💡┃suggestions-web

[midnight shoal]

Suno has a clear penchant to insert heavy beat/drum/synth to any pop ballad-related song, is there anyway for me to exclude that?

I want this song to be more mellow, it's good but overall still too bass heavy

[green vine]

The one thing I would love is the ability to have even a BASIC editing capability. Being able to combine good clips and trim them so you get the VERSE you want the CHORUS you want and the ENDING you want would be so great. As it is I take tons of clips even the 4 minute new ones and pull them into audition to get what I want, and then I am not able to share them with the community. Love that we can upload sounds, but once we generate a clip from that want to take it off the front. For continue maybe we could do an IN and an OUT point? But yeah arranging clips on a timeline and then having Suno stitch it all together and make it sound perfect would be dreamy.

Other wishlist but know its a way off. Being able to do some minor mixing or get stems. I want to be able to use this to start my own music with too and being able to split it apart would be great.

[edgy wren]

Help: looking to separate vocals & stems. How close is Suno to releasing 'stem download' and/or do people have recos to separating? spleeter & lalala.ai haven't been good.

[midnight shoal]

Welcome to the slot machine calls disfusion model

🎰

[green vine]

Yeah I know we are at the mercy of the AI gods for what gets generated, but I would love just a basic online editor, so I can cut together what the muses give me coherently without having to go out of app, because then I can't share it.

[robust jewel]

Suno is working on stems for... eventually. The quality of the output is just currently not quite to the point where it works, as you've noticed. At least one of the staff members asked here before if people wanted stems now, with that quality, or to have them focus on improving quality so splitting it actually works, and the general consensus at the time was quality stems mattered most, not just having bad stems now. It seems like stems will just be Suno including a stem-ripper, not training a whole new style of AI model that generates each stem separately.

[tall slate]

If its possible, please add paypal or Ideal as an payment method. My country does not really use creditcards (Netherlands) and we mostly use other payment options only.

[strong glade]

You really got to sort out this credit reset after each month. I haven't had time to make anything since having to stop a few weeks back for a while. I had over 5000 credits now I have 65. For real. All that money just gone. It sucks. Please sort it out.

[autumn basin]

the engine REALLY SHOULDNT repeat random stuff from before if its not typed in. that is annoying. im stuck at my song right now not able to go on

[autumn basin]

+1 man. that really sucks. you paid for the credits and you should be able to use them as you want

also: how about we follow each other on social media/youtube? here is my cliffhanger:

https://www.youtube.com/watch?v=x4mAJBeDxdg

[serene trench]

Why can't I just buy loans and use them until they run out? I bought 10,000 credits and could only spend 2,500, and everything burned out... It feels like I've wasted my money somehow. The pleasure was gone...

[spare cypress]

Create meta tags for negative prompt and also for accent selection. For example, I am Brazilian and live in Brazil, I am not used to the accent from Portugal or Africa. If it were possible to define this in the first lines of the prompt, it would be wonderful.

[No drums, no guitar][Brazilian accent]

Create a meta tag to define the type of instruments you prefer in the musical arrangement, for example:

[Musical arrangement: piano, cello, harp, flute]

Establish a link between the meta tags, such as having the accent tags relate to the vocal and the "musical arrangement" tags relate to the instrumental.

[modest frigate]

---

I find it a pity and somehow cheating that 50% of the credits are waisted for songs that the customers does not want actually., beause in at least 90% of the cases the second song generated is only what AI wants, not what the customer wants, usually the opposite! It does not even respect basic tempo tags like "slow" in the AI waisted song and credits.

So it would be fairer to generate only one song at a time for 5 credits instead of having us swallow an unwanted second song... or at least to be able to choose one of these two possibilities.

[tepid trench]

I dropped this into known-suggestions recently as it's something often requested, so feel free to put a reaction on it:
#1254418510885687366 message

[rocky nova]

/chirp

[rare obsidian]

The /chirp command is no longer working

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

[modest frigate]

---

Another suggestion to put in the list: add bpm in the song icon. The Ai knows the bpm, so it is not a big deal to display it.
Btw: I've noticed the bpm of generated songs might be not a "whole" number, it might be 82,5 instead of 82 as in one of my songs. A producer complained also that with every extend part added, the bpm changes a bit. I did not check that myself.

[digital spruce]

I’ve noticed the music styles that I can submit not always delivering on my request. It would be great if we would have an advanced option that allows to control the specific instruments that should be used. And also define more specific about the vocals. I’m thinking of amount of artists (solo, duet, choir), clarity and depth (e.g soprano, baritone, etc.), beats per minute, and the chords (being minor, major, etc.)

[spare cypress]

Yes. Look and upvote: https://discord.com/channels/1069381916492562582/1253341197917884530

[tight beacon]

@soft meteor @austere zenith MY suggestion was to add some sound effects to Suno, you don't have to like it but don't stifle other's creativity.

[modest frigate]

---

A basic check for empty style field or a an empty text field would be not only nice and required, but actually a basic programmer rule. So a pop window with a text like "are you sure you want to make a song with no indication for a style- no lyrics ???" It happened probably to everyone to forget some of those and generate futile songs...

[soft meteor]

the thumbs down is not for suggesting something but for providing 0 context

[tight beacon]

I've been asking for months for an Empty Trash feature. Moving them to the Trash doesn't delete them it just moves them to a Trash folder and then you have to delete them from there to actually delete them. Would have been nice to get a button that does the same.

[soft meteor]

you sound like you are smart enough to make your own suno. Do that instead of coming here every few days and disrespecting the developers with how you word your messages

[humble tendon]

Suggestion:
The ability to sort the trending page by language, using the search function is a really hacky way to go about it,
You can search for a char in that language and it will give you some results based on that, but a dedicated language section would be even better (like the one we got here on Discord)
Also when searching, I would like to actually be able to sort by the amount of likes/listens instead of having a random assortment based on the date of upload.

[modest frigate]

---

another suggestion is that the admins and mods should gather ALL suggestions from here to place them in the "known-suggestions" and not only pick up what they think it might be a good one. There are still many suggestions given by many users over weeks and weeks missing there. Otherwise please change the name of that page in "selected suggestions"...

[tight beacon]

Known suggestions was supposed to replace this but hasn't gone live yet.

[tepid trench]

The "known suggestions" are being gathered gradually from this channel by mods and guides. The developers still see the suggestions here as well. I have more I intend to add to that channel.

[desert condor]

Better lower and higher vocal identification levels

[modest frigate]

---

add a song rating for the user with 1-3 stars because the like button is actually or probably not meant for the user, but for Suno. Since there is no other way to mark a good song or a song I want to return to work on yet, the AI might think the song is perfect...but this is not the case always...

[exotic prairie]

Suno! release audio inputs for the plebs and my life, if yours !!

and I don't know, I still think more characteres for the prompts would be good

[visual brook]

https://suno.com/@groovebot have Suno update the Explore page and the Explore page main account to use v3.5

[modest frigate]

---

in the age of AI we still have to manually place tags like in the MS-DOS age! A few definable tags done with a mouse click for things we use a lot would be not only handy, but a normal thing in 2024 ! .. for example [Instrumental intro] a t the beginning or [Instrumental solo] after chorus...

[celest plaza]

we need a Affiliate links for pro users an Affiliate link when other users buy pro or other stuff and get 10% back

[modest frigate]

---

try to make an agreement with splice for using splice samples or at least an option to tell the AI the audio file you use if from splice. Because if somebody used already a splice sample in a song the Suno knows it wont let you use it again and the sample is actually royalty free and is not fair to block other users to use that sample. It should not be "first to come, single one to win" for royalty free samples.

[atomic bloom]

Ok... Suno is king compared to the competition. However, I think I figured out why there's so much background noise. It all has to do with high frequencies. I think that's where the diffusion is or may be failing. Too many frequencies similar to the noise Diffusion process. Just a thought. So maybe use the frequencies of that to diffuse. Or that extra noise is your watermark?

[chilly cradle]

yeah so like i suggest allowing people to upload the amen break cause its free to use

[chilly ice]

For update 4 make the ai voice be more realistic in its way of speaking. Some words and generations that dont use human input to put lyrics makes the lyrics kinda merge with each other and sounds slurring

[wicked hound]

https://suno.com/playlist/57ffd99a-ef14-4c4b-9a5c-9500049be2ff
more homebrewed audio for users to experiment with, i'll keep adding and fine-tuning this playlist

[acoustic mural]

Is there a way SUNO developers can create a queue to let us know if a song gets filtered or not for trending? Also, provide a better explanation of how the logic works. I know it is on the wiki, but I have seen some scenarios that satisfy the criteria and still don't hit trending. I suggest having more visibility on whether a song is okay to go trending or not, like having a banner that marks red if it is not and green if it is okay for trending once the system analyzes if the song is safe. Of course, this can happen when the song is set to publish.

[exotic prairie]

no way they heard me

[acoustic mural]

They listen to everything.

[maiden holly]

I'd like a way to pick a voice I liked and use it again for another gen

[exotic prairie]

didnt expect it would be this soon

[acoustic mural]

This was going to be my next suggestion, but I guess it can be added to the wishlist for now. However, having the option to choose a vocal after multiple extensions would be greatly appreciated.

[maiden holly]

yup or editing and adding to the song easier would be great

say you want to go back and rewrite lyrics or typos ect

[acoustic mural]

Agreed

I know this might sound dumb, but it's important to me. Sometimes I remove items and then regret it, and my trash can is cluttered to the top with tons of stuff. Is there a way to add a search option for the trash can section?

[dense kettle]

Hey Suno Devs!

I'd love to see a "sing-along lyrics" feature where the text scrolls down as the song plays, highlighting the verses being sung at the moment (or even the word). This would make it easier to follow along and sing and make it more fun and enhance the overall experience. Thanks for considering this!

[acoustic mural]

SUNO karaoke?

[dense kettle]

Yes hehe

[raw mist]

Add the ability to download song stems with vocal CLEANLY separated from music.

[sidebar]
I was told here that if\when this was created it would be no better than a stem creation service such as Fadr. To mature Suno however we have to do better than this. Fadr vocal stems can take HOURS to cleanup, if they can be cleaned up up at all (distortion, plosives, pops, hiss, dips\dropouts Ahh!).
[/sidebar]

As a paid user, this feature would be immeasurable. Thank you!

[sage mirage]

It's because Suno does not use stems, at all. It's master 100% of the time, there are no stems to download. This is why it would be 'the same' as how you can do it now, breaking it apart from master

Since that's how Suno would do it internally. Which is why there is no native from-suno solution yet, it's not any better than what is available

The devs know though, certainly

And every request and suggestion is good

[raw mist]

Thank you. Perhaps somewhere in the creation process the vocal can be intercepted. OR be able to create a vocal performance 100% on its own. Same with say a solo (guitar, sax, flute etc.).

[sage mirage]

So thanks for the suggestion, the devs do read this and know how badly stems are wanted by lots of music people

[sage mirage]

Inpainting may get you close

v4 will come with lots of changes, probably. The main one is inpainting, so you should be able to just reprompt over problem areas. Theoretically, only the devs know

Also reportedly being able to upload an inpaint

But again just what I've heard and seen from the devs

[candid goblet]

Is it possible to add a feature where we can remove Vocals. For example I wanted to make an instrumental but forgot to toggle it to instrumental but I like the music but not the vocals ?

[past thistle]

A option to use suno AI or add our own voice model from games, movies or etc.

[robust jewel]

The vocals aren't clean in the first place, so how could you cleanly separate them? Each model aims to be cleaner than the last, but it simply isn't there yet.

[fresh lodge]

Please Suno can you fix this seriously terrible low audio quality that sounds so overdriven and distorted and completely ruining 98% of my song creations, I have wasted again many thousands of paid subscriber credits this month on some otherwise great song creations that sound like trash because of this, totally useable music that sounds like every musical element is peaking hot and running through a tin can as a speaker . Suno really doesn't like Rock or Metal instrumentals for me which is what I like to create. Some start off ok then halfway through bam you are hit with this overdriven wishy washy sound, I am surprised the AI model has been coded to even allow this to be a thing, something is adding it to the music so it must be written in the model somewhere. For an audio tool sound quality surely needs to be the top priority otherwise what's the point? Its becoming so frustrating and as of right now a waste of money to be deleting creation after creation, not to mention 2 creations at a time. With v3.5 a lot of what should be instrumentals end up with lyrics and many with garbage lyrics which don't make sense at all, its just noise with the odd legible word. Still love ya like 🥰 but please its so frustrating.

[tardy anvil]

This for sure. I've been wanting them to at the very least add some sort of visual stamp to the song page after it triggers the filter, labeling it 'mature content' or whatever. I know it's a tricky area as they don't want to reveal too much and have people trying to work around the system to bring mature content to the front page. All I want is to just know I triggered it or not.

[acoustic mural]

That;s all I care if is trigerred or not

[elder moon]

i think it would be nice if you could write the genre separately from the song description

[blazing plume]

Can Sumo add Kazakh instruments to the style of music?🫠

like dombyra, kobyz, jetigen, asatayak

[floral laurel]

The ability to merge two creations and designate which one is the starter and which one is the 2nd half/ 3rd, etc.

[earnest fractal]

I've found the wishy washy sound happens when the Ai "runs out of ideas" if you use the lyric box and only put in [tags] with what you want the instrumental to be I find it turns out much better stuff like [breakdown] [pause] [stutter] [build up] [bass drop] ect

[robust jewel]

Turn on custom mode

[elder moon]

that just lowers the character count to 120 from 200, and doesnt do what i suggested, having a separated genre box from the idea, get me?

[robust jewel]

Put the genre stuff in the genre box. Put the description stuff in the lyrics section. Use [these] to make it not sing it.

You get 120 for genre and 3000 for description

[elder moon]

oh lol when i tried it, it just started singing what i wrote

i mean last time

[robust jewel]

Brackets [] should be seen as information/instruction. If you want to do an instrumental piece with no lyrics, it should be possible to have [Instrumental] as an instruction, but sometimes you have to add stuff like ... to force it to do it.

[elder moon]

i tried "Description: " seems to have worked

[elder moon]

but thanks!

[paper blaze]

gonna say i hate this right here. just let me do one gen when i have only 5 credits left. and also, give paid users daily 50 credits as well

[jovial field]

Hello, you are awesome, I would like to suggest you an idea to create looped music. I, for example, need to make a musical theme that should be played endlessly, that it ends at the beginning and starts at the end. But so far I have not seen such a function, although without it you are just super, but it would be even better!

[copper valley]

Any hope for proper stems, with separated lead & backup vocals, guitar, drums, etc? Otherwise it’s getting pretty close to perfect 💚

[near fern]

as soon as they support inpainting this should be possible

[stray birch]

I'm not sure if this posted since my internet was poor, but booster perks would be neat

[blazing lark]

I have a good suggestion hopefully. A way to take a song done on Suno, and make it a MIDI/KAR/etc file, or a way to remove the lead singer and leave the background tracks. I'd love to sing these songs myself, but hard to remove the lead singer without loss.

[patent glen]

Adding the ability of negative tags would be cool, say if you don't wany any rock elements to the song, or any bass you could add it in the negative

[midnight shoal]

I'm sure their vocal quality would improve if a million dollars worth of GPU cluster just materialize out of thin air

[patent glen]

welp, that's what our subscriptions are funding haha

[modest frigate]

---

beside adding a 3 star rating for any song, and a comment field to any song, I suggest something even more useful than this, which saves a LOT of time for us users: the possibility to mark with a simbol what is good in a song for a later rework: I (intro), V (verses), C (Chorus), R (Rhythm). Because when you listen the song first time and you mark the good song part you don't have to relisten next time from zero. Relistening the songs consumes a LOT of time and it's a pity to throw away good parts of song which might never return again. I have around 6000 songs in my library which many need rework so this would help a lot. But again, I think all suggestions for the user interface are in vain cause no one has ever been implemented in more than 7 weeks.

[wind crest]

i was excited to try out the sound clip uploading but i'm a little disappointed with the results. 9/10 of the generated tracks just started up an unrelated song after the sound clip, and one faded from the sound clip into the song for a few seconds, but i was hoping that the sound clip would be integrated into the general instrumentals/structure of the entire song. it has a lot of potential, but i hope/suggest that future iterations will make more extensive use of the sounds in the provided clip throughout the song

[modest frigate]

---

Suggestion: please hire an AI or a secretary to answer emails. Emails unresponded from a company like Suno is a no go.

[raw mist]

Or eliminate out everything leaving just the lead vocal performance (no backups). OR just the solo (guitar, sax, keys etc.).

[jagged kayak]

Suggestion: Suno should take a cue from Midjourney and have a "Chaos" slider that goes from faithful to riffing. This way you have a way to tell Suno to stick to the lyrics or go ahead and deviate. It works great with images, why not here?

[woeful ruin]

Being able to crop the music and edit in your own music style once the full piece is complete would be great

[rocky nova]

#chirp

[versed wigeon]

military bands missing example: ottoman march

[meager forum]

It would be awesome that if you left the title blank, it would give you a random title, instead of Untitled.

