is it possible to "lock" tempo? When I doing Cover? Some promp for that?

[safe hazel]

If you "thumbs down" a song it will be removed from the Create page, but not the Library

[void spade]

I dont put a "Thumb down" but hope it works

[safe hazel]

That's the only thing I can think of as to why they would not show up in Create but show up in Library

[void spade]

Yesterday every music were shown

[safe hazel]

Well, also, there is only one page of songs in Create, while your Library has many. So your older songs don't appear on the Create page

[polar spade]

Just a reply. I haven't seen a prompt. Many try something like "98 BPM" but it's not seen to affect things by my observations. I'll listen through when I see BPM, especially when they want to change speed, and it seems nothing changes but randomly, if it does.

[untold briar]

You could try repetition "tempo giusto, tempo giusto, tempo giusto", dashes "tempo-giusto", and also may worth trying "guisto, guisto, guisto" without the word tempo, since tempo is used in so many other musical descriptions.

[polar spade]

To avoid the "squeeze" play, that Suno seems to have, to get a large lyrics song inside of 4 minutes and thus sometimes changes speed, many members have noted
segmenting , others, breaking down into elements.

Segment by just doing a verse per extension, for example.

Elements by running an instrumental first, maybe then a cover of that while adding lyrics then.

In all, the idea is to keep suno from being "overwhelmed" with trying to balance everything on a first run and have issues with speed or, chopping up the lyrics or other unattractive processes.

[void spade]

I find them in the Library, was in the second page, but them disappear fom the Create page.

[safe hazel]

#📚┃suno-school message

[buoyant gull]

I has a technical question about the Remix contest, or rather maybe need clarification. I'm wondering if I start with an Extension instead of a Cover will it still flag as an Entry when I submit after tooling up the whole thing?

Remixes would be fun if I could actually... you know, mix? Doesn't seem to be a way to put two chunks I've cropped together then blend them inpaint style yet, but I imagine that's coming soon, after they make marketing money. ... I mean I know you can download the stuff and do it off site but I don't believe the Remix Challenge allows for that since it has to automatically detect it being from the "stems" provided on the playlist.

[untold briar]

Long requested (at least by me, personally) is a "Merge" feature that could do something like that. Fingers crossed...

[untold briar]

It should still flag it, yes.

[buoyant gull]

Well, there doesn't seem to be a limit, just only the highest rated submition so spam it is!

[karmic pasture]

I really enjoy my funk singer in a song and I would like to reuse him for an entire album. A seed from that particular song will let suno get the idea I want that specific singer with a prompt, or it is like finding a needle in a heystack for the AI?

[buoyant gull]

There aren't any tools [yet] we can use to hone the specific voice, just the prompt. Male or Female is usually pretty reliable but giving them specific traits that are reproduceable is hard/impossible at this time. Sometimes uploeading or reusing something with the voice you want in it can reproduce it but no garuntees. No seeds, no CFG, no special Adaptations yet.

[karmic pasture]

I see I am paying big bucks for suno in the future for that singer... sorry, but I see that happening 😦

[buoyant gull]

Heh, proably paying more in time waiting, but time is money.

It works to an extent as does seperating stems and giving the vocals some good clarity over the rest of the mix. This is something everyone needs who uses Suno: https://github.com/riffusion/riffusion-hobby

(cannot spell to save my soul today)
The sound generation isn't so good, but it has a pretty decent free stem seperation engine.

[ancient crescent]

What can we have the way of controlling the velocity of the voice to use the meta tag?
For example,if i want to higher the velocity from the middle of ryrics ,how should i do the meta tag ?

[polar spade]

Many of mine, (I checked) have no visual link back to any previous, knowing that I started with covers on most. As you said, equally a couple I extended, then covered and on one I think I extended once and it sounded great. I made sure I still did a cover to remain in the spirit of the rules.

I agree, it appears the intent is to use the cover feature first... THEN as it says, use any tools available through Suno. So is that a steadfast procedure. Dunno.

So, my presumption is that they are traceable nonetheless, regardless of whether it visually shows a banner in the lyrics section.

My reputation for my presumptions , as they are, often being wrong, I'll bank on @untold briar response over my own.

[polar spade]

Oh. And I started checking other entries. Same showing no visual links to a base Timbaland stem, which is not unusual nor unexpected now, but still head scratching.

Yea. I figure, hey, if you can hear the elements in your final cut, that's probably a sign. But one Bluegrass one I have seems to have no trace in the music, but just a hint of the chord progression. Ahh... no worries... it will get no listens as a bluegrass piece.... but just interesting.

[buoyant gull]

It can get frustrating. When I look at the create screen, I have 4 traks that say they link to each other, but when I try to make the full song, it eliminates the first part.

[untold briar]

Suno hasn't explicitly clarified that you don't need to start with a Cover first to be eligible, but I'd be very surprised if the rules did actually work like that, and I'll try and double check as well.

This is more my personal intuition, or a wild speculation, but I don't think there is a big difference between using a style "country, polka, banjo, female" and [country, polka, banjo, female] in the lyrics. For example, when I want to slightly turn up or down the "strength" of a music style, one thing I often try is copying and pasting the style prompt to the top of the lyrics (to make it stronger) or removing words from the style field, and moving them to lyrics only in brackets (to make it weaker).

[untold briar]

[incomprehensible]

[buoyant gull]

It does spoken word as well, but it's harder to do since it's more tuned to music. I like having prompt tags, but would even more like to see something like low rank adaptations from uploaded sounds.

[untold briar]

A personal favorite workflow is to start from pure hallucination. The simplest method is to just add "vocals" (or "vocals, vocals, vocals") to a song without lyrics to get a whole hallucinate lyrics. Then you can transform the hallucinations into real words, or my favorite, try and refine the Suno "word salad" into something coherent with iteration - let Suno truly "write" the lyrics. A quick example:

This song uses real words, but it's a cover of a song with pure hallucinations https://suno.com/song/b6212236-8934-4353-b65b-917d1bd7bdbc (and the original hallucination is https://suno.com/song/f193f729-cd82-48e9-8454-39e48235e3cf) and the actual words in that song were also composed by Suno in a different song https://suno.com/song/323c2ae8-f1eb-4003-9cd0-cec50afa62bd

[buoyant gull]

Reminds me of how Empire of the Sun creates lyrics.

[nova ravine]

I found the answer to my hz query - use a different distributor

[untold briar]

I've always felt I get the best music starting with hallucinations. This may just be superstition, but it does make a kind of logical sense that maybe the AI model is better able to generate music, when it's not as constrained by having to also get all your lyrics exactly correct in the same generation. So it's maybe plausible.

Even when you want to write the lyrics later, it also forces you to sort of fit your lyrics into the existing rhythm and structure of the song. So the constraints it puts on you, the human, may also help the final result.

[nova ravine]

I'm having much better luck with rhyming poetry, rather than free verse spoken word

[untold briar]

Or Paul McCartney. In the recent documentary, there's a couple of scenes of him working on a song, where he is sort of singing with "hallucinations" and then the real words come later.

Ah, found it, I posted about it earlier with links to YouTube if anyone is interested: #🗣️┃chill-chat message

[safe hazel]

My best harmonies have come from hallucinations. Alex Ayers once said (and I agree with) that it's almost like suno treats the lyrics as music when it hallucinates so you can get more complex outputs

[placid hemlock]

hi 🙂 if i create a track on the free plan but then upgrade to pro and download it.. do i have ownership of that track? or does no plan provide complete ownership of the track?...... what are the ownership rights of a created track, even if i use my own lyrics?

[unique stone]

I see that this question is already being discussed in #💬┃general-chat

[charred merlin]

https://help.suno.com/en/categories/550145-rights-ownership -- a lot of your questions are answered in here, @placid hemlock 🙏

[polar spade]

How or why did Suno compose random whole English words in a song just from putting vocals in the prompts. I haven't gotten complete English in phrase form.

Then you used those words to add to the cover of the babble (Ukrainian sounding) song to get the first song in the list?

[untold briar]

Earlier versions of Suno were syllable salad. Later versions with word salad. Most recently, you can get sort of real phrases, at least with a lot of patience now.

I think Suno is still not quite as coherent as Suno's Bark though. But maybe soon...

[polar spade]

I started about 10 years ago writing with that pattern of just fill words (and even just scat sounds) to get the melody. You try then to "hear" phrases as you write. When I'm at a store and can barely hear music in the background, I love picking out a phrase or melody that's not actually in the song (that I can barely hear if at all).

When I found the hallucinating songs in Suno, this was heaven.

For music, I'll often hear chords best by recording my own written songs (with basic chords) on the worse dictation recorder and play about 50 times at low volume to try to hear overtones, mondegrenes, and even fills from the poor audio playback. I find a great way to open the senses and not be contained by standard music patterns and English. Then rewrite the song and do again. Rinse and repeat about 5 times.

[charred merlin]

I'll always be tickled pink whenever I'm making an instrumental song, and Suno decides to throw a touch of "Prisencolinensinainciusol" into the mix just to keep me on my toes

[polar spade]

@tribal remnant Didnt you mention about mixing Floss and Timbaland lyrics?

I went back to the contest agreement to find the section 8.2: "you agree that a Song Remix will not be considered an Output as defined in the General Terms (of the TOS [sic]) ...."

It seem any manipulation of the stems are for the Contest only, and only within the terms of the Contest. There is no "output" to use for personal use or otherwise. In other words, how may you mix lyrics when there are no permissable lyrics (either from stems or as an output) to mix.

But interpret that as you wish. Just watching your back.

Yea. I should play around with it in that direction a bit more.

I speak and understand "not right" logic. I also have a Masters in "you've got to be kidding me "ology. And did a 6 week stint as a "like a hole in the head" surgeon in Honduras.

[dusk kettle]

Dude that is exactly why I've not been messing with these contests like that. I mean one of us random people makes the song better and it becomes a hit then they try to get out of giving credit and royalties where it is now due and I'm not one for that. If I make something that is a remix for some contest I fully expect to get payment for creating it if they use it anywhere. Only makes sense they should pay out cause it would be no different than if one of their remix artists did it and got paid for it.

[polar spade]

I get your point. From my perspective, I'm not sure I'm contributing anything. Words by Floss or Timberland, music and arrangement by Suno and I sign any rights over by my entering the Contest as a ghost-prompter and promoter. If I win, that's my commission. Lol.

