#📚┃suno-school
1 messages · Page 71 of 1
So right now i'm using the pro version and used the stem break that is included as one of the options when you right click on em. After trying to reduce it messing with the center channel its like the audio is still there. Very faint like an echo but its definitely there. I wonder if it has to do with a surround sound channel or something
With the Pro version you can move the "artifacts" to the correct layer to clean up the stems.
is that done on the site or do i need to do that with a DAW?
Are you using Suno to do the Stem separation?
yes
I thought you were using the demo for SpectralLayers.
i was not
Suno stem separation has a lot of artifacts
ohhh so i'm doing it the wrong way
I can try spectral. i'm not partial to any daw atm
The Bandlab Splitter App gives you free splits. Some on here have said it gives you good results.
SpectralLayers is not a DAW it is an advanced spectral editing software. They just released a 30-day trial with no restrictions. Cubase Pro 13 is the only one of Steinbergs DAWs that has a basic vocal stem splitter. FL Studio has stem splitter in Producer and above, Logic I belive has one, Abelton Live has a Max for Live plugin, and I don't know about the other DAWs.
@spiral bane is this the style you were aiming for? https://suno.com/song/a338b41e-a16e-4ebc-b793-545dc8b21304
Gothenburg groove anthemic aggression sharp riffs melodic scale reluctant heroism dynamic pacing determined rise song. Listen and make your own with Suno.
Yea kind of I want that kind of groove and
Atmospheric and Menacing, Ruthless Precision, epic and orchestral, black Metal, behemoth song. Listen and make your own with Suno.
That lethality
But more melodic 😂
Less ballast beat all the way through kind of thing
Have you tried using ESSENCE EXPLORER ChatGPT to help in creating style tags for Suno? You can also analize artists and song titles to try to get a better result from Suno. https://chatgpt.com/g/g-utsMc0LBR-essence-explorer
nope didnt even know it existed
imma try it out, so you dump your lyrics in and say make tags for x style band or genre ?
It is from the Suno user that wrote this guide. https://www.suno.wiki/wp-content/uploads/2024/04/PROMPTS-FOR-AI-DANCE-MUSIC-🙏mista.lewys_.pdf
There is also the ESSENCE INDEX https://chatgpt.com/g/g-NF3h7vwFY-essence-index
They need to sticky that guide as it has a lot of valuable information.
Pretty much.
neat time to play around
If I am having trouble describing something in tags I go to this tool since the Music Styles on Suno is not just genre descriptions but also vibes.
so i find it easier to use tags in the song instead of the limited 120 character prompt 🙂
anyone have success with work songs? sea shantys? acapella?
alternatively success with getting a song to sound like children singing?
#1256301881966198805 message I made a Prompt Experiment Playlist a while back showing "a cappella" vs the anglicized "acapella" -- they're both very powerfully weighted, but it seems like the traditional Italian spelling leans toward a barbershop style while the anglicized spelling leans toward more of a harmonized choir -- from my observations, it might also include light instrumentals while still vocalizing most instruments and using interesting vocal interplay.
This is fantastic!
what are stems?
Individual layers of a song. E.g. vocals, drums, guitar and so on.
Suno can do a simple split into a vocal and instrumental stem (so all instruments, and vocal only)
OH! I was hoping for something like that!
/chirp
Chirp is retired, and no longer on discord. You can make music now on our website: https://suno.com
If you want to reuse them within Suno it's ideal. But the instrumental stem has quite a bit of vocal bleeding into it because it's hard to do a clean split. Other stem splitters exist that might give better results if you download the MP3 or wav
I just made a stem. was hoping to use the vocal stem only to make a cover, but its not an option.
I would suggest an external application for stem separation.
Yeah not sure why that isn't possible. I had the same issue. Try cropping it to see if you can cover the newly generated clip
Aren't you only able to cover stems on mobile?
I've never tried. Would make sense if it's a menu bug. Actually you can make your browser window smaller/narrower to get mobile view.
I think I remember the harbor talking about this, but I don't have the mobile version so I have no idea
It was on mobile browser, not iOS app
desktop
mobile
Oh I see 🙏
Why every song is jumping right into vocals. I put [Intro] and it's skipping the Intro Instrumental
Has anything been mentioned about some quality improvements?
I think the "cover" sound is better than the std.
Also, "inpainting", ie changing words in existing lyrics, any word?
I can't find a place to ask this.
So I'll ask it here:
Some of my songs where generated off the subscription plan while others where on the plane and I can't tell which ones are owned by me and ok for monetisation.
I'd like to either buy those rights if possible or sort that out with someone from Suno or at least find clarity over the issue.
What can I do?
Is it possible to download WAV on the mobile app? I’ve been having trouble getting it to download WAV on desktop (yes I’m a paid subscriber) but when it first released I was able to download that quality on a few songs, until it randomly stopped letting me
Email support@suno.com about this (name the specific songs) and see if they'll let you have ownership
Did something change about the extend generation or did I just have some really unlucky gens.
I've been extending one song, but no matter where I cut it, it just repeats itself, before moving on to the "new" part.
have any of you successfully used covers to make an instrumental from suno song that had vocals?
Hello! I want to compose a song and I get an error. Why?
the negative prompting options was a good idea. even though I've reached a point where it's very hard for me to like the outputs of suno, i've heard the songs so many times, they sound very alike in many aspects... and I do lots of styles, mostly mixed styles.
any of you old users have some "hack" that is helping you some fresher sounds?
I had some success with the "upload audio" feature, by uploading some of my old guitar and bass riffs
Been getting this for two days whenever I hit "Create" and I'm not sure what the problem is. Can't find -any- information or resources.
@leaden shell @radiant lake be sure not include bands or artis names neither copyrighted lyrics
not
the only word is "8-bit" lol
(and I've tried numerous prompts just in case)
well maybe give it a "bit" mor context (no pun intended) - like "VIDE GAME MUSIC"
its simple, as with the other songs I've composed so far
*video
Such as... trying all those numerous prompts?
yes
The thing I specifically tried, got it.
I've also tried using different browsers, different devices, clearing all web browsing data and clearing the cache.
you don't need to enter styles, try "vibes" or "where you want to use it"
happy 8-bit music / 8-bit video game music
thrilling 8-bit music / adventurous / suspenseful
just checking. Did this message come through?
I noticed that in the lyrics section, there is no longer 3000 characters, it has been reduced to 1250, this is something new, but it still doesn't work
This is not a problem with the words used in the prompt.
The words I used for the prompt worked three days ago, but not yesterday and not today.
And in the troubleshooting steps, I decided to try a fair number of other prompts including ones with lyrics, ones without lyrics, different styles and genres. It doesn't matter what I put in the prompt field, I'm getting:
lyrics too long don't make longer songs, make the "singer" sing them faster for them to fit. you can repeat some parts by using tags like [chorus] and then just usgin the tags
@leaden shell have you tried simpole things like log off?
That was part of the "trying multiple browsers and devices" yes
oh sure my bad
I also get the same error if I try to extend any of the songs I already have
Can you try the following (each separately):
- Change the model to v2 or v3 then switch back to v3.5 (then submit)
- Try using Edge or Firefox
Those were some of the comments I recalled that appeared to solve an issue similar to the one you are having (though they may not have the same underlying cause with how generic that error message is).
Please let me know if one of those worked or if both failed
Literally the first thing worked. I don't think I'd have thought. I've never touched that dropdown, lol.
Thank you for letting me know. It is strange but I suppose what might have happened is that the initial model retained may have an outdated parameter and reselecting fixed that issue
Ooh, makes sense. So I basically switched from v3.5 (which was secretly pointing to nothing), and when I swithced back it was pointing to the correct thing. In theory, at least.
Is there a guide or tutorial or whatever that gives suggestions and whatnot on how to improve the vocals to not sound as robot-y?
I mean after composing the full song, so post processing, not through Suno's AI itself.
That's a simple and elegant explanation for way too many of the recent weirdnesses -- I'd be surprised if that wasn't the case. 🏆
Another Questin: Whats the effect of the mode "cover"? If i used it the ai reproduce my old song but the ai do not add my text oder change my textes oder build in other instrumentals? nothing? Give i the wron pompts maybe?
You could split the lyrics and run them through a better vocal genetor then add them back in?
Oh i'mt not trying this. my subcribe is only 2 days so i try to bring the ai to react to my prompts, first. thank you. i try this way.
Hey, in the prompt can i determine a line, for exemple, i want the music to have a specific line?
Had some luck with "clear (fe)male voices". It's not guaranteed but I think, it helps more often than not.
Yes, if you enable “Custom Mode”, you can choose your style of music and the specific lyrics.
Here is a video demonstration of how you can still use Suno to write lyrics for you based on your suggestions, even in custom mode.
Do you have a video about how I can get suno to pay my taxes for me based on my suggestions too?🙏
Hello, I have a little problem on suno.com with the list of Liked Songs When I scroll through the list, it only goes up to a certain point and then doesn't load any more liked songs. I just see the small charging circle permanently displayed under it. Is there a way to load the list differently? So not automatically but with a number of pages perhaps?
Hello, with the new cover/cropping features, is there a way or trick to keep the core song in tact but just update a few lyrics?
Try to scroll slowly and once it starts to load stop scrolling until it is done.. that is how i have to do it. If that doesn't work i have no idea.
What the...😂
I think you extended from some silent part of song, go extend one second before
I just tried it again and it worked😄
hey i need help with suno
🤷🏽♂️
Can anyone help. Im pro user but the option for 'full song' hasn't shown up for the past few songs ive generated. Not too sure why. Can anyone shed light on this?
Have you looked in the create menu?
Anything specific you have a question about?
Hi, yes I have looked there as that's where it usually is but looks like it disappeared early this morning
I've already tried, unfortunately it doesn't help either. I've tried different browsers, always the same. Reloading the list for the first time goes without any problems. If I continue to scroll until the reload circle appears for the second time, nothing happens. is then only constantly charging.
I've tried the menu drop down under both tabs and it's not showing up. In regards to clicking 'create' to show secondary menu, it's not a clickable on the android version maybe I'm wrong
What Android version?
Hi yes I understand that but these were generated over 10 - 20 hours ago
Yes they're playable
Maybe I should try on a windows browser
12
there isn't an android app though
Sorry i meant the web app*
Sorry how do you mean? These are extensions of a newly uploaded loop. But if you're asking if I had capabilities to dl full song previously then yes
I've only had the account for 2 days. 1st day was fine. Then they stopped early hours of this morning (UK)
Ohh sorry let me check
Oh... sorry, mate.... I thought you had for a few days at least. So you got a full song on your first rounds. Then just check those you did anday ago.
Yes I can see "get whole song" on previous tracks
I do upload audio clips on everything yes mate
And they have all generated artwork/timestamps
Not the specific one I'm referring to but if they do have vocals it's usually one phrase looped (Hip Hop type of samples)
That's was my first move before I found this discord 👍🏼 not sure what the respond times are like but I haven't received anything yet
I will do mate 👍🏼
Oh...oh... and exactly almost 16 hours ago, suno did an update and shut access until 0600 UTC. Ummmm.... 0000EST. Your extensions may have been affected? I probably should have started with that, but I just remembered.
man I am not happy with the way suno has been behaving lately. It is not following style instructions at all, even with exclusions on. It's doing it anyway. This was working so much better in Aug and Sept.
im pissing thru credits now to get one song right. This was definately not like this during the summer
i possibly have noticed this too
Anyone have any luck using the cover beta feature to correct lyrics? It seems like since it allows you to input different lyrics in Covers, that you could fix messed up lyrics changing words with other words with matching syllable counts. But I’ve been attempting this over and over and so far no dice
I am not sure what changed but this has been ongoing since the week of Sept 16th for me. Some songs are coming in at 40 seconds. this NEVER happened before. I must have opened 10 bug reports today alone so far.
Guys, has anyone here found a way to make the vocals "CLEAN", without the noise/hiss in them?
If I do an instrumental, it comes out great. But if I use the same instrumental and add any of Suno's vocals, the vocals come out squeaky, very "robotic". Has anyone achieved results without this hissing?
you can try to put at the top of the lyrics area [Clean Vocals] and then in exclusions add vocaloid, but its hit or miss. Suno's been acting odd lately
Indeed, in my case it sings the tags …
this is becoming very frustrating. I just wasted 100 credits and got nothing that as in the structure i wanted nor did it follow the lyrics I wanted. what is going on here? I know its AI and its not perfect, this is beyond that. A month ago, it was kicking ass, now its really just awful to use.
is there way i can do male and female like a duet, i try type duet, but Ai doesn't seem to understand
this new cover function works great. exxcept for the vocals it is very hard to switch from male to female
maybe they should implant this, like there has been times i got, but then one voice takes over
Getting duets is fairly rare
I've had luck with prompts like Harmony Singing, Mixed Choir and Barbershop (at the end). You can also add female vocals and male voices. In addition you can put it in the Lyrics Box with brackets. Something like [John singing] and [Mary singing]. But there's no guarantee that it will be done.
The best way would be to prompt the song in sections and put them together with the "Get the Whole Song" button. If you got enough credits for that. 🙂
Is there a way to get my own image to be shown in the video with lyrics? I uploaded a new image, but it seems like it doen't want to change
I haven't done duets recently, but in the past I had the best luck with modifying the text formatting of the lyrics to look like musical script.
