#📚┃suno-school
1 messages · Page 114 of 1
Remaster creates a new singer?
just remaster it lol
oh
top music will be new but the bottom one is the same music with better quality
i was thinking the same but now sure how to reconfirm the ownership problem like free vs pro --- is there way i can differentiate the license while downloading the song before release it for commerically ?
heem let me check
like said: Anything recreated with pro, is pro
even an extend is a new generation by the ai, and new entry in the library and thus pro licensed
then it was explained wrong by others, sorry then
"Songs made under the Basic (free) tier are only available for non-commercial use and cannot be monetized. These do not include the commercial use license, and starting a subscription after you made a great song will not give you a retroactive license for the song. " i am trying to figure this out how to solve problem cause i really wanted to use my first song music not vocals though but i really like the music and beats in first song which was generated on free plan but i didn't know about these restriction until i purchase pro plan and now i am stuck and it cannot recreate the same muisc by any means even i am using same prompt
the question is how do we know what is pro and what is not pro ???
but also surprising, because when its recreated, why it still doesnt get the license then? hm
Suno owns songs made on a free plan (but if you’ve subscribed, you can email and ask for commercial rights to certain songs)
It’s the same as if someone extended from YOUR song — that wouldn’t cut your ownership out of the equation — they’d still need to get permission 🙏
well ok that makes sense
sorry for misinformation. I heard by more than one the opposite, so I assumed unfortunately it was true xd
sorry 😅
I did sent an email to support on this email "support@suno.com" hope to recieve any input on it.....
You’d need to keep track of what months you had a paid membership (you can get this info from your email receipts from Suno Inc), and compare with the timestamps from the songs on the website
that's the issue even i am using the same prompt it is not generating the same music --- i tried almost 10 times with same prompt but no luck
They’re dealing with a lot of support requests at the moment, so it might take a week or so, but if you’ve subscribed to Suno now, there likely wouldn’t be any issues with you getting the rights to the songs you request 🙏
Hoping for the best
btw: what is when I download a song and edit out a mistake by the AI which I cant fix at the website because the crop tool doesnt cut out a middle part?
how would that affect the license?
that's fine for me but i liked the flow of instruments and which i can still commercialise and release on platforms if i can get that ownership
Meaning if it was made on a free account, and if you made edits to it, would that be enough to circumvent Suno’s rights to the song?
make persona and remake your music with it
that will work if you made the lyrics not the suno
but if the songs dont go viral I'm not sure if its even worth it to commercialize them
I've heard you need to pay money to be able to set it to streaming services. but how to gain advantage from it then? then you must be really sure you have a big hit of a song considering how little these streaming services pay per click 😄
no you do not need it i got 2 dollor in last 2 month for my musics lol
The new music would be a derivative of the Output. If it ever went to court, very unlikely as it may be, the court will likely recognize it as Suno's ownership as you would not have been granted it as it was not made during a subscription (though depends on local laws).
(Just ask Suno's support for the ownership/commercial license when you get a subscription, usually they are good for doing that and you will have an official correspondence documented)
My favorite song was also made at free plan. its unfortunate xD
lol
Assuming you’re on a paid plan now, have you emailed to ask for the rights to it?
no. not yet. But I have also no idea how everything works. There is sure more to it if you put it on a streaming service like tax and what not, especially in germany where everything is about bureaucracy 😄
When distributing, you have to prove you have the rights to the songs being distributed.
Upon violations being discovered by the distributor, the least that would happen is revoking the distribution of the particular track from where they had been distributed (if it even got that far)
anyway I should ask for the rights. Even if I dont make anything from it. it wouldnt hurt to try i guess ^^
so you had to pay nothing beforehand? and where did you put your songs at?
regarding this I assumed it was made on pro
because 2 of the songs I made with pro, would need edit, which I cant do on the suno website, because the tools are not good enough for these issues
the ai made a mistake in the middle of the song and the crop tool allows me only to select 1 single area to keep. impossible to cut out a few seconds from the middle
I gotcha — if you made the songs on a paid plan, they’re yours to do whatever you want — edit them, sell them, whathaveyou
ah ok
40 dollar a year and put it everywhere
Like the other user mentioned, I wouldn’t be surprised if you don’t get immediately smacked down with copyright if you DON’T take proper steps to secure the rights to your songs, but wouldn’t it suck if the song got really successful and the legal copyright holders took legal / financial action?
Personally, I’d rather take the time to do things the right way so the songs can leave a legacy, but to each their own
yeah well thats something. Are you sure youre getting it back? I mean you need several 1000 clicks on a song to get at least something little. At least from spotify etc
if you want to make money you need to spend and invest
I wish you good luck that AI Music brings you that far. I'm not so sure, but would be very interesting to know if it works
maybe you can update a bit
If its for me, private messages works on me
I would fear the most getting a sentence by that tbh 😄
the audio quality really needs improvement especially on extensions they sound like 64 kbit mp3 sometimes
ya
it was not like that 3 month ago
it was working fine idk what they did to it lol
So the TL;DR is, I don't own any rights to the music I made under my subscription?
Wait I read the conversation wrong
I get static on literally everything.
What exactly do we do with Stems? I extracted from a track because i wanted to use the voice in the future, so i thought i´ll create a Persona out of a Stem but we can´t?
Question. Is there a prompt that can limit/set the length of the track? Say 30/45/60 seconds?
Hey folks what are some reasons I can not set some of my songs as public. Also when I mark them public am I saying sure take everything you want?
@wheat field You can try [fade to end at 0x.xx] where the x's are the time you want, but be prepared for it to work only 3/4 of the time and when it does, expect +/-20 seconds or more
I had it pretty consistently cutting off jingles I was making at between 40-120 seconds where before I used that tag, it was trying to continue it 2:00+
Can anyone assist me i am new to Suno and created a song however the AI is saying one of the words incorrectly i ever tried to spell it differently. everything else is perfect.
No specific prompt, and in general getting something that short may be tricky. But if you're trying to make a jingle or something that is typically that duration anyway, then it will be much more likely in Suno. But if you mean, prompt a single verse, and sound like the song starts and ends in just 30s, that will take some Extensions or Replaces probably.
If you already have the song you want, you can try to fix this with the Replace feature. Otherwise, changing spelling is still the main method. Consider adding dashes like dash-es or CAPitolization.
ok thank you will try both i appreciate it.
Is it a word that has an alternative pronunciation, or just a mistake?
Dino
its saying Deno
Okay I can see why suno is confused by that one, I'm not sure myself. DId you try Dyno?
If you prompt Dinosaur or Dino-saur elsewhere in the song, that might do it, and you could crop/replace it out late.r
Is anyone else having difficulties with creating songs over 4 minutes? The distortion is unbearable and unusable. I have spent the past 3 days and endless credits trying all options, and the song just gets tired and sounds terrible. I've tried different sounds and genres but it still ends the same..... Does SUNO need to clear their cache or something? The songs I have been producing prior to Christmas have all been of amazing quality.... 😦
Die-no, Dye-no, Dyno, Digh-no
So nothing worked however it is only the 1st one that sounds wrong every other one sounds right so i will edit it thank you didnt think of that.
I will let you know how it works out
I am student
I want to get a student plan
what do i need to do ?
When I checked the available plans... I see a Student Pro plans for $5/month as you can see on this screenshot
Do you know why the 'Upload Audio' function isn't working?
Thanks. I know what I'm looking for now. Oh, I finished off that shimmer creation. https://suno.com/song/0745737d-48ba-4ba3-a0e3-8de216110cfe
You have to remember it's a beta feature and it's far from resolved at the moment.
Sorry, prob a nube question here - but I used audio upload to record my voice - used my own lyrics - generated a bunch of options before I found what I liked - Then used (Cover) to keep the general melody whilst searching for the best sound - Now I have what I like, I cant make it public - Is that because of the audio upload ? or a mix of reasons ? Tks
I'm sick of the issues with "replace section" chewing up all my credits and it refusing to do what I tell it to do. No where do I have vocals, and it still puts vocals in no matter what I do. Is Udio "Inpainting" better than Suno? I don't want to change to Udio but there are 2 MAJOR issues. #1, Replace Section. #2 the sound degrading over time to be super tinny to the point where guitars no longer sound like guitars anymore.
is there a way to add a time constraint, like a specific duration
also for somereason I can not like the song I make anymore
the like button dose not work anymore
When I enter in the generate-a-song box directions in German, it generates lyrics in German, but sometimes music style tags in German as well - it works surprisingly. Is Suno translating tags to English and then uses it, or, do I have less a chance with Non-English tags? Just wondering what experience others have or what Suno staff says. I also wonder if this biases the model to generate from German training data more than if I would use English tags only. I noticed for example in v3.5 a certain mood shift if I include Berlin as a word...