I'm a pro-writer by trade for 17 years (a small pitence) to add to my other income jobs, so this is just training and another venue to 1) learn to use and excel at , 2) be an adjunct to any tools presently using, and 3) leave any opportunities open to monetize the skills or product.

[astral osprey]

had this issue and this fixed it thanks

[polar spade]

It's specifically grey. The wording is that the songs will be "ranked" by the judges and artists from a multiple of criteria from listens, likes, promotion, quality, creativity, etc. So there is no ONE criteria. As I said before, if I just cover a stem, then edit it down to a 20 second song with one line "I don't love you", and yet still promoted it to get 300k listens, they need to consider also the impact of Timbaland producing it (as noted in the rules), and at that point, I doubt mine would have the integrity to be considered further.

So let's all have fun. Make our best efforts. Understand it's a wide wide grey area and let them toss the darts.

I'm having a ball at about 45 minutes a day reserved for this, to see what I can do.

[wheat cedar]

When I wrote something like this: "Music Style: Acoustic pop with swift, precise verses, emotive chorus; layered harmonies to showcase the vocal range, 85 BPM." The Suno AI refused to process because of the word "swift." The word "swift" by definition (Dictionary - Merriam-Webster) means capable of moving with great speed. Adjective: moving or able to move quickly; fast. Can someone explain it to me, please?

[polar spade]

Yes. It's pop-ish enough to blend well with lots of genres and mixes.

[charred merlin]

Oversensitive automoderation because that word so happens to be the last name of a popular artist -- I recognize how silly this is.

You could use a synonym for 'swift' and it should work just as well -- fleeting, rapid, lively

[polar spade]

And that artist name is Trademarked.

[polar spade]

You cant even have a Sports Therapy business and call it "Shake It Off" else be sent a letter of injunction.

[languid rose]

Hi there. I just joined your discord server. I'm having a little trouble using the latest update of the Suno app. Every time I go to create, it wants me to record something, which I don't wanna do. I wanna create songs like I used to be able to do. How do I turn the recording feature off? Just asking. I don't know if this is the proper channel to ask this question, but if someone could help me, that would be great. I used to be able to click the create button, and then type in what I want the song to be about and the genre. I can't do that anymore for some reason. Is there a bug in the app?

[wheat cedar]

Trademarked the public word for herself. What a crazy world???????

Thank you to clarify for me.

[polar spade]

Yep. Her name is trademarked. Her last name. few of her song titles. You can't copyright song titles, so by golly, she Trademarked them. 😄 and at least 50 other trademarks and more each year.

[polar spade]

My only criteria in life is to make my mom and dad (both deceased) proud of me. I get by with a squeaker now and then. So I'm good.

[knotty merlin]

i dont know why but the cover function used to actually try to make a cover, now it just gens the exact same song with weird changes to balancin and maybe once every 50 gens it does something new, anyone know whats up with it

i change lyrics and i even change styles but it still persists just replaying the old song

or am i using it wrong

[untold briar]

That can happen with particular songs. If you change the style enough, and as you already did, even try modifying lyrics, you can usually unstick it. Try copying and pasting the new style prompt multiple times, literally fill up the entire style prompt in the Cover.

[knotty merlin]

unfortunatly the original style prompt was almost fully filled up

i often cap out my styles prompt

[untold briar]

If you don't change the style, it very often recreates the exact same song. What I mean is when you try to do a cover that sounds a lot different, make sure the new changed style prompt is also super long, fill up all the space too.

[knotty merlin]

hmmm

alright ill try to give it a shot

though usually i want the old vibe to stick just for 3.5 to polish it out with newer vocals

since these are old 3.0 songs

[glad mauve]

Quick possibly stupid question. I have a really good tune and beat, but the voice isnt the one I want. Is there away to keep the tune but change the voice. I tried cover, I tried replace section but doesnt quite work.

[untold briar]

Oh I got you. Try covering the cover that sounds exactly the same, instead of the original 3.0 song. I bet that unsticks it.

Even cover the cover of the cover. Each one can sometimes unstick it a bit.

[knotty merlin]

yes you can extend before the singer starts singing

usually gives much better singer i find

[glad mauve]

So like start the extend from the very beginning, wont that like cut the time down to like 2 min?

[knotty merlin]

cover the cover of the cover until it uncovers how to make a proper cover gotcha

[untold briar]

exactly

[knotty merlin]

like what you can do is:
say you have a song with a 10 second instrumental intro and then singer comes in

you can extend before they sing at like 00:08 and do however much you feel fits in 2m, then finish the song with the last extend

[glad mauve]

I am actually kind of mad that I didnt think to try that lol its so simple.

[knotty merlin]

been doing it for a while now, when in doubt strongarm extend it out

suno doesn't want to do it the way you want?
bang it over the head with the extend until it does

[glad mauve]

I like it

[void spade]

If i get the Pro version, i cant use the musics i create with the free version, but can i use the lyrics and create new songs?

[hollow hare]

Correct, you cannot monetize the free one. Also, you can ask [e-mail] if they are willing to grant the license for that song once you become a subscription member, I heard people had a good chance of having it granted (doesn't hurt to try)

[nimble raven]

Lyrics are out of sync in video creation. What can I do?

[untold briar]

You can edit the lyrics and then in the "..." menu, Download -> Regenerate Video

[nimble raven]

Thank you. I tried pressing the video generator a few times again, but it didn't sync with the lyrics. Is it because it's a cover song?

[bleak gust]

My subscription is not over and its saying i need to subscribe to use cover song...

[untold briar]

It might just be cached, if the video seems to be exactly the same. After you regenerate the video, when you try to download a new version, it sometimes takes a few more minutes of waiting before the file you actually download from the server is the new version of the video. To test this, you could change a word in the lyrics and make sure it shows up in the video version.

If you are sure it isn't cached, then it could be a bug. You might be able to edit the lyrics and help it figure out how to sync better, but sometimes a song just doesn't sync up. It's like a hallucination.

You could even try Covering the song again, and making a new version that sounds almost exactly the same. Sometimes that can fix it. But also sometimes, Suno just gets confused and loses track of the right words in the video.

[nimble raven]

Thank you so much for your detailed answer. God bless you. ^^ I will try as you told me.

[uncut ore]

Ho guys! After generate a song, you use some kind of tool to enhance the audio quality ? Thanks! 💪

[fair adder]

hey every covers that i try to generate stop at 03:10 no matter what, is this normal?

[olive wyvern]

Hello, is there an official API to automate music generation with suno?

[mild yew]

I kind of know the answer already, and I know it's beta but will we ever have our credits refunded when a "cover" with totally different genre tags comes out almost exacly the same as the original? I got 4 in a row today...

[bitter spire]

Hi guys propably someone startet that discusion but - whats the best way to keep a melody/tempo/lyrics but just change for example I want to have different "bridge" between my 2 parts of song. I tried replace section but it only allows me 30s of selection. And resoults are very different from my style. Don't know if this is possible but maybe you guys have some tips?

[frank portal]

extend

[bitter spire]

but my bridge its pretty "slow" so when I upload a part before bridge it keeps on "going" with beat. Dont know if I am describing it correctly. Are there any prompt words for this like "bridge" etc?

[mystic folio]

Good morning, I have a problem, I haven't renewed my pro.
I don't need it at the moment, I canceled the plan but it won't let me create with the coins I have left

[polar spade]

I think should 'out of sync" be defined first. I've always had my videos skip through as if looking for the words of which the audio is singing, flipping spastically to get to the extension or cover portion where those present lyrics sung were first originated.

So, the more extensions, the more it jumps around. Sometimes Lagging as if physically losing it's place. Acting in a very humanesque manner like a teleprompter with 12 pages to look through. In a sense, my videos have always been out of sync to generally catch up in the end. Regenerated or not. I've never given much thought to the vids to make an effort. But I'll do one with just a one gen process and see if what should be an obviously in-sync video happens.

Am I the odd one out as well and am accustomed to this spastic jitter or is there something I've always not done?

[polar spade]

A first gen was normal on the video all the way through.

Hey, I never noticed much, but also if you just have [chorus] , for example, in the lyrics to save space, the vid doesn't search elsewhere for the lyrics Suno's singing. It just holds on [chorus] or [outro] even if there is singing. This seems to defeat the purpose of having a lyrics video as a "karaoke"-like module to read along with whatever lyrics are sung. I'll concede it's not going to throw text on hallucination babble, but the chorus text is available and being sung audibly. Why publicize the video in that case, to have my friends only say, "but where are the words". I can't ask them to scroll back.

It is what it it, I think , so we just get " [chorus] " to stare at. So it doesn't imply an AI thought process to go find the associated lyrics. It becomes just a computer regenerated video than an AI one. Yea?

I'll toss this to the suggestion box before worrying further. If that can be resolved, to show text of what it is singing (i.e. re-lay text based on interpreting the audio) than show written text (if any) to match the audio from which extension is was first lifted from, then that cures any idea of it being out of sync.

[novel lark]

Anyone got a good way to de-robotify the vocals, can be using other programs

[bitter spire]

Hire a real artist and send him/her a line to sing. It may not be what you expect as an answer but I have done it and it works quite well. You already have a melody line etc so imho this is the best solution. Dont think that you can do much with that generated "vocal"

[novel lark]

frick

been messing with the vocal track in FL studio for over 2 hours

and its got a little better

but yeah

not by alot

I used a different software to separate the tracks

way better

none of that weird echo thing

Suno being ai means it will get better with time im asuming (kinda new to this)

[polar spade]

is the follow through to say you mean to have this vocalist sing and record a track, to be laid over to suno's music portion (stem) while recording at a studio?

Or actually, get the friend to sing a cappella, send you the file, then upload that and generate using the original style prompts?

Which way worked for you that you are suggesting?

[charred merlin]

The vocal stem is still labelled as being v1 -- so you're right, intended growth 🙂

[novel lark]

Very cool, so tommorow right 😄

Sometimes i think the voice in my song is fine, other times it sounds so robotic

In my car its realy good

on my headphones, not so much

[vague inlet]

Wrote a long song, but at the beginning SUNO constantly sings in a distorted voice, or without intonation, and then a normal voice and melody. How to prevent this?

[polar spade]

Another version is to use a styl prompt, like Country, or acoustic country, to get a warm, clean vocal -- oh also, people are putting "clean vocal" in their prompts, dunnonif that does anything -- then first try to do a cover in your original style. This may do nothing and put the chorus/reverb- robotic feel back in.