JONATHAN: Hello I am singing.
MARY: Now I am singing.
As just one example, there are many formats. Look at operas too. If you combine this with other nudges in the style prompt, genres that typically have duets, and the other suggestions above, they can both add a little influence and increase the hit rate.
The other trick is to start a duet using something that triggers it easily, very cliche, and then Extend From that with your real song. So the song history has both voices.
Covers might be good, but I honestly haven't tried any duet tricks with Covers yet.
I tried playing around a bit with uploading a sample with two voices in it, to see if I could force the AI to make a duet with them. But I ended up getting some unholy conglomeration of them mixed into one. But, I havent worked on this enough to give up on the idea
They might need to synergize a bit better. Maybe you could try using the Cover feature to gender-flip one of your Suno songs. So it sounds mostly the same, but with the other gender singer. Then edit the Cover and the original so the verses alternate between singers.
Yeah, that's a good idea. Covers probably has a lot of untapped potential
Oh with Uploads being 120s now, another approach would be to edit all the verses and Cover your Upload. That might transform the crudely edited upload into something smooth while preserving the voices. I haven't tried that specifically, but I have gotten Covers to smooth out clips like that.
Do more generalized simple tags work better, or more specific ones? I am trying to make something like the Estonian eurovision entry with suno and it's very difficult to remake on suno with how unique that song sounds.
There's no general answer to that. It depends. For something truly unique you probably need to push the limit and use more unusual tags, and a lot of them. But of course that can give you something unique, not something uniquely like another unique song...
Anyone figure out how to stop suno from doing a verse the right way, then starting the new verse but going back to the previous verse and merging them together?
Hey, kind of an off the cuff question here but, does anyone have any idea where I can get information on how Suno did their Database to train Bark & Chirp? Such as what tags did they use and for what? what Metadata was used? did they use tempo tags? Is there information anywhere on the more under the hood stuff?
No real information out there, other than very occasional dev posts in this Discord, or maybe on Bark github. But tempo tags we can deduce were probably not in the data, since prompting for tempo has always been slightly-better-than-random-but-barely. Even though you BPM is a feature you could probably tag perfectly since it's easy to calculate, maybe it wasn't helpful to music, who knows.
Do we have a list of tags with their weights kind of calculated? Like for example adding "chanting" will always add chanting to a song. Do we have a list of "this for sure works" anywhere? Also how do I look at the Dev posts specifically? Thank you for your response
There is no list provided by the devs. I don't know the reason exactly, but I can say from personal experience the devs are often less knowledge about prompting the model than heavy users. Plenty of tags that can't possible be official tags are very useful and have been discovered by users.
For weights, I personally estimate the weight of a tag with the following process.
- Generate a few blank style/lyrics song. The most common result will be a specific style of piano music. Learn to recognize that style.
- Add the single word to the style prompt, blank prompts again. If you are still getting that piano music, it's not strong enough. That means Suno punted on the style basically. (sometimes it might work occasionally, it's right on the edge.)
- Add the word twice, like "barbershop, barbershop" and check again. At some point you might notice a threshold is passed and the word actually DOES work, it was just weak. I don't bother repeating more than 4x because if a style word is so weak you have to repeat it 4x, it's usually not very useful.
This is just a quick estimate. There are some styles which if you use by itself, don't seem to be very special. But they have stronger effects when combined with other styles. In V3 one example of this was "field recording". (It still works in 3.5, just not as impactful.) But as a quick "is this word even doing anything" check, learn the spot the Suno style music.
Ahhh, I did not think of that, very smart, thank you
Interestingly, even on the Explore page https://suno.com/explore it's not hard to find tags in there that barely work or don't work at all. It's still worth browsing for ideas, there are many good ones too, but it's no sure thing even for the example page.
Hi, I'm not sure if this is the right place to post this, but I was in my account today and all of a sudden all of my songs disappeared. They are not in my library, not even the deleted ones. Could anyone assist me please?
Having the same issue - I can search up and play old stuff but otherwise the Create and Library pages are empty.
Similar issue, can't use "Get Whole Song" feature either
what the hell is going on?
That's what I saw. I had just generated a couple of songs and was trying to extend one. Suno refunded my credits for one of the extensions and the other one just sounded like noise. Then I got the message in the picture.
"1 out of 5" that was meant to read.
loosing money not seeing the songs....
I see them again. I had to touch the "Liked" and "Public" buttons a couple of times in the Library.
hummn suno is reading whats in the [ ] for some reason again...
as in its singing whats in the [ ]
[Dark] and [Agressive] its singing -.-
aggressive *
not wasting any more credits until y'all fix that
It's so far pretty subjective and random reviews. I liked some things before too but more for the randomness than hitting the target.
but the others are not there
can any support explain this please?
i cant see my songs in library!! Anyone else !!
Yeah, I cleared the window like I did yesterday and it continued to not do that close the browser as well reopen. I’ll just let it sit for a few hours trying to sleep anyway that was previous so I can see it somehow cashed
same here ! wtf, I lost my songs
all of us
Hey, hey, guys. Can you please tell me if Suno can understand and correctly reproduce such a promt?
[Intro - Distorted guitar riff, minor key, slow tempo, with echoing drums] ⏪
Empty streets, the wind puts out the fire
Dead souls wandering where there's no one.
A rotten world, silence, only echoes.
No one's life, no one's fate, only decay.
[Verse 1 - Aggressive vocals, minor key, heavy rhythm] ⏪
Streets, streets, dead and empty.
Blown by the wind, burned to dust and ashes.
Cities, cities, without a trace, without a face.
Death reigns, eternal fear, no escape.
i hope i don't loose my songs, because those are many many many hours on it
If I literally go back and delete my original lyrics, add 1 space bar space and SAVE, suno often acts like the song never existed. This is somewhat helpful in attaching a new extension and not getting that 2nd verse curse, but then that extension is almost just a whole new song so it often doesnt gel right.
But, something to be said to know that suno uses lyrics from the original and every extension simutaneously, or may prioritizes the previous lyrics over the present lyrics.
I had to touch the "Liked" and "Public" buttons a couple of times in the Library. Do this !!
sure i know, but the others
I see them again.
ok me too
For me, I know when Suno is overloaded when it starts just picking random words and making up versus in courses usually happens when I extend a song so that point, I usually shut everything down and start again usually a few hours later and it’s cleared up
Guys, help, who knows the answer?
I was wondering if something like that could be done, a way to erase some of the seams between clips. Suno is pretty good at doing that with extends so it's good to know that it works with covers. I'll try it, thanks!
Are covers supposed to sound exactly the same as the original song even when they have different styles in the prompt box? I am confused because I cannot get a different sounding song even though I change the prompts.
Exact duplicated can happen (sometimes close duplicated with better sound quality, can be a good thing) but it shouldn't happen all the time if you are changing the style prompts. Try changing the style prompts even more - make them longer, even more different words. You could also try slightly modifying the lyrics, eventually the song will change, but it could be the case it goes from exactly the same to totally different. This isn't the typical case, but it could be an unlucky song where Cover happens to not work very well.
First, you try and see if Suno can do your job…, then you tell us the result.
The thing is, your experiment could be the answer for many problems here.
I’ve done it with successful results. Just try and see
you want to split that prompt up
eg from one of mine
[Instrumental Intro]
[Heavy Guitar Riff]
[Blast Beats]
[Melodic Guitar Leads]
[Melodic Death Metal]
[Dark]
[Aggressive]
[Fast-paced]
tho it doesnt understand blast beats. 😂
tho the AI seems to be reading tags as lyrics AGAIN. after clearing etc 😦
wasted 40 credits, on crap again
well im done, might find a new music ai
i spent more credits on songs where the AI read the tags a lyrics than ive made songs that didn't
its has not happened for the past 100 credits
now *
so its not rare its annoying 🙂 so something has changed in the past 2 days, and clear isnt working nor is reopening the browser etc.
It's not a cache issue that you can solve by refreshing the browser, and "clear" only removes the input fields -- it doesn't have any larger effect.
no, clear doesnt do a thing
"Clear" was added to help people from needing to ctrl + a to get rid of all of their input.
As stated by the devs, the biggest determining factor on Suno's output is what you use as input.
I understand it's frustrating getting the same errors 10 times in a row, but it might be an indication that you need to reapproach your methods if you and Suno aren't operating on the same page given the input you've provided.
I saw that Woodhouse had to catch you from swearing at me, but I'm happy to help you figure out whatever your current song is snagged on, if you're willing to take a collaborative approach to it. 🙂 
it wasnt a swear 🙂 it was the nice from of of word with a e at the end 🙂
as i said the clear function didnt do a thing.
😂 😅 I apologize, Woodhouse made it sound like you were swearing -- which I wouldn't blame you for, either 😛
I understand being confused about the "Clear" button if the impression was that it would "reset" Suno's logic somehow, but each song is a separate instance (even songs that are generated simultaneously aren't connected in any way beyond using the same input).
Either way, if you wanted to share the song you were getting stuck on, I'd be happy to take a stab at it later 🙂
Can confirm, Woodhouse tells me it was a common phrase to say "it doesn't do anything", where the offending word rhymes with "night"
British hamster gives a thumbs up -- I immediately recognize what was trying to be said.
i put all 13 generations in the playlist above
All
Good 😊
It did work once for me lol
My only suggestion would be to put more stuff in the style prompt and less tags in the lyrics
Though the style prompt box has limited space, squeeze whatever you can into that and use tags in the lyrics where you want to adjust one or two things.
Not a problem at all man, I don't want you to think you messed up or anything -- there's a lot of stuff to learn, and the only frame of reference we have for a lot of it is other AI generation sites, many of which have some form of "memory".
I appreciate you helping out, dude 
The best way we can learn Suno is together 🙂
This was that is the style prompt in the past has only listened to one of the genres or styles in the list and 100% focussed on that where I have successfully had the AI read the tags and change as required
Only tags in the lyrics section
Worked perfectly
As well as the 3 here
Where season and I were trying to get one of my songs working using the styles section for like an hour? Maybe two and lots of wasted generations during that as well
Dasvon *
Hmm. I have no explanation for that. Pretty sure style tags are meant to have more weight than tags in the lyrics (even then I have had the vocalist sing part of the style tag before.)
Is there any particular pattern to what tags are being sung? Or is it just random different ones each time?
So far it has sung, dark, blast beats electric guitar, melancholy, powerful, uplifting chorus. Those are the ones off the top of my head.
Oh, and melodic death metal
Same or similar tags it doesn’t sing them
Random things I'd try: Get rid of the blank line between the tags and e.g. [Verse 1]
This is in case it's assuming that the block of tags is meant to have lyrics, and are separate to the verse.
Or, combine multiple tags [Blast Beats, Melodic Death Metal] to reduce the number of tags. Never tried this but I'm just wondering if it'd be more tolerant of fewer tags.
Final thing I'd suggest is put your common tags at the start and don't duplicate them throughout the song (only tag what you want to change.)
I’ll try those out. I did have a problem where if there was a comma in the tag it would sing it, which is why I separated them.
Maybe leave that thing to try last if you've had bad luck with it before
I'm cool.
I'd agree with that part for sure.
fwiw -- you're a friendly and enthusiastic helper, and the only resources we have to share are our own experiences -- and nobody can fault anyone else for not "perfectly understanding Suno" in its growing phase.
Please don't be hard on yourself, especially for what amounted to a friendly and collaborative learning experience for multiple people.
Much love from Eastern Canada, hope your <t:1727958660:t> is going well 🙂
I'm chipping into your conversation at the tail end -- personally I'd be more apt to handle it in multiple sections, using instrumental extensions for the guitar solo parts, and fitting most of those metatags as the style prompts for the extensions.
Is that more or less how you'd tackle it?
Agreed it’s a learning but for many of us and it did work for that one song I was having issues with
Yeah my personal approach would be a credit-burner, using the style prompt as much as possible and extending often. But in the long run it probably burns less credits than trying to get a one-shot gen with all the tags/prompts being handled correctly.
Yeah, I’ve already burnt through because of it singing tags at least 300 400 credits and I’m at the tail end of my credits right now even though I only signed up last week 2500 credits in like a week or two🙃😨
I used to think the Piecemeal Approach was credit-heavy, but I'm leaning toward that logic of having a more controlled approach is an investment in surety itself opposed to one-shot gambling through complex prompting.
Yeah, I think I might try later today just like an intro inverse and then if I like what it does then extend the song and the chorus and then extend the song and the next first and then extend the song and then extended song, etc. etc.
I’ll let you know if it works😂
I recommend trying it. Each song can be so different so this method means you can keep rerolling until you get an intro you like, then you extend it and just keep tweaking each piece until you get something you like.
Praying for unlimited credits
Nope didn’t work still at 80 😂
I know you've sunk a bit of time into this one, so total kudos to approaching it from another angle -- this song is going to be your magnum opus.
Let me know if you have issues breaking it into chunks & extensions, I'd be happy to help ❤️
Pro tip -- extending from an extension isn't the same as extending from "Get Whole Song" -- costs no credits to get the full song to extend from, so be aware of the different methods -- both to avoid the issues of extending from extensions, and to know what tools are on your belt.