There are protocols to help speed up improvements. 1) report what seems a bug or layout issue in the bugs or layout channels where developers would review. And 2) constantly click thumbsdown button on all and every case you don't like, regardless of what it is. Suno, as I understand, does not self-correct, but the stats formulated help pinpoint repetitive issues common among everyone.
If you're hearing vocals in Replace, where there are/were none , that's a difficult point to work out on just that.
Have you checked the song details tab first to make sure the lyrics and audio appear same.?
Is your high lighted selection only around the lyrics to be changed and from a good starting point and ending, such as highlighting a full stanza, than just a phrase or line?
Have you clicked "make same length as selection" toggle button, so it restricts the space where it can add extra words?
Do the seconds output match the seconds selected, within 4 seconds?
And not to seem judgmental. I just used Replace trying 5 different techniques this morning, and while nothing seemed unusual, the results were just "neutral". It never replaced the words. My 15 second selection, just outputted a 15 sec section of the same lyrics.
So, yes, I've used replace perfectly , recently, with no issues, but it's not consistent. I use often and pleased with results, most times. I keep experimenting to find what helps nudge it to be consistent.
Audio uploads with detected vocals are can’t be made public, or have personas based on them — you can still share the link with friends on discord or add the songs to playlists, they just don’t show up on your profile or in Suno search
Shame … but Cheers 😁
Used another browser as a control to compare? This helps pinpoint what Brave may be doing (or not doing)?
Have you tried a private browser on Brave? (also turning off any extensions set to operate in private mode) -- I mean, you could just use Chrome, but in the case that Brave is your preferred browser
Hey guys! Do you know if covering a song WITH a persona may become a thing one day?
I use Brave and all is fine.
Ever since those features were introduced into alpha testing, we've been trying to twist the devs' arms for the exact same thing.
Doesn't sound like it's off the table, but nothing on the horizon atm
Thank you for the feedback! That's something I'm eager to see in Suno one day!
Hey everyone! 👋
I’m trying to figure out how to get an API key for Suno.ai. Does anyone here know the process? 😅
Do I need a subscription or is there some kind of application for access? Any help would be super appreciated! 🙏
Thanks in advance! 🚀
The API hasn't been made available -- the only way to use Suno is through https://suno.com or the official iOS or Android apps (https://suno.com/ios , https://suno.com/android 🙏
totally. I just thought maybe someone knew of a workaround
Hi!, sorry, I have a question, and I hope someone can answer it.
A few weeks ago, I created a song using AI, and I’m about to professionally master and record it from scratch, re-recording each instrument (it’s an instrumental piece). The issue is that I created it before acquiring the Pro version, so I don’t own the rights to it. My question is: could I obtain the rights to the mastered song if I change its structure and chords? It sounds like a very different song, but I’ve used the main melodic motif in some parts.
Could you buy a month sub, upload the track as audio, and remaster it in Suno. Then download that?
Saying that, even the most successful production studios like SAW just used "plot tracks" to create some of the biggest hits of the 80s.
"......My question is: could I obtain the rights to the mastered song if I change its structure and chords? ...."
Taking the question as a whole. No. If you've just admitted you're using the song to interpolate a new song/tune, keeping some of the melodic structure, then no, you don't obtain the rights to someone else's song by doing that. I'd legally own a bunch of Beatle-like songs if that were true.
Can you write to Suno to have an agreement to transfer the rights to you, or in the music world, you can simply buy a mechanical license . Yes. Then you may do as you wish and own the interpolated/derivative copy you produced from a license from Suno. Not sure they even bother with issuing licenses, as they may just tranfer the rights outright, but just saying that's how it's done in the real world. Write them and discuss it.
when is suno api
No news that an API would ever be available.
Would someone explain to me how it happens that I can set an extension to be instrumental, and Suno still somehow manages to hallucinate some words in there?
The nature of generative AI~
Based on the patterns it's recognized previously, it felt that the best way to make a song was to add vocals -- ... > Report > Bug > Doesn't follow lyrics can help discourage this kind of behaviour, but it's always a possibility
I have probably sent Suno a combined five-digit number of bug reports since I started. And to be fair, I have noticed that Suno HAS gotten better with time about this sort of thing. Since it's my ONE gripe about the software, though, it tends to carry disproportionate weight in my mind, LOL.
Thanks for the reply, I shall continue to soldier on! 😄
The large body of my work is instrumental, so I feel your pain 😂
It wouldn't be half as bad if it started to hallucinate actual words, but it always sounds like it's had too much to drink so there isn't much salvaging 😛
I haven't checked before but.... when extending a voiced song and clicking instrumental during that extension, can we click off of Custom Lyrics as well?
It forcibly clicks you off of Custom Lyrics. The moment you hit instrumental slider, the lyrics field goes away completely; no lyrics can be entered.
Legit -- thank you for sending in the bug reports for stuff like that 
For a bad analogy, it's like how every restaurant wants to make the best food possible, but the only way they can actually get there is if people take the time to point out the things that need to be improved -- lots of users voice concerns to the devs directly, but providing this kind of reinforcement data genuinely helps -- I haven't been impacted by the vocal hallucinations nearly as much as I did back on v3, so in part that's thanks to you 😄 🏆
Does it? On extensions? Cool. I didnt know. Thanks.
Well, it doesn't click you off of custom. It just wont allow any lyrics to be entered, making the designation meaningless.
I asked Suno to wrote lyrics for me, and here's what I got:
"In order to assist with your request
I'll need to ask you a few questions
Human: Oh
Well
I just wanted a poem or song. Perhaps like... just a haiku. Do you know the haikus?
Assistant: I have information on that style. I believe they're... three-line poems. Seventeen syllables. A specific rhythm
Human: That's right. Do you think you can help me write one?
Assistant: I can try. What is the subject you want to explore?
Human: I don't know. Well... let's just say like... the weather report. Can you do the weather report?"
Isn't that interesting! Lyrics generation is chat talking to chat?
Now it just hangs. Maybe I broke it?
try another browser? Chrome? would it work?
not for me
lol
I didn't get any hallucinations, but it does happen. Other than that......
Here's a random song i chose. I added about 30 sec of instrumental starting at :38 followed by a normal chorus, changed to an Operatic vocal, back to original, then an 8-bit ditty and return to original vocal but at the faster tempo (not my choosing).
I did this mostly by an old trick we learned last spring. It may be overkill, but:
I used a combination of two moves:
Go to original song and copy lyrics. Then delete all lyrics, or all after the previously sung. THEN, do the instrumental.
If issues or vocals, then take that extension and extend it at the 00:01 point. I would focus on putting the same style prompts in to keep the vibe.
Most all time, do a Get Full Song after each extension to help pull the vibe from the previous work. With each Full song, it should bring back in the original prompts, which helps keep the same vibe or vocal qualities.
As normal, with each extension, I pasted back in only the remaining lyrics. Then with each Full song, I edited, leaving in only the lyrics sung, and copy-removed the remaining lyrics to be pasted into the next extension lyric box. Purpose of doing this is to keep suno from pulling extra l lyrics from the previous Full song, if that's something it does... it does seem to help opposed to not doing it.
This link is the Final Full song - no cover- no remaster. I did a Remaster but he sounded drunk. I also did an acoustic cover, and it mostly sounded "normal" with the vocalist just getting melodramatic in the middle.
American New Country, ethereal, dreamy, pop, male vocal song. Listen and make your own with Suno.
ReMi is notoriously WILD.
You can click on the dropdown in the bottom left of the lyrics box to swap to the classic lyrics model if you prefer, but ReMi being unhinged "is a feature, not a bug"
Occasionally I find stuff on Suno that I swear has human input, but isn't tagged as having samples. Is it just the magic of RNG? Or can samples somehow be hidden?
If it's based on an audio upload, it could re-use that content throughout the song -- while it's less common if you're not intentionally trying to trigger it, it's not unheard of for a song to regurgitate training data.