Then alternately, try to stem out the vocals and cover just the vocal. Again, once you change style prompts, it may throw the robotic chorus preset sound back on.

The last way is to manipulate styles by styling the vocal. So as said before , people put "clear vocal" or "clean vocal" in the prompt. I'll try "acoustic vocal", or "country vocal" or "warm vocal" or "crisp vocal" as well. Does any work. Again, I don't know. Just suggesting what comes to mind from observation so far. I'll give these a try myself.

If any of it works, let the room know. Rising tides raise all ships.

[bitter spire]

I just say how I solve it. I am happy with the music composition I get from Suno. Then I hire a real singer (Fiverr). I send an instrumental (Seams) and a vocalist records exactly the same part I send as a acappella. And then you just need to mix your instrumental with vocal...

[polar spade]

Does it have an intro? Did you have [intro] before your lyrics? It may be trying to use your first lines as sort of an introductory segment and so denotes it differently than the rest.

[polar spade]

Cool. Just clarifying. You do that off-site from suno. Ok. Good idea.

[novel lark]

I think it will be hard for me to find people who will sing my parts

[polar spade]

And there are fiverr vocalist who will simply sing while recording to your track, as well, to do , hopefully, a decent worktape kind of mix and send back to save you the effort.

[bitter spire]

Yes, but also if anyone knows if there are any good sites that can replace the quality of a vocal track please let me know. I have found some good for voiceovers but not for actual singing voice.

[vague inlet]

Maybe, but I don't like that without an intro he immediately starts singing, how to make an intro pause?

[vague inlet]

[Verse]
There were troubled times once,
Kings divided lands among themselves,
They brought much sorrow,
And many lives they took.```

[polar spade]

You can try actually [pause].... I've used and feel it kicks in sometimes. But only sometimes. So you may, as an experiment, just load in everything, including a couple of line spaces. Using [end] at the beginning I have found will never end the song, but may drop in a pause or as well.

[Intro]
[Introduction: instrumental]

[Pause] [pause]

[End] - - -

[stiff briar]

Is there any way to change tempo mid song? I've been trying to get something Accelerando, like in the hall of the mountain king. Starting slow and accelerating throughout to a frantic pace.
But no way I can find to control it. Is there a notation, etc. that can control this?

[charred merlin]

You'll have a phenomenally easier time attempting this through placing an extension at the point where you want to effect a notable change, and creating the song as (at least) 2 separate segments that you stitch together using the free "Get Whole Song".

There are differing opinions on the matter, as everyone develops their own "toolsets", but I generally find the more specific of a song that you're trying to create, the less likely you are to create it as a one-shot generation through the use of complex prompting.

[stiff briar]

Ok, I'll try that. - If I run an extension off an extension and then choose 'get whole song' will it pull all peices. Or do I have to get whole song after each extension?

[charred merlin]

"get whole song" stitches all of the used segments between the original and where you're using that option from -- keep in mind that using multiple extensions in a row has a big difference than extending from "get whole song", as Suno potentially can only "hear" the segment that it's extending from -- if you extend from part 2, the extension likely has no idea what the original clip sounds like, so unless part 2 has enough similarities to the original segment, each subsequent extension has the potential to sound less and less cohesive (like photocopies of photocopies of photocopies...)

[stiff briar]

Thanks much.

[charred merlin]

Much love 🤘

[novel lark]

Any way to get suno to generate songs in the same style as a song previously made? Like asif they came from the same artists?

[polar spade]

I just finished trying combinations of prompting and extensions. About 12 gen. Did two 1-sec extension in sequence turn arounds and still it picked up on the original feel but never accelerated or ramped. Just used classical viola in original and tried to extend to ramp up on tubular bells.

Other variations may have gotten a more staccato feel which feels faster, but the Tempo remained the same.

I didn't try off-spelling as "acelerando" , nor did I try something odd like [tempo: ramping]

[polar spade]

Were all looking for the magic formula for that. And avoid having to upload an audio.

I cheat and say I have a band, so that while I try to use similar prompts to keep the song styles similar but vary the songs to feel creative enough, it's in the past given me vocals that sound as if the same or different guys in the band are singing. Yea. A cheat. The more varied the song, the more one can say that's the vocalist simply has a good range of styles when he sings. Hey, you gotta do what you gotta do.

I did get a set of prompts where each Country song would throw out the same "Blake Shelton" vibe. But many times it would not or the voice would change mid song.

[novel lark]

Been trying to use the extension option and just change the lyrics and promt, but pretty much get the same song with diffrent text. or the same sometimes

[frigid orbit]

Are derived songs from love again different that regular gens? I tried making someone an instrumental from one of the segments so they can use it for their song. Just to help them out, but they weren’t able to extend from it.

Is this intentional and did I miss something for the rules of the contest?

[dusky raptor]

Hello, for al almost all the music in one of my playlist, the plays counter is blocked and has been for several days. I have like notification each day but the plays number is frozen... Has this ever happened to any of you ? (Here the playlist :https://suno.com/playlist/c9dbdbf6-9bde-4f0f-bcb6-80b4f0e26cb6)

[frigid orbit]

If not, it would be neat to see this specified in the future.

[charred merlin]

We're only allowed to use the Love Again stems for contest entries, and can't monetize the results, etc -- no idea if specific options have been removed regarding those tracks, but in general, they aren't 1:1 like other clips we can work with.

[loud gazelle]

So - what should I do when I still got no response? 3 weeks and 3 mails to the support and nothing so far (#📚┃suno-school message)... My account is kinda locked and unusable. I was hoping for a fix before the next month to finally get a paid plan... Since mot ppl here are community helpers and no staff there is nothing I can do?

[lunar anvil]

Did you try [e-mail]?

[loud gazelle]

billing and support mail addresses yes.

[lunar anvil]

I don't know what their response times are. Have you checked your spam/junk folder to make sure any replies are not in there?

[loud gazelle]

Yeah of course, I was in contact via the [e-mail] address some time before (with successful responses) in the past and checked my mail folders carefully, But this month - nothing...

[thorny iron]

how can one controll a bit how the track gets an audience? Tags? Names? Pure Luck? Any tips?

[undone basin]

#📢┃announcements message I cant load the app, I am a chosen one... 🍑 iPadOS cannot download this 😄 SAD SAD very SAD

[hollow hare]

At least you can still use the web version

[undone basin]

ya but. If they use to have an app for this why not using the app for this "scene-stuff"

I feel closed out of this feature

[cunning siren]

I wonder how “they” chose from within AppStore users seeing zero connections to SUNO

[undone basin]

nope webversion told me

[cunning siren]

Dogshit it is 😂😂

[undone basin]

😛

[cunning siren]

I downloaded it yesterday without redeeming 😜

[undone basin]

u are from us?

[cunning siren]

No, Germany

[undone basin]

wow what a mess

I'm from germany too

and not even able to download this and have a reedeem

time to fix that

[cunning siren]

First I thought it would be a region problem at Apple 😂😂

[undone basin]

of cause not xD I can't find and chose the app xd

[cunning siren]

Search for the app in the browser

[undone basin]

schon gemacht dude lol

[charred merlin]

👉 https://www.suno.com/ios 👈

[undone basin]

Alter du bist ein Gott

for translation: Damn that dude is a god

Wow great o_o why not at first place? xD

[cunning siren]

??

[undone basin]

u said u are german lol Cant understand german?

[cunning siren]

Doch, ich verstehe nur den Zusammenhang zum Gottsein nicht

[undone basin]

Weil das mit der Browsersuche doch tatsächlich funktioniert hat

Jetzt kann ich Bilder von meinen Fumos machen

XD

[south rapids]

@worthy grotto Any plans to scale up the host becaue this is like the 6 time since yesterday the website goes down 😑 (In Europe)

[cunning siren]

@charred merlin Was it you when we discussed why SUNO does not just create structured MIDI stems rather than producing pure Audio only? I still think that would be way easier

[charred merlin]

The most that we can suggest is to send a follow-up email, letting them know a brief overview of the issue (when it started, when you first emailed, when their last email response was). 🙏

[charred merlin]

Is this theory based on any practical experience regarding machine learning? Or is that how you're reasoning the process would be most easily handled?

[cunning siren]

Well … a score is a structured document basically

[loud gazelle]

I always send them the whole history in the follow up mails... But still - like the 4th weely mail without any response? And the issue began even before, so more than 5 weeks without the ability to use Suno. It just makes me sad.

[charred merlin]

Are we talking about scores composed by humans, or by generative AI?

[cunning siren]

Humans and each score has a well formed structure with rules

Vocals are different tho

[charred merlin]

The issue might come in thinking that generative AI can approach an issue in the same way we do.

While not impossible, creating a song based on the stems opposed to creating it all at once would require 4 separate concurrent generations -- 4x the chances for issues.

I understand the line of thought, but it's similar to suggesting that creating an AI image is easier by compositing segmented images, each of which respects the depth, lighting, etc, of one another.

MUCH easier to create the image as a single pass

[cunning siren]

True but more flexibility because then you can chose the instrumentation to your liking

[charred merlin]

This is the kind of "what if"s that push innovation.

You're right that it would be better, but it's not a matter of "easier to make"

Maybe I'm wrong -- maybe Suno is barking up the wrong tree, and a cool new company will handle things the exact way you're talking about in the near future, or maybe they're currently building that exact tool. I just haven't seen any advancements in the AI market that suggest we're close to that atm.

I'm sure the devs at Suno would agree that would be a really cool idea, though.

[cunning siren]

I believe that would also end the fuzz about infringements because the only thing the model would learn is the fitness of chords to each other in a certain genre or fusion of genres

[hollow hare]

Think of this:
You would assume that splitting vocals from the rest of the music would be easy.
The vocals generated sound different each time, so surely the AI generates it separately and combines it in the end, right?

Well, apparently, not quite. The process how they generate the vocals is integrated into the music is done in such a way that Suno cannot reliably separate it via their stem option, getting lesser results than those from third party softwares with more focused tools for vocal splitting (e.g. UVR)

[cunning siren]

It’s like completing a sentence like “Today for breakfast I had ……”

[charred merlin]

Except remarkably more complicated in application.

[cunning siren]

So exactly what the GPTs are doing 😜

[dense copper]

Why does Suno want to access to my contact list?

I just saw nvm

[keen holly]

Hey guys I was wondering if anyone have difficulties opening the Suno app? mine keep on crashing after downloaded

@icy kindle inquiry above ⬆️

[short fjord]

is there another wayy to subscribe then a credit card ? i mean the prizes going out with paypal and subscribing only with credit card ? O.o

[waxen maple]

does anyone know how to get two different female singer in a song? i tried different things and it is either the same voice or it changes to male.