#😜┃suno-memes message Legit, the hidden 5th membership tier is the best thing they ever made
This is where it gets interesting...
Get Whole Song will give more context for the next extension. If you want consistency, this is good.
If you extend off a short extension you lose some of the context, which is ideal for switching things up.
Not the ONLY way to do it -- but from my observations the extension can only "hear" the clip it's directly extending from.
Each clip you extend from can act like a barrier to prevent the new extension from hearing prior content -- so a guitar solo at the very beginning of a song wouldn't be picked up in a 4-part extension -- but you can "get whole song", then the first chunk is all one section for the extension to "hear".
I know that makes no sense, apologies.
Can be super helpful if you want to force a style change, or have verses that sound similar and build on one another, without necessarily continuing the motifs from earlier in the song -- that would be more easily handled through a multi-extension song than it would be through complex prompting, imho
when i make cover song Is there no prompt for 'follow the original melody exactly? for example [follow uploaded file melody]
like this
has the log in with google button been messed up for you guys.?
it just leads me to this website
Hi everybody. Is it possible to upload a melody made with a synth and generate a solo voice for that with given text?
just curious about how it produces effects like echo, atmosphere, pumping compression. is it just part of the model results or is there some kind of mastering effect applied after generation?
there is also sometimes filter automations... maybe it's top secret info.
As far as I'm aware the audio is generated in a one shot process so any effects would just be based on what the model was trained on.
wow. whatever it is, the pre-masters are awesome!
It gives interesting results, particularly when doing weird stuff like with "glitch" style music. I'm always fascinated by how the results sound
cool. i'll have to try more of that stuff. thanks!
what happened to the cover feature today i uploaded the beat and it took me all day to do it but today it changed the beat
No, it's still working with the usual Google Single Sign-On procedures as usual -- maybe you need to clear your cache as indicated in the page you're linked to?
Failing that, try logging into your google account on another website, like gmail, then try using the google sign-in button, as depending on your browser, it will use the saved login information. 💪
I never noticed the clear button in the browser until your I saw it in your video. I feel like it needs to be at the top of the interface and not at the bottom, but that is just me.
No @flint haven it is not a malicious link. It is just prettier better to use markdown to make links prettier.
I use the GPT ESSENCE EXPLORER https://app.aiprm.com/gpts/g-utsMc0LBR/essence-explorer by @onyx swan who also wrote the Prompts for AI Dance Music PDF from the community suno.wiki. It helps if finding the "essence" style of a song or from your custom lyrics which fit nicely in the 120 characters.
what i am curious about: does anyone know if the suno training dataset is updated or is it fixed as is ,not talking about different models but whatever went into in the first place?
They haven't disclosed the dataset used to train Suno, and aren't likely to comment on it due to ongoing legal cases -- that, and just being part of "the secret sauce".
ok makes sense... but i guess you oldtimers still have a feeling of things that may have changed over the last months 🙂 was just a general curiousity. the diversity in terms of styles languages etc is impressive all the same
You can change the model to V2 and V3 to try to see how the different models have progressed.
One hack is to extend with V3 to end a song when V3.5 has trouble just ending cleanly.
Very true -- #🪲┃bugs-web message one the devs noted how each model should be looked at as its own separate product due to the fact that they are trained from the ground up
I found out what happens when you're down to 5 credits and "stuck there" as a subscriber. When you'll balance out to less than 0 Suno now allows you to use 5 credits to create a song, then you are placed into a debt counter that it replenishes the next day
Also should be noted that features like Crop, Covers, and Extends are focused on improving user interaction and enhancing creative control. They are not necessarily LoRA (Low-Rank Adaption) fine tuning techniques.
Crop saves a bunch of over-extends. it's nice to finally have that feature
but, I think a fine-tuning tool would be insanely useful
#📚┃suno-school 1. Something happened to delete ( move to trash ) option. Some of the songs, cannot be deleted. At least, till you reload the browser window few times .
@prime gale Amazing update! Thank you! Does the 2min length work for cover songs?
#📚┃suno-school 2. Quality of songs get worser every week. Few weeks I can't finnish one song, coz ofvery bad quality of sounds
I'll assume you upload, click extend for the song, then click to cover the song? 3 steps. So, like before, the cover might cut the 2 minute song a second or two short, sometimes?
if i want a slow, chill hip-hop beat or a slow beat in general, why doesn't it give me one when i put "Slow Beat" or "Chill Vibe" in the style everytijme? Sometimes it will but many times it's not and the vocals are super fast. Any way to get around this?
also, when i want to extend a song bc a better beat comes in at 2:45. When i do that and basically want to create a new song with that beat and with the same lyrics, why do the lyrics start at second 0:01?
First thing to try is the simplest. Repeat the words. "Slow Beat, Slow Beat, Chill Vibes, Chill Vibes"
Probably better to also add variety "Chill, lo-fi, lo-fi beats, mellow keys, dreamy, melancholy, boom bap drums" just to go off the top of my head for lo-fi.
So I know crop was never supposed to work on an iPad anyway, but it used to and now it doesn’t.
Hello is there an admin that can help me with my problem please ?
this weird window on the right been for a couple of days there and i cant get rid of it
and whats that for?
it's just the site design
I would say check if your browser is zoomed in more than 100%
I am not experience that issue (though maybe because I am using Google login?).
Do you log in when you are using private browser mode? Or have your browser set up to delete cookies when you close the browser?
hello.
I made some music on the free plan. And I'm currently using a paid plan, and I like the music I created during the free plan. Can't I use the music created during the free plan for YouTube content?
So... It seems that i am getting pretty far with developing methods to silence the demons. >:)
All in post production
As long as you don't stand to make something from it, like youtube monetization
or royalties
Right?
thanks.
Unfortunately, no! Subscribing to a Pro or Premier plan does not give you retroactive commercial use licensing for the songs made with a free plan. ...
@heady fable
Did they do something to suno? Cause vocals have just been absolute trash the last week or so. I've burned through so many credits this month just trying to get something it was nailing quick last month.
And I'm not like a perfectionist or anything. I mostly make stuff for fun or live streamer slop content.
Was this female/male duet luck? Can't see anywhere in the tags or lyric tags indicating one.
Tiento Tango de Malaga, Hip hop, Latin music, rap, accordion, Hip hop, song. Listen and make your own with Suno.
Getting a Network error
suno is being turd again..ALL MY CREDITS WASTED! Whatever I try to do, am always getting female vocals.
I tried all combinations for male singers. Voice, Vocals, singer..Even going into lyrics part and doing the [....] saying just that.
Not to mention, I tested something at random. I made random lryics, and typed in some music type. AND IT CAME OUT! At the one listed.
Am scared if this is the plan, to make us buy better package now!
I wrote in..Darkwave (just that) and made up just wrote random words in the lyric for a test. I got country.
Then I did glam rock, and I got gospel music lol
If I had to guess from the lyrics it was luck and they expected a female vocal all the way through, then the [female] before the final chorus indicates where there was a continue, hoping to steer it back to the same vocal as at the beginning for a finale, and they ended up singing together.
to get a suno duet, i found if i put in the style like "male vocal" and then i extend it but than switch the style to "female vocal" and once it gives me a female for it (usually its 2-3 try) i will then get it to make it into the whole song and after that it will typically stay doing a duet for the rest of the song. atleast in my experience its been pretty easy to get suno to do duet
Quick question - is the voice quality "burned into" the generation seed? Or is it possible that the voice can be better/smoother if I extend before the singing starts?
not sure, but i try to start off with a clean vocal if possible, i found using the previous suno model to start and then extend with the newer version.
Not sure if it's been asked but sorry if it has been already.
How would I go about using specific vocal techniques and vocal styles when creating a song?
So far I've just been adding it under the style of music box
Kind regards
Wen android app?
How to make the covers more like remixes of the songs? Sometimes, the covers sound completely different, not keeping the melody and structure of the song. Any tips?
So covers are limited to 3:08 it seems. It's consistently cutting songs off there.
When they find Kotlin people
I'm having same issue. do they response? do you recommend Udio instead? thanks!
Doing a cover may have better luck at getting a new voice than trying to continue, particularly if you liked the tune you got.
Hi,
I recently purchased the Pro version and would like to know how to remove the Suno watermark in the video. Could you please assist me with this? Thank you!
Not possible.
The watermark is there to let the world know it was created using Suno.
What prompts or method, do you guys use, to establishing a momentary pause?
Example:
I am now singing
This empty space signifies a pause<----
Now I am singing again
Do you guys use any good prompts, instead of leaving a empty space?
If I don't have the text belong within the certain parts: Verse, Chorus, etc
Suno sometimes likes to understand that it's not part of the same section.
I've tried with:
[breath pause]
[sustain pause]
[tension pause]
[Pause]
[Hook]
[Pause] [1 bar]
None of these seem to work
The lyrics-videos you download directly from Suno will likely always say "Suno" at the bottom, but you can edit the videos externally, or download the song & create your own lyrics videos like my friend @mystic flume #📚┃suno-school message
Punctuation, line breaks & spacing seem to work the most reliably for me
Yea, that's what I've been using as well, just wondering if anyone might have a better method
https://suno.com/playlist/41d7b3f6-c779-468c-83b8-b9ceb10ba065
Here's a playlist of all of the "command" type prompts I've tried, vs simply separating the lyrics.
Can you elaborate on what you mean by "better"? More reliable, specificity in how long the pauses are, etc?
Am so irked! I tried eveything. And am not changing the style! As this is style I want!
As it worked before! For months! Am still getting FEMALE SINGERS!
male vocals, neoclassical synthpop, darkwave, Cinematic, 80s Film, ethereal , baroque, symphonic
Like, sometimes I want to have 3 bars of pause, and sometimes 1 bar, and sometimes 17 bars.. But if I write text with a bunch of empty spaces, Suno tends to jump back to these spaces and add them instead.
Let me try this [Wait 1 second]
I just got this on reddit, is this a lie
I covered that in the playlist, and didn't find a single instance of success with it -- but your mileage may vary.
I'm sure, there HAS to be a proper prompt for it
Best of luck in your search 🫡
I personally believe you can't command Suno, you only talk to it and it vibes with you.
the genre sometimes influence the AI to sing in a specific gender.
Yeah, I've trying so many ways to say stop or pause, and it doesn't seem to like it at all
Like saying [guitar solo] might suggest a guitar solo, but it usually works best at a point where a guitar solo would naturally occur in a song, so it's not a firm "redirection" of the song, but a nudge toward a direction it could have naturally drifted toward anyway -- in contrast, try throwing [guitar solo] in the middle of a lyrics-heavy verse and see how rarely it works then, because a guitar solo wouldn't cohesively fit into most songs.
That said, I'm still more curious about what Suno CAN'T do, but that doesn't mean that everything else is completely known and manipulatible with 100% accuracy.
Yeah, that's how I've gotten away with it, most of the times, but this is almost a purely vocal performance that I'm working on, so I don't want to interrupt it with something, which might guide it towards another unecessary outcome
I've been also using interjections, which works, but I just want, silence.
Since we have so many tools available now, it might be beneficial to work on the song externally and bring it back into Suno for touchups now that we can upload 120s?
For instance, you could generate a song you like, strip the vocal stem with bandlab.com/splitter, then take the stem directly into their studio to cut it up and add the spacing exactly where you want -- don't worry about whitespace between the tracks -- then re-upload up to 2m of that into suno and do a cover to "smooth over" the silent whitespace and it may retain the gaps and spacing 💪 🤞
As I usually say, the more specific of a song you hear in your mind's ear that you're trying to recreate, the less likely you'll be able to achieve it solely through complex prompting alone.
I usually do [Interlude][Virtuoso Guitar] to get a guitar solo.
What's the definition of an Interlude?
I haven't played around with it too much, but it looks like that pair of metatags is enough to at least suggest that the pattern should be disrupted, whether or not it adheres strictly to a guitar solo (but still a somewhat reliable way to get a guitar fill at the very least -- worth exploring for sure)
In music theory, an interlude is a short instrumental or vocal passage that serves as a bridge or break between different sections of a musical composition. Its primary function is to provide contrast, transition, or to prepare the listener for the upcoming section. Interludes are commonly found in various forms of music, such as classical, jazz, rock, and pop, and can vary in length and style depending on the musical context.
Key Characteristics of an Interlude:
Purpose: Acts as a connecting element between larger sections, creating smooth transitions or providing a pause in the musical narrative.
Placement: Typically appears between verses, choruses, or different movements of a piece.
Form: Can range from a few measures to a more extended segment, and often includes thematic material that contrasts with the surrounding sections.
Interludes are used to maintain the listener's interest, introduce new motifs, or to create a moment of reflection before the music returns to its main themes.
Can there be multiple Interludes?
Yes
That's great too know, thanks!
ya but it always worked for me now...am trying any style...some I dont even use...and its still all female
Are you saying it is staying female and you want male singer?
What genres are you targeting?
Hi guys, any Intel about coming better stem separations ( instrument layers ) or improved vocals ( less muddy) any soon?
It is best to download the song an split in an external application like BandLab Splitter, SpectraLayers, iZotope RX, or one of the ones that are included in a DAW. You will get better results than what is provided directly from the Suno interface.