[cosmic lodge]

The alternative is to use something like Melodyne but it would be so much better if it was just built in

Actually honoring all of the [Sound Effects] every time rather than when Suno feels like it

[blazing lark]

I tried Melodyne, and it can seperate the sounds, but not make a midi file, plus the voice is still there.

[cosmic lodge]

It can’t make midi files?!

[blazing lark]

not that I found yet.

My last resort is pulling out my keyboard and make files.

[opaque gull]

In the library and maybe any other page can you show the page as "page 1 of x" instead of "page 1" so that everyone knows how many pages already exist.

[thin ermine]

I may be better off not knowing...

[kindred fable]

Folders - that would be neat. Would help a lot

[spice rover]

I think it would be cool if we could upload a reference song for the ai to use for how you want songs to flow. Certain times I generate 20+ songs trying to narrow down the flow to get it as close as possible. So having a way to upload or give a link to qa song you want to reference would be cool!

[cold valley]

maybe a shuffel function for playlists? that would be great.

[robust jewel]

Suggestion: Maybe consider moving the known-suggestions channel above this one (and changing the order of the name around so it clearly says suggestions), so maybe people will consider looking at it before coming to this channel to ask for Paypal, stems, editing tools, negative prompts, control over song length, or control over number of songs generated for the second of third time that day.

[wooden lintel]

Being able to type in a specific page number to jump to would also be nice

Would it be possible to get the Random Music Style button back along with the suggestions?
I got some really nice combinations with it

Having the option for Prompt go to Custom before making songs incase we'd like to tweak something that the bot comes up with instead of using another set of credits to tweak stuff

[teal perch]

🔥 suno developers, you're the best🔥

[tepid trench]

Did you upload someone else's song?

[teal perch]

It turns out, it's hard to call this kind of music a song.
The full version from the author never came out, it's been 6 years.

[tepid trench]

If it's someone else's song and you don't have rights to it (even if it's just part of a song it still belongs to someone else) you're not meant to do that. That's why you have to agree to this when uploading audio for the first time.

[teal perch]

I did, otherwise the audio will not load.
About copyright, this genre of music is dead. Its audience is 0.000001% of the world's population.

Music from big name artists that may have copyright issues doesn't interest me.

[tepid trench]

You'd still be breaching the Suno Terms of Service by uploading other people's work though.

[robust jewel]

And violating copyright law

[teal perch]

Accordingly, I was most looking forward to this cool feature. So that the neural network could finish a musical short preview of a track that was never finished by the author.

[visual brook]

give the Feed an RSS output for RSS readers/easier notifications - do the same for Trending somehow

[teal perch]

I don't think it's a big problem. On the contrary, many authors experience creative burnout.
Because of this, many compositions remain unfinished.
Thanks to suno ai, writers can now generate full versions of their tracks by pushing back from the initial idea so that they don't have to completely come up with the full composition themselves.
More to come when the neural network learns how to finish unfinished compositions even better.
I am on the contrary very happy about such events.

[wooden lintel]

Why is this discussion happening in Suggestions?

[tepid trench]

Of course. The original creators of a piece of music can benefit from using Suno in this way. I've been using it on some improvised pieces of my own to get ideas from it.

But, it is a problem if you are breaking the Suno Terms of Service.

[teal perch]

The only problem is if I start making money from it.
But I don't do that sort of thing.

[tepid trench]

As far as Suno's terms are concerned, even uploading the audio in the first place is a problem.

[teal perch]

The circumstances of my actions are being investigated.

[teal perch]

I understand that, but how are you going to control it? I mean, anyone can upload any song they want.
You need to build in a copyright checker. Like youtube and other music services do.
Of course I'm not good at this.

At this time, as I understand it, there is no copyright check on uploaded audio recordings.
And therein lies the problem.

[robust jewel]

Or, you know, when people show up and publicly announce what they did...

[teal perch]

I was merely expressing my approval of how well suno managed to finish a specifically unfinished composition.
What difference does it make?
Anyone can do it, and distribute suno ai generated audio files, it's even worse if they try to make money from it.
Since there is no copyright control on the part of suno ai.
What can I do if now suno ai doesn't know what death step is, suno ai doesn't know how to create such a powerful sound.

[tepid trench]

It is up to the user to ensure that what gets uploaded complies with the Suno ToS. There is a copyright check but some things can slip through it (e.g. if it is not well-known.) This stops the most obvious cases where someone might upload part of their favourite musician's work.

[teal perch]

It turns out suno ai is trained specifically on copyright-free music?
Is there any way to add new musical genres for learning?

[tepid trench]

As far as I know, the only way for a new genre to be included in the model would be for the developers to include it (similar to how it lacked any digeridoo instruments until recently.) So you could suggest the genre(s) to be added here.

[teal perch]

The genres are very specific. In fact, we can say that they are already dead genres (they are by design). Really experienced authors of such genres don't release new tracks anymore. Because of the fact that almost nobody likes such music.
https://en.everybodywiki.com/DeathStep

https://en.everybodywiki.com/Minatory

So too are there combinations of these two genres with each other.
For a fiercer sound.

It would be really cool if you can find copyright-free training materials.

[teal perch]

But there are individuals like me who really need this kind of music like air.
There's very few of them.

[sweet cape]

i need music like water with every meal

🙂

"breathes heavily to create steam"

[wooden lintel]

Why did you post your generations in here anyway?

[teal perch]

It's not generational, it's a demonstration.
The 1st is the original 41 second version.
The 2nd is what suno ai did to her.

[sterile slate]

request/suggestion: Explicit or NSFW section. Also! A social media page attached to our profiles, similar to twitter or tumblr. Maybe that's waaay too much but it'd help keep everything in one place haha

[spare cypress]

Hello Suno team. What if you offered the option to download just the instrumental part of the song?

[tepid trench]

This is often requested and requires separate stems (which Suno doesn't use.) It's likely they may offer post-processed stems for this, but from what one of the developers said recently, currently they aren't "clean". It'd probably give the same result as using something like Ultimate Vocal Remover.

The best way for them to achieve this would likely require a completely different set of training data and lots of changes to the way the model works (plus additional compute time) so I would expect getting the "true" instrumental of a song probably won't be possible for quite a long time, and would only work with whatever model is actually trained on stems.

[spare cypress]

They could train the AI to be capable of reconstructing the audio with only the instrumental part. When I say reconstruct, I don't mean taking the audio and separating it into tracks; I actually mean having the AI understand the arrangement, musical notes, and then generate an instrumental as close to the original as possible.

[modest frigate]

This is how I would do the song management, visible in the song line, so the user can know also a month later how the song was and which part is good without to have to relisten. The flags could be set with the mouse in the desktop version. Of course, one could add even more flags like "S" for song saved... or "F" for song finished or "L" for lyrics (redo lyrics)

[tepid trench]

Not sure how difficult that would be but it would be cool if they could figure out a way to achieve this. As far as I'm aware it just knows how to make sounds, with little/no info about the actual musical composition. I could be wrong though.

Interestingly, sometimes Suno can repeat parts of a song with some instruments or vocal missing. It's incredibly difficult to make it do that though. I spend a lot of time trying to convince it.

[modest frigate]

---

well, it took me 3 days to find that pdf with the meta tags or the wiki page and I dont even know how I got to it.
Why is such a document or a link not included in the HELP page?
How can be the most important piece of documentation for suno missing from there ???
Is there a link to the wiki on the help page?

[tepid trench]

The likely reason for this is that the only official documentation is what's on the website's Help/FAQ pages. Anything else, such as the wiki, is an unofficial community effort that the Suno staff are not involved with. That said, I think it would be nice if there were some "community resources" of this kind linked to from the site so that they are easier to find.

[brazen dawn]

Hello Suno Team

I guess users already requested the "Download only instrumental" from a full track with lyrics. It would be useful for people who want to sing on top but follow the AI-generated voice as a guide.

So, I will make a more tricky request... What about giving us an actual UI with few functions like:

• A horizontal timeline where we can add markers to control the arrangement of the song.

This would allow us to place and move sections like intro, verse, chorus, bridge, and outro along the timeline. For example, users could drag markers to define where each section starts and ends, making it easy to rearrange the song structure.

In addition, the UI should provide 7 main layer tracks where we can generate sound separately based on what we already have:

1. Melody
2. Harmony
3. Rhythm
4. Bass
5. Vocals/Lyrics
6. Accompaniment
7. Post-Production

We should be able to generate sound for these layers individually. Also, it would be great to:

• Generate only parts of the song, like selecting one of the arrangement sections and regenerating the sounds for that section.
• Be able to reconstruct the song if we want to remove a section or add a new one.

Of course, each change should keep the current style of the work without breaking the style or timing, unless explicitly requested by the user.

Each layer should have its own explanation tab where we can describe everything to Suno for that layer.

Being able to export all of it together or as 7 separate layers would be the best.

This level of control would make your platform much more powerful and flexible for users! It would be extremely powerful!

If interested, I would be happy to show you an UI/UX Figma prototype for this idea, even though it's much more of a long term project and it will take a while to implement.

Let me know if you would like to see how I'm imagining your platform evolved in few years 😊

[tepid trench]

Pretty sure thats basically what stems would let you do. Which would likely require a completely different model and training data set as described a few posts above: #💡┃suggestions-web message

It'd be great to have this level of control. I don't expect to see it for a while though.

[brazen dawn]

Yes I've read it in the known suggestions after posting.
I guess my suggestion just goes a bit more into the whole idea on how to control and manage those stems.

I've called it layers because I have a graphic design background and I'm not so much into music compositing, but Suno just made me so interested in it.

[brazen dawn]

Yes it's definitely a long term project 😅

The markers, to better control sections timing and cutting for the [end] issue should instead be possible to implement already in this current status of the model right?

[final forum]

Has it been suggested to have a trainable AI vocal based on your voice?
Is it possible through Suno, or could it be in the future?

[verbal ivy]

Is there anyway to download the video and song without the lyrics on the screen? Or is it something that might get added in the future?

[nimble ice]

Thats a great idea. I hope the developers can see how that would help with more data points to reference

[signal dove]

Hello
I would be most happy with a development where the completed song could be downloaded in a format in which the music and the vocals are on a separate audio track, so that both the vocals and the song can be compressed afterwards. The only thing that would be more serious than this would be if the instruments were all in separate tracks, but this is frankly not that simple. Obviously, this would be a paid option, it wouldn't bother me, I'm still in the pro package.
Keep up the good work and we look forward to the next development!
Hi

[modest frigate]

Suno is proably the only company which dares to hide the user manual so the user has to find it somewhere on the net. Kind of "hide & seek" game for the users. It is like buying a complicated machine without user manual. I don't know the reason for that, but this is unheard ! Is there any official explanation or does this simply go hand in hand with other abnormalities I have noticed?

[modest frigate]

you must be kidding! They did not care even for the bugs, for basic things missing or for a user manual link... they would only do this when other competitor would implement it, so I sent it also the competitors.

[nimble ice]

🤣🤣 that's one way to force development

[modest frigate]

---

I improved the song management added in the song line, so the user can know also months later how the song was and which part is good without having to relisten the song (which so time consuming!). Song management is missing yet completly in the software. The flags could be set with the mouse left click in the desktop version. No serious musician uses a smart phone to make music with suno. It is a piece of cake to implement this. Could a mod add this in the known-suggestion page?

[jagged trail]

Could we get some organization for the library added in? Like delete all in the trash bin, and possible folder structures for organizing various types of music, and other uses?

[stone cobalt]

Is this something similar as udio has the management that they generate the first section and then you can add intro, individual sections like verses and outro? If yes, that could really enhance the performance of the Suno's engine and hopefully Suno could increase quality of the music at least on 256 bps for mp3 files. I mean why should Suno generate whole song when user needs like 30 seconds to get idea about the style and then just decides if that is good or not... But if you are interested, they miss one thing important for me. When I want to add next section of verses, I have no idea if the words will fit to music because I don't see the music arrangement for the section like time signature (e.g 4/4), tact or beats (number of the tact and it's length - count of notes) tempo, rhythm... I mean: The length and number of beats in music is rhythm. Rhythm is a fundamental musical element that determines how notes are arranged in time and how long they last. The duration of notes and their organization into beats create a rhythmic pattern in music. So I need to see the notes or the beats so I can add the words. If there would be a tool which would display this portion of music, I can find out what words will fit. In my language (Czech) we use a lot of long vowels, if pronounce incorrectly, the song is very odd. So we need to select carefully the words. ... In conclusion, if we could view a certain portion of the song and then to view the music rhythm, pattern, then we can write own lyrics to fit the music.

[modest frigate]

I did not work with udio, I dont know what they have. But how can song managment be missing completly in such a software, when a user has thousands of song every month! It is like missing the knife & fork at the restaurant. Why they did not care at all about this from the START ? Not even a search function was there from the start!

[robust jewel]

Out of curiosity, Grabby, why are you even here? You seem to really dislike Suno and have a personal hatred of some of the staff (even if you don't know who they are), so why do you stick around?

[tidal star]

right, wanna backup what noli said. plss can we have organisation folders for the library, and can we delete all in the trash bin?
just some quality of life improvements would be very nice!

[modest frigate]

Version 1.3 of the song management suggestion added in the song line, so the user can know also months later how the song was and which part is good without having to relisten the song (which so time consuming!). I added also some other important things were suggested already like bpm and song key. The flags could be switch on/off with mouse left click in the desktop version. No serious musician uses a smart phone to make music with Suno. It is a piece of cake to implement all these functions, I am a programmer and I know this.

[stone cobalt]

This is my suggestion regarding process of creation of song. Make it more steps which would help to get better control over quality of the song. The user could also provide the complete lyrics first so the AI has idea what we will be working on.

[sage smelt]

You could learn music theory and harmony theory, then you have everything you need... you can learn an instrument but if you don't know the basics of music, it's useless to write songs. The rest is creativity 🤷🏼‍♂️

[modest frigate]

please think before you write: I talk about the meta tags and the user manual ! What does have to do with music theory ???????????????????

[sage smelt]

If you can write sheet music you will know how to use meta tags 🤷🏼‍♂️

If not... ChatGPT 😉

[modest frigate]

OnkelMao, so you like complex things that comes with no instructions ???? maybe cause you are from China, considering the nickname... 🙂 I dont even know how someone can defend or find a reason for hiding the meta tags and genres list from the users !!!

[worn atlas]

#😜┃suno-memes message

[tepid trench]

Have a look at what other AI services provide for comparison - for example Udio - to see if it really is as unusual for a music generation AI service to not have tag documentation, organised song management, a way to go to a particular page etc.

Guess how you navigate to a particular page of songs there? 🤔

[modest frigate]

Man Furrybeans, what the heck is such a was of thinking ???? Should we aim the stupidity and laziness of others or the common sense, the logic and the intelligence ?
How can you defend such things with such arguments !

[tepid trench]

Just saying how things are.
I'd like the UI to be better. Id like there to be better/official documentation.
Suggestions have been made.
One day maybe these things will happen.

[grave fulcrum]

Why extending fron own audio there's a 1250 character limit for lyrics????

[finite trellis]

The ability to separate music from voice already on the generated song

[young bloom]

Can you guys take back the auto-removal feature for the create page? Recently ive had to get on and search through 20 pages of my library to find something i worked on a day or two earlier. It makes the workflow take so much longer.

Can you guys allow all the entended clip images to stay in the "whole song"? Whenever you go into the edit details section, and hit cancel, it erases all of the images except the currently selected, and i've lost a lot of cover art this way. And since Suno doesn't utilize seeds, it is impossible for me to get a similar image.

Also, can you add new kinds of feeds related to discovery? The Explore page is awesome, dont get me wrong, I use it all the time. But there are top 10 worthy songs out there on Suno that are buried under thousands of other songs that never get seen. Not many of us have a following on social media or connections within this community to have our content heard.

Also, voice chats for discord. Most of us would really like that. Like, unbelievably.

[stone cobalt]

I'm removing this content, because as you guys say later, I was wrong assumptions on udio.

[tepid trench]

I'm not sure that's correct. You'd need to ask them or their community (Discord) if you need clarification on it. As far as I was aware the terms are very similar to Suno's.

[tardy night]

I'm using Udio as well since is much, much better at orchestral music compositions and quality. There, even if you are on free account you "own" every generation but you have to credit Udio either in title, description or in metadata. And now you can earn extra credits by providing feedback to audio generations. Which is awesome! With paid account you have more credits (duh), early access to new features and you don't have to attribute Udio anymore. I wish Suno would be the same in that matter.

[young bloom]

The extra credits thing is wild, i wish

[tardy night]

Both! Ownership of free generations and extra free credits 🙂

[young bloom]

I mean, im not sure if ownership of free generations could be a thing, but the extra credit stuff? That would be wild for people

[tardy night]

Not to mention that it was Udio song that went for the first time in history of ai music mainstream

[young bloom]

Which one

[tardy night]

I found out recently

[tardy night]

BBL Drizzy

[young bloom]

Wait. THE BBL Drizzy?