[sage trench]

Hi, the campaign to invite friends and earn 500 credits is valid?

[ripe sleet]

All my distributions getting rejected by platforms, do we know of tool to detect if it's AI generated so I can pass through it before uploading for distribution?

[charred merlin]

I haven't had a single song rejected. There's probably a number of things to consider, like whether your songs were all generated using a paid plan, if you extended from anyone else's content, who your distribution partner is, and whether you provided all of their required information when submitting your track to them.

[safe hazel]

The last person in your situation had to go directly to Mike B, the community manager, to get the issue resolved. Maybe send him a DM.

[cunning siren]

You get a reason for the rejection?

[ripe sleet]

I have released a good chunk of songs till now but recently they started rejecting stating it's AI generated, also changed distributor from amuse to ditto same problem, seems like platforms have started rejecting them

with ditto although my release worked partially but youtube content ID didn't work and apple rejected it completely

[charred merlin]

I'm not familiar with Ditto or Amuse, but if they're not willing to host AI content, the focus is less on how they're able to tell that it's AI generated, and to look for distribution-partners that ARE accepting of AI content.

https://support.landr.com/hc/en-us/articles/23716156904983-Distribution-of-AI-content I'm a fan of LANDR and DistroKid, and haven't had issues with either 🙂

[cunning siren]

You do some work on your pieces before transferring?

[ripe sleet]

Yes I do remaster using bandlab and that's it, also remove dead silence that doesn't fit

[cunning siren]

Obviously watermark stuff, you may try to apply some reverb to it (just slightly)

[safe hazel]

If they don't accept AI generated music then find a company that does. Don't try to trick them into accepting it.

[ripe sleet]

Would try with LANDR for now, thanks for the suggestions, have a great day 🙏

[merry sonnet]

From Germany

[burnt flame]

They should have a suno ai app for android too

[cunning siren]

All full compliant here to whatever some people decide 😂

We definitely need more French attitude here!

[sweet wigeon]

What's Landr's Fair use policy? Do you have link pls? So after I read the link it seems that they support only human content + optionally modified with AI. I wonder if Youtube also has this policy or more strict.

[charred merlin]

https://www.landr.com/acceptable-use/ While I'm going to direct any further follow-up questions regarding their policies to LANDR themselves, they're accepting of Suno content -- it's when there's zero human involvement that it becomes a problem.

There's obviously a grey area, and holes to poke all through the matter, seeing as how people can just put minimal work into a Suno song to the point where it contains a negligible amount of human direction or composition, but they're one of the few companies willing to put themselves on the line to allow you to explain yourself, instead of coldly cutting ties.

If they believe there's abuse, you can request a live audit to show them your song-making process, and since they're more focused the artistry side of things, I'm sure they'd understand the human elements that go into the curation of the few Suno songs that we pick out of the mix.

[royal rampart]

How can I determine whether a song I created was made under a subscription that allows for commercial use? @final hazel@wet minnow

[eager condor]

Okay, I'm trying something wildy new: a brief radio show hosted by a male and female, one introducing a short rock instrumental and the other introducing a short house instrumental. For the first time, I can't even get the vocals to separate into two people, and then there's only about 3 seconds of music before the spoken word kicks back in. Any ideas from the crowd?

[untold briar]

Maybe try focusing entirely on spoken word radio voices, and either Extending or Replacing the later section with the music.

[eager condor]

I can do that. I just hate myself, so I wanted to do the hard thing

[vapid nymph]

I've been using distrokid and haven't had any issues.

[eager condor]

Plus, nobody thought the 4 minute mile could be done until someone ran it, right?

[lofty willow]

I'm sharing this here because I think it's a problem. But I ask myself why?... everyone should collect donations and better get a psychiatrist for Suno, even if it's totally the wrong thing, I think female part is awesome.
https://suno.com/song/e395be48-c6f7-48ef-ad46-183349f9cac1

[charred merlin]

Could you elaborate what the issue is? If I had to infer, you're upset that the song doesn't follow the melody and timing of "Let it Go" from Disney's "Frozen"?

[hollow hare]

These are some of the successful entries my test results. More often than not, it did not work out well. If you want a playlist of all the generated tested results, let me know:

• https://suno.com/song/8141b517-0f1a-484a-bec6-688f73065ac9
• https://suno.com/song/8f545d00-58f3-4a9f-a778-b0e4a1456802
• https://suno.com/song/8456aad8-7f18-4aea-be07-bd2bcae790c3
• https://suno.com/song/8141b517-0f1a-484a-bec6-688f73065ac9
• https://suno.com/song/a2017943-bb38-4ac4-8db1-b6ef27a78244

Dang autocorrect changing well to we'll

[lofty willow]

I'm definitely not upset. On the contrary, I'm even a little amused and surprised at the sound of the female voice and the power.
But the problem is, I entered the entire German song test and if you translate the lyrics and compare them with what she sings, the error becomes audible...lol...and it was just a bit of a gimmick, a parody of Let to make it go.

[eager condor]

Honestly, some of my results are so bad it's hilarious. I've got about 20 seconds of intro, and I've had one that was nearly 3 minutes of simlish

I hate that I got near perfect results from v2, but they sound like robots 😭

https://suno.com/song/d6986707-306a-4aa9-b465-b16e404a9516

[south rapids]

I know the feeling all to well.

suno.com
Host
Error

There it goes again, 8th time now?

[eager condor]

I see you ran into the same issue as I did, where it constantly refused to separate the voices for you

JESUS ROLLERSKATING CHRIST
https://suno.com/song/194490b8-cdda-46b3-ad0b-53751d16a451

I apologize to anyone who listens to that and subsequently can't sleep

[worn flicker]

I know this was said long ago, but I wanted to see if a question was asked already.
For the entry contest, I see people simply using the words and making a whole knew song. Am I mistaken to think the contest requires doing a "cover" the original song using Suno or take a branch and extend, but in the end the words must remain the same and use the original content. Now if you can get the content to do some crazy stuff that is fine, but the song still must be at least listed as part of the entry as a cover or extended from, did I make all that up...lol?

[stone tundra]

His mustard ligotane with a towel. Perfectly valid points she's making there.

[slender pine]

How do I make suno not repeat the same verses ?

[eager condor]

I think I've actually got a few great intros now. Time to write a rock song about whipping an ass with a wet towel 😆

[slender pine]

Like I want it to sing each leach one time

[stone tundra]

if you insert [Break] here and there between verses or around floating choruses, i find it helps

Kinda gives it a chance to have a think for a second

[slender pine]

I tried putting just one verse

it did it two times 😂

[weary dune]

can you judge my ytb account ( only song from Suno ) pls , my account : 21xxsong

[stone tundra]

Ahh are you trying to do a short song? If so, maybe try [Fade Out] and [End] at the end?

[slender pine]

Like a 1 min song max

something like that

[stone tundra]

What might work as an option to have a playaround with is to have it create the part you want with the lyrics, then do an extend from that point, set as Instrumental only, and do [Fade Out] or [End] (or both! I usually do lol) in the style settings?

You could take the final result and put it into a DAW and do a manual fade out or cut at the end

also as an option

[slender pine]

I don't have the premium plan 😭

[stone tundra]

Ah not sure then. I have the big account lol haven't messed with the freebie for ages

[slender pine]

Broke people problems XD

[stone tundra]

I get it for sure. Broke people here too but figured I'd cut out a game to do music instead rofl 😉

skipping 2 meals this month is totally valid, Serene... you need to make more musics... hehe

[slender pine]

[mental jungle]

v3.5 has a tendency to drag a song out too long, repeating your verses/choruses, and doing weird endings even if you tell it [End] or [Fade Out]. It's seems to want 3-4 minutes songs if it can find a way.

I regularly have take a v3.5 song and Extend with v3.0 to get the ending where I want it.

[eager condor]

i literally just got the $10 plan because I needed the crop feature for what I'm doing. I've already justified spending more on dumber shit to myself in the last week, and honestly i can afford $10. the 70 the other day, probably not so much

sorry woodhouse

[tranquil dune]

How do I make the song have clear vocals

[toxic oyster]

Suno android when?

It's fine that i have some invite to ios app, but i'm not this type of user

[eager condor]

So I did an oops. I got the Station intro done. I wrote the song and got it sounding great. But I forgot to create the song as an extension of the intro 🤦‍♂️ Is there another way to attach the two?

[hollow hare]

Not through Suno at this time, you will need a DAW like Audacity.

If the transition is not smooth, you could try making a bridge the two by:

• if you have a sub, upload a clip of the joined portion and try replacing the transition
• or extending the earlier part to make a suitable (instrumental) transition part

[eager condor]

Thanks. I was hoping to have the two as a single piece so I could have a smooth video made through Suno 🤷‍♂️ lesson learned I guess

[sacred tendon]

Hey, I am trying to create music with Suno. I added this in the "Style of Music" - Genre: Trap/Hip-Hop
BPM: Medium-fast (90-110 BPM) for a balance between anthemic and energetic
Style: Incorporate bold. Is it the right way to do it?

[untold briar]

Lately, I just focus on the voice style and then try to clear up the vocals later with the Cover feature. As long as song is short enough to be covered, this usually works well.

[untold briar]

Just write a list of words in the style, separated by commas. "trap, hip-hop, fast, anthemic, energic, incorporate bold" for example. You may need to repeat some words more than once to make them stronger. You can also add styles in [brackets] in the lyrics, usually in a similar format, but there it is best to use shorter lists or else the singer is more likely to sing the words instead.

[untold briar]

"BPM" with an exact number almost never works. The best you usually do is try to find music styles or descriptors that are strongly connected with particular BPMs. Just using the word BPM will influence the music style somewhat, so you may want to try it just in case you like that effect.

[sacred tendon]

Thank you very much. I appreciate it. I also got some feedback in general chat. Sorry if it is repetitive. But it is always nice to get feedback from multiple different experts.