I know, I already do that
That wasn't the answer I'm looking for..lol
No, we don't know if Suno is working on a better stem split.
Hi ! Try this: [Brief Silence], sometimes, it's working well .
Do you have any examples per chance?
have you tried using 'male-voice' in your style section?
I was able to get pauses based on how words were spaced in the prompt. https://suno.com/song/de9ca997-3731-4497-8c65-d1fd3abc9027
Neoclassical Trance, Glitch, Stutter, Modulation Variation, and Hypnotic Loops song. Listen and make your own with Suno.
I don't have an example right now, at the moment, I know I've used it sometimes. If I find one, I'll put it here! Otherwise, yes, spacing out the lyrics can also create a break! 🧐
The fact that they're explicitly calling it v1 suggests it'll be improved at some point, but all internal information is shared in #📢┃announcements #🚀┃updates or #🔔┃social-media -- anything else is speculation or rumours.
Unless it is V1 forever 😉
Not even good Rumours like Fleetwood --
unless it is v1 forever
Very true -- naming schemes are as much an advertising technique as anything else, but it seemed like an intentional choice.
A lot of new features are locked specifically to v3.5, so I wouldn't be surprised if the existing features have better support in v4 -- but again, entirely speculative for me to think that
How can I get the music to stay at the same volume and intensity when vocals come in? The default is to quiet the music on verse and I don't always want that.
@charred merlin I was wondering why when i do want to do a cover wit hthe same syllable it shows me the new lyrics but does the same song wit hthe wrong lyrics that i modified?
Hello guys. I need some help creating a song by my exact input in the Style of Music.
I wanna create a Japanese 8-bit game song which is grim and dark but suno always generates a rather happy song. How do I put more emphasis on a dark styled song?
Something like this? https://suno.com/song/92a908e5-5311-472f-b754-b35cef19b30d
8-bit, Dark, ethereal, gothic, minor scale song. Listen and make your own with Suno.
I dunno if minor scale does anything. but you definitely don't want a dark sounding song in major scale.
I’m having trouble logging in with my discord on my phone.
It shots me to a page that’s just white and nothing else
This is nearly perfect. It just needs a japanese setting now 😄
Thank you
Here I am again, with another question.
First time attempting to use uploaded audio. It's a short (25 seconds) amateur radio speech in a foreign language (slavic). I wanted to use the first 13 seconds of it as intro for a song, just out of curiosity.
The problem: Suno ignores it when I try to extend that segment into a song. The most it carries over is about one second of sound.
After spending quite a few credits trying to make it work, I decided to ask here: am I doing something incorrectly? I tried using [Male Narrator], [Intro] tags, and no matter what I try, the result is the same.
If I don't use any lyrics, I get an instrumental song, uploaded audio segment is ignored.
If I use lyrics, only the lyrics are used, no uploaded audio segment.
You have to use "Get Whole Song' to append the upload to whatever suno generates from it
I could have sworn there was some overlap when I extended my previous songs, but you know what, yopu might be right, because this is speech-based, not music-based.
sometimes suno will repeat portions of it in its generation, but if you want the actual sound clip in the generation you need to Get Whole Song
But usually it does this more the longer the clip is
Can I post feature suggestions in here?
Thank you,. yes, that did it. All is good now 🙂
Sorry to keep you busy 😬 I also happen to have a problem using Suno lately
A notification of "System Error" pops up whenever I try to generate any music.
Does anyone know what could cause this?
If it helps, I also recorded it happen.
People who have had this error have said they needed to clear their cache, clear their cookies. Also try a different browser.
oh, okay ... I think my cache and cookies get cleared automatically, but switching the Browser I can do.
Thank you for the quick advice ^^
Aaand it works like a charm! It appears, the cache and cookies were in fact not deleted automatically. After handling that, everything was back to normal. I didn't even have to change the Browser. 🤩
You can contact support at billing@suno.com for Discord account linking issues 🙏
Covers is still in beta, and there's still a lot to be learned about it.
I'm still experimenting with it, though maybe the question could be opened to the guides, moderators, or other helpful community members 🙂 💪
I'm guessing it's not likely going to come up as frequently, since the reported instances have dropped, but I'd seen a couple weird instances where it seemed like simply changing the model selector was enough to fix it -- one user suggested it was as if the v3.5 selected in the dropdown menu wasn't actually pointing at anything, but re-selecting a model was enough to fix it.
Super elegant explanation for the errors, and it costs no time or effort to test.
#💡┃suggestions-web #📱┃suggestions-ios #🚧┃suggestions-layout are browsed by the dev team occasionally, if you want your suggestions to have impact 💪
Hi guys, sorry can someone point me in the right direction. I'm having trouble extending instrumental songs. The song only extend by 3 seconds and these clips sound like the end of the song. Not sure where I'm going wrong, has this happened to anyone before?
Can you share any more information about what you're extending from?
Are you extending from silence, or a quiet part?
I'll try it. thanks.
Thank you, I was extending from the end of the first generation but that sounded like the end- hence I now understand why the extension was an extension of the end. I'm new to using suno, really appreciate your help.
I'm not sure if it would be helpful, but I made a simple video tutorial for using the extend feature a while back.
Welcome to Suno -- if you need any help, feel free to ask any of the guides or moderators for help (people in blue), or consult the knowledgebase at https://help.suno.com 🙏
Thanks for the link- will watch that now.
Ugh, actually -- don't watch that video -- I was demonstrating a couple different things at once and I feel it's more liable to confuse 😅
I'm gonna take it down in a few as motivation to make an updated one that coherently demonstrates what I'm talking about 😂
Hi, does anyone know what I should do when I edited the lyrics of a song, clicked save and played the song but it still has the old lyrics in it? Should I wait for it to process or do I have to click something else?
If you're talking about the displayed lyrics, Editing the lyrics only changes the way it appears directly on suno.com -- if you're on a paid plan you can click "regenerate video" to force the download / discord embedded video to reflect the new lyrics and uploaded images.
it won't change it in the song though
you might be able to get a Cover to do that, but it's extremely unlikely without changing the song up
I've been trying to find a prompt that helps clue Suno in to do that, but I haven't had much luck
is there way i can download the audio and use it in a new song with the new lyrics?
you could maybe download the stem and use that as an upload
you could cover it that way
where i can doi this?
should be in the 3 dot menu
I have a gen of a song I've been trying to change lyrics on with a bunch of hallucinated gibberish that I'd love for it to realize is gibberish and fix up with my new lyrics
think it might understand [Vocals Retake] in the lyrics field?
or [New Vocal Take]
A little joke thing I made...since suno is on the fitzs me voicing my rage
I admit making this random thing, made me laugh and feel better
lololol enjoy! I mean no shade..its just a joke to myself, and some friends..I wanted to share here too ❤️
is there a term for using gibberish as a guide for actual lyrics later on?
yeah, I'm done with that for now
you know what, the hell with it. I'll wait until we get real inpainting
getting sick of the monkeypaw effect
There is no inpainting to fix lyrics. That is a feature request we have all been asking for. The only way to fix lyrics that get messed up is to extend at the timeframe just before it happens and rewrite the lyrics starting at the next line down. It will use the instrumental and voicing in the extension.
yeah, I know that
there isn't yet
it's been pretty well-hinted that it'll be a part of v4
I lucked out on one, and got my biggest changes merged in, with only a tiny sound drift, and a vocalist gender swap
she delivers the same performance as the male did though
Are you using Covers or is that same performance from one of the generations/variations?
A few style variations, 2/3 parts up to about 00:15, the rest is still up to change, so we sharing mp3s. Minimal tags and styles:
Trombone, Keytar, Bongo, Roland Juno, Sovietsynthe, Tuba-Didgeridoo
Bongo still aint in there, but I might just continue with drumkit.
Maybe it doesn't know what Bongo is. How could you describe it to match the style you want?
Idc tbh, this direction is fine.
Question. When I extend a song it only goes to 2:00 minutes and when I click 'Full Song' on it, The song gains a total of 9 seconds. Is there a way to fix this?
Extensions are only up to 2 minutes each.
How long is the original that you are extending from?
10 sec
But the extended songs have a 'Get Whole Song' option
And that's where it only goes to 2:09 total then cuts off
ohhhh
wait
Then, is there a way to create a song with my 10 second clip and have it be the full song?
When I use 'Extend' is there any way to force the a.i. to use specific lyrics?
There is a Crop feature recently.
I used that to get the 10 second part I liked
You can "Public" that 10 second part, it doesn't need to be a Whole Song. (In the early days, it did need to, but not anymore.)
What does making it public do?
Also, I have another problem now
I want to extend a song into the outro but it goes back to playing the intro
It just puts the song on your Suno profile. Optional. You can share the link to a non public song by yourself.
Hi, I wanted to ask if with the pro version it was possible to modify the lyrics of a song already created and obtain a new song with the same tune and the same music but with the new modified lyrics
There is a feature that sometimes is capabale of doing that called Covers, but it is in beta and many times does not get the desired outcome
Whatever lyrics you type in the lyrics section is that the Ai will use. You might have to try a few times or move your extention point a little for best results.
Try using the cover feature. If you don't get everything the same with the new lyrics, you might still like the new result, which tends to be better sometimes.
Hey. If you upload audio and have a distinctive melody, let's say from a guitar piece or piano, and write lyrics to do a cover, does suno recreate the melody over the lyrics? I thinks it's done that before, but I'm not certain. Anyone having success there?
so how do brackets work? Can I do more in-depth instructions than just [choir] [solo] etc or are there preset things for that
I don't think it reacts to [drop bass before each verse] for instance, though I see lots of songs with phrases and sentences -- but from listening to the songs, I don't see a direct cause and effects.
As far as I know, just cleverly placing the [metatags] around your lyrics seem to give Suno a starting direction to go with what follows next. It may still be a bit random, where such randomness may skip over what your intent was-- and frustrate us -- but the [metatag] still is supposed to give Suno a branch to crawl out on before randomly sprouting off on one of a thousand random twigs.
My problem is that when I make Covers, it doesn't keep the original tempo or key. I know that this is not the purpose of the feature and sometimes not possible, but it would be nice to have an option to adjust it.
Many suggest, if tags are also weighted, then repeating them might get results, than to type only once... such as
[End]
[End]
[Fade out]
.... might be tried just to get suno to end the piece.
Alternatively, you could also have a thermostat-like setting after upload to change how much the original song changes. This could also be fine-tuned, e.g. just instrumentation, style, etc.
My only observation, as a prem sub, is that I can run about 30 generations and eventually, one of them is different than the others, and sometimes, different and meets my intended direction. That's so vague an answer, but Covers, still is in beta, so there no discerning answer that would help, I'm afraid.
You may wish to address any suggestions to the developer area https://discord.com/channels/1069381916492562582/1260263379243761776
Yes, that's how I do it, but it might be more useful to help more professional users. Quite a few of my professional musician friends use suno and they also have in their heads broad strokes what they want to hear back 🙂
This has a yes and no answer. For the melody, you could start by doing an extension of the original upload with your lyrics. It is possible to create a song with that melody. You can then create a COVER from the song. You can also go straight for the cover. I've had luck with replicating the melody
Thanks. I'll just have to experiment with it more on during my free time and keep notes.
I can’t see the trash section
how to shrink the size
Do they also delete songs when I’m paid coz on free account it deletes songs once I try to keep making songs after a certain point
they disappear off the Create page, but you can find them in your Library page
You need to make a playlist, and put them in there - Or they will disappear
'bout how long does it take for the updated video to be available once you click "Regenerate Video"?
this is on a "Whole Song", but it should still work properly, right?
Is there anything that would prevent a song video from being regenerated?
Sometimes the length of the song has an influence on the processing time...
it's only 2:33
I've tried to regenerate the video 4 different times already
I can tell immediately because it's got all the phonetic helpers for the lyrics in it
Are you looking at the video you downloaded from Suno or the video embedded elsewhere? I know there are issues with embedded videos due to caching
downloaded from Suno
Hmm, wonder when that had happened. I will see if I can poke someone to look at it (though likely they won't until the weekday business hours)
I'm trying it now. Did the first Regenerate maybe an hour ago now?
Hm can you let me know what browser did you use and if it was on mobile or PC
Brave, and I'm on desktop site
does the video generation have any content filter that would prevent it from being generated?
it's a general TOS popup
LOL!
I know, right? If you have the balls to actually listen to my stuff, and you like it, download it
hey, you oughta see the ukie ball PFP here... total lolcow in how hard he rees
To whoever needs to read it, when regenerating videos, make sure that you're waiting until the video finishes processing before you attempt to share or download it.
You think an hour's a long enough wait?
the only other explanation is that there's some kind of moderation preventing the regeneration
Try refreshing the page and see if the video icon is still spinning.
I genuinely wouldn't be surprised if the video regen has some moderation on it to prevent people from using the same content that would get blocked when being entered as the original lyrics.
Stop spamming please. And for the record, the one who won to be in New York, is chosen by Flosstradamus himself. He was not the winner of the grand prize, he only was invited to be at the show.
Says who?