[tardy night]

Yes! It was generated on Udio. There is a song on youtube. You wil find it.

[young bloom]

No. way.

I had no idea

[tardy night]

The owner is making a music video now to it lol

Anyway, lets keep this channel clean from this topic

[young bloom]

https://tenor.com/view/dr-dis-respect-doctor-strange-yip-yep-yup-gif-25698074

[tardy night]

Why not? Udio is already doing that. It's a plus for users who can't afford subs. They have much fewer credits to work with and limits, but with time, slowly but steadily, they will finish their tracks. I'm always for support for those less fortunate.

[young bloom]

Im in support of that too, however i would imagine that their ownership terms are a part of their business model.

A business model change isnt always easy to apply in scale

I dont really know how difficult it would be for them, obviously, but i can only guess.

[tardy night]

I would love to see that anyway. Maybe in the near future!

[young bloom]

Absolutely.

[modest frigate]

Hi, is Udio better also at other music styles in your opinion regarding sound quality or composition?

[tardy night]

To be honest, I enjoy more Udio at this moment. Suno AI was my favorite and its still good sometimes but I'm waiting for v4. I hope its gonna blow my mind!

[modest frigate]

does the sound quality also goes down at Udio when extending a song?

[hybrid crane]

Let's keep it Suno here.. Udio has their own Discord

[tardy night]

Exactly

[hybrid crane]

I asked to keep it Suno here.. and you are still posting a full talk about Udio?

If you want to give a suggesting, I advice to not make the suggestion 99% about Udio

[stone cobalt]

Send me personal message I will send you example what I have done yesterday.

[tepid trench]

@stone cobalt Suno songs are private by default (so they are not "published" automatically.)

[tepid trench]

To avoid any doubt or confusion, on the Subscription page, make it so that when a button is pressed to activate a subscription the user is given a short summary of key billing points with regard to payment and credits before proceeding to the payment. For example:

"You are about to activate a yearly Pro Plan subscription
Payment for $120 will be taken immediately
You will receive 2,500 credits on the 6th day of each month
Unused credits do not carry over at the end of each subscription month"

The information about credits not carrying over is at the bottom of the page and users often seem to not scroll down all the way and don't see it. Yearly billing is offered as the default and sometimes people take this when they don't mean to. Having a summary like this should make it very clear to users what to expect.

[stone cobalt]

Regarding my diagram, where I suggest to make the song creation multi-step process. I suggest that user could configure his account if he wishes to continue with the design as it is now, or to change it where is starts with A) Chorus creation, B) first verse creation C) bridge creation. So when he will click the Create button, then he will add the expected lyrics no matter how long they are. This does not generate content. Next step is the settings at step 1, step 2, and step 3, with various prompts and specifications. But the AI will choose default content from the lyrics which user pasted at the beginning. So AI can choose a Chorus lyrics from text where is not Chorus defined at all. Or can select first verse and modify it if needed to fit rhythm. So the user when will click on a button Lyrics Lenght or Mood, all should be pre-set by the AI based on the preparation process. But if user will do a change, it is necessary to play small sample like one sentence (10-15 seconds). Vocal range can significantly change the feeling of the song, imaging Vocal range of Mariah Carrey - even that user cannot specify her name he can prompt like "remarkable vocal range" or just "remarkable range"... Step 2 should include the add bridge...

[modest frigate]

Hey Radek, this is the best idea I have seen so far here! Congratulations ! This is a genial idea ! 👍 I did not have time to think about such a thing but you are right !
This is the only perfect way of composing without having to throw away parts which are good but they will never come back again !
I have so much songs where only a part is good and you cannot get it to put it back in the song ! And btw I have just listen to a piece from the competitor of suno and I am astonished by the audio quality with the crispy sound ! Wow!

[robust jewel]

What the... I just had a flashback to Julie when I read that. I had basically forgotten about her.

[modest frigate]

---

It is also very annoying to loose the text you write with indications for Chat GPT. Specially because I use that text more times to generate different lyrics.
So a back button like on a website would be good to have. So in the first window you place the indication and in the next you get the lyrics.
So it be good to be able to return to the indication window and to have your text still there.

[robust jewel]

Just so you guys know, there is a way to embed tiktok stuff on Discord so the video can actually be played instead of just being a picture with a play button on it. Currently, the announcement just expands the picture if you click on it.

I believe you just change tiktok to vxtiktok in the link

[patent sonnet]

Add a playlist showcase channel to the discord?

[gloomy surge]

#🐣┃suno-showcase

[shell kestrel]

I just opened Discord to ask about an Android app and saw that there will be an announcement soon. I would love to listen to my Library stream in the background with a proper music player on the phone which doesn't stop when the screen turns off or when another app is in the foreground.

[low oracle]

Make it so I can remove the lyrics and make it an instrumental

[nimble prawn]

Add PayPal support ffs. I want to subscribe but can't cuz I don't have a creditcard. Why is it taking so long. I asked weeks ago. Udio does have more pay options yet Suno only accept creditcard smh

[near fern]

Please write your news directly here on discord, or even better, on your website. Don't force your customers to use apps like twitter or tiktok just to read news....

[robust jewel]

All the information in the announcement was in the announcement, except for the video not being imbedded. The news so far was just "something is happening tomorrow."

[sly blade]

ngl udio sucks compared to suno. it's like it's only good with some genres. Suno is way more consistent

[violet bison]

How do u made your own songs? I keep trying to but it says this

Won’t let me click create

[robust jewel]

You haven't given it a prompt

[violet bison]

?

[robust jewel]

You have to tell it to do something

[violet bison]

I did 3 times

I put my lyrics in and I put custom and still nope

[robust jewel]

Ok, well, in that screenshot, you are at 0 of 200 characters...

If you are putting a prompt, do you have credits?

[violet bison]

It says 0-

Credits

[robust jewel]

Are you on a free account?

[violet bison]

Uh idk I just sighed up I’m new to it

I sign up with google

[robust jewel]

Sounds like you are. If you didn't subscribe and pay for credits, you're on a free account. You should get 50 more credits in a little while. Free accounts get 50 credits per day.

I believe it will be in about an hour and a half.

[violet bison]

Dang- oki welp thanks 😊

[west elk]

Credits refill every 90 minutes!?

[hybrid crane]

Free credits will refresh every 24 hours

/refresh

[rare obsidian]

Free Credits Refresh

Refreshes at 0300 UTC+0 (<t:1713236400:t> in your local time)

[west elk]

Ah, ok, it's just that the deadline is approaching. Got it, thanks!

On a free account, do you accumulate credits if they are unspent? Like, say I am away from my computer for a few days, do I come back to several days of credits, or is it capped?

[hybrid crane]

no.. you cannot save up

[west elk]

Ah, I see

[hybrid crane]

It's a company in the first place, the 50 credits per day are meant to try it out. But if you want to use it seriously, it's best to get a plan anyway.

[west elk]

Yes, I understand it's a company. However, I know that I am not (and realistically, will never be) a "serious" user. So I wanted to understand how the free tier works.

[river notch]

singing suggestion(given it before)

https://www.youtube.com/watch?v=8n4gS4MGetM

[shrewd pike]

I’d really love to be able to create a radio station of my songs to embed into my website. Could that be possible?

[sage mirage]

If you are a paid subscriber you can do whatever you'd like with your songs

If you mean somehow leverage Suno to do the song playing for you on your website, that would be in the API realm which does not have a timeline on official support. As of right now there is no official API

[mighty ledge]

Hmm, a feature i think could be more open unsure how it could work, but a section like a drop down menu that lets you select some voices and then assign them to certain parts of a song would be a ideal feature.

[shrewd pike]

Yeah I know. I’m on the higher plan, I do have a radio station too, but I’m creating more songs than I have allowance for on my radio station plan 😔 I can’t embed a playlist either since it only plays the first song. I’d like a radio player for embedding from Suno.

[leaden marlin]

there are many types of tracks/songs that you really just cant replicate/describe effectively to someone or an AI in words
like, say i want something like kanye west's On Sight. that's impossible, literally, i tried, it just makes me bass boosted EDM. if the AI were able to take into account the full song, and instead of extending it and incorporating melodies and parts from the OG track, it'd take the general sound and feel and do something new with it, that would be a game-changer and would make OSTwork much less a pain in the bum than it is now.
a little extra button next to extend that'd say "make similar" would be groundbreaking

[gloomy surge]

So the problem here is that Suno doesn't recognize specific songs (or maybe it does, but it doesn't recognize it well enough)

Which is why it's not giving you the style of "On Sight" because Suno has no idea what that is

Also the AI already does it's best to make it sound as similar as possible to the original song, it'll probably work better whenever the next version releases

[fair ginkgo]

Have a view option showing different takes of the same prompt of different track with the ability to cut the audio on each track so you can create a composite or all the different takes, or, the ability to upload a full song.

Support uploading a full song. I edited a song in a DAW and wanted to upload it for the contest, but I didn't see a way of doing that.

[leaden marlin]

Agree. When the new Suno rebranding coming up later, you can change when the time comes.

Not only the role icons tho.

[leaden marlin]

you misunderstood me. i described it so hard it reached the character limit

[shell valley]

Sometimes i wish to reuse the style not the prompt. A new song which sounds like the first one. Not only the same genre, like from the same artist.

[novel compass]

The announcement looks exciting. But when will be we able to randomize the playback of our playlists?

it must be a bug at this point.

[jagged trail]

So aside from the possibility of folder system for music, and adding QoL stuff like deleting all of your trash bin with a press of a button, Could you also consider allowing us to be able to buy points for other people on there pages, theres some artists that im pretty sure are unable to fund there own enjoyment on the site, that id like to hear more of there creativity come out, would be much appreciated.

[autumn basin]

you should definately not lose your credits you paid for... i lost 3.5k credits. its like you buy sometjhing at the store and if you dont manage it to run out within time you need to give the product back..... come on man... i want them back...

[wanton hamlet]

Would be nice if you could set specific instruments into reverb/echo.

[autumn basin]

i would buy some more now but i dont because i know im in holiday this month and it will result in another loss....

[merry sigil]

It would be nice to add a "Seed" to Suno so that sometimes you can go back to liking the same style that was once used with different text for example

[wanton hamlet]

how about a music equalizer, remaster once song to a V2 (including boost to vocals)

final suggestion for the day, how about a 3-4 sec preview before hit master it and pay the creditz ? (wishing for free-for-all)

[grave crest]

this is why Udio now gets my money and im more invested creating with that, cause clearly suno does not want to have billions into music by offering different payment methodes? and i really want suno to be able to have me as a subscriber

[stone cobalt]

"i want something like kanye west's On Sight" Do you mean Suno should copy the music? The only way to do that is to steal the music and train the model, which is illegal.

[wicked hound]

https://suno.com/playlist/eeae2397-9e90-47b6-bfa1-4eb62a24d824

I have a playlist with dozens of unique and clean songs that I bet the Suno team might feel is a little more representative of the spirit of Suno

[solemn bolt]

can we get a notification when someone follows us so it's easier to follow them back?

[young basin]

===============================
| Big PROPOSAL 🙏 |

We ask you - the entire community - PLEASE - make the music divided into 2 tracks :
+ vocals and
+ music

  thank you very much 🙏 🙂

===============================

[rotund lava]

I got an email I should tidy up and delete some unused stuff 🙂 I want to suggest a bulk edit feature so you can delete multiple items at once 😉

[robust jewel]

Stems are a known request and we have been told they are on the roadmap. The way the AI creates things, they aren't currently an option.

[robust jewel]

Control-click to select them each individually or click one and shift-click another to select everything between them.

[rotund lava]

Oi..thank you 🙂 Would never have thought of that in a browser window 🙂

[robust jewel]

The way it works right now, it won't repopulate the list when you remove multiple, so you have to reload the tab eventually. It also will stop thinking you have a second page, because it thinks the first one isn't even full.

[shell valley]

define the length. I try to make a short sound, and get over a Minute. Need only a 10 second jingle.

[royal phoenix]

PROPOSAL: negative prompts
example:
Prompt: fantasy battle music
Negatives: guitar, piano

Result: fantasy battle music without guitar and piano

PROPOSAL: upload audio reference and generate similar song but no the same

[wanton hamlet]

Perhaps a temp/direction slider, alternate between ai artistic freedom to slave of prompt from 1-10

[pseudo lake]

Pretty sure an Adherance slider is being considered, though when and if it comes to fruition is another story

[hollow plank]

Roles to tell people what's our music taste

Like example

Synthwave, rap, R&B, Hip-Hop, Regga, Pop, EDM...ect

[steel fractal]

Highlight the specific prompt word/phrase that is being moderated.

[frank swift]

please put shuffle option for playlists!!! SHUFFLEEEEE

[short vigil]

I would like to have the ability to upload audio for a specific instrument and let suno make the rest of the song for me with my melody and notes.

[hybrid vine]

Can i make a part of the song going faster then the others?

[graceful ruin]

Why just 2 tracks? Rather proper stems thank you very much...

[oak narwhal]

Suggestion:
Is it possible that Suno could generate the Sheet Music, or even just tabs, for the music it generates? It would be amazing to use it as a foundation, for figuring out how to play, what you've generated.

I suspect that a lot of the musicians using Suno, would greatly appreciate this feature. Not to mention there'd be a huge demograph of musicians, that could then use Suno as inspiration for what to play with their song lyrics.

[blazing lark]

Not to mention that's easily convertable to midi.

[pseudo lake]

Whoever asked, 100% worth it. Just spend the $8 instead of buying your next pizza, and you'll realize how much better Suno is than pizza.

[wary hazel]

I want the style of music to include famous people like Eminem, etc., to include his style of rap and other styles of famous singers and musicians.

[pseudo lake]

Great news-- you can already do that!

Since we're not trying to outright plagiarize their work, we can't just simply say "Eminem" or "The Beatles", but we can describe their musical style. For instance, depending on what era of The Beatles you're looking to emulate, you could try "psychedelic brit-pop, wistful acoustic, melancholic harmony", and try experimenting with different descriptors for that style of music-- with careful use of extension points and a bit of practice, you can create almost anything you want!

[wary hazel]

I understand that you can't plagiarize, but is there any way to write down, for example, Eminem? it's rap, no one can really repeat it differently, either Michael Jackson or Verba, a band from Poland, etc. I would really like to do something in their style, but not plagiarize. they have good flow, I love this style, and if I combine them, it will be really something

[pseudo lake]

Part of the issue with that is that the AI hasn't been trained on what specific artists sound like, so if you can describe Eminem's music or rapping style, you can recreate it, but saying "Eminem" doesn't mean anything to "a musician who doesn't know who Eminem is".

[small cedar]

Hey, I would like to have only the instrumental of a song created in Suno, possibly in different versions. So that I can sing on it myself. Or is that already possible?

[pseudo lake]

Not specifically in Suno (though using "a cappella" instead of "acapella" has been found to give good results for generations), but if you search for "stem separation", "openvino" or "uvr" in the tech-discussion channel, there's been lots of discussion on the matter 🍻

Personally, I LOVE Audacity + the OpenVino AI plugin, super easy to use, and I find it gives good results. Best of luck! ❤️

[prime rivet]

I'd also love a feature where you can recapture the same instrumental or revise the lyrics over the same instrumental.

[indigo ginkgo]

I would love in instrumental to be able to generate loops, if you can get that to work (i know generating tiles in stable diffussion is possible, so loops should not be impossible either) that would be great!

[stone cobalt]

In real, various sections of song may have various tempo. But I don't quite understand what you mean by slave of prompt. I think having a song where the tempo is really constant is strange. But the same thing is with rhythm. It is good to make song which is not monotone... So you need to change tempo and the rhythm sometimes. I mean, normally you play the verses with same tempo, but Chorus or bridge or intro or outro may have different tempo and rhythm. I guess the property name could be tempo seed and rhytms seed. Too high value would lead to odd results. Too low like 0-1, could result in very monotone song. In fact Suno created such song for me when I have specified what rhythm it should play, so the result was very robotic and unempathetic.

[glass trellis]

Hello Suno peeps, is there anywhere we can email the Suno team for recommendations and ideas?

[tepid trench]

This channel is where to post suggestions for the team.

[ebon vault]

Yeah. Of course iOS the lesser of the total market of mobile phones. Why work on a proper application while your web uix is garbage for the global market share leader, Android, with 76% over iOS at 21%. Maybe you should design an app for Harmony with their 3% market share before Android and iOS since you seem to think it a better idea to go for the lowest hanging fruit.

And if the Android app is some garbage iOS to Android port with the standard weird nonsense buggery that comes about with that. I'm out 100%

My suggestion is that you focus on the total market share of the global market when designing a mobile app while you have slacked on your mobile UIX web.

[small cedar]

Thanks, I'll take a look at the channel. I had tried various voice remover apps but the quality of the output was poor.

[ebon vault]

Cos currently the mobile web is horrendous absolute garbage.