[sacred tendon]

I want to add this informtion while creating the music. But cant find any suitable place to do it. Is it better to do it later with extend. Intro: Slow build with a suspenseful beat drop to hook listeners.
Verse Flow: Slightly conversational but rhythmic, to match the swagger in the lyrics.
Chorus: Strong, repeatable lines, almost chant-like, for virality—especially “some text”

[untold briar]

That type of natural language description is too complicated for the AI to really understand and follow. But... just using a lot of [descriptive words in brackets] in the lyrics does change things in some way, so it is sometimes worth doing anyway. Just don't expect the song to exactly match what you describe. The general "vibe" will probalby match, but it's more like you had prompted [slow build, suspenseful, beat drop] [conversational, rhythmic, swagger] [strong, chanting] and so on. Those words the AI does have a sense of, and if you use them in the lyrics at the location of the chorus or a particular verse, do have some influence.

["some text in my lyrics"] will probably confuse the AI and it is likely to sing the words, if they match the words in your lyrics. Now I personally love doing that, but just a heads up.

[untold briar]

There's no simple answer whether it is easier to do it with an Extend. It is sometimes though. Probably more often than not.

[ornate belfry]

are there any discord bots that will play suno playlists?

[sacred tendon]

In the chorus, I would like the AI to chant the text "Supasexi Stacc". How can I tell it?

[sacred tendon]

Thank very much for the insights.

[hollow hare]

Edit: actually, no, this is likely what you do not want, hm...

You can have backup vocals by using parentheses, so something like this might work?

(Supasexi Stacc)
(Supasexi Stacc)
(Supasexi Stacc!!!)```

[unique stone]

@tribal remnant regarding your recent message in the suggestions channel where stuff from previous generations seem to be creeping into new ones: This might be a bug where some earlier input isn't cleared by the browser. If you have this happen again, try doing a page refresh to see if it stops the behaviour (or try doing this out of habit when you want to make sure you get a "fresh" generation.) It shouldn't be necessary really but like I said, seems likely to be a bug.

(thought I'd follow-up here instead of the suggestions channel to avoid cluttering it there)

[sacred tendon]

Thank you very much!

[unique stone]

I expect it's probably the part where it converts your input into whatever is being prepared to send to the server, but I could be wrong.
Apparently this can happen with negative prompts if you turn them off but don't empty the box.
If you're familiar with the dev tools in your web browser you could monitor what requests are being sent to the server to see exactly what it's requesting to be generated. Though that can take up a big chunk of screen space to look at.
The server side shouldn't have any recollection of what you did before though.

I asked one of the devs about the behaviour and was told that it's not intended and more likely a front-end bug.
And one of the mods posted this earlier: #💬┃general-chat message

Just thought it'd be something worth trying, see if it avoids the behaviour you're experiencing

[worn flicker]

[Mid-tempo] worked sometimes for me, and with style Capitalize the words and maybe think of fusions like Hip-hop Trap, instruments and things that I tend to add is ambient, melodic, or just my personal preference Epic Hip-hop Trap, or Epic Trap Hip-hop depending on which part you want to have more elements of, capitization is key for style and vocals Male Vocals, Female Vocals, Duet, Overlapping Vocals, Strong Male Vocals, and so on...

[eager condor]

are extensions limited to 2:00 or have i done something terribly wrong?

[unique stone]

Extensions are half the length of a part 1 so for v3.5 where part 1 can be up to 4:00, yes extensions are limited to 2:00

[eager condor]

dang. I've got a 0:24 part 1 and a part 2 that gets cut off mid-lyrics

thanks for the help

[charred merlin]

You can continue to extend -- ... > Create > Get Whole Song on one of the extensions produces a complete, "seamlessly stitched" version for free.

Extensions seem to only be able to "hear" the clip they're extending from, so it can be helpful to "Get Whole Song" occasionally to "lock in" the earlier sections you want to keep, and keep everything sounding like part of the same song.

[leaden lark]

man, this song is quite something, i'm having to keep on jumping from 3.5 to v3 just to get the structure right

[charred merlin]

Is this related to the snippet you shared earlier?

[leaden lark]

new song, that song is already done

just a little look into how i had to do it

[ripe sleet]

Has anyone recently released AI song to apple music like in last week or so, just want to confirm before paying another distributor only for my release to get rejected

[quick garnet]

Hi, is there any guide on how to extend songs? I looked up a few on youtube but none of them worked for me, Thank you.

[untold briar]

There is this https://help.suno.com/en/articles/2409601 and there are new features now like "Replace" that can be used as a way to Extend a song as well, from the middle, not just the end.

[quick garnet]

Thank you! I will go check it out.

[untold briar]

That article mentions this, but I would emphasize that the time you choose to Extend From is very important. Sometimes one specific time just keeps causing problems, but one or two seconds earlier, everything works great. I strongly recommend listening to your song and choosing an Extend From time that makes sense (after a verse or a line ends, for example) and not going with the default which is the very end of the clip. And trying different times.

Also, Replace is SUPER FUN if you are a paid user. And you can use it a lot like Extend...

[quick garnet]

ahh, ok, i see the issue now, was extending from the default recommendation. I shall try your method. Thank you again.

[untold briar]

If you are on a paid plan, be sure to try the Cover feature. I think the best modern Suno workflow is to generate the overall song first, and then "remaster" it with the Cover feature. This lets you focus on the music, and not worry about the voice being super clear, because you can (usually...) get the voice clear at the end by covering your own song in a way that keeps the music the same. This only works if your song is long enough to cover, which is a time I can't recall off hand, but something like 3:20 or 3:40 max length.

[lime thicket]

I haven't released in the last month, but I haven't had issues with distribution to Apple in the past. I do master my tracks after downloading them from Suno, though, so that might help. Apple can be touchy with sound quality. I'm guessing you haven't been given a reason for the song rejection? Perhaps you can get in touch with your distributor customer suppor to see if Apple gave a reason for the rejection?

[sacred tendon]

Hey @untold briar , I have created two version of my track one with male vocal and other with female vocal. Will you be interested to hear the tracks and give feedback?

[untold briar]

Sure, but I'll listen to the tracks tomorrow and get back to you. Caught me just about ready to sleep tonight

[sacred tendon]

No problem. Sure

I will dm you

Thank you very much in advance.

I have sent you a friend request.

[untold briar]

My DMs are open (to anyone) but I have not been adding people as Friends recently, just added way way tooo many when I first started using Discord, and need to cut down the list and reset 😄

[sacred tendon]

I am getting this msg though...no worries..you can sleep...we will touch base tomorrow.

[untold briar]

Hmn, let me double check my settings.

[untold briar]

Ahh, I think it's because your Discord account is set to friends only. I can't DM you and it says that. 😄

In any case, I will be around

[ripe sleet]

It's because of AI usage and only started happening since about a month It's been perfectly fine before

[eager condor]

Finally got my SUNO FM track done. Had to upload it to youtube though. Lemme know what you think. This took 20x longer than it should've. ADHD really got me a few times.
https://youtu.be/00OQLTSiHLg

[ripe sleet]

That's why I want to know if anyone has been able to successfully release on apple music in past few weeks

[thorny vale]

Hi! I found this API for Suno on GitHub but it seems unofficial. Is it okay to use this API, or is there an official one available for developers? Thanks!

[quaint holly]

Why did suno ruin Music? (This looks bad)

Why is this generated in low quality?

When I first used it, the quality wasn't there and since I've been using it for a few days this has been happening

[quaint holly]

with every song I add this happens and I like the song but the quality has completely ruined it 😦

Completely confused and unthinkable

[quaint holly]

Bass completely bulged

[unique stone]

Suno do not offer an API.
All Suno APIs that you find are unofficial. It is against the Suno Terms of Service to use them and your account will likely get banned if you use them.
Please edit your post to remove the link to this unofficial API.

[quaint holly]

anyone help?

Another Example

[unique stone]

This is happening with every song? Can you post a couple of Suno links to affected songs? Might help us (me or someone else who might have other ideas) to diagnose it.

[quaint holly]

Each Fragment of a given note (which is consistent with suno), i.e. Without the Artist

I will upload a random file, note: It has just been generated

one sec

model version 3.5

one sec

Another Example Just generated

1:07

Like someone made crackling sounds

1:45 can be heard better

and this happened after the Update

[unique stone]

What is the link to the Suno song on the website?

[quaint holly]

https://suno.com/song/86476fa9-588f-49cd-8862-9e6307b44419

The same Song

[unique stone]

Ok so I see this one has been extended from an uploaded audio which seems to already be of fairly poor quality

[quaint holly]

Yes, because I wanted to extend it ❤️

[unique stone]

The quality of the uploaded audio determines how good/bad the next part sounds

[quaint holly]

I tried to reduce it somehow, but even AudaCity can't cope with the squeaking and crackling sound that hurts my ears

[unique stone]

Cover can sometimes make things sound better, though that's a paid feature. I have a very noisy guitar song that Suno generated, covered it and it reduced the noise.

The original upload sounds like it's an audio clip generated by Suno. I would've suggested extending from the original Suno generationbefore the sound became that bad (if it is a Suno song.)

[quaint holly]

I tried but it's even worse

I have already generated over 10 songs recently, even extending the generated song again does nothing

[unique stone]

Generally once a song starts degrading in quality, it's hard to recover from that with extensions. I have a lot of bad sounding songs because of this.
New generations (Part 1, not extensions) should at least start better.

[quaint holly]

It doesn't matter whether you generate the Original Song again, there are still disruptions every time

well, it generates me in part 2

[unique stone]

So that's an extension. How does part 1 sound?

[quaint holly]

No, the first part is not available, I don't see it (Part 1), what I will send is Uploaded and Part 2

[unique stone]

Oh, part 1 is the upload?

[quaint holly]

I don't know

Upload what I did (Original sound) and generates only in part 2

[unique stone]

Yes so the problem is that you're extending from the poor-quality audio still

[quaint holly]

And how to fix it so that it generates from the Original file and its Original quality

[unique stone]

What is the original file?

[quaint holly]

Better quality

What I extend it is after generating the note breaks down

[unique stone]

That sounds bad to me, at least up to 33 seconds. Theres a tinny/ringing sound all over it. This is the kind of thing that will make extensions sound bad. Where is this originally from?

[quaint holly]

Track generated by suno (Before Quality Problems )

[unique stone]

Can you extend from an earlier part of the original Suno clip?

[quaint holly]

Yes

[unique stone]

I'd try choosing a time before it starts to get noisy. I've sadly got many clips like this, from v3, where I chose to continue anyway and it's hard to recover when it starts to sound bad. The quality before an extension influences the quality of the next part, and it will never get better with an extension, it can only get worse or stay the same.