Then you didn't read right
no. there were two 'big' winners: One got the 25K top prize, the other goes out to NYC
I mean, I had my own issues after reading the methodology, but I simply just didn't enter
The cash prize winners are not known yet
If you been around for the other contests, you should have noticed that numbers aren't everything. Some people were really suprised they won since they didn't have the numbers to back it up. So I think it's absolutly fairly judged.
I'm gonna do Unicode replacements on letters in what I suspect to be the words triggering the filter. If that works, then I'll 100% know that there's moderation on it
You can check the home page and see the list of SOS2. Check for yourself.
Pass that to Flosstradamus.. He picked him out.
Please stop making unkind remarks. Everyone's entitled to their opinions, but let's keep things respectful and kind.
No only the winner who was in New York City yesterday.. Which was Teemuth
Not the cash prizes yet
our... criticizer seems to have the performance winner and cash winners mixed up
Reddit's a trash site anyway, so...
Because Reddit echoes the opinions that you find more favourable?
(looks like the issue wasn't echo chambers after all, it's differing opinions 🤷♂️)
Bringing in opinions from outside platforms isn’t helpful here. Let’s stay on-topic and focus on respectful conversation.
...so who is that in your PFP?
looks kinda familiar
yeah, so... it still won't regenerate
it's either broken, or something in the Displayed Lyrics is preventing it from happening
Yet that "no hate speech" part seems to be ignored
yep, and all he's been doing is coming to attack whenever I try to say anything
I don't need to see the history between you both to see you are not buddies.. Which is fine, but then just let each other be.
but uh, just so you know... that rule applies to the Discord, not the site
I made a song he disagrees with, so now he wants me shut up... and likely permanently
Yes I am only here to make sure the discord rules are followed
Why is it that you can post hate speech links in the discord though? It seems to be against the rules
Okay.. if you both do not stop this now, I will add a time out on you both.. Your choice.
Can you respond to my question?
So... you mean someone can just attack someone else and get them timed out for it? Good to know
You both are attacking each other..
And I don't need your opinion on this @brazen yoke, else you give me no choice to let you join them.
So now I'm not even close to the mark with style prompts. "30s sci-fi movie" gave me death metal
What is going on?
And not like once. I just wasted about fifty to a hundred credits
Likely the model does not understand what you mean. Try looking at websites like https://rateyourmusic.com for a list of genres so you can provide the correct subgenre.
I would use 1930s as well instead of 30s (or both). In addition to movie, you may want to try 'cinematic'. But definitely look up the specific subgenre you want
In addition, v3.5 seems to be more tuned towards more popular genres. Hence if you are trying for an uncommon genre and having no luck in 3.5, generate an instrumental piece in 3.0 (better at understanding more genres) then Extend into 3.5 for the better song structure
Here is an example of an output from 3.0 that sounded like futuristic jazz at the start (as an example):
https://suno.com/song/21972385-bafe-48fc-9d2e-5980a2798e96
The other 3.0 track: https://suno.com/song/8131fbd7-c7a7-453a-8ed7-f6cbcc1da957
For comparison, here is the 3.5 output with the same lyrics and styles:
https://suno.com/song/265df54c-2e80-4718-8056-be07cb6d74a7
https://suno.com/song/ad7cdf5c-c144-4eb2-a8af-03d0a9d03a92
How I can use natural sounds? Like rain, wind, birds?
Is there any list tags what really works? Its funny that tags works some songs but some songs not works at all
Yeah did all that
they need to fix the annoying robotic voice that try to sound natural
they key missing in the puzzle ..
Thanks for the advice
people, what's up with Suno lately with intros?? I throw away 95 songs out of 100!
They are songs after songs that make me incomprehensible intros, with the words eaten, or without an instrumental, or he reads the words to me in Italian with a horrible accent. And going forward, he always mixes up the words of the chorus, or repeats the same parts.. and he doesn't do it once but almost 100% of the time. When it goes well, there are bugged words and I have to throw them away anyway. And every time I try to change the words, style, settings, prompts, position of this and that.. start from scratch, with new songs.. and always the same problems. It's useless to pay for so many credits if they all go to hell
If its just the intro sections, I usually do a form of instrumental intro and a chorus saying something like "oh-oh-oh". Either the bugged part will eat that section and not the actual lyrics or you'll get a nice intro 🤷🏼♂️
Been trying to make something similar to "Making of Cyborg" but can't get even close, any suggestions?
https://youtu.be/eFsRiT6e2EU A cyborg might be difficult to start with, maybe start with something like this and work your way up 💪
Build a simple robot with a Raspberry Pi and Adafruit's awesome robot chassis kit. See the full guide at: https://learn.adafruit.com/simple-raspberry-pi-robot
Visit the Adafruit shop online - http://www.adafruit.com
Subscribe to Adafruit on YouTube: http://adafru.it/subscribe
Join our weekly Show & Te...
...thanks? 
Hmmm.. what does that have to do with Suno?
It's non-sequitur humor.
Where do we go to contact the Suno team? My card keeps getting declined. Can't sign up
E-mail billing@suno.com
thanks
it's really annoying that you set female voice and female chorus and the voices are often male...
ello ello, i think im having the same issue as above. my card keeps getting declined and i cant sign up.
do i also email the same above email?
||billing@suno.com||
I am having problems with covers. If I want it to be in a sub genre, say, take a rock song—-> heavy metal, 95% of the time it returns the same song. How do you do a cover in a sub genre?
if i have an audio of my voice can i use it to become a voice in a song?
Well it finally happened. Suno blocked an upload of my own song because I have previously registered it with ContentID on YouTube etc.
😭🤣
How so?
All of sudden I'm not getting any kind of audio output lyrics or music or anything in between and it seems I'm just wasting credits with no results whatsoever just empty white space no music nothing
-
Are you sure that the audio is not muted, volume set low, or browser silencing the tab? This scenario happened before to someone who had their audio muted.
-
Are you extending from a portion of a song that has silence? In that case, try extending just before the silence begins
I basically had to log out and log back in to get it to work properly
Why do you add more letters that are not written?
What do you mean? Can you provide an example to clarify your question?
For example, the composition of a song would be:
VERSE 1
CHORUS
VERSE 2
CHORUS
END
CHORUS
then suno adds this:
VERSE 1
CHORUS
VERSE 2
CHORUS
VERSE 2
CHORUS
END
CHORUS
adds the same VERSE 2 - CHORUS, I don't know if I explained myself.
Ah, usually the song structure is the issue (Suno likely thinks it was incomplete or did not flow well). See the following: https://discord.com/channels/1069381916492562582/1261146873331318835
I think some people simple decided to manually extend before the repeating verses (though that may need [break] or [instant switch] to avoid the repeats). It has been a while since I looked into that scenario though, opting to regenerate
So I can put [stop repeat] in the song and it will no longer repeat?
One more question, when you use "crop song" when you crop, you will hear the cut, would that be very bad? I want to subscribe for that reason but I'm not sure if this option is good.
Just had Suno steal a track from me while I was listening to it and it was 🔥 wtf??
I'm trying the cover feature, I had an account that when using Cover it would actually try to replicate the melody of the original song. Now I have another Premier account (so far, I'm using the free cover songs) and none of the covers even resemble the original melody. Anyone has any idea how come?
I feel like thumbs up/thumbs down marginally affect new covers and it gets new covers to slightly do more of what I want but I'm so far away from the actual original melody, I don't know what to do
I just tried the old account and the cover feature works correctly, it gives me a cover with the original melody (and some where it just gives me back the original audio with no changes). It's the new account that's borked for some reason (maybe Suno is intentionally giving me bad results until I'm out ot free covers on the new account?)
Just trying cover for the first time - it definitely seems really inconsistent.
Sometimes it does what it should and changes up the genre, sometimes it will just completely make a new song, and then other times it will literally change nothing at all.
When it does work it's really cool but it just seems like RNG on whether or not it will. Thanks for giving 200 free covers otherwise it seems like it'd be a waste of credits rolling and hoping it works properly
Yes that's how it was for me with my old account, 1/10 it would get at least one good cover with the actual melody, I'd give it thumbs up and it seemed it understood what I wanted more of.
However with this new account, I'm 200 covers deep and zero with the original melody (giving thumbs down to all bad ones).
It's very vaguely keeping the cadence of the original vocals but everything else is randomly distant from the original song, different melodies for vocals and instrumental, different instrumentation I just tried my old account and by the 4th attempt it gave me at least 2 that resembled the original melody, both instrumental and vocals
Hm yeah that's odd I have no idea why it would be account based if it is. I'll see if thumbs up/down help train it to get better results on my end.
Honestly a sliding scale would to control how similar or different the cover is would be really nice
It has nothing to do with your account. It feels random because it is random.
Having it create a completely different song with no essence of what it's covering is more than just random though, it's just not working at that point
Yes. Old account gives me a good cover at the 4th attempt or right away. New account I'm at 200 covers and none have the original melody.
Only thing I can assume is it's account related
The main advantage of paying for a subscription (aside from ownership of your songs) is you have a ton of credits to keep iterating until you get what you want. The stronger your structure, the more likely you are to get consistent results, but that's also a matter of iteration and trial and error.
The whole point of Covers is to make a new version. It can and does make similar versions, too. I use it all the time. I just did a whole live album of songs and they are all very similar to the originals, for instance. It just takes effort, time, and if you don't have either of those: luck.
This indicates a problem with your approach. What's your process?
My process has been identical on both and I have gotten good results right away on my old account, none on my new account.
Describe the approach / process.
Upload original audio as wav, paste the lyrics of the original song in the lyrics box, put 1-2 words for style (usually the genre, in this case "ballad").
Do the lyrics have tags?
"[Verse 1] [Chorus]"? Yes
Which is why there should be a sliding scale to control how similar or different the cover is. That's how Udio does it and it's much less of a credit sink because of that user control. I am literally just covering a simple jazzy instrumental song from v3 into something more metal and it just 100% changes it to something unrecognizable most of the time. I'd say 1 in 10 attempts it actually sounds like a real cover.
And you're using the same prompt?
Yeah
If you want, post links to both attempts on different accounts of the same upload, and I'll look at it to see if I have any suggestions.
This song has lyrics so I can't make the links public
I can try some instrumental now though
Oh wait nvm, I finished credits on the old account
Will have to wait to tomorrow to specifically try that
I see. Well, as always my big advice is troubleshooting. Eliminate possible invisible characters by re-typing the lyrics manually or pasting them from a different document. Try with and without tags.
Verify Suno interpreted your lyrics correctly. Try uploading the WAV again. Etc.
Ok so I'll give you some comparisons, hold on
First suggestion is to paste your tagged lyrics into the displayed lyrics for your upload. Not 100% convinced this does anything, because Suno seems to stick to its interpretation, but it's something I do and since I'm having better results than you perhaps it's a good idea.
Old account:
Original wav:
https://suno.com/song/e544797a-08f5-4faa-bf75-c688d1c99265
Cover (1st attempt) - The recurring glitch of Suno spitting back the original song unchanged
https://suno.com/song/0a122733-633c-4ac1-bb32-f1ce9bd21443
Cover (2nd attempt) - Very good cover
https://suno.com/song/e1371954-5f0d-4613-9032-a60ab99da5e5
New account:
Original wav:
https://suno.com/song/4b78f3b1-b88a-4c7c-a94a-8fa6b2b120f7
Cover (1st attempt): - not even close to original melody
https://suno.com/song/67b93b8c-81d3-457a-8b30-f4dc31620b81
Cover (2nd attempt): - not even close to original melody
https://suno.com/song/6528850e-3431-40ca-baab-1de42c36a40e
Cover (3rd attempt): - not even close to original melody
https://suno.com/song/3309a28a-9c2c-45e2-8bd6-131543192346
Cover (4th attempt): - still far away but some vocal cadence sounds a tiny bit like original but melody very far away
https://suno.com/song/0bca03c4-06c8-4c8b-81e3-bd9fe9ac5f26
Cover (5th attempt): - not even close to original melody
https://suno.com/song/9e6ff90d-1a25-4c84-8c91-c4555d8e0ff8
Cover (6th attempt): - not even close to original melody - whole different genre
https://suno.com/song/49ac104b-a1bd-4c8f-b8ff-f09c26f6a545
Cover (7th attempt): - not even close to original melody
https://suno.com/song/b9ee77ba-6f3e-4e02-ab63-bf7c5f0d0db4
Next suggestion: leave genre ("ballad" etc) out of the style prompt. Use words like "remastered" and describe the vocalist, and indicate specific instruments.
All of these simply have "ballad" as style, same lyrics. As you can see, my old account by the 2nd cover already got exactly what I wanted. The new account, I'm at 200+ covers and nothing close to the original. It just keeps doing these covers where it kinda sticks to the genre but it does its own melodies rather than the original melody
I've also tried doing that. I've tried re-uploading the wav 3 times on the new account, and doing covers from there. Nothing seems to change
I will try that
That's actually what I did on this attempt for the new account, you can check the lyrics on the original wav link
The interesting thing is that the new account is not even doing the Cover glitch of spitting back the original song untouched, it just completely sticks to doing melodies off of its memory
If I had to guess, each account gets a seed and this new account just happened to get a very awful seed that doesn't even try to lean on the original song on barely any aspect to do a cover
Now I'm out of the free 200+ covers zone and using my credit to do new covers (lol)
Just another datapoint to contribute on my end - one of the most common occurrences I'm noticing is that the cover will simply just change the key of the song (usually just up or down a half step). Same vocals, same instrumentation, same tempo, but a small key change. This is despite listing completely different styles.