[ebon vault]

Also try inlining [Verse 1][Vocal Solo] or [a capella]

[glass trellis]

Thank you, also does anyone know tips for the lyrical portion? I've had written a ton of lyrics for demo songs but the delivery of the vocals and instrumentation isn't quite right. I'm not sure if it's the way I'm writing the lyrics or what... Also also, does anyone have tips for getting the most out of the styles portion? I'm having some difficulty finding the right sound/style for songs after inputting things into the Style of Music portion.

Are there any apps that can isolate different tracks after the song is made? Can you isolate the different tracks in Suno?

[ebon vault]

Same as I said above, try inlining things.

For example, if I'm trying to utilize vocals as vocal samples for a turntablism style song.

Lyrics...

Lyrics

[Vinyl Record Scratching]
Sample

Lyrics

Within the lyrical sections

Also Just because it does not work, the first generation does not mean that it's not going to work. It's more likely to work but it doesn't always

[low adder]

iOS App Suggestion: Add Lyrics, no way to view lyrics while listening to song

[ebon vault]

@glass trellis Here's an example of it in action- https://suno.com/song/ba1c66ad-38e5-48f2-af9a-ffaec633d86e

If you go to the end you'll see where in the style prompt. I stated vocal announcer, and then at the end I put it inline... So while announcer was in the style prompt, it didn't activate until I set it inline

[visual brook]

wrong chat for my last message

[tepid trench]

This kind of question is best asked in #📚┃suno-school (this channel is for suggestions on how to improve Suno generally.) With regard to isolating the different parts of a song: Suno doesn't do this at the moment. There are online services llike https://vocalremover.org/ or apps like Ultimate Vocal Remover that can help but they won't give perfect results.

[ebon vault]

Funny story, non ML/AI removal systems work best because while to us the vocals may sound somewhat natural, apparently they aren't the same to the voice is used to train the instrument removing vocal extraction AI. So they come out sounding extra odd lol

[glass trellis]

I've tried using the brackets for whatever section I'm writing in, however it seems to skip things like [Instrumental Intro] and [Instrument solo]. Also the way the songs are structured sometimes do not work, for example when I put [Chorus] or [Pre Chorus] the songs tend to sometimes bi-pass the brackets and make the instrumentation change to make them sound like a verse or bridge. Is that due to the length of the lines for sections?

[ebon vault]

Sometimes it does completely ignore. But other times it probably is acting on it in a way that isn't obvious.

For example, if the intro might be just very short because there's no lyrics which are what it may be basing timing and cadence on. Lol

[glass trellis]

I'm primarily using Suno as a tool for song ideas and demos, and so far I'm loving everything it's provided!!!! After making 10 songs I immediately went to pick up the yearly subscription for 500 songs per month. Is there a channel specifically for tips on cleaning up/refining the song making process?

[ebon vault]

And I really don't know why it does what it does. I have not really done extensive testing due mainly seed dynamics...we can't use reuse same seed so we can't test what changes come from what prompt mangling we do lol

[fresh lodge]

I will give that a try, thank you for the tip.

[ebon vault]

Hahaha yeah I've been paying for months cos really even just for kicks it is fun. But also cos it is very interesting the things you can get it to do. But lets take this to #📚┃suno-school or #💬┃general-chat since really this isn't the place for anything but suggestions.

But really it should be a forum type which would make more sense, than chat

[twilit hedge]

The first part Suno can do already, the 2nd part where you want Suno to create a song using your melody and notes? not so much...

[candid knot]

I think they should focus on reusable singers/voices rn and negative prompts. After that they can keep updating and adding whatever they please. I need those right meow.

[junior sundial]

there are lots of fake Suno apps in the iOS appstore -- please get these removed so people don't get scammed

[near fern]

okay, that news was quite disappointing 😴

wake me up when something cool comes out (e.g. inpainting)

[finite rampart]

I wish there was a we can go back and Edit the songs we have made, for example i mispelled 2 words but can't go back to fix it, on what was the "perfect" sounding song.

also choosing the voices would be dope

[sage mirage]

inpainting is coming with v4

So you should be able to correct mistakes like that

[mossy nexus]

Ya know...with the song length increase to four minutes, it'd be nice to have an increased character count.

[robust jewel]

Big news! (Not really!) It's been over a month since we were told V4 was cooking. I'm not one to rush things, preferring things be done well instead of fast, but an update would be nice. Has anybody checked the pot to make sure it isn't burning to the bottom?

[sage mirage]

v4 has been talked about for monthS

Probably will be a large thing than 'just' a new model

It's still being worked on and the devs are active and all that jazz

[sleek abyss]

I would like to suggest a "remaster" feature. To let us take older version songs and up them to current model quality, or to do a musical version of a paint-over or remix process like the image-gen AIs use, new prompt modifying the older gen. Probably would take more than the standard credits but would be VERY handy.

It's best to implement something like that early as its only going to become more handy as quality, coherence, etc, improves.

Some kind of equivalent to inpainting would be sweet too. Select some sections of the timeline and reroll them.

[last gust]

My biggest suggestion that I could really use right now is the ability to shuffle our playlist or library

[arctic herald]

Suggestion: in the app when recording own voice and seeing output, remove the first part where it’s only the voice which was recorded, just skip straight to the generated track instead of having it appear after own voice replays

[wanton hamlet]

sorry for the misunderstanding, Im not well versed with this particular ai wording and settings, I was referring to temperature as in some other chat bots randomness etc Im just a layman

[cloud schooner]

Suggestion: I don't know anyone else asked this. Will you guys have the app available for Android users and other smart phones too? I don't have an iPhone while I've been enjoying using Suno on my desktop computer's browser but I would love to use it on my android phone too.

[tepid trench]

An Android app is planned 🙂 this is mentioned in the announcement ("stay tuned for android") -
#📢┃announcements message

[cloud schooner]

That's good to know. Thank you! 🙂

[modest frigate]

---

The possibility to add an intro to a song "done" already would be a great thing. Because some songs start right away with the singing, no intro, nothing, specially in V 3.0.
Udio can do that if I understood well.

[hasty osprey]

[Intro] or [Instrumental opening] or just whatever, improvise

[tepid trench]

Most reliable way to get an intro at the moment when starting a song is to begin with an instrumental, then extend with lyrics from the point you want the vocals to come in. Being able to later add an intro or parts before the existing song would be great though (I have some old v2 songs that just start somewhere in the middle that I'd love to extend backwards.)

[hasty osprey]

Let's introduce daily tasks like liking, fallowing other creators and their songs, introduce comments. Rewards credits. This would help with people actually starting to explore and interact, creating even better environment for all of us

[modest frigate]

it is almost impossible to get the same sound as Suno produces with a DAW .. and besides that is time consuming

[modest frigate]

good idea. In 3.5 adding [instrumental intro] in the song lyrics works. V3.0 ignores that many time.

---

is there any document which explains what meta tags has been added in V3.5 comparing to V3.0 ?

[modest frigate]

lets better first make Suno eliminate the known bugs, add basic missing things in song management, add the prompts list in the help page and also the difference between 3.0 and 3.5 regarding metatags ...I have never heard in my life of a software with millions of users not to include a link to the most important documentation or to postpone bug fixing for MONTHS! These are mandatory things!

[hasty osprey]

Idk it was 1 google away
https://www.suno.wiki/faq/metatags/

[feral aspen]

The lists of meta-tags is not a fixed known list, not an instruction manual. Instead it's a list of things people discovered and tried. It's more like a list of Stable Diffusion prompts. The devs occasional mention they tried to get some tag working (even [chorus] at the start wasn't very intentional) but all the fancy stuff is community discovered and documented.

[solid sage]

Have the ability to keep the beat/music but generate new vocals for it

It would make duets heck of a lot easier

[modest frigate]

that was a good joke Jonathan 🙂

[solid sage]

Honestly this format always works good for me.
[Intro]

[Verse]

[Chorus]

[Verse]

[Bridge]

[Chorus]

[Outro]

[End]

[modest frigate]

very few songs have only one verse before chorus, there is not enough time to build tension in the song

[hasty osprey]

I know how confusing AI models can be, but in reality you have to experiment, it's like a casino on generations. You won't get what you want in first try, or you can, all on luck. It's easier to find what works and what's not yourself or just ask others, that's why we're here. Rome wasn't built in 1 day.

[solid sage]

It depends on the tempo

For my slower acoustic style it works

[modest frigate]

it depends also abut the verse lenght....but still

[feral aspen]

Well now that Suno has a search, it's easiest to search and look at other's prompts to start

[solid sage]

And it depends on line numbers

I usally do 8 lines for chorus, 6 for verse, 4-5 for bridge

[hasty osprey]

And prompts for song style. Check other creators, check their lyrics and prompts or metatags they were using. Create your own version with the help of others

[feral aspen]

If you find yourself annoyed that the songs are too traditional or poppy, also try using no tags like [verse] and [chorus]

[hasty osprey]

I always get inspiration by looking at others, especially when their prompts and tags worked well. Experiment.

[solid sage]

I havent needed anything too advanced. My styles mainly a slow acoustic country

[modest frigate]

I can't believe the mods find excuses for ANY abnormality from Suno! Of course there is also google to find the metatags.. it took me 3 days !!!
I will start to gather all these funny excuses... it will be a good comic book, specially for developers.

[verbal ivy]

could we get a local timestamp of when a generation was created

it would make it a lot easier to keep track of what im doing

[hasty osprey]

This platform is for creating and experimenting with SUNO model. Not to keep popping bangers each round. It's at least not yet what it's capable of. It's a play field where we all play and share the results. Imagine this as NightCafe image generatorion platform, everyone experiment, laugh or praise other creators work.

[modest frigate]

That is a good idea and observation! I was just trying today to find out which songs I made at 3 am and could not find any

[modest frigate]

🤣 🤣 🤣

[hasty osprey]

Idk what's so funny 👀 some people just feel like they gonna be the new top artist with the help of AI. It's not happening unless you put the work, gather information, grind your lyrics, tags, get prompts right etc etc. Do your research, learn and try again if you failed.

[pseudo lake]

I understand your frustration regarding the documentation issue. However, I want to highlight that this is the second consecutive day you’ve expressed significant dissatisfaction about this matter. Given the tone of your messages, it might be more constructive to approach the situation with a collaborative mindset.

As Jonathan and Wings of Pride have mentioned, the community documentation is readily available, and it would be beneficial to utilize those resources moving forward.

[reef grotto]

Things we must have in the future updates:

1. 'Lock tempo' feature to keep the same tempo
2. 'Lock key' feature, so track would play in the same key
3. 'Disable vocals', because I sometimes get still vocals, even if I'm using 'Instrumental' switch. Even with different no-vocal tags vocals appear every now and then. Sometimes it takes 30-40 different generations from me to get a generation without vocals

[modest frigate]

How come you did not want to see that the lacking of a documentation link was only ONE of many issues in my comment? I have just replied to another user who wants some new stuff. But any normal thinking person should know that lacking of mandatory things and bug fixing should have the highest priority. And even the think with the missing link you missinterpreted. I did not said there is no docs, I said only that any company makes a link in the software to the doc and does not hide it in a wiki page wich is not linked in the software

[pseudo lake]

I’m well aware of your other posts, I’m simply choosing to ignore them.

Many, if not all of the points you’re raising have been noted in #1252298754787377225 and #1252292428258480209 .

It’s not official documentation, but again, the suggestion to link to the community wiki has been noted.

…how did they phrase it to Walter in The Big Lebowski again?

[modest frigate]

ignore me, cause I ignore you... I did not tell you where to write, how to write, what your wrote... so stop chasing me

I don't write here mandatory things to be noted somewhere, but to be finally done, implemented.
Only writing them somewhere did not change a bit in more than 8 weeks...! so, urge Suno to do also something other companies do without user complaining !

[tepid trench]

This is a channel for suggestions by all users. You want documentation and UI changes. Other users want new features and have their own opinion of what they feel is important.

Out of interest, which other AI services document their supported tags?

[modest frigate]

which other software company hides important information for the user on unlinked wiki pages ?????

[pseudo lake]

They aren’t hiding anything, they just don’t have official documentation.

Legitimately asking so comparisons can be shown to the devs, can you share an example of a similar service where the supported tags have official documentation?

As an olive branch— I understand it was frustrating having to spend 3 days looking for documentation, only for that frustration to be compounded further by realizing the only documentation was community-sourced.

Sharing actionable information like this would be WAY more helpful than making repeated complaints. ❤️

[modest frigate]

Who cares about what other Ai music company do? Look better what software companies do ! I pay SUNO and there is no link to the OFFICIAL wiki page from SUNO !
I did not even know udio or other to that point! What kind of sick argument is yours ? Should I applaud now that I had to search 3 days to find the suno wiki page ????
Should I applaud that another stupid company exists which does the same ? Should I applaud that a company which say is the top of AI does not have official documentation?
Who saw something comparable in the software world ???? This is sick!
And the missing link is only ONE issues...the other ones are the known bugs staying untouched by more then 8 weeks now ! Instead of fixing bugs they develop an IOS app !!! Man, that is sick. It would cost max. 10 days to fix all the UI bugs and at the most 15000 usd ! But who cares about millions of users ? nobody !

[pseudo lake]

who cares what other AI company does

You do, seemingly?

[tepid trench]

Regarding documentation, you said: "urge Suno to do also something other companies do without user complaining"

Suno is a generative AI service. To be a fair comparison, you would need to consider what other generative AI services offer. If other similar services do not offer such documentation, why don't they?

[modest frigate]

the more you mods defend with such childish arguments the grievances at suno, which are unheard, the more obvious it becomes ! users dont care WHY a company does not have documentation, why they dont put a link in the software, what other companies do, why they cannot fix a single bug in more than 8 weeks! They only wonder, why the normal and the common sense is missing in Suno. This is only what they care! And many they pay for that ! Should we all applaud now such abnormal behaviour?

[gritty prism]

With all due respect, but as soon as a “company”, or “society” or “service” (in short, a group) begins to apply monetization to itself, this automatically obliges it to take care of the service for its clients - who do not voluntarily donate, but pay for a certain type of service... Isn't it?

[pseudo lake]

Not quite sure what to say, you make comparisons to other companies, then get indignant when asked to list the companies so we can find a common ground.

[pseudo lake]

The service IS being taken care of, seeing as how we can use it.

Feature requests (regardless of their level of complexity) aren’t the same as bugs, though.

[modest frigate]

haha, theharber, you might think that udio and suno are the only software companies in the world 🙂 never heard what "software companies" means... 🤣 you think, Udio is the standard in the "software world" ... 🙂

[pseudo lake]

This is getting close to spam, if you’re just going to blatantly refuse productive conversation and spout the same talking points…

[tepid trench]

Of course. And they provide the service as described on the website. For example, the Pro Plan provides the following.

[modest frigate]

stop commenting continously my comments than! You spam and you want to have the last world... I blocked you know forever since I never read a common sense argument so far..

[pseudo lake]

Never said Suno and Udio are the only companies, just asking for you to give an example so that this can resemble anything close to a meaningful discussion.

[gritty prism]

I understand you, my dear, additions and errors are two different things. But - subscribers don’t even look at the “road map”, or at “working on mistakes”... All we see is either ignoring or “moving” into a different direction from the conversation. Those. - even some “conditions” of the tariff plan are far from a “public offer”. Even I, using the “Premier Plan”, did not get an intelligible answer from your “support” answer - regarding payment for the service from other countries. They answered me in the spirit of “show the card and pay” (like a complete idiot). That's what I mean, I hope you understand me.

[pseudo lake]

100% agree regarding the support response— even having an estimated turnaround time for a support / billing email would be very helpful as the service grows so rapidly!

Definitely room for improvement ❤️

[gritty prism]

• so we can find out about this seemingly “small” problem, for some reason - only here, not “in the main news” - do you know what I mean? 😉

[pseudo lake]

(I’m not sure if you meant me by “your ‘support’ answer”, or a response elsewhere from Discord or email? Either way, could you be able to share the content, either here or in DM? I’m sorry that you were made to feel uncomfortable, though 😦 )

[gritty prism]

I believe that this could be due to the fact that I live in the “wrong country”... And I, for example, have to find ways to renew and pay for your wonderful service, despite the fact that “we” have been blocked almost everywhere , and "condemn". (Although, I personally have nothing to do with this, and I don’t have the slightest desire to develop this topic. I work, and at the same time I make music.)

[pseudo lake]

🫂

Thank you for sharing— I’m sorry that all of the global red tape has to affect so many things.

I know it’s not an immediate resolution, but I added my upvote to https://discord.com/channels/1069381916492562582/1254232844767854665 as I noticed that the suggestion didn’t have much attention.

I’ve seen some people mention different companies for services like this, such as n26(?), but admittedly many of us in North America are aware of the limitations but perhaps not to the extent.

I hope you’re able to continue sharing your music with the world ❤️

Daizy has suggested PayPal & ideal, but please let us know if there’s any other services available in your country that could be added to the list for consideration.

[kindred hull]

Is there any way to include a negative prompt? For instance, I would like to request an instrumental WITHOUT piano (in image AI Gen, it would be "-piano" or "-not piano". Is there any way to do this in Suno?