[quaint holly]

I also have many (One of them)

(It may be quieter at some point because I was gluing a couple of passages together) and I didn't notice that one of the passages was quieter)

[unique stone]

That starts out sounding ok. I would set the extend time to before the sound quality starts going bad.

[quaint holly]

2:55 (interference)

4:02 kilp pieced together from re-generated fragments not extending those with interference

I fell in love with that moment of time 4:02

4:02 even AE (Adobe Audition couldn't do anything about it)

[unique stone]

Apart from seeing if Cover will "regenerate" it at higher quality, I don't know how to fix this. It's not a simple thing to do 😦
The last time I had this was with a very poor quaality vocal part that I ended up taking into a music program, layering some synth sounds on top of it, and then reuploading it and telling Suno to sing the same words over it. Because the added sounds made the quality sound better overall it avoided the poor quality.

[charred merlin]

Yes. I release content regularly.

Like I shared earlier, there’s nothing wrong with AI music, it’s a matter of your distribution partner.

[frail grove]

Site down

[ripe sleet]

Thanks for confirming would go with landr then, distrokid requires separate payment for youtube content id 😀

[charred merlin]

Congrats on your impending release! 🙂

If you'd like, I can DM you some information about my latest release -- screenshots of the submission form, the resulting promolink for the my most recent upload, etc 🙏

[ripe sleet]

Yes please also do you have a pipeline you go through before releasing like remastering steps or other tool that improves the quality
I'll be great full, appreciate your support 🙏

[charred merlin]

LANDR provides tools for AI mastering, which can help ensure your audio levels are optimized for radio playback, but an alternate would be https://www.bandlab.com/mastering. My mixing & mastering process usually sees me giving a quick pass through LANDR's automastering, then separating the stems in Audacity /w their OpenVino plugin, and using LANDR's instrument FX plugins & the base effects from Audacity to inch closer to my envisioned output

[frail grove]

Mastering over sunos pre mastered tracks that are compressed and limited, double squashed and further degraded

[ripe sleet]

Do you have any other tools/process that you use?

[frail grove]

I use spectral layers split the stems apart, import them, quantize them in cubase, learn the tracks on instruments, recreate the drums partially in toontrack superior drummer and partially on a live kit, use sunos vocals as backing vocals re record the tracks. Its been an excellent tool for writers block

Low quality cant really be upscaled unless you add something thats 48k or 44k. But sunos quality isnt that bad

[junior sun]

the referal links have been removed again?... every time i have someone to give it to its gone by the time i want to use it... i bet removed cause of that ddos the other day? feels like it disapeared around then?

yeah... dont got an iphone. sadly. android user 😭

[worn flicker]

Hey I know what you are talking about. It is like a flanger is getting added to the song as it moves on, I'd have to look in my history but I had to delete a ton of songs because this sound would sound like waves loudly going over the whole song

https://suno.com/song/a726655d-5430-4227-8b39-023478337f93 this song is a good example. It was after this that I started putting things like flanger and reverb in the negative prompt

As the song moves on it gets thicker by the end it is over the whole thing

It seems to have

Let me dig up the song after

[charred merlin]

-edit - I missed the context of asking if negative prompting "flanger" and "reverb" works the same, kindly disregard the following ska
https://suno.com/song/c8ee585b-89d2-40ff-962a-d4abdf54d295 Turning a rock song into a ska song using covers & negative style prompts -- helps to hedge your bets with negative prompts & try to describe what you're trying to avoid in a number of ways.

[worn flicker]

Bah I have too many but there are many things that I see happen and only last for a period of time, as if all generations are in a bad seed area for a bit

That one happened to me a couple days ago

I can tell you things to try, mine cleared up but I can't say why. I stopped putting industrial or any type of prompt that would call for the flanger or heavy Reverb effect.
Next, switch up your song pace, [Tempo Change]

It's in beta so...I can't say that

You mean the negative prompt right

Ah, I would agree if it wasn't a tool for everyone, but not everyone knows what Reverb does to a song. Hence why it is in tags

Lol...agreed

Try the negative prompt, I'll show you a screen of my songs deleted where I started negative prompting

[charred merlin]

Grandma uses Simple mode ( or Suno Scenes if she's on iOS)

[worn flicker]

I started negative prompting anything that could cause a weird overlay when I don't want it

[naive quail]

I'ma tell @delicate idol that you called her Granny!

[worn flicker]

@tribal remnant I follow you and you seem to make enough to notice that when generations get weird that they stay that way for a decent amount of time, as of now, almost every song I make one of the two will be way off the mark. Especially in Covers (since it is in beta I say nothing)

Almost like a bad "Lot number" like seeds are on some increment that changes over time. It would make sense to keep people hearing different styles of the same prompts at different times. The same way things would move to that generic electronic alternative pop Rock sound

Since it is generative AI, they should allow for control of the seed. Rather than just extending a song with the same seed you are looking for.

[worn flicker]

I had asked about the contest yesterday probably during your night hours, about people doing contest songs without actually covering or extending parts of the original. Did you read that one?

[wintry hinge]

had anyone understood how to use the Covers with vocal recording? No matter what I try seems like the output is not related to any of my recording? Shall I put something specific in the prompt? Any tutorial?

[worn flicker]

I don't know about vocal recordings, but covers in general for me are extremely hit or miss and all over the map. Unless you use a script and that changes everything

Given they are in Beta, not much else to say.

[wintry hinge]

How do you use it? what is your original song?

[worn flicker]

Use covers?

[wintry hinge]

do you use your song you created with suno and keep the lyrics in the prompt?

[worn flicker]

Oh, I got ya

[wintry hinge]

They have an option to record yourself in the app, so I tried it and the output is nothing but a mess

[worn flicker]

Yes and no, I tend to leave them, but I believe I get less effect when I do, so I'd say remove the lyrics or remove instructions and leave the lyrics only

I don't use iOS, but I have zero experience with upload audio. That is all the app is doing.

[charred merlin]

I'll read through the contest rules, but I was under the impression it needed to use the provided stems. I'm not wildly familiar with this contest though, so the support team would likely be able to provide answers with more confidence on the matter.

[worn flicker]

I have never been able to get in touch with them, that is how I ended up here when I did

Choppy during covers or songs?

Or you mean too many instructions...

Oh, definitely. I honestly will be seeing if they still have any job opening for developing when I graduate in the spring. I see them now, but I have one more semester left.

I notice that when using too many tags during specific speed songs causes the tags to get Sung out loud

Especially Building Intensity and Four-on-the-Floor

I DM'd you

Yup, that happens often to me, but it is because I like to make certain parts longer and there isn't much Tempo control so I have to use beat changes

[charred merlin]

When did you reach out to them? afaik, they're only active Monday - Friday, <t:1730206800:t> - <t:1730235600:t>

[hybrid vine]

Is there a way to have two-part chorus (each part having slight melody differences) sung entirely the first two times, but the third has the first part instrumental only (a guitar solo doing the melody), If I put lyrics, it will sing them, if I just put a metatag, it will play an interlude that doesn't follow the chorus melody, then the second part is sung like the first. Is there something I can do to force it to play the melody on the first part, then sing the second part only ?

[worn flicker]

What timezone? If you don't mind me asking question that I should have read the answer already

[charred merlin]

It should have formatted to your local timezone, but I marked the EST business hours, since they're Boston/NYC-based

[worn flicker]

Oh OK, I kept ending up at the knowledge base

[inland wharf]

is it possible to merge two (different) pieces of music?

[worn flicker]

I wish Suno did this. Then we could insert pieces without external tools or extending a song from an intro

[charred merlin]

I'm not convinced there's a way to "force" Suno to do much of anything, and feel like we suggest everything to Suno -- the lyrics we input are strongly suggested, but still can be interpreted differently than expected or outright ignored.

My general advice is that the more specific of a song that you're trying to create, the less likely you'll be able to achieve it as a "one-shot generation" solely through the use of complex prompting -- this might be more easily accomplished through the use of extensions, since you'd feasibly be able to prompt the [chorus] structure, but not provide lyrics / use "cascading slashes" (//\//\//\//\ for example) to mimic the fact that the section should be interpreted somehow, but without providing lyrics to be sung

[hybrid vine]

lol it sung the part in first and made an instrumental second xD

but at least we're getting somewhere lol

[jovial trail]

Anyone have some advanced tips on how to master the stems from Suno with something like Ozone 11 to make them sound wayyyy better? I've done a lot, from like creating bass and kick tracks equalized out, to adding Neutron 4 to the vocals and Antares Duo on stuff like that - but it's like things just remain a little too crispy. And the changes don't really make much of a difference... any tips would be much appreciated

[worn flicker]

Is that why I see people with arrows and other insane things in their lyrics? Does it actually listen to those things, I know there are key tags that work sometimes

[charred merlin]

Apropos of nothing, to follow up on the cascading slash idea, I feel like the old-guard of Sunovators would concur that the blocks of symbols like those used in @unique stone nightmarish classic, "Pigs Might Sing" (https://suno.com/song/197cffc1-da6e-4fb6-afe7-362b62f0c84b) are the trick to forcing suno gallucinations.

[hybrid vine]

wtf

[graceful condor]

Are there any suggestions on how get to female fry vocals like the band "Kittie"?

@charred merlin always has something good.

[charred merlin]

That's my opinion, yeah -- the artists might have differing opinions on what they're intended to do, but if those results can't be replicated by anyone else using them, it's sort of a moot point 🤷‍♂️

I'm operating under the belief that they operate as "prompt noise", as @naive quail puts it -- I shared a rough example of prompt noise in #🗣️┃chill-chat a while ago relating to image-gen, if you wanted me to find that link?

[hot trout]

I've been encountering a random error when attempting to change the name of a track. It tells me to check the size of my uploaded image and try again. This is odd for two reasons.

1. I am not changing the image at all, so what would it matter anyway?

2. The image set is just some random one Suno generated; I don't play with the images as I simply don't care. Doesn't Suno know how to generate images of the right size for... ya know... Suno?

Once it throws the error, it prevents renaming completely, not just on that track, but anything generated from it... I can crop, extend, or cover until my heart's content; the name is LOCKED. Attempting to generate new images to use instead, or upload images, result in the same error message when trying to finalize the settings. Any help in figuring this out would be appreciated.

[charred merlin]

I'm not immediately familiar with trying to affect those vocals from Suno, so I'm gonna play around with this, but Kittie is from my hometown 😄

[graceful condor]

I've been playing around @charred merlin , but I'm gonna run out of credits.