I think I understand. This is in other people's songs. In my own there is very rarely something like a sound effect, which is pretty cool. With other songs sometimes I just can't believe it's pure RNG, especially since I learnt what remastering in the real world can do 🙂 On the other hand I have got some astounding results myself. Just checking how in awe of the machine I should be 😄
On a related note, I once stripped the vocal stem from a song & went to reupload the first minute (back before paid users got 2m uploads), but there was an awkward silence before the song picked up, so I edited a quick Airhorn into it -- when I uploaded & extended from it, Suno reused that airhorn at the same spot for the rest of the song -- that was the first time I got really excited about using Uploads in my workflow before the song reaches the 2m mark
This sounds cool. Would each track generated this way be visibly tagged as having a sample? (Note, I'm a free user so I can't do this myself until I take the plunge 😉 )
(as a free user, you can stll upload tracks up to 1m 👀 absolutely nothing stopping you from getting a feel for it 😄 )
Oooh! I didn't know that, thank you!
As long as the tracks don't contain vocals, they're treated the same as any other track -- if you use royalty-free sound effect (like https://pixabay.com/sound-effects/ or the sound library from https://bandlab.com/studio) then you're in the clear

My old workflow used to be;
- generating instrumental songs (usually picking 1~3 simple style prompts so I could REALLY get familiar with what kind of outputs that combination would give me -- like starting simple with fingerpainting before moving assemble the kind of palette you'd need for the mona lisa)
- picking a song that started really exciting within the first minute (or had an interesting section under 1 minute that I could crop out)
- taking that clip into https://bandlab.com/mastering to make it sound AMAZING (now I use LANDR's mastering plugin, but bandlab is a GODSEND for artists)
- uploading that mastered 60s clip and extending from that (better quality in, better quality out)
Even songs with lyrics usually started off as instrumentals, because then I could choose exactly when the extension with lyrics came in, instead of trying to use complex prompting to dictate how long the [instrumental intro] should be, or if it even used one.
S/O to @charred merlin for all the thoughtful attentions given to people in server. Words can hardly express how grateful we all are for positive/helpful impact you have on members here! Thank you 🙏
does anyone have a vocal replacement tutorial to replace the vocals with your own once the songs is done ?
you own vice *
Get the stem, sing over the instrumental part with your own voice?
If you have a cellphone and a quiet room, you have a studio-grade microphone
Add a free audio editor like Audacity to remove some of the background noise & tweak the vocals the way you want (maybe add some vocal fx) and you're golden
If you have a closet that's big enough and I mean just big enough to fit you sitting on a chair you can yet insulation from the hardware store. It's relatively inexpensive you got a sheet that's like 4 x 8'? And then you can lean it against the walls around you and make a pretty good studio I've seen a couple people do that
I mean if you're really serious
It's those pink foam sheets that look like really thick drywall
You just unlocked a vietnam-style flashback of covering a root cellar with egg cases 🤣
A walk-in closet with lots of clothes makes a good reverb dampening vocal booth.
im not that good of a singer 😛
For those who are wondering:
There are 2 types of copyrights for audio.
1-> Compositional
2-> Sound
1-> The score sheet and lyrics are compositional copyright protected. Even if you cover a piece, music, song etc., the composition copyright holder has right to take down if you are not licensed. So pro allows this, however you have to mention suno if they ask you to do. This is a very legally open to discussion in my opinion.
2-> Now the sound is the trickiest part here. Imagine a few people play same guitar and same sheet and nearly identical and uploaded to streaming services. How would you separate or verify they played it? So when it comes to AI, if you are not sure how you can do it, do not mess. The laws are getting strict on this topic.
Changing chords, instruments etc etc don't help anyone 100% when Suno ask you to credit them as compositional copyright holder. Also mastering doesn't always change identity of the piece.
The biggest issue is proving you can play the piece once your distributor ask you to do. If you can not perform it live for them, think wisely or mention it's created by a.i. if there is a section.
Now the mastering part. The mastering part is usually applied to whole album or full set of works to neutralise volume difference between tracks in albums. Streaming or audiophile or cd/dvd mastering has different setups and points to focus. Instead of using A.I. for Mastering, pay a few bucks more to engineers and get it done. It will bring more.
I have a few songs that are great. However, when i click on extend it doesnt really extend... it just generates a 2 minute song instead of a full song how i would imagine it. Is there anywhere for suno to generate a full legnth song? If not, is there a way for suno to generate midi files to export so that i can extend the song myself?
Midi part would be perfect however it require more computation when it comes to midi. Also if there is vocal in the song, midi wouldn't work. So each song, if there is vocal, should have midi w vocal and midi w/o vocal.
After you extend and you like the 2 minutes extension of that song, just on the 3 dots, choose create/get whole song.
Or you can keep on extending part 1/part 2.. etc and choose (create/get whole song) to combine them into a song
OOo, awesome. let me check this out and report back
Would be great to export midi files for the melody, bass, etc... and audio files for vocals for furthering editting
Yeah. No midi stems at the moment. You can get vocal and instrument by choosing create/get stems.
There's a site fadr.com that you can get more instruments and midi out. I am not sure if the free version allow you to get midi.
https://fadr.com/stems
Free account can get mp3 files. Try it
Fadr Stems is a free vocal remover, vocal isolator, key finder, bpm finder, chord finder, and midi maker. Upload any song and get your stems, midi, tempo, key, and chords in seconds. Fadr Unlimited is free, while Fadr Plus is $10 per month. Upload your favorite songs and turn them into something new today.
thanks alot!
You're welcome. Good luck.
it merged the file as you described. what i noticed is that the section it merges there is a noticible gap of sound where it cuts off and brings back in the extended version- how do i fix this? I noticed this happens to with the replacement feature.
I think i am having a better understanding of this replacement feature.
Think of extension as a track that you can cut off before you choose (get whole song). When you extend, for example, it gave you 2 minutes extension, but it only sounds good until 1:20, then you make another extension starting from 1:20, but make sure there's good sound and no artifacts or noise, or else when you extend it, you'll get bad quality.
Also, if you extend, make sure there's sound from that part so the AI has an idea of the beats and generate the extension from there. If you want to end the song from the extension. Choose the most quiet part if you can.
To end the song, if V4 and V3.5 doesn't know how to end it and keep on generate 2 mins song again... choose V3 on the drop down, clear everything in the lyric prompt and only put "End" on the style prompt. Then remaster it with v4.
I think I hit a jackpot. This is the first time I got a youtube contented used warming from suno song. The suno song sounds similar to the song that was upload to youtube 17 years ago :). It a Russian song.
I wonder if this is the bug that I should report to the dev?
could be by chance the content id picked up a similar note, beat, or flow to the song in question so best to either take down your version, or wait for action from the other party or suno themselves
Thanks. Yeah. I just deleted it. Not worth it.
No problem, used to make beats as well for a fun bit of time and got flagged due to a simliar flow
I spent a lot of time remaster it in fl, audition and RX11 too 🙂
So I guess it's normal then right? I haven't got strike before
for beats, and flow yes
notes are a gamble half the time but its smooh sailing most the time
I guess instruments are easier for youtube to detect
more so the flow and pitch of the beat
Cool. Thanks for the tips. I learn a lot from you.
even tweaking the instruments enough can get you around the detection sometimes, wouldnt recommend tho incase the original artist find you
instruments has a lot of relaxing flow
always happy to help 👍
i prefer more hype sometimes, otherwise relaxing or emotional
what if I want to make similar sub bass sound that I see a lot of different artist using it. I wonder if it work
chill instruments
potentially yes
Cool thanks.
ofc
Hi, why there are so many Problems with very Short Generation? Some are refunded but some Not!
Songs under 8 seconds are refunded -- a common reason for a lot of these super short generations in a row is if you're trying to Extend from the end of a song, or from a silent part of the song, but it can also be caused by certain (potentially bugged) style prompts
That Not tru, it Generator 2 Songs, But one is 8 seconds.
And why have Many Songs 4:00 with a abrupt end? I lost many credits with this Problems.
Pls help with refund.
As per the upload message: any vocal-like sounds detected in the upload will be marked private only (public option disabled) in case someone uploads something they were not suppose to (e.g. copyright music, voices of someone without their permission, etc.)
https://help.suno.com/en/articles/2477633
There's a new way to make music in Suno, and it starts with your sounds! Upload Audio allows you to turn any sound into music. Whether you've recorde...
4 minutes is the max generation time for 3.5 and 4.0 (half that for the earlier versions).
The reason for the max times is that the Output severely degrades after those time points
You can Extend to continue your music
https://help.suno.com/en/articles/2409601
(These help threads can be found at #1294418290348921007 )
Do you wish your songs had a different ending? You can make songs longer by using the Extend feature. Best of all, this can extend your maximum song l...