[tulip steppe]

Make a button where you can upload a song or choose a song you made and choose a instrument and it would write tabs and stuff.

[kindred hull]

Ups, sorry! It seems that I was on the wrong forum. However, maybe my post can be reconverted to "Please add negative prompts functions"!

[tepid trench]

You should be able to edit it 🙂
You can also react to this suggestion to show support for it: https://discord.com/channels/1069381916492562582/1253341197917884530

[kindred hull]

I have reacted to the suggestion. It's funny that the example provided it's also about the piano. I guess there are too much piano in Suno's instrumentals!

[silk gull]

is it possible to get the ai my voice and let them sing a song for me in my voice?

[gritty prism]

maybe this has been considered (bringing up and re-reading the entire thread is hell), but there is a serious lack of a simple editor for trimming the “rudiments” of songs that haunt everyone permanently... Especially at the end of the song, I will modestly keep silent about the “beginning” ...

[tepid trench]

This would be useful. As it was asked for by several different users I posted it as a "known suggestion" recently - #1252773477157699664 message

[gritty prism]

Yes! You are right, because... It would be possible to add a bunch of other useful small functions there, such as “crop, copy, normalize, etc.”

[tepid trench]

I think normalisation happens automatically but I could be wrong. Just simple cutting of the unwanted parts of the song is a good start. Maybe eventually it will be more advanced but anything to get rid of unwanted song start/end would be better than nothing 🙂

[gritty prism]

For simple editing, I use this https://editor.audio/ - simple and accessible. This is exactly what is missing in SUNO.

[visual brook]

The new Suno discord icon deserves an #📢┃announcements post - sleepy me before my coffee took a full 5 seconds to find the icon had changed in my lists! Terribly inconvenient

[arctic whale]

I love the new Icon. It pops nicely. ^^ I kept overseeing the darker one.

[tepid trench]

I can see how the stark change in icon would be confusing but if they posted in the announcements channel and you weren't already looking at the Suno discord server.... 🤔 😅

[visual brook]

I have notifications on for that :
😉

[visual brook]

It does look better upon further waking up

[tepid trench]

It could do with having a darker outline or shadow around the text IMO

[arctic whale]

I have everything dark around it. 😎 Love that they changed to a brighter color.

[waxen hinge]

Im not sure if this is a bug or something, but as a dane its hard to get the lyrics right a lot of times. The AI will be confused from the lyrics, and will either make the voice sing in Danish or Norwegian because the languages are similar.

[pseudo lake]

I almost missed this comment because of the ensuing unrelated chaos, but this is a PHENOMENAL idea. Even if it were a small amount of credits, like an extra 10-20 a day (as not to massively skew the current distribution), that would really propel a lot of users who are on the fence to take a more active role.

[acoustic pike]

Male/Female Voice generating problem still do not work.
If I Write male voice in music style box it generate female Voice.
Please fix this.

[scenic forge]

Hi guys

is it possible to add in functional of making song, choosing voices of other characters as any other AI song creators offer? and also adding menu with uploading your own audio voice stream for creating songs ?
would be cool to see that fixture here

[robust jewel]

Try Vocals instead of Voice. It seems more reliable.

[acoustic pike]

It Means that i write 'Male Vocals'

[rocky wigeon]

Please, no one take this personally, but this suggestion is for the Discord itself. Can we please allow the #🪲┃bugs-web channel to be just a place to note bugs or what we think MIGHT be bugs we have experienced without a laundry list of responses as to how to "get around" said bugs or what to do to "improve" on the outcome of the fault by burning through more credits to "try" workarounds. If I were one of the actual developers, I would not want to attempt to scroll through all the posts present to try and find the actual bugs people are posting. I'm afraid that may be the case as to why some things may not be getting addressed. If someone is looking for help suggestions, is that not what #📚┃suno-school is for?

[acoustic pike]

You mean to say like that??

Person 1,
person 2,
person 3

[scenic forge]

something like
male voice 1
female voice 1
male voice 2
female voice 2
yep

[tepid trench]

Valid point. Both this channel and the bugs channel do get cluttered so things can get missed (I'm guilty of contributing to that clutter too...) Eventually #1252292428258480209 should end up replacing that channel so it should become more manageable.

[acoustic pike]

Use
[male voice 1]
[Male voice 2]
[Female voices 1]
[Female voice 2]
In the lyrics verse and chorus.
Hope it will works.

[robust jewel]

If people actually read known-suggestions and known-bugs, it would also reduce the clutter a lot

[pseudo lake]

I agree in theory -- in practice, without anyone helping redirect comments to #1252292428258480209 and trying to elucidate misunderstandings on somewhat researched topics helps prevent turn these channels into echo chambers.

I hear you though dude, sorry if it feels like suggestions are being "whack-a-moled"

[scenic forge]

oh its already added in menu? but i talk more about hud where i can choose it in lists, would be more usable, than just type in prompts
its about comfy interface 😄

anyway thanks for tip

[twin hatch]

Idk if someones already said but Id love to be able to make an instrumental version of an existing whole song so that I can sing karaoke/cover the lyrics

[tepid trench]

Yup. Lots of people want this.
https://discord.com/channels/1069381916492562582/1253172964728377455

[acoustic pike]

You want to choose the vocals of singers that what you want. 🤔

[scenic forge]

yup

[acoustic pike]

I want to my voice on that generated song 😂

[scenic forge]

yes that what i exactly ask too, add list with choosing other characters which i want to have in my generated songs
and also adding menu where i can upload "my voice stream" for using in generating song

[twin hatch]

Another thought I had was to be able to use a small selection of emojis to lead the direction of the vocals but idk how feasible that is.

Like if I want a part to be angry I could just preface that line with a 😡 or some other tag/command

[acoustic pike]

Thay will add slowly. Like Upload voice sample.

[scenic forge]

i hope it happen lol

[pseudo lake]

Another upvote for https://discord.com/channels/1069381916492562582/1252789683176345641 coming from me-- I'd love more reliability over that, or even to be able to train my own vocal model, similar to how I think ElevenLabs does it?

I'm guilty of forgetting to upvote the good ideas for visbility. I stink.

[acoustic pike]

waiting for the wonder.

[acoustic pike]

Why not try [Aggressive Voice] in your lyrics.

[pseudo lake]

Or stack synonyms; yelled, aggressive, shouted, angry, growled
I'd love more research into it it, but "voice" has the most common weight distribution, so I find including that contributes to common mishaps, like getting male vocals when you wanted female.

[acoustic pike]

You have to write it in lyrics box with verse and chorus

Try this
[Verse - angry]

[Verse - male voice - angry]

[pseudo lake]

You can include the suggestions I gave in metatags-- [verse] is just another metatag that tries to suggest song structure, but it's as much of a suggestion as anything else

A lot of songs don't use intro, chorus, verse, bridge, outro etc, I personally only use styles and "thematic" prompt decorations (skateboarding, sleepy) as metatags to add extra injections of style prompts, or to place specific style prompts at strategic locations in the song.

❤️

[acoustic pike]

It is the technical Problem.

[pseudo lake]

Yeah, but that's just AI in general. But keep an eye out as things evolve! Look at the difference between the first versions of Dall-e just a year or two ago, compared to now!

[acoustic pike]

Yeah I also talks about it

[hasty osprey]

Yessss

[short root]

suno generate an image as icon that identify each generated music . We can change this icon using an more aproprieted uploaded image but, when we download the respective video, the new image uploaded is not used. The video uses the previosly generated image. I sugest generate the video using the uploaded image.

[north jasper]

Can the ios app make WAV files available for download?

[topaz siren]

guys, should I type "rnb" or "r&b" into the style of music? which one will suno detect?

[obsidian gale]

I would second this!!! Because I want to create my character’s discography

[hollow harness]

Android app please

[pseudo lake]

#📢┃announcements message

stay tuned for Android

[topaz siren]

I would like to to have a feature implemented where you can disable the lyrics display when downloading a video format file. Having those brackets and all can be really immersion-breaking and confusing to some. Ty!

[grand moat]

I think a list of newer submissions, or a public playlist people can submit to, specifically requesting feedback or a rating, would encourage communication. Especially if giving detailed feedback could earn extra credits, or xp towards a level ranking system. Like how Newgrounds lets people gain levels, which magnifies their future impact. Even a daily or weekly Reddit submission thread. Give feedback, and gain a reputation as a helpful member of the community.

Another reference would be the “who should I start/sit/trade” system on r/fantasyfootball. Members ask, others contribute opinion, and those contributions are voted on. There are makeshift leaderboards to let others know who has earned trust over time.

[hybrid crane]

It's comming 😉

[elder valley]

you recently added a feature to upload your music to the site, right? and this music can be extended. from that moment on, I had an idea, why not make a feature for the AI to remix the user's music?

[lilac lake]

Quality upgrade will come to inst?

[midnight shoal]

Suggestion: give Suno a picture and it can generate a song base off that

[inland portal]

+1

[strong gazelle]

I can’t find the options in the new iOS app to “extend” or “reuse prompt”, if they exist please let me know where, otherwise I present the quite obvious suggestion of adding them

[heady olive]

I only need to bother you once! How the hell do you do Intros?

[pseudo lake]

Yeah, not all of the web app features are present on the iOS app (I'm still recovering from the lack of the ability to toggle between v3 and v3.5, but the stitches are still healing so I can't talk about it again)

[jagged kayak]

SUGGESTION: IOS APP - Allow casting for google and apple. I want to listen to the tunes around my house.

[short vigil]

Request: The ability to download STEMS for individual instruments in songs we generate.

[short vigil]

Sub-Request: This might be required in order for the above to work so, I'll request it too:

The ability to have more granular generations for pro and/or higher tiered users. i.e., when the song is generated, it generates instruments individually and then pieces them together for you for a more clean sounding song.

[zealous otter]

There should be an option to create/render the vocals and instrumental separately

[lofty parcel]

The number one thing on my Suno wishlist is to be able to reuse seeds, I hope this can get added at some point.

[half lotus]

Any way to type normal words from the english dictionary and not trip off the copyright system?

If not, id like to suggest that we not blanket statement words because its limiting to song creation. The prompt was "a nerd who fell in love with his Ai" and it has nothing to do with whomever "nerd" is.

[pseudo lake]

#📚┃suno-school message I'd say that's a good submission for the list they're gathering of content being mistakenly flagged for moderation

[leaden marlin]

I would like to see a differentiation between liking instrumental, voice and melody - sometimes one of the three is great

[true coral]

suggestion: any time you give me female vocals when I specify male vocals, you refund me twice the credits. Thanks.

[empty rover]

bpm and gradual bpm changes please

[rain pike]

Would like the ability to use an uploaded sound clip as a reference for a generated song, as opposed to use the "extend" feature on it.

[austere night]

is there anyway to make the ai not confuse post-hardcore for screamo? i know that screamo is a sub genre of post-hardcore but i don't always want to use screamo when i want post-hardcore

[austere night]

i sometimes get male when i want female

[ornate drift]

are any news for v4.0 out currently or at least infos what will be improved next?

[true coral]

you should get 4 times your credits back.

[drifting jay]

Hi Is it possible to add the possibility of making API calls from the moment we are on the paid subscription?

[sage mirage]

There is no official API at the moment nor has one been announced. The devs know people want it and it is on the list for the future, but as of right now: No

Any use outside of the website or official apps (of which there is only the iOS one at the moment) is a TOS violation

[sage mirage]

'soon'

[ornate drift]

XDXD

[sage mirage]

I know it's a trope to just say 'soon' but v4 and a new UI is all being actively worked on as well as a whole host of other tweaks and features

[sudden mist]

This is a simple request. Make it so I can search my songs on the app. So much is good about the app but not being able to search specifically within my songs is annoying when you have a thousand songs to scroll through

The web version does this. It’s annoying the app doesn’t

[strong sentinel]

it would be really nice to have a vocal stability slider a la elevenlabs, I have to invent really absurd genres like glitchtime ragwave and use [Glitch] headers everywhere if I want to destabilize the vocals, which is kinda necessary to get emotional expression. I notice the more human a voice sounds in suno, the less emotionally resonant/expressive it becomes. the really crazy robotic voices can do awesome melodies

[vestal lagoon]

Congratulations on #11 on the apple app store. Are you guys planning on making an android version? If not that's my suggestion.

[sage mirage]

#📢┃announcements message

"stay tuned for android"

[fallen sleet]

I would like if there was a way to play the songs that I generated from the oldest to the newest, then I will play my music that will be there while I'm playing or outside of Chrome.

[junior flicker]

i know this is probably not possible but ios 15 support please

[cosmic terrace]

about playlist , having possibility to add songs directly from album\playlist : such in a playlist press select songs and it open a small bar on the left or rigght side were you can tick songs you want to have in it press okay\enter\confirm and songs get placed in that spefific playlist 🙂

[vague harbor]

Guys you are way down at the bottom on the app store. I passed a ton of suno impersonators to find you. And i had to come here to see your icon to even know what app to download.

[noble canopy]

Bring the app to Australia 🇦🇺

Plz

[twilit yacht]

Hello , Thanks for the great suno, I would like you to add Egyptian music and Egyptian shaabi music ( local music) also I would like you to add Egyptian dialect because it's hard to creat Egyptian songs with arabian dialect. Thank 😘

[leaden marlin]

We need a way to be able to end a song and make an explicit outro at a given timestamp. So many credits lost trying to find a good generation to close a song that's not 4 mins long (like want to be able to force like 10 second outro or something lol)

Oh, and completely unrelated; Allow playing random order in playlist! :^)

[zinc saddle]

Agreed, @leaden marlin I’d take the ability to generate a special extend that’s just 1 second of silence. Set it just below the Extend option for all the songs so users can just select the time stamp and clip it short.

I have found I can get 20ish second endings if I shift the extension to V3 and just type [end], but clipping a song immediately would be nice. Even better if it didn’t cost credits of course.

Being able to resort the order in a playlist would also be a great feature.

[leaden marlin]

(can't make a thread in hear evidently, so I'll try to keep it brief / not super-offtopic veering lol.)
But that's def useful to know, thanks; Had no idea you could mix models like that.

|| right now I've been spamming a combination of [fade out] [ending] [silence] to mixed effect. lol ||

[zinc saddle]

I just discovered it. I think the official help section was updated in the last month because I don’t remember seeing it there when I first became a member. I was doing the same, OR taking a version I like and making an Extend with the same lyric content as the extend is limited to 2 minutes. Most of the Psalms I’ve generated so far for the project fall under 2 minutes, so before I discovered it I ended up tossing out a fair number of good generations.

[leaden marlin]

Appreciate it || And work you're doing. Followed! ||

[zinc saddle]

You're welcome and thanks! Note: sometimes switching versions can cause a drop in the audio quality. Always double check the final product after it's merged.

[keen crypt]

Why is there no explore bar on the iOS mobile app?

[elfin trench]

It's stupid but generate sounds

[pseudo lake]

Not all functionality is present on the iOS app— but a lot of users want all of the web app features on mobile, so I’m crossing my fingers for us!

[viscid vale]

Can we make the suno app compatible with Vision Pro its really just a tick in settings?

[leaden marlin]

Could you guys make Suno stop randomly making songs metal when metal and rock are not mentioned anywhere? If I just mention an instrumental it immediately turns the whole song into a metal song 9/10 times.

I'm legitimately at the point where I'm considering not renewing because I was getting such great hits a month ago and now I can't get even a single one.

[stray birch]

having a profile banner would be cool!

[kindred hull]

I have seen that some people are having issues with ending/fading out songs. Audacity may be a perfect fit to solve this. It's free and easy to use (at least for those things). If anyone has issues for ending/fading out songs, please feel free to contact me.

[vivid matrix]

320kbps really need to be a quality standard for downloads at least if nothing else for people paying for a sub

[tepid trench]

There's WAV downloads already at least (which gives the highest quality that the model can output.) I expect MP3 bitrate will increase as the quality of models improves.

[brazen moss]

Is it blocking China now

[brazen wing]

SUGGESTION:

• Be able to make album, so we can put our songs in albums, its easyer for us and those who want to listen
• sometimes you have a really good tune, but the lyric or promt are just not working, be able to use the tune over again and redo the promt and or lyrics
• wordpress plugin radio station like plugin for WP, connecting to suno playing diffrent genders or artist from suno site

[olive frigate]

Hi, does anyone else want the option to hide prompts from public. i.e. I spent a month finding the combination of Styles of Music that fits me, but if I make it public then anyone else could take my work.

[pseudo lake]

I can't express how much I want this feature, and as far as I know it's a possibility.

https://discord.com/channels/1069381916492562582/1253328110934425714 In the mean time, throw an upvote onto the known-suggestion 😄

[olive frigate]

Thanks! first post, I did look in the known-suggestions. I appreciate it! Upvoted it.