[hybrid vine]

I managed to get Michael Jackson vocals by puttin [Michael Jackson] as a metatag in the lyrics.

maybe it will work with whatever artist you want ?

[charred merlin]

Are there any specific prompt chains that have worked better than others? That way I have a jumping off point

[charred merlin]

I think Lumine discovered that resizing the image using the slider seems to resolve that?

[hot trout]

One thing that I have found is very useful for getting specific vocal styles is using "vocals sound like real name of singer". Trying to use a BAND NAME there gets denied every time in my experience, but doing it that way always gets me something CLOSE, anyway.

[hot trout]

Thanks, I will give that a shot.

[worn flicker]

I know exactly what you are talking about. I have done a lot of image generation

[graceful condor]

female fry vocals, female metal vocals, jane angry vocals, screamo jane, judy screaming angry...I honestly didn't keep track, but that's kinda where I was going. @charred merlin

[hybrid vine]

Just to be clear about the Michael Jackson stuff, it's very close to Michael, but you can hear it's not him, but the tone and way of singing is very close, and I even managed to get "hee-hee" working

[hybrid vine]

I'm always denied when I put any name in the style prompt, so I do it in the metatags, and it seems to understand. Someday it might deny too, but for now, we can have fun lol

[worn flicker]

Yea in [Sounds Like: ....., ...., ....] part at the top

I get varied results, but I do use the production portion of the SONG_DETAILS stuff

[hot trout]

Huh... it may be just because I tend to use more obscure artists when I do it. I'll have to test. I haven't had a name drop get denied yet, but it would be interesting to see where Suno draws the line.

[hot trout]

I use that style of prompting too sometimes, but I have gotten to where I seldom if ever need it anymore; I have my normal prompts so locked down that I can really reliably get stuff in the style I am seeking almost every time.

[charred merlin]

Okay, perfect -- imo, it's the exact same effect as when you add keywords into your prompt that don't show up in the resulting image, but make SOME sort of change if you were to use the same seed between images (i know, the further we get into the actual implementation, the less direct of a comparison it becomes 😂) -- I don't doubt for a second that the symbols make an impact, and like in the Furrybeans song I linked earlier, sometimes they're the catalyst for some kind of hallucinatory result, but beyond simply looking pretty and making people spend more time talking about their results, I don't believe for a second that anyone is using them with any form of intention that is being respected and output by suno.

[worn flicker]

I have never had a single thing rejected either, but I haven't ever even tried due to how much drama, justified drama about AI copying their voice.

[hybrid vine]

to be perfectly honest, I made a MJ-like song because I was playing around with Sonic 3 music, which he reportedly worked on, and wanted to make a song out of a level music, sung by him. Had a pretty good result, but impossible to make public inside Suno of course

[worn flicker]

Right along with the knowledge that the dataset the whatever you are prompting was trained on. Like when backgrounds generate with specific descriptions, like singer prompts a background depending on the model, a person's name without instruction will put them a place they are at the most in the dataset

[hot trout]

Okay, update... It does indeed fix the fact that I can't assign a new image, but assigning a new image does not, it turns out, fix the underlying problem of being unable to change the name without throwing said error.

[worn flicker]

Even with the image generator on Suno it grabs some lyrics from time to time

[hot trout]

I understand that, but because it never gets it right, just something close enough with the same vague style, I sometimes used it in my early days. Don't need that trick anymore, as I can usually get a desired vocal style through specific prompting.

[worn flicker]

I'm sorry, I was still talking about generation and I forgot to reply to theharbor

[old light]

The replace feature is great, but with Japanese lyrics, even if I select the part I want to replace, the lyrics don't show up, so I can't replace them. Does anyone know a solution?

[worn flicker]

Look at the bottom. Every time I extend or replace those lyrics have gotten thrown to the bottom of the song after the original lyrics

[charred merlin]

I LOVE this comparison -- just because you say "jazz singer, cybernetic villains lair", there's still a reasonable chance that "cocktail bar" is more strongly weighted in association with "jazz singer", resulting in the typical background you'd expect for a jazz singer.

[worn flicker]

Correct. So if your lyrics match certain parameters then they will go in that direction

Like unprompted style with long run lyrics tends to go to speed rap because it is not normal to sing that many words In an extended singing way

I do think they should implement weights like images though, like (black long hair:0.8) or ((long black hair)), like caps vs non caps in style

[charred merlin]

I get excited chatting about this kind of stuff, because as beautiful and straightforward as it would be to talk about strict and explicit documentation on how to use Suno reliably, unfortunately, there isn't a whole lot that has 100% concrete reliability when you factor in that every part of your input has the potential to skew the output.

In contrast, we CAN clearly document all of the peculiarities and "expected unexpectedness" inherent in working with generative AI, and use that as a "negative roadmap" of sorts -- not necessarily spoonfeeding us what we need or want to know, but outlining what kinds of straightforward, route-one approaches are less likely to give the desired result.

[charred merlin]

You took the words out of my mouth -- and for a wordy dude, that's no small feat.

[worn flicker]

Pot meet kettle...lol

[hybrid vine]

Don't know if I'm able to post a youtube link, but this is what I got with Suno putting [Michael Jackson] in the metatags. You hear it's AI, it's not perfectly him, but close enough in style and voice (couldn't replicate his choirs, though) : https://youtu.be/lkaLlohZlZg

[worn flicker]

I don't do it as a personal respect for the artists. I am not trying to be Mr Ethical or anything, but personally if I ever went pro when I was making music, I would not want people copying my voice and turning it into a Spotify release generic nothing in everything

[hybrid vine]

oh I'd never monetize the voice of someone else, this is purely for the fun of it

[worn flicker]

Like I said..lol it is a personal thing, I have generated pictures of many famous people doing things they would never do...like James Hetfield dunking a basketball or Jenna Bush with a beanie rapping at a camera. It is all in good fun when it is personal

[thick estuary]

Help! I generated a Timbaland remix I really like, but I cannot add it to a playlist, or download it. I don't want to lose the generation. Can someone suggest a fix?
https://suno.com/song/69cd0ee5-806f-40cb-948c-7e2136e730a3

[worn flicker]

This conversation flared up my computer nerd and it becomes a problem...lol. I have problems putting it down. I must go for now, but I will be back...even though I should be doing other things.....

[thick estuary]

sorry to jump into the middle of it. Is this the right place to ask for technical help?

[worn flicker]

Not official developers or help desk style but from experienced people

[charred merlin]

Those aren't too far off from where I'd start, though it can be difficult suggesting "a better prompt" because so much of how they work depends on the style prompts and lyrics that it's being used with -- "screamo vocal fry" might readily evoke that tone in a metal song, but might not have the same weights when prompted alongside a love song

[thick estuary]

I am experienced with the program but have never seen this before. It seems "stuck" everything just spinning in the menu

[worn flicker]

I put metal and Growling and the opposite vocals in the negative prompt almost every time I don't want the scream. It loves to scream

[charred merlin]

Looks like everything is good -- it's available for download from everything I see on my end.

[thick estuary]

Oh cool. Thanks for checking! I love that there is a supportive community here!

I'm gonna clear casche and try again

[charred merlin]

@thick estuary -- mind if I shoot you a DM?

[thick estuary]

not at all

[modern obsidian]

this new cover feature is what can be defined "The fork in the soup"

even if you paste the original prompt into the "exclude style", it doesn't care at all.

like, taking a rock song, putting in the prompt "piano ballad", pasting in the exclude style the original prompt, it will still return you a rock song

LMFAO

what's the whole point of that?

making people try 129038219038 prompts to rapidly drop down all the credits?

[unique stone]

Were you uploading audio of the Sonic 3 music into Suno? Looking at your YouTube channel this does appear to be what you have been doing.
Suno's Terms of Service require you to own the rights to audio you upload. So these would be Suno ToS violations.

[naive quail]

It may have been rendering still. The mirrored image on your embed suggests that. Some options are grayed out / unavailable until the track is fully rendered.

[lyric sonnet]

What is the best way to change lyrics in a section of a song? Tried 'Replace Section', it completely ignored the new lyrics and kept the old

[thorny cobalt]

Hello everyone! I am new here… but I have been trying out Suno almost from the start…. Initially I had great results but it’s been a few months that I am experiencing a decrease in song quality. Anybody having this issue?

[vapid nymph]

I have. It seems like Everytime a new feature gets added quality gets wonkier.

Managed to mitigate some of it by messing with different workflows but it's kind of a pain when things that were fairly consistent one month, gets janky the next and has to be tweaked.

[thorny cobalt]

Is it better for the subscribers? I am currently on the free plan….

[vapid nymph]

Mmmmm.... I'd say yes but mostly because you have more chances to re create and tweak something till it's right where you want it

[charred merlin]

There isn't a quality difference between free and paid users, just features and the amount of credits / concurrent generations (plus commercial license on paid plans)

Not meaning to dismiss PoxToxic's observations, but just to posit my own, I'm getting a grab bag of audio qualities, same as I ever have. While the frequency varies depending on what I'm using as input, I wouldn't say that I'm getting "reportably bad" songs any more frequently than I was a few months ago -- the biggest change has been what I've come to expect from Suno -- since I've already had my fun "mucking about", I have a greater expectation of "actionable" songs to work with, but my satisfaction of the song isn't the same thing as the song quality, so to speak.

[thick estuary]

I was able to extend and then crop the continuously (4 days!) rendering track. It is now on the playlist. Thanks for your help!

https://suno.com/playlist/e5009f8b-4192-4b24-8791-2f71843d0552

I see that most of the Flosstradomis winners rendered out full songs. Although never implicitly stated, does that make a difference to the judges?

[charred merlin]

I don't know if it's an outright stipulation, but it stands to reason that songs with more work put into them would implicitly sound different than Suno songs made without that level of human intervention or direction.

Even if a song is just an original segment and a single extension stitched together, the fact that the artist specifically chose THAT section to extend from will have a butterfly-effect on the potential outcomes, opposed to it just being "a really good one-shot generation", regardless of how much creative effort was put into writing the style prompts or lyric-section /w metatags.

Just knowing there's a hint of human-driven collaboration makes the song more interesting, imo. We already KNOW Suno can make radio-quality songs with little-to-no effort, so after a certain amount of trying WILD prompts and having Suno "creatively interpret" my prompts, I stop caring less about the prompts that people use to create the song, and try to pay attention to the artist & their choices in picking THAT particular song out of the mix to extrapolate from and polish.