Yes, I vaguely feel like that happens when you thumbs up something. Suno at that points uses that as a "good" cover and repurposes some aspects of it for new covers. However, on this wav upload (this was my 2nd wav upload), I hadn't thumbed up a single one and I was downvoting them all to try to make it understand I wanted the original melody
I don't think the thumbs up or down affects song generations directly. At least, not in the way you mean. It seems like that data is curated and then used to train the model (in some way I don't yet understand).
I'd be fascinated to learn if your account seed theory is accurate. I highly doubt it, though; what purpose would it serve that a uniquely generated random seed for each generation wouldn't? I'm not discounting it entirely, of course. There's still a lot I don't understand about how Suno works.
I had this problem. To 'fix' I used the cut feature to trim one verse and a chorus with no surrounding instrumentation. It then generated unique results. I think it pulls solos/intros etc as part of the song (esp. if melodic) and then it ties its hands.
I can't subscribe on the site
how long does it usually take for an email back?
nvm i realised its the weekend, so it might be slower
nono that wasnt it, i solve my own issue i had a spelling mistake in the email recipient
Dear developers. Please make it possible to choose your voice
Hello, I keep getting this error message : An error occurred.
A system error occurred.
caant create any songs.
Insufficient information. Do not have an acceptable method of payment? Nothing much you can do AFAIK
Card being declined? You may want to contact your credit card support line if you think the problem is with your CC or email billing@suno.com if you think the issue is with the system they are using
Please post that at #💡┃suggestions-web but do be aware that there are other services online that do vocal swapping which you can do. I think the terms included RVC model? "Vocal swapping with my own voice" should be enough to search for
Try changing your model to v2 or v3 then back to v3.5, that solved the issue with a few people
Yea no I'm pretty sure it is the system for billing. I've sent an email just waiting for a reply back but yeah, seems to be a known issue so it's all good
Ah, I see. It currently is the weekend so they likely will not reply for a few business days, depending on the volume of emails they get
as a pro subscriber, how long can the upload be? in the free tier its 60 seconds tops
120s
#🚀┃updates message -- just recently announced, bottom of the post I linked
Thank you!
the video finally regenerated. Seems 24 hours was long enough
Why does Suno ai just refuse to follow instructions
Did the person working at machine learning forget to train ai some music theory
One more question, when you use "crop song" when you crop, you will hear the cut, would that be very bad? I want to subscribe for that reason but I'm not sure if this option is good.
My account can't create stuff again... For weeks, first the same "old" bug where the Create Button didn't work, now this small message box pops up. I can't toggle older songs to public as well. Seems like my account is kinda "locked"? Does this happen to anyone else?
Delete your cookies, clear your cache, try a different browser. You can also try changing the model used in the upper right drop down bar from v2, v3 and back to v3.5. These are all ways people have reported as fixing similar issues.
One more question, when you use "crop song" when you crop, you will hear the cut, would that be very bad? I want to subscribe for that reason but I'm not sure if this option is good.
Crop lets you cut the start and end off a clip, keeping a selected part. It sounds like you might think it works the other way (delete a section, keeping the start and end) - that's not possible at the moment.
Once you have cropped a clip you can then extend it and it behaves just like any other clip.
After trying the steps Zothmire provided, if you’re still not able to use the site, you need to contact support@suno.com.
On a note of personal curiosity, I’d like to know if changing the Suno model from the dropdown in Custom Mode fixes anything
I'll try and give feedback again. Alr tried switching between Libary / Explore and such. But not changing the model. My cache is always cleared when closing the browser - and I tried the latest Chrome and FireFox ones. Let's see if the "Model" switch can do it.
Still the same... I guess it's Support-Time again sigh...
Good morning
Every song I custom write, isn’t being sang, it’s being spoken, as if someone is talking fast. Tried different lyrics with different types of music and it’s the same result.
Why are my generated covers audio volume keep going high and low
Hello. my cover function doesn't function
So last week my distributor declined a song i uploaded, which i had created on suno. I don't hide that i create my music with AI, so it's not like i'm trying to trick anyone. But apparently my distributor has a policy against AI generated content. I'm not sure why it just now became an issue when I already have several AI generated songs distributed through the same distributor. I also don't know how they flagged this song but not the previous ones. Has Suno begun adding some watermark or metadata or something to identify the music as AI-generated?
There's an inaudible watermark. I don't know what this contains or whether the tool to detect this is available to other organisations. Suno songs have certain audio artifacts present as a side effect of the models - there are efforts being made by third parties to detect AI songs based on this.
There's been a lot of noise lately from artists and labels wanting AI-generated songs banned. Could Suno have given in to the pressure and begun assisting distributors and/or platforms in detecting AI-generated songs?
I don't know tbh. I don't expect this is something they'd want to do and I doubt they'd give in to such pr sure easily. Their watermark is meant to be used for protecting against people claiming that their songs were made by a famous singer, rather than detecting whether AI was involved at all.
I also agree i don't think it's in Suno's interest, but the labels have so much power that I wouldn't put it past me. Thanks for your thoughts
Instead of some massive shakeup in the industry, which you're only able to infer from a separate company acting differently, it's more likely that your distributor has always stated they were against AI content and just so happened to clue into the fact that your recent upload used AI.
In my opinion, a distribution partner is exactly that -- a partner -- if they don't want to distribute your AI music, they're within their rights to refuse it. The fact that you've managed to sneak however-many songs past them despite the ToS you agreed to shouldn't mean that you're allowed to continue blatantly misrepresenting your uploads to them as non-AI generated, as they're expecting the uploads to be.
Hi guys! 🇫🇷 👋
I have a question, maybe someone will have the answer...
For a few days now, the cover images,
are cut off at half height of the portrait format.
Not that it affects in any way the functions of Suno...
But still, visually it is very ugly...
There, I wanted to make the remark.
As I wrote previously, I'm not hiding the fact that I've generated the songs with AI. I haven't "snuck" anything nor done anything "blatant". Kindly read properly before you begin berating me; you are being unpleasant.
You're also being unhelpful. I'm asking how my last upload was flagged, and the fact that "they clued in on the fact" as you put it is obvious. I've done nothing different from my other 20 releases which went without a hitch, i'm assuming that plus the fact that i am honest about creating the music with AI in my artist bio, should've clued them in before too, but it's only now that it's become an issue.
Wasn’t meant to berate or insult, apologies if it came across that way.
The details of Sunos audio watermark are unknown, but as it stands, your distributor has directly expressed that they don’t want you putting AI content in their system.
In regards to being actionably helpful, I’d advise that you contact your distribution partner for clarification and ask why your earlier AI content was allowed (or maybe you’re happier without them knowing that content is AI generated, because they might remove it?)
Alternatively, you can look into a distribution partnership that aligns with your values, like one that is supportive of AI content 🙏
You need to reach out to billing@suno.com or support@suno.com to request account deletion.
**Thank you for the info. **I guess that's perfect insofar that I can include a rant about how fucked up it is to not include that 😂 They probably boast with their huge user numbers everywhere 🤮
I love the extend option I use it frequently. I wonder.... how difficult should it be to extend from the front? As an example adding an intro. I have songs that start... suddenly but in a way that makes little sense. And I'm afraid they would change into something less enjoyable should I extend that close to the front.
They're a bunch of low-skill hacks that are scared about being replaced. Problem is, because they had an essential monopoly on music creation, they're being listened to, despite being whiny children about it
Try a Cover instead. It does amazing at giving a good intro when a song lacks one
Can't do covers. Not a subacriber.
Anyone know of Extend issues?
On some songs it seems to only give me a few seconds
Maybe Let the OG song end and hit extend and use some kind of editor?
Seems to be happening on intrumental songs for me.
I read that audio uploads can now be 2 min long. I don’t understand what the benefit of an audio upload is? How is it used to make a suno song better? Thanks!!
i noticed 2 different ways suno acts:
- It counts the entire "original" song you extend (if it is < 3 minutes) and only generates up to 4:00 or 4:30 total time (say, your original song is 2 minutes, and you extend, it generates around 2 minutes)
- or it just takes your input from extend and generates like 4 minutes of new material
Hmm I am getting like 2 seconds of extend and it refunds me.... maybe
but thanks for the response
oh I got those for very "distroted" songs too, where it basically just gave up and couldnt think of anything
Yeah I am out there on Experimental
Sometimes it helps to modify the input in, for example, audacity and upload it again, no idea why
Are you extending from the end of the song or a part that sounds like it is winding down? Suno follows the trajectory the song is already on, if you extend from silence, you'll get silence
Not sure it can keep up with me 🙄
@safe hazel thanks ill give it a try
I am a newbie
try making the extension at a part where the music is still going strong
Cool will do
Wish it gives you the same graph as it does for Cropping
Like continue from here
yeah that would be useful, but the cropping editor seems to be a rather new addition, maybe they will add them to the extend aswell if we make enough noise 🙂
@safe hazel thanks that got it I think
It should give you the stats on a playlist like Run time etc.
An audio upload can set the voice or instruments for an Extension or Cover. It needs to be an audio file of sound that you have a right to use. For instance, I compose my own music, using my own voice or a voice file that I have the right to use, or instruments that I have in my DAW. I upload my composition and Suno will use that to help create a full song.
Now an uploaded audio file can be 2 minutes in length, an improvement on the 1 minute length previously.
can I "redo" a song in the same style of a previous song by "extending" it from the point of 00:00 so that it generates a whole 4 minute song in the same style as the old result? Or is there a better way of doing this?
are there any tags that makes for sound effects in a song
Pls provide the function to "ban" certain words, now my songs are "Wishsper" everywhere, "Tranquil" the moment when unnecessary, "Tangled" with a breath a few times
Thx FreddoBee for a very clear and helpful answer to my question. I really get the possibilities of an audio upload now 🙂 !
Best bet might be to just write your own... Suno's generation can only do so much
I dont know if this question makes sense to people but if u crop a song, does the model forget the parts that were cropped so it increases the chance of getting more of what u like in an extension?
You have to extend songs from a point where the Ai can have a grip of something. Extending from 00:00 has no context. But extending from the end of the song has the whole song to work with.
BUT!
The longer into your EXTENTION starting point, the shorter the output. Find a comfortable starting point for your extensions. Work with the "Extend from: 00:00" point and eventually get the whole song from the CREATE Menu.
You must experiment a whole lot to get something worthwhile.
This is a tricky question. I've results when I requested DJ Scratch in one of my songs. You may have to try a few times.
There is exclude styles to exclude what you don't want in a song. Have you tried that?
YO folks!! for your consideration, https://suno.com/song/ecb67056-ac34-4b54-82a1-eb9d6b6359e8 Snap into me
catchy, synth-pop, ambient electro, bubblegum bass, gentle female vocals, digital song. Listen and make your own with Suno.
The cropped song is now an original song. You should start from there. Are you having problems?
OK!
Suno Showcase is also here
https://discord.com/channels/1069381916492562582/1083129341870878780
Let's try it again 😆
its in there 😄
Extensions work better when you want o extend sections of a song to GET the WHOLE SONG later... So, do it by sections. It works best with the Extend from: 00:00, by selecting your extention point.
Trying to extend the end of a 3 min 40 secs can only give you a few more seconds to round up 4 mins.
But If you are looking for a longer song, that is good too. You can hope for up to 2 mins in ver 3.5. In that case, past your lyrics in the lyrics section after clicking extend. Then chose your EXTEND from time
The crop tool currently works with any audio, including cropped audio. What do you have in mind here?
I may be wrong, but technically, it won't do anything at [00:00]. It will just remain with a blue ring. You have to put at least [00:01] for it to register.
Help me. How to make a narrative without music at the beginning of a song and then an intro and the beginning of the song?
Yes! you have to extend from somewhere..., not from 00:00. That's why I said it has no context to hold on to.
I guess I was "extending" that inference that by having no context to hold on to, it won't even function. Teamwork! Hazzah!
I see. 00:00, nothing to hold on to.
00:01, something to hold on to 😄
does past songs get deleted automatically in suno?
I repost:
Hi guys! 🇫🇷 👋
I have a question, maybe someone will have the answer...
For a few days now, the cover images,
are cut off at half height of the portrait format.
Not that it affects in any way the functions of Suno...
But still, visually it is very ugly...
There, I wanted to make the remark.
My songs were all the correct length for a while as long as I included 'outro' and instrumental. Suddenly, I keep having abrupt endings for no reason... is this suno trying to motivate me to sign up for a paid subscription so that I can drop the ending?
The paid plans and the free plans use the same models. Purchasing a paid sub won't change your output at all, just what features you can use and how many credits you get (and song ownership).
It's just strange how all of a sudden, all my songs have a few seconds extra. Right now I'm cropping them off with an audio editing software but I can't help noticing that the subscription version of suno now allows editing/cropping.
They didn't intentionally make the model worse to force people to crop the songs
There are so many lyrics errors that it's inconvenient to use. Most of the songs I created today have wrong lyrics, so I can't use them. I don't put the lyrics in order and just do whatever I want. It's getting worse. What do I do?