[wet inlet]

I really would want the upload audio feature to download longer audio

[strong swan]

i suggest that the profile embeds show the actual profile pictures we set for the profile. https://suno.com/@mrtree

[sage mirage]

I am vehemently against this, as a counter suggestion. A prompt is only a small fraction of what makes a song and it's silly to try to hide that. Even if we all had the exact identical prompt all of our songs would be completely and utterly different.

All hiding a prompt would do is make it so new users couldn't learn from anyone

Which then runs counter to the entire point of ~Suno~ itself.

Any prompt I use is not my 'work', the output is.

Like trying to prevent people from using specific musical instruments or something.

As for a suggestion of my own:
Show me what songs I have are a part of playlists in the player somewhere. So I don't keep adding the same song to the same playlist without having to go all the way to the playlist itself to see

[olive frigate]

I dont think its stopping anyone, its more like intellectual property, since I made the prompt, if I choose to share it then thats fine, there are many sources for learning, I used the exploire tab as an example. I can make a sound that is unique and reusable.

[sage mirage]

Right but it becomes an ideological one. I do not think anyone owns 'prompts'

Just as a general rule

Suno is all about collaberation, sharing, community driven stuff.
If you decide to share your song on Suno, you share the prompt as well.

If you do want to not share it, don't share your song.

[olive frigate]

you dont own prompts, you own the combination in order to produce a sound

[olive frigate]

Thasts whats happening

[sage mirage]

Then there is no suggestion since the solution is right there. If you want to keep your prompt sercet, you can.

[pseudo lake]

This isn't really helpful. People are welcome to share suggestions.

[sage mirage]

Right, and I shared mine as well. 🙂

I am a user just the same

[pseudo lake]

@olive frigate Thanks for sharing your suggestion, I'm emphatically in favour of expanding the ways that users are free to share content on Suno.

As it stands, we're forced between choosing to keep our songs private (at least on the Suno social media side of things), or making every aspect of them public for dissection and remixing.

People who enjoy the collaborative / remix side of Suno are free to do so, but you're by NO MEANS using Suno "incorrectly" if you want to keep your creative musical process personal, I'm very much so the same

[sage mirage]

Understood, but if you share a song you are sharing your 'prompt' already since I have ears. It is, after all, how we are all making songs with Suno already

If you do not want to share your song, do not flip that 'Public' switch

And share it on other social media where there is no prompt to share

Again, in my opinion

[pseudo lake]

Dude, there's a difference between listening and making the songs contents public for remixing within Suno. If someone listens to my songs and feels inspired to create something similar, I'm honoured -- less so if they simply click "reuse prompt." You're right, I'm free to not share, which is why I don't.

I know people are very passionate about their preferences, but I like hearing novel ways of approaching the issue.

Sorry if I struck a chord with you.

[sage mirage]

It's less struck a chord and more how I just feel about it. We are all free to suggest things, my suggestion is just as valid as yours. I am not attacking you, apologize if you feel that way.

[pseudo lake]

You told both me and @olive frigate that we're free to not share our songs in response to suggesting that the idea could be handled differently. Not very community-friendly, especially in a forum designed to foster new ideas of approaching things.

[olive frigate]

It feels like you are saying its about sharing and community but also don't share if you don't want people to see your prompt, so then don't share your music either.

[pseudo lake]

I already do, we're simply trying to discuss different suggestions for Suno. You don't need to agree with all of them, but likewise, I appreciate you sharing your passionate feedback about how public content should be handled.

[sage mirage]

I was suggesting a solution to the issue of not wanting to share your prompt, you can share your song, right now, without sharing your prompt. It is already possible.

You guys want to have the share-on-Suno aspect as well, I get that

[cursive shell]

It would be very cool if you can choose a part of a Song which sings for example a female. Or you choose a part who is a rap part?

Or is this still available and how it works?

[sage mirage]

You will most likely get there with inpainting, not possible yet but coming with v4

[olive frigate]

you can sort of, check out the wiki, https://www.suno.wiki/faq/metatags/voice-tags/

[cursive shell]

Okay nice

[sage mirage]

As suggested already, you can kind of do it with novel cuts and heavy prompting

[sage mirage]

It will work as you have the entire generation and you can re-prompt over a section

All we have to go on is what the devs have said, so I do not have first hand experience

But, theoretically, you can just paint over and reprompt in a song you already have

Fix lyrics, add solos, change words being sung, change genders, all sorts of things

Inpainting is coming at it is going to be a part of the v4 model, probably with a whole host of other things Suno will do at the same time

[midnight shoal]

I don't know if this is possible but here goes: Explore section could help us find the kind of song we want to put in prompt, by turning Explore into something like a Prompt Analysis section.

I put in song prompt, Suno does it magic and analyze the prompt, then give me the exact Genre and Subgenre I'm trying to describe in Prompt section of Create.

This can cut down prompt experimenting massively and reduce credit waste

Instead of a modal wheel as it is now, Explore could be visualized as a tree branch or fish bone structure

[pseudo lake]

I love this idea! That could be helpful for people who have a style they want to emulate, but need help thinking of specific style prompts.

Frequently we get people asking how to recreate specific artists' styles, but they aren't musically proficient enough to break down "Eminem" into descriptors of his rapping or instrumental style, so this could help bridge the gap 😄

[olive frigate]

This is cool, like an evolution of sound

[midnight shoal]

If Suno suggests a genre, it can be further expanded from the parent genre

[midnight shoal]

Yeah the 'knowledge gap' is why people struggle to create the right song they want, including me

More than once I burned thousand of credits and frustrated because I couldn't generate the song in my head, but later that day I found an exact melody I want it to do in Explore section, just by clicking on things randomly for a few hours

[pseudo lake]

This ties into the conversation @olive frigate and I were having earlier. If someone likes my songs and feels inspired to use this type of feature to try to create a similar piece, i would be honoured! But it could feel like they didn't appreciate the hard work or effort that went into my specific creation if they can just click "reuse prompt"

[pseudo lake]

Marginally related, but I would love if clicking "What will you create" on the explore page populated the style prompt with the genre you had selected at the time-- sometimes I forget "afro-carribbean electro swing-core" in the couple seconds I stop hyperfocusing on it 😛

[sage mirage]

And it is impossible to recreate someone's song from prompt alone. Extend from sure, which I would give a lot more credence to making that an optional thing since you're actually taking someone's output

[pseudo lake]

Not a 1:1 recreation, but taking inspiration.

[sage mirage]

And we've all been inspired by all the music we've heard before

I've suggested a sort of fingerprint for songs so if someone does 'extend' from a song, it will always remember where it came from. If I extend from someone else's song, that song should always point back to the original person

1 second, 2 minutes, whatever. The initial generation is remembered

You do not need permission to extend from anyone's song, public or not. I'd like the examine that more directly as well.

That should probably be opt-in on Suno

[pseudo lake]

That would be neat if they could do something like that, I really don't know much about digital watermarks.

Regardless of what elements are made public, sharing for listening vs sharing for collaboration would be a helpful distinction to make.

[midnight shoal]

AI Forensic is gonna be a huge field in the future

Disfusion model is a slot machine so I think this can be classified as a chronic addiction

[sage mirage]

I mean once you take something out of Suno who knows. Within Suno though, surely, we can keep track of where songs original came from

And since everyone is after likes and subscribes it just seems like the right thing to do.

I shouldn't be able to see a song of yours and extend from and just erase your existence from it, you know?

[midnight shoal]

I suppose they have to implement some sort of fingerprint frequency in future song generation for 'tracing back' purpose

Like how noise canceling headphone just play a subtle frequency to dim out all the noises

[sage mirage]

I mean you can also get around that simply by extend -> download -> upload

[sage mirage]

What I'm suggesting is even more simple, nothing about fingerprints or anyhting and simply a tag that follows it

[olive frigate]

On other platforms you can choose if you want to allow your video/music to be collaborated with.

[sage mirage]

Yeah see this makes a lot more sense to me vs. prompts. Your song, your output, that is yours in a very direct way

I should have to ask your permission to mess with it, it's yours

Alas~

[winged haven]

An option that allows you to download just the instrumental or just the AI ​​vocals

[sage mirage]

That isn't how Suno currently works, it can't give you that output because Suno generates the whole thing at once. It doesn't have the vocals or music in seperate things to give you.

You can do it if you download the song and run it through various vocal remover apps on the internet though

Still, valid suggestion

Devs have talked about offering that internally with Suno, don't know how close they are to offering it though. Last I heard they weren't happy with the outputs

[winged haven]

I've done it before but it doesn't give great results.

[sage mirage]

It really depends on the song and whatever you are using, yeah

[winged haven]

This is why I proposed this option because I design it is hard to do but for example I uploaded my own instrumental then I extended with the AI ​​which allowed the AI ​​to sing on my instrumental. The problem is that on these vocal remover sites it is not of very good quality

[sage mirage]

and that is what the devs have said about their internal way this works. Suno would have to do it this way as well

Suno doesn't use 'stems', which is actually what you are asking for

It's a start to finish master

Vocals and music are 1 thing the whole time

[winged haven]

I see, thanks anyway

[sage mirage]

Like I said, it still good to suggest it. I was just explaining why it isn't a thing right now

Always suggest everything, the devs do read these channel and take things seriously

[winged haven]

So it might be possible in the future I hope.

[sage mirage]

Very much so, yeah. They just have to get it right

[winged haven]

yepp

[short vigil]

I would like to request a feature:

The ability to upload a 1 minute audio file of me playing an instrument, be it drums, bass, guitar, piano, etc. And ask suno to create a track of one or more instruments to accompany it that I can mix together later. I do not want to extend from what I make; rather have AI create a separate track that follows the timing and key of what I've provided with a single instrument.

[wary hazel]

I would like it to be possible to choose the style of music, e.g. "chorus - female" or "verse 1 - male" and have designated people sing.

or in a song, designate the words which are sung by the woman and which by the man

[lyric plank]

Can we get some sheet music? (So's we's can play/perform it ourselves?) In other words it would be nice to be able to print out the sheet music of the songs we make so we could perform it live (or in private where nobody can hear our singing 😅)

[glass echo]

I just want a way to cut the ending and add a fade out post effect for when the AI goes on and on never ending.

[vast sun]

Button for creating instrumental music from a track you have already created with text (sometimes background audio sounds amazing)

[jagged kayak]

SUGGESTION: Allow the cleanup of lyrics. Often times, when you do an extend, the lyrics get all messed up. This also happens when Suno goes off script.

[glass echo]

There are some nice happy accidents sometimes when it does go off script.

[wintry gazelle]

Android when

[prime rock]

Re: the model. 3.5 is often giving me great intros, then it gives me a vocal which is very generic and not really tied to what was established in the intro. Like two different songs grafted badly together. And it never really returns to the style of the intro.

[stray bone]

I think a lyrics help/discussion/ channel would be nice for those times when you get writers block. Writing their own lyrics is the one area that gives the user some sort of creative and meaningful ownership of the final output. Writing lyrics for songs in suno is the equivalent of putting an IKEA table together (I made that table). So I think it would be wise for Suno to support that aspect of the creative process where possible.

I guess its not so much suggestion...

but I put it here anyway.

[dire nexus]

A French channel, so that the French can speak :/

[jagged kayak]

Suggestion: Showcase Discord channels should be organized into Genre's to slow down the feed and make it more focused for what people are interested in. Right now they scroll by so quickly you don't have much time to see them if you are away awhile. Plus if you contribute you would reach an audience that's interested in your genre and for a longer period.

[fresh lodge]

Suno should have a song reject option so that when you reject a song because the audio quality is distorted, seriously overdriven, wishy-washy, terrible autotune on vocals, vocals on what should have been an instrumental, vocals that aren't even legible words, or because it hasn't followed the prompt once you reject it, Suno then automatically deletes that created song and refunds the credits used to create that song.

Audio quality should NOT be a gamble or luck of the draw or take hundreds of credits before you get something workable, it should be mandatory on absolutely EVERY single creation, no matter what style or genre. At the end of the day a lot of us are not using this for free and are paying subscribed customers, why should we have to pay for a non-useable creation when we have no control over the audio quality of the created song, 98% of my tracks are ruined because of this really bad, crap audio sound which is an absolute shame as many could have been bangers, month after month I am just throwing away money because of one of the most important single factors of Suno "Audio Quality" doesn't work far more than it does for many of us.

I really struggle to understand how these various audio issues are even coded in to any of the models for them to be added to any creations and more importantly, why hasn't it been fixed already, enough people are saying the same thing day after day. Come on Suno please up your game, millions of us love this, but the audio quality needs to be paramount, it's an audio creation tool.

[sacred breach]

I don't know if what I'm going to ask can be done...but if not (yet), please Suno team, take note!
I think it could be very interesting to have consistency? continuity? That is, if I want to record 3 songs with the same group-voice, is there a way for the singer of my first song to be the same as the singer of the second and third? it is? thank you very much!

[pseudo lake]

https://discord.com/channels/1069381916492562582/1252789683176345641 This has already been noted in known-suggestions, but throw an upvote on it to help let the devs know it's important to a lot of people! ❤️

[worn atlas]

Go here https://discord.com/channels/1069381916492562582/1154779533627621518

[worn atlas]

https://discord.com/channels/1069381916492562582/1154790975240155247

[grim frigate]

Hey to the great Suno dev team! As soon as the output quality will match a proper bitrate (v 4?) please start thinking about how to give the possibility to remaster old created songs (v.1 onward) to the better quality. Thank you!

[neon harness]

You've come really close now! 👏🏻👏🏻👏🏻 Now that's the "point" in question. (in quotation marks) Tell AI from SUNO: "This is the style I want in my music."

[twin marsh]

increase the maximum song length from 4 minutes to 5 minutes

[shell quest]

Uploaded Song Analyzer : I'd like a feature similar to sonoteller.ai which tells you the genre subgenre mood theme instruments and BPM for any song. This would help me create similar but original songs.

[olive frigate]

In addition to this, I think it would be nice to create a lora/embedding using your own voice/generated voice (perhaps with selection) but I can see how this might leave space for people with questionable ethics

[final echo]

it'd be neat if we could combine the two tracks that get generated into one track

[cold vine]

Piano cover generator!!! even just "instrumental" I want to be able to give an audio or pick a song I like that suno has created and then re-generate just the instrumental or a piano cover of it which also converts the vocal (molody) into some instrument melody!! it be so good for karaoke and stuff

this doesn't even have to be the exact instrumental, as long as its a proper 'cover' of the song as an instrumental... only thing i've found close to this was (pop2piano) and that wasn't great

[glass barn]

song combine (insert 2 song clips) options: fade in to or fuse

[keen crypt]

Why not design a comment function?

[verbal ivy]

Hi, i kind of sort my songs by liking, disliking, or putting them public. Could we get batchjobs like "remove all disliked", "remove all not liked", or "add all liked to playlist" 🙂

[stoic ingot]

It would be very helpful if we could select a playlist during create. Helps organization a lot

[young basin]

Please add - IF 10 credits are missing, but there are only 5 credits on the balance (for 1 generation), then PLEASE = let 1 generation happen!

Thank you 🙏 🙂

[white python]

cant find any solution to below and would be great to see these , thanks
-if u use your own text for generated song that u do like but notice there is a mistake in text... Is there a way to keep the music style intact and just fix typo or replace the word etc ?
-Noted if u try specifying male, female it won't follow. Example [verse1: male] and later [verse2:female] Suno will not follow the structure ....
-Is there a way to do male, female mix in verse for example [Verse:female] I'd live to see this feature ([male voice]: feature) I'd like to make it work ([male voice]: work)

[untold grove]

have you tried [female singer]?

[radiant egret]

it would be awesome to be able to edit the lyrics of songs that have been generated without changing the melody or the delivery of the lyrics (crucial!!)

[hybrid crane]

Something I found on X, and agree with:

[indigo ginkgo]

I'm just gonna say this again, I want to be able to trim my songs play-time on suno, that it plays from a certain time to another time, in ms

[pseudo lake]

I'm new to exploring these kinds of things, but I think that in order to actually file a copyright violation you need to attest that you're the copyright holder (or a representative)?

[hybrid crane]

Yes, but it would be helpful to avoid songs on trending who are cleary copyright violation

We now do it too, but we send it to the devs directly

[pseudo lake]

https://discord.com/channels/1069381916492562582/1252773477157699664 have you upvoted the known-suggestion for this idea, to make sure the devs know it's important to a lot of people?

[indigo ginkgo]

I did not, but I did now!

[pseudo lake]

Wicked! There's a lot of things on there I think people don't know about, or there'd be thousands of upvotes on some of those features 😛

[indigo ginkgo]

Yeah, but for me it doesn't need to be full editing stuff, I would be happy with two textboxes that allow me to set start and end-time in ms. which is a lot easier than creating full blown editing. I think with most feature, you know what you actually want and what would be enough. Be starting with what would be enough, you get a lot more features, quickly

[grim frigate]

another suggestion: make the expand button configurable, Expand forward or backward (to get a proper song intro), and how many minutes/seconds to expand (it doesn't always have to be more than a few seconds)

[kindred flare]

For a paid service, I’m not sure I like people being able to “Extend” my songs.