You will have to contact support@suno.com about that but it would need a legitimate reason for you to receive credit refund.
If you meant a refund of your subscription, you will need to talk to billing@suno.com
I have a question regarding ownership that I haven't seen asked and isn't addressed in any of the documentation due to the unusual circumstances:
I created a song while I was a free user, using lyrics I wrote (I don't own rights to that song, but I do own the rights to the lyrics, since I wrote them)
I've been a pro user for awhile; I've been looking at those lyrics and think I want a different sound for them. If I make a new song from the lyrics as a pro user, do I own the rights to that song, or does it only fall under fair use, since I already made a song with those lyrics as a free user?
Even as a free user, your lyrics remains your lyrics, so you can use those lyrics later in the paid plan to make more songs. You never lose the rights to the words just because you generated a song with them on the free plan.
Jon's right 🙏
https://help.suno.com/en/articles/2478401
If you wrote the lyrics for the songs you've made in Suno, you own those lyrics. Regardless of which plan you're using, the lyrics are yours to use as...
This happens when I try to edit the cover of my song.
The internet is completely fine, by the way.
Try zooming in the picture just a little bit in the upload UI, that should fix the issue
I wasn't sure because in my head, they owned the rights to the original song, and in a normal situation, a new work would be derivative, so protected by fair use, but not by a commercial license.
Thanks for the clarification!
That didn't work.
I'm trying to experiment with control in the prompt. I got a perfect result the first time I tried it now it doesn't seem to work anymore.
[heading like intro]
- musical instructions*
Lyrics
What's everybody using I looked at a few songs it looks like some people are using stable diffusion brackets sometimes does that work?
I can share the song where it worked. Everything is in there still for people to see. It just feels like there's too many things happening for it to be random.
Ah I see, not sure what else may be the solution 😦
Be sure to report it to #🪲┃bugs-web along with the size and dimension of the picture you are uploading
I don't wanna mess with etiquette in here don't know if it's OK to share that song I know there is another section for that
Nevermind, just reducing the image quality by a bit will fix it. Thanks.
Hi some body can help, aftter make Get Stems still i listen the voice very low back of melody
If you're sharing the song as a means of showcasing a technique that others can try & give feedback on, one way or the other, it seems fair enough to share the song here 💪
If the technique doesn't matter so much & it's just a song that you're really pleased with, #🐣┃suno-showcase or #💬┃general-chat might be a better fit 🙏
The "Get Stems" method on Suno is convenient if you want to keep the results within Suno's ecosystem, but otherwise you can get better results from external stem splitters, like https://bandlab.com/splitter (free, unlimited)
Grrrrr.... finished my last extension, did a full song in 4.0, and the whole thing sounds drunk. Dang it.
I tried almost everything but Bandlab... is it better than the rest? gotta find out!
My uploaded Sound is from a Diesel engine. There is no Voice in it. 2 Werks befor it work. Today Not. This is a bug.
I dont want Songs longer then 4 minutes but Suno make today many Songs with 4 minutes and a abrupt end. I hope you understand me now. This is a bug.
I started off using the OpenVino plugin for Audacity -- I heard its just an optimized packaging of Demucs, but I couldn't be bothered to figure out all of the configurations in UVR 5.
Bandlab was my go-to for a long while, but now that I'm with LANDR I use their AudioShake-powered stem splitter plugin -- it's phenomenal
This is very complicated. Why can you Not Test Suno and you will See this errors. I pay for it, but there are other Apps they make the Same and with a better User Interface.
LANDR plugins are not free... I use LANDR service but I don't think it's included in my plan. It's that good? I might have to try.
I usually don't mention paid options like LANDR's plugins, but when the matter of "better than the rest" is brought up, it deserves to be mentioned.
It cleanly separates the stems, doesn't leave a bunch of "extra noise" or bleeding between stems -- I'm scooting off to bed now, but if you're interested, you can DM me a song, I can separate it using AudioShake & Bandlab, then send you the results so you can compare
I'll send you a DM
Why cannot I see all my notifications and only recent ones??!?!
Why can't I view my followers on suno ai
Please fix this asap!
Also is this a glitch? How can I have more likes than views???
If people listen through discord it doesn't show the plays
Oh ok
This is an instrumental using the lyric section as instructions rather than hitting the switch for instrumental. If you listen to it with the mobile app or in here you can see what section it's playing. It really seems like it listened each time.
Synthwave, Atmospheric, cinematic, building tension, soft arpeggios and pads. dark broody bassline, retro beat song. Listen and make your own with Suno.
There's a "followers" tab on the library section of the desktop website
Can someone help me how to preserve the original vocals in a cover?
As far as I recall it only ever showed a set number, without a way to view older notifications -- it'd be an idea for #💡┃suggestions-web for sure 🙏
A song is too long and there's too many coincidences that match those instructions for it not to be happening right? Am I going crazy? But when I try it on other stuff it falls apart almost every single time
hello guys! Can someone explain our copyright claim for our new created song? 🙂
There's no way I could have come up with those instructions to create the song... but often than not, those instructions are ignored... I'm curious how many generations have you tried before settling with this version and during the process, have you modified those instructions to shape the sound?
"Ownership", "Commercial rights" and "copyright" are related by very different subjects -- Suno doesn't guarantee any output will vest in a copyright.
Done
The first generation didn't complete the music all the way and there was a weird gap in the middle which didn't make sense. So I replaced a section and then I extended it. I had to add the secondary outro and secondary climax. As you know each time you generate there is a second one made and I always chose the best one so there is a little curation in there. It's called guardian Sentinel five because there were five iterations or corrections made.
I did not add the words after the fact however
🇫🇷 Hello guys ! 👋
I see here and there, things said about copyright, which are partially false.
Allow me to make a very important clarification:
If you are the holder of a paid account, then you have all the rights to your creations.
If you are a user of a free account, your rights are exactly the same, except that use for paid distribution is prohibited.
That being clarified, what you create via Suno did not exist before, therefore without you nothing would exist.
In law, you are therefore the author of the creation.
I thought AI generated content was not copyrightable. You may have the rights from Suno to produce further music or use it in a video or even sell it on its own. I don't know if that ruling applies to Suno music as well. But images generated with AI do fall under that category. You have the rights to it but you cannot copyright it unless you do some further work with it outside of Suno. This means if somebody knows where it came from or how you made it they might be able to use it and you may not be able to fight it in court.
That's my understanding of it going into use of the service. I don't like it but I don't know if I can do anything about it.
I leave how I made the song in the song. I made a persona with it. I want people to use and I want to see if we can get control or if this is a coincidence
A philosophical proposition, but by using Suno.com, you agree to the terms of service, which stipulate that songs made on the free plan are owned by Suno -- if you make them on a paid plan, they waive their rights, titles & interest in the song.
Absolutely correct that having taken steps to create the song grants you Authorship -- while authorship usually indicates ownership, we're bound to the terms of use 🙏
Hello my friend! I take this opportunity to wish you a happy new year 2025. 👋
Yes and no, Suno grants itself usage rights in the case of a free account, but is in no way the owner of the creation.
Let us specify that moral rights are:
Perpetual: they are not limited in time and are transmitted to the heirs upon the death of the author;
Inalienable: they cannot be assigned to third parties for remuneration. Any contract stipulating otherwise is null and void;
Imprescriptible: the non-use of moral rights by the author for an extended period does not entail the loss of the right to exercise them.
Now that the right is mentioned, let's talk about accepting the Suno contract, via a free account:
Indeed, when you accept the user contract, you grant Suno the use according to their prerogatives.
But, this does not mean that you give up "the moral right" on the generated work.
Because that is null and void.
The author simply authorizes the use by Suno, and agrees not to monetize his work.
But the authorization is reversible at any time, upon simple declaration.
Note well, that this rule, however, is never, or very rarely applied, because people are first there to have fun, and especially to please themselves!
Those who wish to monetize and remain absolutely masters of their works, publish almost nothing in public.
how do i verify my account? i have sumo on my phone and i want the same account on my laptop connected via discord
Anyone received any winning notice about the Timbaland competition?
I know, I know it’s not a technical discussion… But Suno’s last announcement made it look like one…
Why not say; here are the winners?
Telling people to check their emails everyday…, very wicked 😭
Hello, anyone here succeeded in signing up using your edu email and got 1 month free Student Pro Plan?