[steady kelp]

yeaaah I hate to be that guy, and I know I'm late to the party, but super TOS violation to upload audio into Suno that isn't yours. The very first pop up modal you see before doing it is you agreeing that you either own or have exclusive rights to the material you upload. yikes

[graceful condor]

@charred merlin I finally got the scream I was looking for!!!!
https://suno.com/song/23398c28-403a-4bb3-a825-742a6a5ce941
Thanks for the help. I think it was the styles messing up when I was trying to get. AI prompts are a mystery to me still.

[polar spade]

Would there be a difference in cut and pasting the prompts or typing them in individually? The obvious answer is, duh, why should there be.

The real answer can be more subjective.

Another thing to add to my test list.

[polar spade]

Maybe this is obvious, but a newbie tip. Not only use tools in Suno, but if using ChatGPT, do t just ask it to create lyrics for a song, say, "epic song about vikings" , but ALSO, be intricate in details of the metering of the line, syllables per line, atmosphere, emotive feeling. In addition, tell Chatgpt (or whichever you use) of what not to have in the lyrics -- "don't be sing-songy, don't alway have end-rhymes, frequent internal rhymes, "be conversational and casual" (or) " have a melodramatic feel"

Doing this can keep what is intended to be , say, an epic viking song turn into a pop song by suno because Suno may see 4-lined 8-syllable sing-songy rhymey-whimey verses and think ,hey, this must be a pop song.

[covert edge]

just curious if bypassing things like 'you're dead' with 'you're ded' could lead to a ban? i understand and respect the rules and don't want to influence any system that bypasses rules however subjective my views are towards the song

song example https://suno.com/song/4413f173-875b-400c-ac92-59c98bd3b9af

[polar spade]

I'm just a member, same as you, but in my opinion Suno phonetically sounds out the text, and compares to its bank of sounds to relate it to words and phrases which in common have some sort of meaning. This could explain why it seems to understand and use lyrics as well for driving the direction of a song. Certain combinations fit an algorhythm which are weighted ligh to give a specific response.

So whether it's phonetically spelled out or spelled correctly, it wouldn't make a difference, to a point where, once it's more implied than purely sounding similarit might not catch it. Then it's a test to see if you get filtered anyway.

The argument on the other side is that it knows words like "bass guitar" and knows what it should relate to, but may still pronounce is as "bass" as in the fish and not "base". So who knows if it always reads ded as "dead" or just a misspelled "deed".

I had a 20 minute talk with Gemini on the pronunciation and it never pronounced it correctly.

[polar spade]

So, I didn't give any answer worth two cents. But just adding to the conversation.

[covert edge]

that was more than an answer, i appreciate it greatly it actually made me think more about my line of thinking and questioning

[charred merlin]

Also just a member, but I'd imagine there's a reasonable grey area.

As long as you're not using phonetics to bypass the blocking of something that one would consider reasonable to block.

There aren't "Common sense and good taste" police on Suno, the automod tries to act as a first-line-of-defense blocking heinous AND innocuous all-the-like for better or worse, but as long as you wouldn't ask someone to delete your Suno account if you were hit by a bus, I would imagine you should be safe to use it in order to express yourself 🤷‍♂️

[polar spade]

I would even add, it may depend on the context of the words and phrases surrounding the "sus" word, too.

I keep pointing out a song called " wanna new girl" that I wrote. It had no bad words, but gave an objectifying position towards women. The subject of the song was written as a idiot kind of bigot who couldn't see that he was the problem. It got blocked. So I cover wrote it into an environmental song called "wanna new world" changing 70% of the lyrics and it went through as a completely different song.

[worn flicker]

I say it depends on where you are copy and pasting from, but very unlikely. If you were using a web based document editor and the characters are using a different encoding, remember this is all a "technically" scenario, then when you paste the result there may actually be special characters that don't exist the same way in the encoding of another page. However, this is just using a perspective of what you see isn't necessarily what a computer sees. Web browsers automatically use localization to make encoding adjustments, but technically there are ways that a copy paste can be interpreted differently. Since you said you are testing it.

[polar spade]

Cool. Any thoughts help in the process. Thanks.

[covert edge]

i have been thinking about these things but i have limitations with my technicalities, i generally try to force a good result without prior knowledge but i realized that some things work better than others but not really being able to understand the complexity of what's really going on so i end up wasting a lot of time and credits, it's good to have a more insightful perspective on it as i've mostly been trying to learn by listening to other peoples music but you've put it into words that actually had an unexpected benefit for me hahah, it is appreciated

[polar spade]

🎶 Don't ban, don't ban, Don't ban so close to me 🎶 -- the Ban Police

[worn flicker]

I made a Marine Corps song and many know the saying One shot one kill. Since it is a very common Marine saying, it is in my song. I have never gotten flagged for anything.

[covert edge]

i wouldn't know too much about phonetics to be able to do that but as i learn, i have seen one or two dodgy things myself and kinda ignored it but this makes sense as to why things happen when they do

[polar spade]

I don't think my song are so esoteric as to be misunderstood, but that doesn't help knowing what I can and can't write. I'm usually so bubblegum and sanguine that there's no problem. But then there's that one I write where My Hyde takes over I guess.

[worn flicker]

I have not noticed any censorship personally. I have read about people using workaround but I haven't ever even been warned. Did this happen more often in previous versions? I didn't start until it was at 3

[covert edge]

it's interesting to try and acknowledge where the middle should or could be, i'm personally a fan of perspective.. even ones i disagree with but a.i seems to be more strict than anyone i've ever met in my life haha

[worn flicker]

When it comes to AI (Chatbot/Natural Language Translations) they are wonky at best. I have been creating things with a gpt many times and in the same conversation that the AI itself gave me the result. Later in that conversation it will tell me it can't display it because of protections. Even when I tell it to look in the conversation and that it came from IT how could it be copyright protected...nope still won't do it and I have to begin a new conversation.

[covert edge]

many a time have i argued with gpt about its contradictory nature, i wont lie

[worn flicker]

Oh...haha. yes even knowing that it does not care or even begin to have the ability to care. I have argued for way longer than anyone should

[polar spade]

Contradictory AND overly agreeable. "....You're absolutely correct I can be contradictory at times. Would you like to continue discussing in that direction?....."

[worn flicker]

After having conversations with people that fear what is being called AI these days, obviously they are meaning natural language chat bots. I explain how they function as of now with word prediction, but still will get sucked into an argument with a gpt simply because I have always hates when things didn't do the one thing they were designed to do...

[frigid orbit]

I thought paid plans received 50 a day after you run out. Has this been changed?

[solemn lintel]

I just ran a song I really like, but the quality came out all noisy, like... white noise and static and just poor quality. I tried running a couple covers and they also had a similar issue (though one was a little clearer). Any suggestions?

[knotty nest]

Same here, I'm getting garbled messes or blank songs and losing my tokens!

[solemn lintel]

Yeah.

[worn flicker]

I dont know the official reason, but I personally notice the AI get "lazy" during different peak hours. The latest issues I have had is a flanger/wave/overly Reverb Heavy type sound gradually increasing throughout the song, making it sound bad, I have an example higher in this channel. So do others. I find that specific styles do it more than others, ones that would normally have the Reverb effect more than other styles.

[solemn lintel]

As I've said before, it would be nice if we could flag songs as "terrible quality" and trash them and get our credits back.

[worn flicker]

You can report them

[solemn lintel]

Huh, weird!

[worn flicker]

Not that you will get credits back

[solemn lintel]

But reporting is like "inappropriate" not "bad quality"

Oh! Under "bug" OK!

[worn flicker]

I haven't done it myself. I just have read people saying to flag or report poor quality songs as feedback

Yea, bug

[solemn lintel]

Ah this is great. Not sure it will make a difference, but I've wanted to flag these specific issues in a lot of songs.

The admins will get sick of all my flags now! ^_^;

[worn flicker]

It is a bug regardless of how. It is not supposed to be happening and it isn't just you

[frigid orbit]

A lot of this has to do with your pompts.

[worn flicker]

At least you did it correctly and not like other people are doing just using the words and calling it a remix. You actually follow the rules of the remix. Sounds great. Did you use jog wheels or anything, timecode records or "insert your physical device here"?

[worn flicker]

Yup, I completely had to abandon industrial for now because it is the worst for me. Other than the style, if you aren't using Reverb tags or anything it should not be getting progressively heavier throughout. This is an example from above, the user uploaded the audio originally and you can listen to that and it doesn't have the issue, but the extension does with synthwave as the single style prompt. https://suno.com/song/86476fa9-588f-49cd-8862-9e6307b44419

[safe hazel]

I would be surprised if it got blocked for the underlying meaning of the song (objectifying women). As we have seen from others, you can make a completely bigoted song and it doesn't block it. Even if you have lyrics praising hitler. I dont think the moderation is looking at song meaning at all. Just words, and the number of times those words are used.

And that same person has prided himself on his ability to put any slurs he wanted to in his songs, as long as he spelled them phonetically

[full field]

@neat nexus@untold briarcan one of u pls help me with suno. when i download the song its putting a different name then what the song is titled. i tried extending and redoing it and did the samething

[neat nexus]

Weird. What are you using to download with?

[full field]

pc

this is what it should be titled

i renamed it from this but the download is still showing this

and theres a different picture on suno then the download

[wet minnow]

you have to regenerate the video from the download submenu

[full field]

?

[wet minnow]

that will take over the changes in the download and in the embed (links) you share

[full field]

are u kidding me... wtf now im pissed off

so now what am i supposed to do cuz i cant sub?

i've never had this issue before. it always changed the name in the download too

@wet minnowthis is bs that they allow u to change the details and all but u have to fricking pay to make changes in the download

[hollow hare]

Just make a music video. The simplest thing you can do is plop an image with your music and, well, done.

There are many videos editing softwares out there, several for free. These includes ShotCut, KdenLive, etc. I think noisee.io is still active and can generate you three 1-min clips per day for free (be sure to read the tutorial otherwise you will end up generating just 30 sec clips)

For an online music video editor, there is vizzy.io

[full field]

huh? im sorry but what does this have to do with my issue?

[hollow hare]

You were complaining that you cannot change the title within the video because you cannot regen because you do not have a subscription.

This is an alternative to, well, buying a sub and regenning

The video generated occurs when the music is initially generated with the initial title and lyrics.

Changing the title or displayed lyrics will not affect the video as it would be too computationally expensive to update every time someone makes a change, hence the Regenerate option