There should really be a toggle for that on each song. Like “Allow Remix: On | Off” etc.

[pseudo lake]

I 100% agree, and I believe the concept has been brought up amongst the team before, though there hasn't been anything definitive one way or the other.

That being said, the "all-or-nothing" implementation as it stands is a major sticking point preventing me from sharing my music on the 'social media' side of Suno.

[honest moss]

I would a option that modifies any audio upload, and don't just extend it, i'm making a concept series of a reunion tour from a band that I hear.

[pseudo lake]

Your own content, or just from some band?

[honest moss]

Any audio upload, even it gonna be my own content or no, like extend function.

[pseudo lake]

Some people love the "remix culture" and are vehemently protective of it, and while I respect that viewpoint, I don't think the view should be forced on others.

One argument that I've heard is that it would be harmful to new users because they couldn't take an existing song and simply study the tags to make their own song-- but on the other hand, we discourage people from simply using the prompt "I want a song that sounds like The Flaming Lips", and try to encourage them to use style prompts to describe that artist's style. If they aren't able to figure it out on their own, they aren't prevented from making great music on Suno, and they will figure it out through experimentation or through community engagement.

If anything, I think it would foster more creativity because people who pour a lot of time into their songs (and feel protective of not necessarily "the prompts", but of the heartfelt attention to detail and finetuning) would feel more comfortable sharing their works and inspiring others to see some of the "more professionally curated" songs that the platform is capable of producing, instead of simply highlighting the suno-celebs.

[pseudo lake]

I interpreted the first suggestion you made as making edits to someone else's song, which highlights the complications in simply releasing that feature without "safety nets" to cover Suno's own butt.

That being said, audio-to-audio (like img-to-img) would be a really neat feature

[honest moss]

Well, my principal language is Portuguese, so I apologize if you didn't understand it.

But thanks the help

[honest moss]

Like audio-to-audio

yes

so thank you

i know the terms of use

and as a user I have to respect it

[pseudo lake]

Apologies, not meant to accuse you ❤️ Just explaining the kinds of things the devs need to consider before "giving us the keys" to drive on our own 😛

[small junco]

This seems like a simple feature and hopefully it would be to implement it too! : Whilst currently we can download audio of individual songs, it would be really handy to have the ability to download a whole playlist too (eg it creates a .zip folder of all the audios in that playlist to download all at once).
This would be useful and more convenient for people using playlists to create 'albums' or group many songs and styles together etc.

[delicate cedar]

I'm not crazy about letting someone just take something I worked hard on and slap a few seconds on it and make their own "version" of what I did. I'm not even truly comfortable with someone extending from the first thirty seconds and making something new with a sound I rolled the dice and tweaked at to get either.

But...genuine remixing? I'd be really interested in such a thing as long as there were some constrictions that both parties had to follow for them to work. There's need to be a fairness thing so that if someone takes the time to do a remix I couldn't just be obstinate about signing off on it if others agreed it was a good remix. But I'd like to see a situation where I would get credit for my original and who ever it was that was remixing would get an acknowledgement of some kind for bringing a new style or outlook to the song.

I used to genuinely hate remixes. I wanted the song to be the song. Then I heard some really great ones on those old cassette maxi-singles and fell in love with the way people could reinterpret a song and make it even more than it had been before. Years later I grew so much in this area that I came to appreciate mashups and edits and so on...

So I'd love to see a point where real actual genuine remixing was possible as long as stuff was credited and it wasn't obnoxious and so on...

[pseudo lake]

Well put, both sides of the fence considered 🙂

As it stands, there's definite room for improvement to satisfy a larger amount of users, or to give users more control over how they choose to share and collaborate on their work.

[somber granite]

Easy-to-implement suggestion:
Include the title, image and author of the song in the file's metadata, so that when downloaded, it appears in the music players.
https://tagmp3.net/

[small junco]

^^ I was actually thinking of writing that alongside my playlist download suggestion from earlier! Especially for the people who do like using suno to make 'albums', having the ability to have metadata included and the ability to customise metadata (eg adding your own album name in so it shows when you download) would be really helpful.

[hasty kite]

For those of us with thousands of generations in our library, it's hard to find past stuff sometimes. I'd like it if we could create folders within our library to better organize our work.

[pseudo lake]

Private playlists can help with organization (you can use common selection tools to highlight multiple songs to make it easier), but it still leaves the original in the library as well (deleting the song from the library removes it from all playlists) -- I've been pushing for 'workspaces' or alternative organization methods as well, hoping to see something soon 🤞 ❤️

[hasty kite]

Great point about the private playlists as a workaround for now! But, yes, a workspace or alternative organization method in the future would be amazing. Fingers crossed!

[warped ridge]

will we have an easier way to pay?I'd like to pay for this, but I don't have credit card.

[pseudo lake]

https://discord.com/channels/1069381916492562582/1254232844767854665 There's a lot of good known-suggestions and known-bugs to upvote if you want to increase their visibility to devs

[robust jewel]

If "offensive content" is a reportable offense, I will never have a public song on Suno ever. "Well, I'm offended." is such a low, and vague, bar to pass that you can never know where it even is. That kind of unexplained "violation" is terrible. Even if people want to follow the rules, they can't, because the rules are made up on the spot. That said, being able to report blatant copyright violations seems like a good idea. At least that can be confirmed, when it's a well-known song.

[hybrid crane]

For me it's about the copyright 😉

Not offensive..

[robust jewel]

I kind of get where the offensive one is coming from, but not the execution of it. Like, I have seen people make songs that are 50% racial slurs and stereotypes, so I get why people would want to report that. I just can't stand when there is no explanation of what "offensive" is, but it can be reported and punished completely at the whim of whoever sees it. The copyright one is far less abusable, and I'd probably have flagged a couple songs that made trending since it was added.

[pseudo lake]

From my understanding, it's to flag instances where offensive content is generated without the user's desire. Like if you get swearwords in a song that you're attempting to keep clean -- sometimes lyrical hallucinations happen. This kind of reporting helps narrow down those instances when they're troubleshooting or refining data for future models

It might also flag the song for a human to check for an extra consideration as to whether it breaks the agreements the user made in the ToS and if the song should be removed.

[velvet salmon]

Please Suno Team, can you add the possibility to comment on songs? Some songs are truly amazing

[sage mirage]

A long time ago this was brought up and the devs aren't against it on any technical level, but it becomes something that has to be moderated

as in, how do you prevent people from posting the most awful of things

Then they mentioned maybe emojis as a sort of compromise, then that was wondered about how many eggplant emojis people would do

or the peach

[pseudo lake]

Something like the iOS message reactions would be an easy workaround for that issue -- just give people a limited number of interactions or responses. It only becomes a problem if it's hastily implemented.

@velvet salmon I've seen a lot of people suggest this recently, and it's a clear sign that the Suno website is thriving as a social media platform 🙂 ❤️ I know there's a couple songs that @tepid trench has made that I'll never be able to get out of my head, and just giving it a single "Thumbs Up" doesn't express how much I love them 😛

Thanks for sharing your ideas 😄 🤘

[sage mirage]

This bit of conversation might as well been another lifetime with how stuff has changed

For the longest time we were all convinced we would never get the ability to upload to Suno

It just made sense, why risk it if you were Suno

and, welp

That was wrong

Same might be said for comments and more social stuff

[pseudo lake]

🤘 👍

[sage mirage]

The Upload thing came out of nowhere too, seemingly

[pseudo lake]

hrm, yes.

It'll be interesting to see what the devs think are good ideas and decide to implement -- I guess we just have to keep giving them crazy new good ideas and different perspectives to see what comes out of it! ❤️ 😄

[sage mirage]

[patent sonnet]

I would still like to transition two existing songs into each other like parts even if it's two separate songs.

[jaunty moat]

I noticed that when you extend a song from somewhere toward within the song, and add new lyrics, it lists the old lyrics and the new ones when you "get whole song" it treats the lyrics

it treats the lyrics like you "extended" the song from the end of the song.

You should be able to edit the lyrics. I don't know how hard that would be given it has to put them in sync with the song in the video download, but it would be really handy. Especially when you have to use phonetic spellings sometimes or when it makes up its own lyrics.

[sage mirage]

This is coming with v4 and inpainting

Simply inpaint over the ones you want to fix

Granted we don't know how that will look practically, but it IS coming

[bold cedar]

But when? I loes my Hope.

[sage mirage]

🔜

More true than ever

[keen spruce]

jersey club music style not working

[hexed gale]

HI. Are you planning an option to save each track (channel) of a generated song seperately for musicians ? For mastering process.

Thank you Falank ! I totally agree with your suggestions as well.

[tepid trench]

Separate instrument parts are known as "stems" and it is often requested to be able to save these individually - #1253172964728377455 message

[sage mirage]

I am also not a fan of my point of view being forced somewhere else

[chrome quarry]

it would be nice to have the option to disable vocals in tracks, occasionally i will get a song I really like but the vocals are tinny and sounds like crap, or ill select instrumental and still get disembodied vocals or hums and it would be nice to just turn that off and download the instrumental in that case.

[tame swallow]

With ability to Extend the song, there should be an option to "Regenerate" only a selected sections. As an example, say you love what you ended up with, however from 1:50 - 2:15 the AI got a little wonky, but everything after is perfect... As of the moment the only option is to completely regenerate everything after 1:50 and often times you lose what you liked before.

[robust jewel]

That's inpainting and is expected whenever V4 releases.

[kindred flare]

Seems to be mostly ignoring the "Instrumental" toggle in 3.5

[small galleon]

Is there a way to feed it a track and have it build vocals on top?

[pseudo lake]

The toggle isn't a hard rule, but suggestion -- I often keep the lyrics box open just to use 1-3 metatags per extension.

I make a lot of indie-rock, and I'm very familiar with "oooh woooo oo-oooh"s, I think people who simply use "rap" or varying similar vague style tags get gibberish rap, but it's just a matter of placing an extension point before that happens.

[pseudo lake]

No current way to guarantee the same output as input, but if the uploaded audio is a sizeable enough sample, and the musical structure is coherent enough for Suno to pick up on it and continue, it's not impossible to get more or less the same beat with vocals added on top.

Definitely possible if you really care about it, but hopefully in a future version we'll get something like audio-to-audio, or perhaps inpainting will allow you to modify uploaded tracks 🤞

[small galleon]

Savage. Great work

[lofty hatch]

"add to playlist" is sorted by last date modified (aka the last playlist added to is at the top)

A setting under create that lets you set a playlist that will auto add the thumbs up 'd songs to.

[tardy girder]

Would absolutely love if there was a way to put in a seed. (Like in stable diffusion.) Or a way you could actually steer the music and tempo. I've been messing around and I can sometimes get good results by putting in things like 4/4 or 3/4 or 90 BPM or something. But it would be nice to have a little less randomness (or at least keep the same sound and not have it go wild when you extend and it sounds like the old one for 2 seconds and then... WAAA WAAA WOP WOP WOP... Goes insane. 😄 )

(also would love to have a way to download all or put a checkmark and download all the ones I want.) If there's a way to access the API I could build myself one in python. 😄

[lofty wraith]

need to lock down the bpm with uploading audio, i can stretch it but its just a quick fix, its always 1 bmp or so off though!

[small junco]

☝️ 100% would definitely love to see an option to choose a time signature and/or a bpm and/or a song key ( as optional inputs you can add when creating new generations ) to control the output more. esp for ppl who like to create things in suno as inspiration for making real tracks as a starting point (me! 💡 )

[daring patrol]

you dont, god bless free speech

[sage mirage]

That doesn't extend to private enterprises like Suno

Their site, their rules

[daring patrol]

Please add an INTERROGATOR so we can upload songs and get a prompt from them, like instruments, tempo, etc.. It'd be great for people without musical literacy

[sage mirage]

I think one of the devs mentioned something like that before

Never promised or even talked it up

[daring patrol]

yee i was suggesting on a philosophical level not legal!!

[daring patrol]

That'd so cool

[sage mirage]

It was something in casual conversation a few months ago

[daring patrol]

I was going to upload a tame impala song but the copyright warning got me worried so I didnt

[sage mirage]

Yeah don't do that. Suno takes that very seriously

For obvious reasons

[daring patrol]

Yeah I didnt want to copy the song I was just trying to get the instrument names theyre cool and im a crass so idk how to describe a song like that

[sage mirage]

Usually the wikipedia article on the specific song will give you some tags that are useful, or you can ask your favorite text LLM

[daring patrol]

mybe i can train a model, are you guys hiring:-)

[sage mirage]

I'm not on staff, just ~some guy~

but
https://jobs.ashbyhq.com/suno

[daring patrol]

thanks lad!!

[prisma aurora]

This is the way

[fallen zinc]

i need a slider to prevent AI from changing tempo all the time

I love suno (for real... it's changing the game) but at the same time I feel enraged wasting tokens and not being able to use great parts it's producing because of the constant tempo changes

[quaint lynx]

Are we gonna get something like random seed exposed in tha UI? It's hard to understand what any of the tags you add do if you can't experiment with them and a fixed seed.

[real compass]

is there any chance we'd be able to get an option to re-render small sections of songs? For example one verse I generated is basically perfect, but a single word was just skipped over, so the whole thing is worthless

[small junco]

v4 will supposedly have inpainting for this very reason! should be comin soon 🤞

[silver crane]

hello, can you guys please support Arabic text when downloading the video, the lyrics appear as unreadable symbles as it does not recognize the Arabic characters..

[jagged kayak]

SUGGESTION: In library you should be able to FILTER on all THUMBS DOWN clips and have a SELECT ALL. MOVE TO TRASH.
There's no easy way to round up all the pieces of songs after your done extending and push them to trash. You have to do it one at a time. It's time consuming, and prone to accidently deleting things you need.

[zinc void]

Suggestion more creativity thing needs to add in such as :

• Instrumental Button : When you have a Song/Music with Lyrics , Generate Them out Without Lyrics and have Instrument
  Translated : You have a Lyrics that in another Languages ,we change it to another languages
  References Music : When Creating A Music , find some References Another Music/Songs
  Remix Ver. : having a Remix ver when u have a lyrics song , change the Instrument

[pseudo lake]

Common desktop selection methods work (ctrl +a, shift + click, ctrl + click), so you can move multiple songs to a playlist or trash

[patent nacelle]

Suggestion for both browser and iOS : possibility to create multi-language songs or change lyrics in an existing song. Sometimes, we can need to remake a song in another language (example: I wrote a song about 9/11 in French, but since the events took place in USA, it was obvious to me that I needed to make an English version, so I wrote the English version with the same structure, and used Suno to make the song from those lyrics, now I'd like to have the same song with the french lyrics, but it doesn't work). It would also be cool to be able to use a previous song, so we can input different lyrics with the same structure, but in another language, or if we need to change the lyrics (for example I have a very cool song I really love, that sounds very Irish, but talking about a town in France, It would be nice if I could reuse the song but have new lyrics to fit the music better and do something else with my song about France)

[wheat temple]

Assuming Windows desktop, ctrl +a, shift + click, ctrl + click works in Windows folders but not in a browser. I'm assuming jimerb was referring to the browser interface. Also, a filter feature in the browser interface as jimerb mentioned would be very useful specially as playlist or libraries grow in size. Having to manually click from page to page through hundreds or thousands of pages is extremely cumbersome.

[pseudo lake]

I wasn't saying his filter idea wasn't a good suggestion, I was just giving suggestions on how to manage things as they currently stand.

but not in a browser
Not sure what you mean, this functionality absolutely exists on Suno? In your library, hold control and click multiple songs, or click a single song then 'ctrl + a' to select all on that page.

Not saying it's not tedious to manage as it stands, but it's helpful to be aware of.

[true coral]

Would be great to have a specified folder or heading for uploads!

[wheat temple]

Here is what ctrl-a does in the browser. It selects the text in the periphery.

[pseudo lake]

Feel like there's a degree of user error going on here.

You haven't clicked on the song field to put focus on the songs -- you've clicked the UI wrapper and then hit ctrl + a.

[pastel zenith]

Maybe make it so it has drop variation so the 2nd drop is different from the first (referring to dubstep, DnB and that)

[wheat temple]

You're right. I missed the step of first selecting a song priuor to ctrl-a. The filter function would still be useful. Having to manually page through hundreds or thousands of pages is still impractical even with ctrl +a, shift + click, ctrl +.

[pseudo lake]

100% agree we need better options for filtering or searching through songs -- whether it's simply to delete all at once, select ones made with a certain model, from a certain time frame, yadda yadda yadda

Yeah, mass organization is pretty cumbersome when you can only select a max of ~20 at a time

[reef grotto]

could you just stop messing with v3.5 every day?!

Like... one day I make 6+ minute track and it sounds great throughout the track. By great, I mean the audio quality

the next day, using same prompting logic, even similar keywords, same genre, the outcome is totally fcking different, audio quality is horrible etc.