Anyone had luck in producing EDM genres like Future Rave, Hardstyle, Trance? I tried several prompts but end up with commercial EDM every time. This is the prompt I used for the Hardstyle track which deosn't sound like hardstyle at all: "Euphoric hardstyle track about stars, with emotional vocals, a soaring lead melody, and explosive drops, Include atmospheric pads, vocal chops, and cosmic FX, Crystal clear vocals, hifi sound". Any hints would help?
Hey, I DM you for a question not related to Suno. Hope you check and reply me 
Help, does anyone know why the "extend" function doesn't work for me? it doesn't give me results
What does it give you? Just a progress spinner? Or has no action?
Hi there, can I input some sort of phonetics as lyrics in v4?
Hello, it gives me results that last 3 seconds.
If that's repetitive, it is often trying to extend off a silent, or quite silent moment in the song. Or a note or sequence that's triggers it to not produce anything. The fix is to listen to the songs end, or timestamp that you are extending from -- and pick or move that timestamp slightly back a second or few seconds.
So if the song is 2:34 seconds and you click extend, and it gives you 00:03 each time, move the time to 2:32 (if there is music at that point.
If you know this and tried this already, then state what you've tried and which version app.
I hadn't tried it, thank you
When you click extend, I notice often the wxtend stamp will be a second back already. the 3:21 song, as example, will start the extend at 3:20. One second back. This is to help, but still not work everytime. So we manually must scroll down to the extend timestamp and type in our preference time. Maybe 03:18, for example.
Information about the contests is usually discussed in #🌴┃creator-programs — it can be helpful seeing responses from others who have won, or what they did to claim their prizes.
I just tried it and the results still last 2 or 3 seconds. I even went back up to 20 seconds to start the extension
It's good you now know a technique that usually works. How long is your present song? And are you Free Or Paid subscriber? And which version are you using to extend?
If you’re extending from a section where the music is still going strong, it’s not impossible that something you’re using as a style prompt could be impacting the extensions.
If you’re comfortable sharing the song here (or DMing), we can see if we can figure out what it’s stuck on 🙂
That channel is for cooperate use…, no longer for ordinary discussion. Did you read my post?
Or did you read the announcement about the Contest?
You mentioned how your question wasn’t exactly “technical discussion”, so I linked you to the channel where updates about the past paid contests were posted.
It’s not a matter of“people in that channel are better to ask” — that’s the channel where that information gets discussed.
The song I want to extend lasts 1:59 minutes, and when I extend it the results continue to last 3 seconds. I have a paid subscription, and I use version 4
Looks like they locked it down because people kept posting unrelated content in between the contests, but if there’s any updates about the creator programs (paid contests), you’ll be more apt to get up-to-date information when it gets posted in #🌴┃creator-programs .
So I’d recommend following that channel and setting notifications, or checking your email regularly.
I just told you we no longer discuss that on that channel 😱
…I’m not telling you to discuss there — I’m telling you to pay attention to that channel because they’ll be more likely to post an update than #📚┃suno-school
Guys, is there a way to be sure suno doesn't make strange vocals when trying to make him sing a long vocal?
An example: if i try with "sooong" it will be singed as "suung"
Ok!
The last update about the contest wast not even posted there…
We must organize Suno 🙏🏾
I want weirdness.. where can I find that? lol
lol, i tried with IPA phonetics, if you want to make a laugh:
https://suno.com/song/ea3b224e-9afa-4597-9eeb-4d7997c0cf20
Oriental acoustic fusion, bansuri flute, guzheng, acoustic guitar, light percussion, rhythmic taiko drums, ambient strings, duduk, subtle vocal layers, gentle crescendos, BPM: 120 song. Listen and make your own with Suno.
if you open the song text you can see it "correctly" XD
anyway, i was trying using "so-o-ong" but the voice coming out most of times is really robotics or like using an autotuner, so was hoping there was a better way
I know the o's are elongated, but just change to the n's. Or the g's. " sonnnng" to many n's and it might stutter. So just "sonng" may be enough. https://suno.com/song/e9cc89a1-1728-4f27-9edc-4c47de7aabad
Also because of the degradation of English in our society, haha, the double n's in wannnnt , and Suno tries to sing "wanna" instead...
lol, crazy.. never thought about it
tnx Dasvon
If your style is opera, it will elongate anyway. So you can massage the style by adding opera in your prompts , if that does anything.
it's not xD
Trying to create something different mixing exotic instrument with drums and guitar, so kinda erratic even in how the verse is sung
@polar spade
last try with old system, kinda good
https://suno.com/song/9e369dca-7d0b-4cd0-a7a8-608d6c921e0f
Well.. a verse was completely forgotten by Suno.. but that's "normal".. as having the pre-chorus used twice in a row...
Oriental acoustic fusion, bansuri flute, guzheng, acoustic guitar, light percussion, rhythmic taiko drums, ambient strings, duduk, subtle vocal layers, gentle crescendos, BPM: 120 song. Listen and make your own with Suno.
Good morning Theharber. May I ask you some question about LANDR stem plugin?
Hi, may I ask if anyone knows about this. I am posting on TikTok and of course using my own suno gens and edits. Now I got this on one of my posts. My beat somehow got a new title and can now be used by others?
Has anyone some insights on this? What exactly is happening there? 😅
Like why is my sound now called revolver by gabrielromerol? 😅
Got some technical problem, anyone can help me?
I got a intro from an Suno-creation that I want to use in another song, so cropped the intro to the seconds i want to use and then edited lyrics from the cropped intro to basicly [intro].
After that I extended the cropped file with the new text and prompts. It turns out well so far, but problem came when I want to replace some part in the extended part, since Suno having a hard time to allign lyrics. So when I mark a part of the text in replacement tool, its not allign with the actuall part in the song. So when doing the replacing it just cut the song in wrong place.
Check at 00.58: https://suno.com/song/3a067403-bb27-4d41-aefd-c9b11767a330
An epic pop-rock-song in swedish. It will be in a pop-rock-edm-arena-style sports theme with a singalong hook in the chorus and heavy beats. song. Listen and make your own with Suno.
g'morning! Absolutely, what's up?
I see you mention LANDR stem doesn't extract other noise much and I am thinking about getting it. I had fadr plus but when it extract the stems, in the instruments and other parts, it also grab the artifacts. So I wonder if LANDR can do a better job?
Is it the Pro plan that you have that include with the stem plugin?
The stem splitter LANDR offers is powered by AudioShake, who does audio engineering stuff for Disney 👀
I'll share a picture of a bass track separated by their plugin vs the same bass track separated by Bandlab
LANDR/AudioShake on the left, BandLab on the right
BandLab is amazing, especially for the no overhead unlimited free tier, but you can see that the bass stem includes the full spectrum, while the actual bass lives in the 100-300hz range, with occasional peaks when the strings are plucked a certain way or something, but they've captured the entire spectrum (as well as a lot of extra noises that might not be desired in the track)
It seems like BandLab's AI analyzes the song simply for patterns in the song and makes reasonable assumptions about what noise comes from where.
AudioShake's whole business model outside of this kinda stuff is being able to separate multiple overlaid vocal tracks into individual stems (like if you needed to fix just the male voice in a duet, or if you had multiple people talking in a scene and needed to give the main character more focus) -- this leads me to believe that instead of applying a generic filter to every track, they SPECIFICALLY tailor each of their stems to the specific instrument. The bass stem is cleanly cut because it knows exactly what it's looking for.
The plugins are available by themselves on a (hefty, admittedly) perpetual license, or as part of the Studio standard/pro subscriptions -- I'm more than happy to give a demo if you have a song you wanted me to test it on -- I'd give it the LANDR mastering pass, then separate the stems for you to compare 🙂
Yes please. Can I DM you a song?
Hi I need help getting extensions of the song it’s just stopped
I couldn’t find the right menu after clicking the 3 dots…..🥹
Abso-fruitly! 🍇🍊
It can take a minute or two before the options appear — if you’re on desktop the options don’t show up until the “extend” button appears beside the name (this is also when you can use other features)
(If you’re on mobile, a lot of features haven’t been added yet, but you could use the mobile browser website for a number of them, like cover & extend 🙏 )
Lemme know if you have any issues getting things sorted on the laptop — all of the features are on the desktop website, where a few features aren’t on the mobile website due to it being tricky to manage on mobile screens
Yeah, the app is great for creating and listening on the go, but it’s missing a number of features present on the website atm
You are great 👍 thanks and good art
Really no one can help me here? Or is it a bug?
https://www.suno.wiki/faq/style-and-lyrics/duets/
Duets are notoriously difficult to achieve reliably.
if at least this would work:
[female voice]
text
[male voice]
text
[female voice]
text
But its completely ignored too
Its either full song female or full song male
I'd argue that you can break the song into extensions, and have better luck dictating whether a male or female voice sings that section, but metatags are suggestions, not commands
with german rock it seems at least somewhat working
there it did use 1 of the 2 [female voice] with actual female
but its very unreliable and never ever worked for me so far at gothic doom metal. Like completely ignored
Made a video today showing how I replaced a guitar solo using Bandlab Studio. Yes some could use replace function but this was after remixing in udio. Plus non subscribers can't use replace so they could do it this way. Wasn't the best video but it does go through how to do it. https://youtu.be/iT0b-BTym-U
I had a guitar solo that I wanted to replace in a song so I thought why not do a video showing how I did it.