Well, I’ve being a subscriber since April… I mostly cropped some of my songs in Cubase before the crop feature…
Some songs end properly, some are in the wild, Wild West 😁
The crop feature has solved the annoying extras
I still do not know what is the exact issue you are having, I am assuming it is that the AI is repeating verses?
The typical causes of that is poor song structure. See https://discord.com/channels/1069381916492562582/1261146873331318835
Or is it hallucinations (vocals singing words that you did not expect)? Usually Suno tries maoe things flow so, if your lyrics did not flow right, it may modify (e.g. add or skip words) to make it sound better
In these cases, the usual solution is:
- Make your lyrics flow better. Some people generate an instrumental part 1 then extend it with lyrics as they have an idea of how the song will be sung
- Make sure your song rhymes and is sufficient (e.g. similar syllable counts)
- Is the issue happening at the end of the song where everything repeats? Add a few more verses so it tries not to do that
The rhyme of the lyrics is accurate. The problem is that the phenomenon of ruining the flow of the lyrics by going back to the first verse in the middle while singing the second verse after the first verse is over is over. If one or two of the 10 songs were like that until yesterday, most of the songs are like that today. Is Korean a problem? I'm afraid to add lyrics now.
When I look at the songs that caused problems, I often end the first verse and sing the second verse only to return to similar words in the first verse. That is why I repeat the first verse once more. Then the second verse is much longer than the first verse. I think the progress of the song is influenced by words more than by the melody. I ignore the melody and go back when I see similar words. For your information, the lyrics I write are Korean lyrics.
basically using the "crop" functionality on the "extend" when Creating a song, as in, crop the song just temporarily, of course you can just upload the original and crop it with the crop tool and then reupload the original song a 2nd time if you want to use a different crop, but "being able to actively crop the extending song" would make extending so much easier 😉
try to fix the vocals... they get a bit "dusty"
Hi here and there! 😁
I am writing songs in Belarussian language and facing some issues.
In most cases Suno starts singing with a Ukrainian or Russian accent or making stress in words in the wrong place. In addition, Belarusian is missing in the language filter on the main page.
Is there a way to officially add Belarusian language to the application?
Well, I’m going to answer the question myself then.
When you add an image, AI-generated or other, through the song details section,
a setting is available to crop it before validation.
By default, it takes up the entire width of the provided frame.
However, for a portrait, it is preferable to expand the frame vertically until you achieve the desired visual.
For example: For a 1024x1792 portrait, by default, the image will be cropped by about 50%."
same voice
Has the Covers model been updated lately? That's pretty much the only feature I use in Suno.
Just last week, I used to get music productions that matched my instructions. Today, I've flagged pretty much all the productions I've generated as they no longer match my style instructions. Specifically, they don't respect my BPM instructions, they always get off at 3'30'' as if the songs are now clipped at that length whether the song is complete or not.
I used to be very impressed by the quality of the productions but today I feel like you guys have made such a giant step backwards. Can we get the previous model and rules back online, please?
can you make the voices sound clearer? some words are barely understandable, especially in my language
Hi! You can try: Clear voice or Clean voice ...
I'm far from done, but hit this vocal. I wanna use it 
anyone know a good way to switch from a rapped verse to sung chorus?
over 5000 songs feels like now iam prof. Dr. Suno aka Me 😉 Stems not workable for remixers or is it marketing for get more users. Suno s you are fantastic ...but there are very much to do 1. stems min 4 Tracks. Quality useless. 2. Maybe a Mini DAW implementiert. 3. A Label account or premium with releasing otherwise lets talk woth my Dance All Day Kontor Munich FEIYR. OR i open Suno Records and the top get signed. Greetx Frank www.emunmusic.com
the premaster very fine to master exellent
I have real difficulty figuring out how to size imagery so it crops in a predictable manner. The preview box when uploading an image is in one format but the image on site is displayed in another format. This can be a real pain if you have included type that gets cropped away.
It would help if the preview box on upload accurately reflected how the image will display onsite.
I definitely would like the same, one for the #💡┃suggestions-web if you did not mention it there yet
when will you guys create an API so that us devs can auto generate songs?
The long standing response to this question has always been:
We don't have a publicly-available API at this time, but it's on our roadmap.
Let me ask you a question though...
Why should Suno release an API?
People can use the AI just fine as it is now.
How many songs can be on a playlist?
I've got over a hundred in my 'completed' playlist but it's getting harder and harder to navigate... 😅
Really Mine seems to stop playing on 10 songs.
I'm playing around with duets and it seems like you can get a fairly reliable generation with alternating vocals. But you can't really control which vocals sing which verse to a great extent and it's hit or miss if you'll get a verse with the vocals overlapping. Oddly though, it seems that no matter what I do the first verse of the song has female vocals, even if I keep telling it to have male vocals.
hey SUNO Team! I have uploaded my own thumbnails for my songs but, when I download a video, it's not add my artwork - any ideas? @flint haven
you have to regenerate the video ✌️
ok cool - (I am new! 🙂 what do I click to regen? SCRATCH THAT - I see where 🙂 TYSM!
Got ya!
Stop telling it to do anything.
You should start asking it nicely...
On the style section:
clear-male-voice, male-voice
really?
Just keep adding them until you find out.
You have to regenerate the video
Go to your library and navigate on the top right. Also use the search to find your song. Make sure you give your songs useful names
Like I said, I'm trying to do a duet where the male voice is first. It just always comes in second, and the female voice is first. Every time. It's just kinda weird.
Is anyone else having problems with Suno? I'm trying to write a song and it keeps saying: A system Error Occured
That's what I end up doing more and more now. But it's frustrating. Some twenty or thirty songs ago it was easier to just scroll the list. Now you're fighting the auto loader or whatever it is that populates it.
Delete your cookies, clear your cache, try a different browser. You can also try changing the model used in the upper right drop down bar from v2, v3 and back to v3.5. These are all ways people have reported as fixing similar issues.
. . . And refresh browser. Might as well.
Looking like the server is having issues.
Yeah!
Sometimes it is a browser issue. Have you tried other browsers?
Yes. seems about the same more or less.
I have uploaded a thumbnail for a song then hit summit for the close to a week all I get is network error. I have tried restarting my computer and browser, tried different browsers, even different computers on different internet networks. I think the last time it worked was Oct. 2nd or 3rd.
It seems like part of the time when I do a cover, its returning the same as a regular prompted song with no elements of the Uploaded Audio. I hope that can be made to work better, seeing how its in beta and not a permanent roll of the die kind of thing.
I feel like having a variable Denoise option would be what I need but I'm not sure if it works like that, since I'm thinking along the lines of Stable Diffusion.
Someone had mentioned that when they had that issue, they resized the image (via Suno) just a bit and it worked (likely an undeclared resolution limit?).
Try doing that and let me know if it worked or not 🙂
If I want only a song 30 seconds long not any longer is it possible?
Well, 3.5 does not like to [End] but 2.0 and 3.0 can. So you can do a short tag and just [End] it off in one of those two versions
Alternatively, you can just crop/trim 30 sec out of an output using a DAW (or crop feature if you have a sub)
who do you talk to for member support
For official support, please contact:
- For billing-related inquiries: billing@suno.com
- All other inquiries to support@suno.com
If you have a question that you believe any member of the community can help answer, just ask here 🙂
i can't sign in on my phone but i can sign in on my computer.
so i guess i need to email them about that huh
Alright, here's a test question, hopefully I'm not doing it wrong: Can the Premier Plan let you upload songs to Cover that are longer than the 2 minutes you get with Pro (specifically 4:30 in a perfect world), or is it limited until it can hopefully be worked on in the future?
@hollow hare that worked sort of, had to cut off the title and cool border. Might have to see if I can change the aspect for the image.
Thank you for confirming.
Currently, free plans can upload to 1 min and subscribers to 2 min. It is unknown for now if a future update would allow more time or not (reasons not to include increased chances of copyright material being used which is against Suno TOS)
Oops, sorry, misread that.
Replace with cover and likelihood of it being increased may likely happen, albeit probably dependent on the model released though I would not expect it anytime soon
Nah, I think you got it right on the money the first time, & my response is "Crud, didn't even think about that last bit..."
Reason I'm asking is because a friend made lyrics AND a basic instrumental track for my series' theme song, so I'm not sure if I need to somehow Frankenstein together clips somehow to just give it a voice singing the words... though the heavier rock vibe isn't unwelcome either. XD
Hm you can use a DAW to join them together, just ensure you have a nice instrumental portion/transition at the 2 min mark +/- 5 sec on each side to make a smooth transition
Then you can upload each portion (first and second half) and cover it, though consistency may be an issue...
Yeah, I was afraid of that last part. Sorry for the n00b question, first day on the site, but what's a DAW?
No problem, do not be afraid to ask questions 🙂
A DAW (Digital Audio Workstation) is a program that allows audio editing, recording, and other related manipulation
The most famous and simplest is the audio editing software Audacity, an open-source program with lots of tutorials available online
For a list of other DAWs or audio editors, see the "Recording and editing" section at https://en.m.wikipedia.org/wiki/Comparison_of_free_software_for_audio
Gotcha, third party, makes sense. I figured I'd probably have to cross the bridge, just wanted to ask in case I could be as lazy as possible. XD
I have a few more questions, but just for the record, thank you already for being SUCH a fount of knowledge & help & all that! An explanation would've been fine, but you go above & beyond to give links, & that's just awesome for someone who has no idea what they're doing! ^u^
Unfortunately Suno does not yet have a way to merge separate multiple tracks (well, those that were not extended from that is) so that is the only solution at the moment.
The message I was typing was summed up with the 🤷 lol
This is what I had:
I personally cannot say, I only ever used Audacity (I'm basic like that). Hopefully other people will be able to answer that question - you may want to consider asking that in #offtopic as well
No problem; I appreciate the input either way. I've recently (as in, an hour ago) discovered I have access to a crapload of royalty-free samples through my distributor and I'm carefully considering dipping a toe into creating short clips that I can then upload to Suno and use as the basis for new songs. It seems awfully complicated, though. We'll see whether or not I stick with it.
Just be careful. Mind you, royalty-free does not mean you are given the full rights/permission to use. Read what the license(s) and the Service TOS says and keep a record (e.g. PDF printout) of those documents in case you ever need to use them.
Yes, I understand. In this case I used "royalty-free" as a shortcut; I have full licensed rights to use these samples as part of a subscription.
Understandable, I used to do the same (and kept track of what I used in a spreadsheet)
Used to? Meaning, it didn't end up working very well or you just moved on to other methods?
I don't really use my computer too much so I haven't done anything non-mobile related for months
Heck, I even used ShotCut for a while on my phone for making a music video. A much more painful experience than what it would have been like to get my lazy butt on a computer and use the softwares on there (albeit my PC is so slow these days so it may have turned out to been faster but certainly less convenient...)
hey guys can be posible in the future to have this way of making lyrics for suno songs check this https://www.youtube.com/watch?v=Dp85NmrfuAI from https://videobolt.net/author/harchenko/typography-lyrics-2?themeId=59948996&startTime=2.188415
Videobolt Lyric Videos (20% Off Discount): https://bit.ly/lyric-videos-sd
🎶 DistroKid – My Favorite Music Distribution Service: https://bit.ly/distrokid-sd
*I have the Musician Plus plan
This is not a sponsored video, however, I may receive a small commission if you use these links :)
►Let's Connect
https://seandivine.com
Instagram: https://i...
Set the stage for your song with our dynamic typographic clip, where words dance in tune to the beat. The modern, procedural frame changes and a striking torn paper effect add to the edgy aesthetic, inviting your fans to sing along. Customize your creation by choosing from a variety of fonts and colors, and launch a ready-to-publish video that v...
any suggestions for how to avoid having the creepy robot voice talking at the end of my song? I've had a few songs that were great .. and then get totally ruined at the end by having a robot recite the lyrics in a scary voice 😄
Hard to avoid but possible to deal with:
- if you're subscribed you can use "crop" to cut the end off
- otherwise you can try extending from just before the bad part, switch to V3 model and have the style as "end", turn off instrumental mode and put [end] as the lyrics
Can I upload songs that I made before my sub now without any problems?
Upload to Suno? Or upload Suno songs somewhere else?
Like Spotify
Ah. The songs would need to have been made during a subscription. You can e-mail support@suno.com to ask for permission otherwise.
And if i cover it after my Sub?
Or re-use the promp
Reusing the prompt would be fine. Covering I would expect is a derivative of the original so would be something else to ask support for clarification on.
Okay, thanks.
That means I can upload all the songs I make during my sub, right?
Yep. Anything made during a sub is yours to do with as you please
Can i use the Cover Images too?
Is it possible to generate songs always with the same voice?
I know you can get stems and the vocal, but the quality is shit. Is there a way to get it super good? Another ai to improve the extracted vocal?
@tame forge yes images too
I use bandlab to get the stems and then master the song myself
Also hey harber
Here's a technical thing I'd like to see. A way to edit a songs inital lyrics to fix Grammer and missing words. Whilst keeping the same music. It does mean the song would have to re-edit the rythms maybe... but... in my case it's missing one word that takes a single beat that wouldn't change thr song's tempo
Ex: in a song I wrote "we knew he would Dedicated and kind." Missing word..... be.