And I'm talking especially about the audio input feature, where I upload first 60 seconds produced by myself and then continue with prompts

it feels like the v3.5 is changing on a daily basis, sometimes to the better, then to the worse

it's exhausting

[hybrid crane]

They do not. The model is static once done. Quality goes up and down, predictability goes up and down. This is normal behavior. This was me in december:

[reef grotto]

mkay

but what causes that?

[hybrid crane]

Now I understand better. This is the current state of not only Suno but other services have it too. Make a long audio with Elevenlabs and the quality degrates too

The solution is being worked on. This is new tech. Still a lot to discover.

Every model gets better though

[reef grotto]

yeah, eagerly waiting for v4 🙂

I've made so many good tracks, that need heavily spectral audio editing later to make them sound somewhat good

[fallen zinc]

i dont believe it... when 3.5 came out it worked fine but now, 2 weeks later 70% of my generated stuff keeps speeding up and I'm just burning tokens

and i have tried with different uploaded clips too

[reef grotto]

same issue here

[fallen zinc]

something must have changed

[sage mirage]

They are following your arc 3Daizy

[hybrid crane]

I believe the developers, and I believe my exprience. My logic says it's normal behavior. ChatGPT still lies too. AI model is static once done, and AI is unpredictable. If you ask all the regulars here, they will all tell you the same thing.

[fallen zinc]

well one way or another ... it seems; it is what it is...

hopefully newer models will fix it

[hybrid crane]

Yep... and if you use it as much as I do, you will change your opinion

[leaden marlin]

bruh that was a detailed suggestion that got insta deleted. if i would have known i'd copy it so i could shorten it.

could a moderator send what was deleted to me please?

[leaden marlin]

oh hey sorry could you help me with this?

[hybrid crane]

too many enters (empty sentences) was the problem (more than 3 lines)

[leaden marlin]

ah i see, was unaware of this thank you

[shut cypress]

OK, this is more of a suggestion down the line - I notice for every song, a WAV, MP3, and Video are made. I try on about 10-20 versions before I choose one - which means 19 other versions don't need an MP4.

Why not change that to a "Render at click"? I think you'd save a ton of time on server render time, with not dealing with songs that will never be used. Same thing with MP3 - the WAV version is the most important part to this. Compressed audio and video options can be selected and rendered when the user asks for it.

Every MB counts.

[queen wigeon]

A hopefully simple thing I would like is a trim function. From time to time I get a track that's perfect, except that it comes back after the fade out or whatever with something random. To me this is a finished track, because it's trivial to trim off the extra stuff at the end before uploading it to YouTube or whatever. BUT I'm not going to make it public on the Suno website simply. Same issue applies to many of the songs with my uploaded sound at the beginning.

[sage mirage]

Been asking for this for a bit

[slim warren]

I want them to make a feature to where you can make the same song but in a different language. It would be cool because what if you upload it and people around the world who love ai music has a hard time understanding the song? It would be such a helpful feature

[true storm]

please let suno go into producer chapter tier and allow ableton extractions.... like soundful.com has. it would be game breaking

ableton file export ... trust me

follow soundful.coms model on that and you have a grand slam

[sage mirage]

Suno doesn't use Stems by default, it's a whole thing. Suno is master only

[true storm]

copy

ahhh

so theres no way to stem... shit ok

[sage mirage]

It's still a good suggestion

[true storm]

ugh is theree anyway

[sage mirage]

Nah so how it will have to work is the sorta janky way of breaking the stems off after

It isn't the same though

And Suno does want to internalize it

[true storm]

please im begging you

[sage mirage]

I'm not a dev

Just a guy

[true storm]

this will change the game

big time

[sage mirage]

They know

[true storm]

soundful.com do a collab

ask them for their model

work out a deal

[hybrid crane]

Devs want it too.. it's just a matter of time (stems)

[true storm]

i can middle man it

iive been with soundful and dj.studio and kaiber.ai since inception

[sage mirage]

heh

[true storm]

dj.studio is doing massssssive shit right now

[hybrid crane]

If you think you can be of assistance in collabs, you can always write to [e-mail]

[true storm]

i can def make the right introductions

im a 1987 fire rabbit with so many bad habits

i get shit done for this wor;d

world

L F G

[hybrid crane]

Then pitch yourself and who knows 💪 .. Just send Suno an email 😉

[true storm]

i already applied for a position

[hybrid crane]

Good luck then!

[true storm]

yesterday

do you work in house?

im in NYC

[hybrid crane]

I am a moderator not staff.. They work inhouse yes

Then you are in the right place 😉

[true storm]

always was 😉

0cool 2.0

im dumbfounded by this amazing platfform. im a music producer and engineer over 15 years and ive been djing night clubs 25 years lmao

nyc and miami and international

this platform broke the world. im not gonna lie. are they being sued by the big labels?

theres so many stories i heard i hope it lasts this is amazing.

[jagged kayak]

Ideally I'd like to filter everything with thumbs down, select all. Delete.

[frigid surge]

the labels realize they are being replaced

[feral aspen]

I can never predict the playlist ordering in the add to playlist dropdown. Sometimes I feel like it's trolling me, like, really the playlist I'm about to add 10 songs to, is at the very bottom of the list, number 30???

[robust jewel]

The only guess I have as to what might cause that is... Have there been songs added and then removed from the playlist? If there have been a total of 30 songs, even if there are only 10 right now, then it makes a little sense how that would happen.

[frigid pebble]

There were already proposals to add folders to organize generated media. When there are thousands, it's a nightmare to navigate through them.

So far, I begin to use internal-purpose playlists for that.

So, my proposals:

• supporting idea to introduce folders (with several levels of subfolders) to store the media
• please allow free form tags, to allow grouping media by several criteria
• in the Library/Playlist listings, please display which Playlist(s) and tags does the media belong to

Thanks.

[fossil parcel]

Hi, Last time I used my local language, it was pronounced correctly about 60%, but now I think you've updated to a newer version and it no longer works. Could you please make it work again? I can help with pronunciations of special characters and words. It would be helpful if you could do this so that our community can create more songs.

[feral aspen]

As far I am aware, nothing should be different with language support. Have you tried with the exact same prompts you used last time? What language is it?

[keen drum]

Can someone help me identify the musical style of this song? What style of electric guitar is being played and everything. Thanks. 😁

[restive oak]

How do I extend a song without making another song?

[wicked hound]

would love for a way to "detach" our uploaded audio from the "whole song". that way in instances where you only want suno to sample the audio and not necessarily use it as an intro would be a lot easier to share 😄

[tidal star]

hii, can we have more customisability for our account pages and such?

like profile banners, choosing how you customise playlists, and maybe a display of songs chosen by the creator

just anything to make profile pages pop out more!

[daring goblet]

┬─┬ノ( º _ ºノ)

[warped finch]

i've noticed sometimes that when suno extend songs he also removes vocals really clearly, maybe a vocal isolator would be cool with suno algorythm

like that instead of removing the frequencies of the vocals, it recreates the instrumental

[kindred hull]

I've got one (very simple) suggestion.

If you go to the custom mode and generate lyrics from a prompt, the song/library will not show the prompt used to generate the lyrics.

This behaviour is understandable, because the user that the option to modify the lyrics, so it is NOT a bug.

However, having the initial prompt may be useful for the user.

Maybe Suno could show the initial prompt adding an asterisk (*), a different color or any other visual cue to show that the lyrics generated from the prompt may have been modified.

However, for simplicity, it may be simpler to simply show the prompt as is.

[old halo]

Is there an opportunity in suno ai to add specific chords to the prompt (like: Am, E, G) in order for the song to be generated with them? What should such a product look like in tags?
Hope this is the right branch for question like this...

[hybrid crane]

you can do with upload audio, tags doesn't work. It does not know chords or notes.

[old halo]

Thank you, I did not consider this option, I will definitely try it now! Although, it would still be much more convenient to do this through text, eh.

[hybrid crane]

Still much to discover for the developers. Like learning how to recognize chords and notes. it's easy to come up with ideas, but how to teach AI is a different and often puzzling task.

[cyan acorn]

make it be able to add ai vocals to songs, that would be cool tho

[solar cliff]

For ending a track, a feature where we can pick a time segment, and face out the volume over a selected period of time (3 secs, 5 secs, etc etc), would be a great help to end a track instead of relying on [Outro} or [End} as a way to stop, and gives us a little more freedom on how a track finishes. Maybe something we can do after we select "Get Full Song" to chop the last part of a track out when they go for an extended outro, or stops for 5 seconds, and starts up again.

[leaden marlin]

I would suggest adding a Karoke option? Suno already auto makes the videos with our lyrics. I wonder how hard it would be to take your songs and just earse the vocals.

[tepid trench]

That requires stems. See: https://discord.com/channels/1069381916492562582/1253172964728377455

The only way to do this at the moment is with a stem splitter service/app.

[leaden marlin]

thanks

[worn atlas]

Add a tab for Full songs

[weary bough]

how can i add lyrics to my beat i made in SUNO? when i try to extend with lyrics, it just does another instrumental

damn, it won't add lyrics to my beat i made

[weary bough]

why can't u change the cover art? I mean u can but when u download it, it's the old one

[rapid hatch]

Hi there

Is it possible to have a feature about enhanced or clean the music sound after releasing? Because when i got a great song, then I must remake the full instrumental or the song to have in HQ quality. Even if I used an other software to clean the crisp and the distortion (like too metallic sometimes) then I am losing some instruments quality ( it’s like fading) . So I would pay more or spend 10 credits to enhanced or clean and get a clean STEM after that

[feral aspen]

The Suno devs mentioned they're working on stems, for sure (no timeline though). I'd pay more than 10 credits for a song too, if it was even a bit better (or even just unusual in some way, like a negative prompt).

[rapid hatch]

Well, hope to have that before the end of this year with the 4.0 version. It will be so great.

[hollow stag]

./chirp

[hybrid crane]

/chirp

[rare obsidian]

The /chirp command is no longer working

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

[hollow stag]

/chirp

[rare obsidian]

The /chirp command is no longer working

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

[hybrid crane]

/help

[rare obsidian]

How to use Suno

Hi there!
If you’re ready to dive into creating music, head over to Suno.com.

For more detailed information, check out docs-guides-resources.

If you have any more questions, just ask in #general-suno where you can always find our community, moderators and guides to help you with most questions. 😊

[hybrid crane]

@hollow stag ^^

There is no bot anymore on discord

[tawdry wave]

Please share your workflow.
What do you do?
What do you use to improve audio quality?

[tawdry wave]

at least suno.com had the output in midi, like Fadr or Chordify

[rapid hatch]

Hum , well, I am using vocal remover because SUNO doesn’t have these features yet. So I can get the STEM , instrumental and voice separately. After that I recreate the entire song with my own voice because some of the vocal parts are crispy. So I blend my voice with auto tune with the SUNO female voice when it is possible. But hard work and time consuming. And remaster the entire thing. And because SUNO female vocal is high pitched so I must change the pitch and drop to -1 like , Bb minor semitones.

But I would be happy to have that in SUNO directly. Like this I can work in the song directly in SUNO. Because each time I must transfer from SUNO to vocal remover and after in auto tune app then I am losing quality. 😔

[tawdry wave]

OK.
What vocal remover do you use?
And the instrumental? What do you do?
Convert to midi to apply VSTis plugins?
Which VSTis?

[rapid hatch]

I am not using MIDI files, not my thing, so I can’t help you if you are using MIDI files. And I am using Vocal remover, you can google it

[pseudo lake]

https://ultimatevocalremover.com/ ?

[rapid hatch]

that the name of the software VOCAL REMOVER, and auto tune app on iPad.

[rapid hatch]

Yep this one

And remake the instrumental with GarageBand and Logic Pro but time consuming. And the final result is 80% close but not exactly the same. SUNO sound is more round , like a Dolby sound in a theater for example, when Logic Pro sound is more clear with less reverb

[tawdry wave]

You use the UVR then.
Why did you choose this one?
Have you tried other online options or software?
Like Moises, Fadr, Izotope RX, Spectra Layers, Melodyne, etc...

[rapid hatch]

Yes I did, but I am using exclusively iPad. So others are difficult to use on iPad. And I have an IPad 12.9 inch 4th generation

[novel compass]

We need a randomize/shuffle button for playlist playback.

[real copper]

Yes please for the love of God let us just 'select all' on any given page of songs, or even better just let us batch download on a search term

[pseudo lake]

select all on any given page
click a song then hit ctrl + a

[exotic harness]

please let us purchase pro with paypal or giropay

[brave canyon]

Can you select if it's a man or womans voice?

[pseudo lake]

https://discord.com/channels/1069381916492562582/1254232844767854665 Upvote the known-suggestion to let the devs know it matters to people ❤️

[pseudo lake]

Not as a guarantee, but there's many tips on the community wiki https://www.suno.wiki/faq/metatags/voice-tags/

[halcyon vale]

To enhance user experience and prevent unnecessary spending of credits on songs that may not appeal to us, we propose implementing a feature utilizing artificial intelligence (AI). This feature would generate two distinct song snippets and present them through a user-friendly interface. Upon viewing these snippets, users would have the option to choose between them based on their preferences and musical taste.

The UI would prominently display the two snippets side by side, accompanied by brief descriptions or tags to provide context. Users can listen to each snippet and evaluate whether they enjoy the style, melody, and overall feel of the song. If neither snippet aligns with their musical preferences, a straightforward 'Generate Again' button would be available.

This functionality ensures that users have a practical method to preview potential song selections before committing credits. It not only enhances user satisfaction by allowing informed choices but also promotes efficient use of resources within the platform.

[solid sage]

#📢┃announcements message

Make mobile easier to use. I play through songs in my browser and half the time i accidentally click on thr styles of something i wasnt meaning to click on

Or i click on a song in my library to listen to it and it jumps up a few songs and begins playing a different song

[robust jewel]

Suggestion: Color themes setting. Different people are going to like different looks, but the new 1980's brown look is an... interesting... choice. Would be nice to be able to choose, for those of us who aren't huge fans.

[brazen arch]

if there was one thing suno shouldnt have changed, it was the design. the design was very clean before, why overhaul it?

[limber vine]

My overall feedback

[wooden igloo]

Is there a way to return back to old view?
Everything is too big...

[onyx wasp]

I would suggest being able to create stems! I’d love to create a backing track for guitar practice, maybe even going as far as separating lead guitar from rhythm guitar

[pseudo lake]

Seconding the idea, I'd love if it were possible to use www.suno.com for the reskin and old.suno.com for the classic view

That being said, it's completely subjective and not meant to dismiss your concerns of the sizing, but the reskin has really grown on me over the past couple weeks of testing.

[wooden igloo]

The covers look nicer, font is good... but for me, buttons are too big, and browser is freezing in some cases. I use "song upload" and edit/image a lot. All locking the browser.

[pseudo lake]

Have you cleared your cache since the update went live?

tech issues aside, valid feedback ❤️

[wooden igloo]

Yep... tried in safe mode, incognito... Not working 😦 . Will try in Safari.

[wooden vector]

Can we get a widened display box for the images of our songs. Right now it seems to cut off a lot of text and/or important info on our images.

[frosty cave]

suggesting having more details like: bpm or key

[delicate cargo]

I'd also really like to suggest a glowing effect for new songs that have not been played yet.

[fair cipher]

For me, I dislike the new design of the profile. Before, I could just look down the line to see how many listens, likes etc everything had. Now I have to zigzag all around.

I also don't know why they keep having to make the create button bigger. No one had any problems finding it before, when there's literally a create menu on the screen. The only people who had problems were the ones who thought chirp still worked and didn't know about the website to begin with.

[short dune]

Art cover looks somewhat large, but its a improvement from the old one for me personally. The font choice is really questionable. The new logo is a downgrade. Color profile is weird too. Just full black with no accents to make things pop. Weird orange thing on the right hand side and odd choice for buttons.

[grave pecan]

Yall need to compress this 1.5MB 4K image loaded for every visit: https://suno.com/Aura-06.jpg

I don't want to know how much this is costing you in server billz

[soft meteor]

Yes make them square again please... that would also save some space

[leaden marlin]

Train suno model with more genres like hardcore, dubstep, rap

Also would be great to download stems

[fair cipher]

People are gonna have different visual preferences but I'm trying to focus on things that are more inconvenient now than they were before.

[wooden vector]

I don't mind the tiles being rectangular, but if you click play on a song it should display a square as you say. Also, I like the overall aesthetic, does feel a bit Netflixy though.

[grave pecan]

There's janky spacing all over the place now, but the biggest gripe I have as someone with poor eyesight is that text is so tiny and squished

[soft meteor]

Mhh than they need to change the upload picture so it gives you a preview of what will be cut off

[tepid trench]

Related: It seems excessive to load the full size images for the genre categories and the user-uploaded artwork just to show them on the home page.

[wooden vector]

Or that yes. But if it stays this way I have to go and change most of my pics and that would be so tedious 😔

[grave pecan]

Yeah my custom artwork is 2MB and I saw it's loaded from their CDN.

LMAO

holy 6MB? Are people uploading 8K RAW and you're just hosing them on CDN without resizing?