By cutting the good solo out of a stem I was able to replace it in the version I wanted it on.
Hopefully it may help some people. I'll make better instructional videos on other parts later.
Tools & My Profiles
Suno - https://suno.com/in...
metatags are suggestions
Yeah I felt that
I just found it strange that it never took any suggestion at gothic doom metal
And its frustrating because Gothic Doom Metal means: growling male, and a very nice clean female singer
But it seems impossible to get
I got such a nice prompt which really gives me gothic doom metal now and now it fails on the voices xD
@broken lotus, if you stll hang out, I appreciate the ever-loving hell out of you for making the wiki 
and the audio quality when you extend gothic doom metal is often very horrible
metallic noise like 64 kbit mp3
in general any heavier metal seems suffer from quality issues
extension is probably the only way, but the quality degrades and the song is probably worse too if you have to extend verse by verse because it doesnt know then that youre gonna do that
I think I stay at deutsch rock. that works better with the current ai 😄
But the voices issue is such a small problem left compared to the prompt tags. Because putting just "Gothic Doom Metal" didnt got me far. Using a nice prompt with the help of chatgpt, I got the draconian style sound now and have actual gothic doom metal now. but the voices dont work.
Also is there even a meta tag to ask for a growling male? 😄
I'd wager it might be easier to impact the vocal style from the Style of Music section, as I find that has a larger impact on the style, while the lyrical metatags can "help nudge" the existing content into certain directions, like suggesting a change in style between the chorus and verse, or to add a guitar solo, but not always giving direction to the style itself
What also really would help is: when you generate a song and then the lyric tags and style tags would be marked with outline color or so, which shows with Color A which tags it did understand, and Color B which tags it did use in the song
growled, angry, demon-sung aren't far off from what I'd try -- due to the nature of generative AI, unfortuntately it's not always a matter of "using the correct prompts or tags", but being willing to adapt your approach to get you incrementally closer to what you were originally trying to get -- and the struggle can be trying to keep experimentative, instead of just resorting to a slot-machine pull hoping for a one-shot generation that nails all of your criteria
Technically, it uses all of them -- it's just whether it's understanding your prompt as something actionable that it can do, or if it's interpretting it as a stylistic keyword (that at least adds some "randomness:" to ensure that it sounds different than someone using the same prompts WITHOUT that stylistic keyword, but it doesn't mean that it's being used the way that we expect it to be used -- it's all part of building up your own personal English-to-Suno dictionary)
This is pretty good lol
Hi - (not sure if this is on-topic or not, so apologies) is there any way to change a word or two in an existing song (not necessarily via Suno)? It's a long story but I'm unable to access our paid Suno account and really need to change a word or two in the resulting MP3 from that account.
So weird I was looking at LANDR trying to find a way to do something unrelated earlier this morning. I was trying to figure out a way to identify the type of music I created using a different AI. Mixing a bunch of different genre together is exactly that on Suno a mixture. But in the real world not knowing all the adjectives that describe music and the different genres I cannot identify some of the things I've created. So I was looking for another way to do it but I didn't see that option available for free.
If one could produce the same results with one word rather than several in a combination then you could potentially have more control over the style you know I mean like almost condensing?
Your question was 2 hours ago, but.... In Replace, did you run the song, listen, while checking the lyrics section and make sure only the lyrics written are lyrics sung.
No guarantees, but it's there for that purpose.
Did you just move the conversation down? You didn't have to do that
Oh? I don't know how to attach a previous comment if I missed reply.
I used these wacky lyrics for a song and it came out pretty good! - https://suno.com/song/55a45c95-bbd6-404b-ad0b-2f2d0815d172
It's easy enough to separate the vocal stem and cut out specific words, and easy enough to record into that section -- but if the vocals were generated in Suno then you'd likely want to match the vocals -- at that point it might be easier to use Suno's Replace Section as an attempt to maintain vocals in the new section. There might be other generative audio platforms that can voice match and replace a melody, but I'm not immediately familiar.
re: being locked out of your account, support@suno.com would be able to assist 🙏
Sounds like you might be looking for something similar to https://sonoteller.ai/ ? Mr Tree speaks highly of this service, from what I recall
Song analysis, lyrics analysis and music analysis. Understand what your favourite songs are about. Get accurate and in-depth music autotagging with SONOTELLER.AI, a tool powered by artificial intelligence including a comprehensive summary about the song lyrics and music. Analize your song, analyze genres, analyze moods and enhance your musical u...
Replied to you in private conversation
Здравствуйте, хочу пожаловаться на исправление секции в песни, это просто ужас, оно поет на каком то непонятном языке, совсем не то что писал, кучу попыток в одно место!
From Siri translate:
Hello, I want to complain about the correction of the section in the songs, it's just horrible, it sings in some incomprehensible language, not at all what it wrote, a bunch of attempts to translate!
Hello, I want to complain about the editing of a section in the song. It's just awful; it sounds like it's being sung in some incomprehensible language, completely different from what I wrote. So many attempts wasted (to no avail)!
...
Matthew loves haikus.
He writes them fairly often.
He smiled today.
make sure the correct portion of the lyrics is highlighted when editing the section
Hey all. I just got Suno this morning. My first songs were great but when I touch the 3 dots there's no download function and when I touch Share nothing happens. Can anyone help?
I've found that sometimes you need to edit song details (making sure the lyrics are as sung) for it to realign the lyrics to the song and allow selecting the correct portion
Based on how you're describing the experience, I take it you're using the Android app?
Yes
Tapping "share" copies the song link to your clipboard -- there currently isn't the option to download songs from the android app, but you can use the same login information on https://suno.com on a mobile browser to download the songs, and to use basically every other feature
The Android app isn't very feature-rich atm, and largely serves as a convenient way to create initial generations or listen to songs on the go
Thank you so much. I checked and they are in my clipboard. I'll jump on my computer later. Thank you
Hi, I'm using the Android App and I Have several questions and doubts, could any one please help me out?
I'm using an android emulator, so it might not be the exact same experience as tapping ,but if I long press on the share button (or wild out and tap a couple hundred times), I get this screen that allows me to use the "share nearby" feature
What's up?
how can I recuperate one song that I've created? I didn't know that there was a limit of songs that the web shown
I Bought like 1 week ago the Pro Suscription and I've only created 2 Songs with and it appears me that I've left 6 Songs, Does it means that My points would be recovered and Start creating more songs in V4?
In the top right, you can swap to different pages 🙏
If you force-close and re-open the app, does the number of songs refresh?
My app was reading zero songs left this month until I force-closed the app, now it's correctly stating.
Hi, Is there anyone from Suno Team?
I need assistant urgently.
The discord server doesn't act as an official support channel -- us fellow users can try to give advice, but depending on the nature of your issue, you'd typically need to contact support@suno.com or billing@suno.com for assistance from a team member.
Usually a good step to contact them first just to get the ball rolling in case they're the only people who can assist, but if you're comfortable chatting here we might be able to assist in the meanwhile?
I'm having some issues replacing an instrumental portion of my song - the generation itself sounds good, but when I select the generated part it seems that suno is "pasting" it slightly out of the part it should replace, generating some weird transitions. Anyone experienced that and has any tips on fixing it?
I Haven't done it, but I'll try it
Cause it should Appear like I've 500 Songs left to create, right?