Listening to the song idk 16 times I can here the missing word... and also know it wouldn't affect the song overall. Would love to just go into the existing song add "be" and keep everything as it is... because it sounds great and that one missing word makes the song unreleasable to the public in my own opinion.
The rumor is that v4 will have inpainting, which would be able to correct that error
Just to take note, Suno doesn't know what a "glass harmonium" is. I keep getting an accordian.
To be fair, I also don't know what a glass harmonium is.
Hello, I created a song and was listening to it. While replaying it an "error occurred generating clip" and the song disappeared. Is that song recoverable?
anybody dealing with suno just not singing the words you put in? i've burnt well over 6000 credits just trying to get the chorus right and suno just keeps being drunk
No, sorry 😦
Ouch, okie
Is there any way to prevent that from happening?
Nothing that I'm aware of. 🙏
You could reach out to support@suno.com for more information if you find that it's happening regularly.
Does anyone know if you can create songs in bulk?
Whenever you create a song, the lyrics & style prompt fields stay filled -- free plans can make 2 generations at a time (4 songs), and paid plans can make 10 generations at a time (20 songs)
Short answer -- click "create" multiple times.
It's that glass bowl contraption invented by Ben Franklin. You wet your fingers and play the glass bowls like you would a wine glass.
That's my routine. Click twenty times. Listen. Delete 35 gens. Tweak the best of the 5. 😊
It thinks a hurdy-gurdy is a plucked guitar apparently too. 🤣
What a stupid robot ... 😓 I'm gonna go use google for an entirely unrelated reason now.... 😓
ha I just had my line deleted cos I called myself by a slightly rude name!
just me learning the rules of the room
not to be spammy but
HEY ALL quick question is it possible to use my real voice for these ? i mean i did it already and it kinda worked but kinda not. i am on the free version just wanted to know if the paid version offers this service ? or any way possible of it
so I can create a script to generate millions of songs everyday without needing to script it on my web browser. Duhhh
Song generation api, and all the ways to manipulate it to generate music and other capabilities from a program.
What possible purpose would you need to script a million song generations in a day?
Since no one responded yet, I'll say Suno doesn't emulate a person's voice exactly. It purposely tries to recognize voice from other noise/sounds to likely pull some nuances like baritone, gravelly, tinny, soft, and so on and give back something that could sound like a distant aunt or uncle of yours. With instrumentals though, you can hear exact copies of passages thrown into your covers, here and there, or an assemblance of tempo, style and sounds which compliment the upload or recording.
Certainly, if all you do is hum or do-da a melody, voice only, you can possible get back a vocal that somewhat complimentary to yours, but I have gotten anything exact yet in my trials
Heh, what timing dasvon 🙂
You can with Audio Upload (and a very good dose of good luck) but note that the song will not be able to go public on Suno's site but you can still spread that link.
(Suno deactivates the option of the song going public if the uploaded audio is detected to have vocal-like sound for liability reasons, e.g. in case the person does not have the right to use the voice being used).
There is no difference between the free and paid users in terms of the model, they use the same thing. The differences lay in a sub-only feature like Crop, longer Cover time (2 min for subs, 1 min for free), and of course, the amount of credit and commercial use.
Alternatively, I would recommend looking up vocal swapping/cloning services if you want to use your own voice, especially into the RVC model (RVC = Retrieval-based Voice Conversions).
I believe there were methods to do it yourself for free/very cheap in conjunction with Google Colab?
It has been a while since I looked onto this so I am not sure what services are the best but you can look that up yourself, esp. with AI reviews on YouTube. Possible useful links:
#💡┃suggestions-web message
#🗣️┃chill-chat message
Yes, I forgot the "good luck factor". It's what helps monkeys eventually type a novel. But I'll take it.
Personally: I would like to have an easier way to automate song generation, sync with my own storage and have cool things like my own radio while coding. I understand the challenges that comes with an API, costs and possibility of abuse and stuff. I just would be happy with any form of API, even if limited or with limited copyrights. It would also be fun to play with other types of integrations. Speaking for me, I don't think suno SHOULD release one, it is more like a fun nice to have feature for some folks. Still a great AI anyway.
That's my whole point. Abuse is already happening with the copycats that exist. So given that we've already seen what could happen, why should Suno make it easier for them?
Why not just use the radio station feature? That should get you a series of song similar to the one you start with, right?
why would you want to auto-generate songs? Sounds like a recipe for bland nothingness if you ask me
swirling neon city lights as far as the eye can see.... eurgh
I did not know that, that is too bad.
The radio does not give me much interesting results, and also little control for what I would like to do, with effects, fades and other stuff to make a seamless thing.
Well, if you do not know the language, or invent ones, then the weirdness actually is not weird 😆 . Lyrics are more a tool for me creating interesting sounding stuff, I rarely use english.
Im using the bark google colab to generate voices, is there a way I can have it just auto generate random talking without having to type out a sentence for the model to speak?
You will have to set up your own pipeline for that. E.g. get conversations from an LLM (like Llama) and pass it along to Bark to generate.
I have not worked with either software so I cannot help you with that directly but hopefully that gives you an idea of one method to go about it
Yeah, I couldn't do that. I prefer to actually write my own lyrics anymore. Just flows and sounds so much better
hello everyone ... can someone explain the difference between the new covers and the old way using suno
and when doing covers for certain songs it duplicates the exact same song even with some prompting
it's still in beta, that's the main reason. Just down vote the duplicates.
good to know. Do you know the difference between the covers and the older way to do the songs if you upload a track. I noticed that the previous way did better with isolated vocals and covers works better with full tracks.
I don't think there's really a correlation, They are two separate things.
Is it possible to use a song as an mp3 in a edited video without watermark as an subscriber
i cant really find any notes as to what the covers does vs the original. any idea where i can find that
Yup -- the video watermark is only included when suno makes the video for you -- you're granted general commercial usage for songs made on paid plans, so you can use the songs in videos without giving attribution to Suno
#📢┃announcements message
There might be something on https://help.suno.com, I haven't looked but it's still in beta -- as Ayxo noted, they admit that it returns the original sometimes, and ask for 👍/👎 ratings to help improve the feature for when it gets released in earnest.
@charred merlin@outer token thanks. Yes i was always using either an isolated vocal before covers and now im using the full track so im pretty much getting the same results of both, hence why it was confusing to me.
i have tracks that i own from demo stages to final so ive been testing out on some unfinished demos
Thank you 👊
Any reason the ai generated one song but charged 10 credits?
And don’t refund from from a failed generation like it usually does
So there is no way to let the bark model just ‘hallucinate’ speech without giving it specific text?
Any tips for generating one instrument? Like if I wanted to make a loop able riff to add to a song I've made?
Any good words for that?
It's becoming irritating that the "cover" mode doesn't really follow instruction.
I try to keep the female voice of my original song and add -male vocals -male voice in the exclude section, but it replace the voice with a male one more often than before.
Also, if I want a "live" recording, it works good, most of the time, but sometimes add a lot of "blah blah" talk when it just plain replace the song with something I never heard before.
Report feedback to Suno. It is beta for a reason.
I do wonder if they'd be able to let you produce a song with separated stems from the get go.
Literally be able to denote the guitar before generation so it has a separate stem.
1- I do /bug it
2- It is beta that costs credits (that wastes credits)
3- Is it forbidden to speak here ?
One mustn't read a message and assume ridicule.
I just told you to report it to suno. Then I mentioend it being beta.
At no point did I throw any negativity.
Did I throw negativity ? I just answered.
Hi, when you download the video it comes with the “prompts” used creating the song. Is there anyway to get the “Displayed Lyric” only instead?
You'd likely need to contact billing@suno.com so they could look into the matter for you.
Regenerate the video
It can, but not immediately. Sometimes it happens sooner, sometimes it happens later.
I've invited my friend to suno. She registered, but I've not got the credits
The invitation program stipulates that the invited person must make 10 songs as well. (I have yet to receive clarification if this means that the person needs to make 5 or 10 generations)
which genres should i use to do a song that sounds like charli xcx "brat"
Here are some subgenres of the song from RYM you can try playing with: https://rateyourmusic.com/release/album/charli-xcx/brat/
You may want to add some synth instruments in the style as well
Only after she will generate 10 songs we both will receive our credits?
I still did not get clarification about that yet but I would assume so as the worst case scenario based on the wording given
Likely it was still loading the music, it happens time to time
Hmm might as well write up a quick FAQ for the audio for a future copy-paste for myself
No, it affected all music ont he site and I was not generating. Other site's sounds were working. Went for about 30 minutes of trouble shooting prior to commenting, but it resolved very quickly after with zero efforts adminstered.
Ah, strange
hi can someone explain to me why is my instrumental music only 1 min long and when I try to extend it adds only 5 sec of nothing?
Thanks!
Are you extending from the end, where there is silence? If you extend from silence or when the music is fading out you will get silence since Suno follows the trajectory the song is already on. Try extending from a point where the music is still going strong
I tried it added only 16 sec of the song, I don't undestand this
I wrote in prompt make 4 min long
You can't prompt for a length of time. Are you suing instrumental mode? Try using lyric mode and put in meta tags and a lot of characters that cant be sung (like * * * * *//////////// etc)
it produced werid noises
So I've got a song I really like but it was generated with 3 stanzas per verse instead of 4. I've rewritten it, and I'm just trying to inject the new stanzas via extending the song. But when I do that, Suno ends up completely changing the tune of the song and ruining the feel it's already got. Any suggestions?
Omgomgomgomg
Is there anyway to do API access to automate song creations? I have a niche need and API access would be a game changer.
I want to do everything with a python script rather than the browser interface.
You could write a crawler with python that log in automaticaly and does all the things.
Until we have to do "Are you a robot" shit XD
So it won't work?
And lo and behold Suno releases a feature that might do it...
It was just because you asked so nicely 😱
Steam
I think so! ^_^
And the inpainting extravaganza begins, this is going to be incredibly helpful to clean up those areas where the system was just a little too AI’ey
You also can 3.5’ify small sections of your song instead of having to completely replace it, which will be awesome!
Well, I seem to have broken it already... it basically didn't work, didn't add anything new outside of garbling a bunch of my audio.
that's why there's so many free credits
no more trying to get alternate uses of Covers...
I keep trying it, and it just... doesn't work.
Changing a word here or there is OK, but if it's not the same length as the original, it just barfs.
How do I produce thunder sound?
New feature is cool but a few things I noticed
It has a lyric cap of 3000 characters which can be a pain. Also you can't use this on any parts as it'll make the replaced section and the part the full song. You can't then bring it back into the rest of the song
lets see if the replace section helps with the fact the ai repeats lyrics that it shouldn't 👀
Oh, yeah, it does that all the time, rather annoying.
I had another song a while back where it randomly added a spoken interlude, but half of it was just scrambled gibberish. Gotta cut that out too.
One other thing I'd really like to see, but I just don't know how this would work from a copyright perspective... often I'll want to edit songs I've made (and also have a couple songs made using another tool before I discovered Suno), which I'd like to be able to bring back into my Suno library. But certainly they are more than 2 minutes long.
ok idk why but this inline regeneration is kinda awful
I just want to change a few lyric pronunciations
but it basically generates an awful new section of the song that doesn't flow at all
oh, weird, it cut off the intro to my song too after selecting a replacement section which wasn't anywhere near the intro
how to contact customer support?
yeah... I'm getting all sorts of audio glitches
any tips for getting the same rhythm and flow?
all my retake attempts really change it up too much
Not sure. I can't even get it to that point. ^_^;
I'm just trying to remove a scream in the opening and it does remove the scream but keeps putting lyrics that I did not type in
I'm trying on the original, I'm trying on the Cover.... nothing's working
Hi, I'm trying to extend an instrumental and it's doenst extend anything
come the hell on, Suno... this shouldn't be so goddamned hard
Just three tiny edits. Why the hell can't you get it right?!
I mean, even a goddamned COVER got it more right than Replace Section has
so, I just savaged the hell out of the feature in the feedback form
and why have I lost 200 credits?
I had 475, now I'm down to 275
I've done nothing but try Replacing Section
goddamned Suno monkeypaw... we get inpainting, but it's god-awful
I just lost 1000 credits trying the replace section feature?? Didn't the announcement say you could do 1000 sections free?
same here, but only 200
It actually caught the same rhythm as before, but decided to just plop Verse 2's lyrics instead
this needed to be in beta for longer, I think
it worked for me on one song. just needed one word replaced and it did it and didn't change anything else.... so it's like 1 for 48 so far
I can't get it to do shit
this is the exact feature I was looking for... but it sucks
me too... but other features have had some issues in the beginning but improved. hopefully this will get addressed asap
Just checking — you’ve watched Franklin’s tutorial video from Loom showing how to get the best results?
I went back and got the exact lyrics prompt from when I first generated it
I agree it’s not at the “perfect inpainting” phase you dreamt of, but just wanting to see if you’ve checked what the staff have to say on the current release state
my issue is it absolutely refuses to catch the same rhythm of the lyrics
are you changing entire lines or single words?
single words
hmmm... well, like I said, it worked once for me... I'll see if I can get it to do it again