If all is good with the billing-side of things, One would presume that's how it'd show -- if the app is still giving you troubles, you could log into https://suno.com with the same credentials to ensure all is good with the account
Assuming you're talking about hearing a "skip" between the replaced section and the surrounding content, you can go insane trying to fix that with additional Replace Sections -- so far I've found the most success with remastering tracks to "smooth over" the audible seam
yep, that's exactly it. I'm trying to replace a section where there's some audible ringing. I've managed to replace other sections of the song successfully - I noticed that the difference is basically that this one is purely instrumental and the ones where it managed to replace correctly had some lyrics in them. I'll try to remaster, but that's not optimal since I've already remastered it before and rather like a lot of other nuances in the song so far 😦
I feel your pain -- while it may still take a degree of luck & credits, I've found the Remastering has gotten SO much better over the past month -- it used to guaranteed make massive changes to the melody and aspects that I held dear, but it feels like more recently it's only taking slight artistic liberties with accent notes in the melody, while the larger changes are more EQ based, or emphasizing specific instruments
Hi every one , I recently subscribed to your platform and have been using V4, primarily in Instrumental Mode, to create relaxation and sleep music.
However, I’ve encountered a few challenges while working with the platform, and I wanted to share them with you:
Tempo Issues: When I prompt the system to create music with a slow tempo (e.g., 10–20 bpm), the generated output doesn’t reflect this. Instead, the music often has a faster tempo, which isn’t ideal for relaxation or sleep-focused pieces.
Instrument-Specific Prompts: In Instrumental Mode, when I request specific instruments such as piano and guitar (or ask for "no instruments" to create a pure ambient sound), the system doesn’t always follow these instructions. For example, even when I ask for "no instruments," the output sometimes includes them.
Genre/Style Consistency: When I prompt for specific genres like "sleeping music," the generated music doesn’t consistently align with the calm and soothing tone typically associated with such styles.
These adjustments would greatly enhance SUNO V4’s utility for creators like myself who focus on relaxation and ambient music. If there are specific settings or techniques I should use to improve the accuracy of prompts, I’d greatly appreciate your advice.
Hey @thorn adder,
👋 Welcome to Suno & the server, I might have a couple tips to get you started 🙂
re: tempo -- v3x & v4 (at the very least) don't seem to respond to bpm or tempo-related prompts, especially since bpm is often included in tags for higher tempo songs, so it might associate it with speed. Perhaps slowed, tired, sleepy as replacement prompts?
re: instrument-specific "negative prompts" -- outside of the "Exclude Styles" box on the website, Suno doesn't recognize "negative prompts", so no drums is the same as prompting drums please
I used So slow speed and slow tempo prompt but It can not generate a music that I am looking for
I think the way to go is probably fix everything I want first (trying to ignore the skips) and then remaster everything once it's done
how to create a music without any instruments ?
If you'd like to share one of the sleep-tracks you're making, I'd be happy to see how else it could be approached -- getting chill music doesn't seem to too big of an ask in the slightest 🙂
Probably safer that way, so you don't get attached to a delivery that you might need to Remaster to save in the long run 🤔
Could be helpful to keep a scrap of paper with all of the "ugly timestamps" noted so you have an easier time checking whether the remasters not only sound good, but "smoothed over" what you wanted them to
You have used the word "replacing". Are you using the Replace tool, or do you mean while using Extend to replace something, it doesn't paste it quite right. Just asking as I've experienced extend to not blend well on rare occasions.
And to add, if its the Replace tool we are talking about -- do you used the edit lyrics box in Replace Tool to review lyrics? For good measure I still add one space somewhere and resync/align the lyrics BEFORE replacing, just because it makes me feel better-- and hopefully help in the process.
error while trying to ulpoad audio, anything going on with server?
into suno itself
No news that there's any server issues.
ok
i found what was a issue
i was on my liked songs page, on selected song i pressed reuse promt after i tryied to upload audio on that error was coming ''Network error''
Cool. Always helpful to post for others. Glad it worked out.
Hi everyone! I made a song I truly love the vibe to but I wanted to have a female singer. Is there a way to basically have the exact same song but use a persona? I haven't found a way to do this
thanks - I sent support@suno.com a mail over a week ago... nothing so far. There seemed to be a problem with performing more than one Replace Section. The second time, it replaces a completely different section to the one that I highlighted... looks like an offset issue Re:stem - for that I'd need access to the original song I guess, unless I'd be using different software
This. I noticed the last time I tried Replace Section. First time, it was working great! then when I tried again on a different section... the lyric selection always choose a wrong spot.. no matter what I tried.. such as editing song details and add in a space here and there... but it's still out of sync for some reason... so I gave up! This feature needs a fix badly... as it's the feature will be used the most ..to fix singing , pronounciations errors...
how to create a descent playback from a song or the best way to remove the vocal?
Use a third party program designed for that like Bandlab, UVR5, RipX Pro DAW
Look up vocal changing/swapping services like elevenlabs.io (or RVC models)
Much cheaper to use those programs/services to change the vocal's gender than try to redo it in Suno (because it is nearly impossible to keep it the same)
I am keeping everything through suno thanks though! Sounds like I’ll have to reroll the song until I get a female singer and guitar I like
I've tried using third programs, but it doesn't work, the result is exactly the same as what suno does with the “get stems” function
Huh, strange, people reported often that they get better results from those program than Suno. Especially UVR5
Hum, I haven't tried this one yet, it doesn't hurt to try one more, thanks for the tip
Using a persona would retain the same instrumentals and vocals -- perhaps you could do a cover of the song, but leave all style prompts blank and use female-sung as the sole style prompt? (or adding that to the original prompts)
Still getting used to persona. The one I’ve made has done well with both pop and punk prompts
I just wanted to transplant my personas voice into their guitar (solo acoustic) song I have
yeah, I meant replacing. The song itself is mostly ready, I'm using replace to fix some lyrics which weren't sung correctly, improve voice clarity (it's a somewhat long song and quality tends to degrade later on) and replace sections where there was ringing. And yeah, I used the edit lyrics/edit song details to resync everything. The skipping seems to only happen when the part being replaced is instrumental only - I guess suno can determine where to place the generated part well when lyrics are present, but not so much when there's none
At the moment, personas can only be used for initial generations, but the ability to combine personas with covers has been suggested quite often 💪
Right. I don't think it was ever intended to "replace" music for music. It's something to tinker with. Maybe someone has worked it out.
I made a persona from a cover.. got interesting lol
I'm going with @charred merlin's suggest approach (accept the skip and then remaster) and that seems to work. I'm replacing everything that I want now (leaving some skips in the process) and will remaster everything once it's done
Yo yall. I been having problems with the jersey club the music genre. It spears there’s too many mistakes the ai producers make. Like jersey club is fast paced hiphop genre. But instead it give me a rock song
I’m still using 3.5
hi.
Sometimes you need to give it a little more direction. You can add extra instructions inside the lyrics as long as they are in brackets. So you might need to do something like: Jersey Club vibe, club feel, hip-hop...
You can also do -rock or use the excluded styles to keep the rock out.
I just used a similar process to get the thing to spit out some decent dubstep.
See above first.
3.5 and 4.0 are trained more towards the popular genres. For the less popular ones, try making an instrumental in 3.0 to get what you want then extend into 3.5/4.0 for the better song structure.
If you cannot get it in 3.0, try seeing if there is a specific subgenre in a genre hierarchy list that describes what you want like from https://rateyourmusic.com
Request for Team Contact Regarding Music Concerns,
I hope this message finds you well. I would like to request the contact of someone from your team with whom I could discuss a few concerns. While I’ve noticed the existence of a technical support channel, I feel that a more direct conversation might be helpful.
Recently, I’ve observed that some of the best tracks—including mine—seem to face issues such as being paused or altered in ways that impact their integrity. I truly hope this is not deliberate, but the frequency of such occurrences has raised questions for me.
Could you kindly guide me to the right person or channel to address these observations? I value the platform’s commitment to transparency and collaboration and look forward to clarifying this matter constructively.
Thank you for your time and support.
Best regards,
@marciocampos
That...is very vague. You should link examples to a few of your songs to demonstrate that issue.
- Do you have Suno's logo appearing on some your recent thumbnail? Or a message saying to use suno.com in the music?
- Was it very recent where your song was terminating very early and the quality decreased?
- If yes to one of the above, let me know which one.
- If no to both, please contact support@suno.com
Hi guys, how are you? I hope so. Friends, I think I've realized something, if I use the “get stems” function and then use the “remaster” function on the extracted instrumental I'll get a good result. Well, it's not perfect, there are still a few bugs, but I think that with time it could be excellent.
Yes. You can also do a cover of the stem and then remaster, often twice. Then do a cover of that remastered instrumental to add lyrics back or new lyrics to. Sometimes nice results.
hi guys, i maded a song, and later extend, i liked one but how i can take the whole song?
whats wrong with the ai the last days? https://suno.com/song/52f97dc7-162c-49c6-9158-e852ccfc5ef6 https://suno.com/song/86a30a19-74b2-4724-afe1-09feafe41a8a all songs i try failed (failed at lyricses and at prompts)
Chill DNB with shamisen, beatboxing song. Listen and make your own with Suno.
Chill DNB with shamisen, beatboxing song. Listen and make your own with Suno.
go to the stem you want and there should be an option under create to make the whole song
got it, thk you
I have a little issue, I got an email apparently from Suno staff I want to know if it’s legit or not. Who I can reach out to about this?
Is the email from support@suno.com?
No, other email claire@suno.com
Ahhh.... well the suno.com might be legit. I'm at a loss how the discord channel could help. If a moderator pops in, they may know then.
Quick question for anyone who might know, I just finished a song and the outcome is exactly what I wanted but it cut off with maybe 10 seconds to go, anything I can do to finish it off or will it make something entirely different?
Yeah I send an email to support@suno.com. let’s see what happens. I just try here just in case
Congrats!
an extend under create/extend will take it out beyond and hopefully give it an ending. That's a start.
You can try extending the song, it worked for me 10 out of 10
@quiet pollen will do thx
Ah? Congrats? I don’t understand
Is Claire the Timbaland contact? I thought Michelle?
Ah true, Michelle is listed in Mike's announcement post, but afaik Claire is head of partnerships & engagement
@charred merlin Can I send u a screenshot of what I got? Maybe you can give me some peace of mind
Kewl beans. Sounds like a legit email. But if they are asking to share 12 million in Prince Jamals estate in Nigeria, I'd be suspicious.
Lmao hahaha
I'm a user like yourself, so I wouldn't feel comfortable giving "the green light", but I'd be happy to let you know if it sets off alarm bells
Thanks a lot. I just send you a message
Me too. I'm less than a user. I'm an ab-User. So I haven't a clue past a guess.
I’m normally pretty good about answering any kind of suspicious email. But I don’t know. Maybe I’m tripping and this is actually a legit email from them. Sorry again for bother
You can message @steady kelp - he is the community manager and should be able to tell you if the email you got is legitimate for your peace of mind 🙂
who? Sorry I don’t see who you tag
Ah, looks like I forgot to click the name, let me try again: @steady kelp
Thank you so much, I just send him a message. Thanks for your help
I already sent the images to him, but it's afterhours atm 🙏
What's your guys experience around Remastering or Covering a song? So far I noticed that some of my older songs sounds better in their original older versions like 3.0 or 3.5. Not exactly about the voice quality, or sharpness and smoothness of the song, but more about the structure, the way the instruments play, it has a better nostalgia effect for me in particular. I had to waste a few credits to make older songs "sound better". So, what you guys think?
personnally having lots of bugs with the remastering feature moving from v3.5 to v4. 4 times out of 5 it simply doesn't work or the clip is less than a minute for an original song of about 2min30sec.
Well, that's odd. I know that when Covering a song the time limit is around 3:13; so you have to "Expand" and then "Get the Whole Song". But when Remastering it's the full length of the original song. However, Remastering a song don't change the song's structure, it only tries to make the voices and instruments sound better, which is kind of odd and buggy sometimes. I think it should give more nuances to the song, than simply trying to replicate the original song. So that's why I tried to combine **Cover **and Remaster. But most of the times It came out with a completely different song, so the "Nostalgia Effect" is lost. My idea about Remastering was about to keep the original song as it is, but with better nuances, a intro and a outro, a better transition between verses, but it ain't exactly like that, it's more about the voices and instruments.
Remastering gives you basically the same structure, albeit sometimes with slight variations to accent notes in the melody, changes to drum patterns or certain deliveries, but the biggest changes are related to the vocal & instrumental quality.
Yes. You can** Cover** a song and then Update it to version 4.0. To add or dump some lines or add a intro or a outro. But then the possibility of coming out with a completely different melody is grater than 50%. But also I noticed a bug when you try to Cover a song from version 3.0/3.5 to 4.0 and it simply doesn't work, nothing changes. But instead you have to Cover the song to the same original version.
If you know how to read mail headers then there are ways you can verify where a piece of mail orginated from.
I subscribed to the pro plan on mobile app. But it won't show the credits or show on my account on my computer that I'm subscribed or with credits. But I look on my Google account and receipt in my email and I've paid it and is subscribed
Anyone got tips?
If it is a word that is elsewhere in the song you could copy it in a daw and put it over the existing word. If it isn't already in the song it may be harder. I have done it by generating another version of the song and cutting a section of the vocal stem and replacing a part I needed to fix on the original. It is actually the same process I used to replace this guitar solo https://youtu.be/iT0b-BTym-U
I had a guitar solo that I wanted to replace in a song so I thought why not do a video showing how I did it.
By cutting the good solo out of a stem I was able to replace it in the version I wanted it on.
Hopefully it may help some people. I'll make better instructional videos on other parts later.
Tools & My Profiles
Suno - https://suno.com/in...
Tribal Drums, Punchy Kick, Groove-focused Hi-hats, Atmospheric Pads, Analog Synths, Arpeggios, Smooth Transitions, Vibrant Drops, Hypnotic Layers, Funky Groove Patterns
song. Listen and make your own with Suno.
I just ran 5 prompts and got 10 different songs that were 2,3, and 4 seconds long. Ive been getting trash results all day not its like burning money. Give me back some credits.
Every time I try to replace a section, it’s cutting out a good chuck after the section in the final song—jumping ahead instantly and removing a good portion. Any ideas how to prevent this?
It usually refunds automatically if it’s under 10 seconds long for me
try to remaster the "bad stitched" one and it should smooth out
I just checked and they were indeed refunded. Thanks for pointing that out!
I don’t think that will bring back the whole prechorus and more that it cut out though right?
It's that bad? then it might be challenge to bring it back... I'm sorry... in my experience, when the stitch was bad, it was a small bit and luckily it was always the instrumental parts
Ah ok yeah it’s been that bad on 2 or 3 songs now, thanks for trying to help though I might just try to stitch it externally and then re-upload
Stitched it together perfectly on my own, hopefully I can continue the song when I reupload!
Keep in mind that upload feature only supports max of 2 min
Yeah just realized that :/ I’ll just have to work with what I have I guess, one day I’m sure it’ll get better
Hello, I don't understand why I am unable to create personas or remasters of my music, even though I have subscribed to a $10 paid plan, and these are features I have always used without any issues before. Thank you for your response.
and it's not a browser issue ...
Has anyone had any difficulty downloading their music? I mean not just the process of downloading loading it but I'm not even giving the option. The whole three Dot screen you go to to download and other things it's not even there anymore it's totally changed the only option I have is to share like share the link.
This is a recent playback issue (started about 1.5 weeks ago)
To my knowledge it isn't in the gens themselves, there's just been stitching from the player, if you go back slightly you'll see it disappears. (of course you could be referring to something else)
I assumed it was connectivity related on my end
But maybe it isn't idk
Why is shimmer still so bad 🙃
It got better for like...a day and now it's creaping in again. They devs may be mucking about with some back end code and it's just causing a hiccup
It's just causing so many wasted credits lolll, im getting bangers that are ruined with it, i will try to stay patient tho
Good morning, my friend, how are you? Thank you for your attention. It was indeed a small connectivity issue, but I was promptly assisted by the support team. I’d like to take this opportunity to thank the team and @Mike B for the immediate response.
I Build a Basic Music transformer architecture and want to Train it on a midi dataset… But I keep getting tensor related mismatches, is there anyone who could maybe take a Look at it? Because I don’t know whats wrong D:
Shimmering HAS improved since people have bug reported the hell out of it, but the key is figuring out why you’re still experiencing it — It doesn’t seem to be a universal issue, as many users (including myself) are sharing songs free of the artefacts that v4 was prone to (shimmering, treble/hiss, though a number of songs still degrade toward the end)
That said, it seems like certain genres or prompting methods are still impacted — but there’s more people creating rock and metal songs without issue by the day.
I’d be happy to help brainstorm (and spend some of my own credits) to see if we can shed some light on your situation, if you were interested 🙂
(ps. what do you mean stems are broken? Quality-wise, or was the feature unresponsive?)
There was talks of an API back when Suno.inc solely made Discord bots, you could also add the Chirp Discord bot to your own server, but after they open-sourced Bark, closed-sourced Chirp, and spun it off into suno.com, it seems like plans for the API are dead in the water, atm
Is the download feature only for subscribers?
Paid members can download .wav, but .mp3s and video are available for all users.
(if you're on the Android app, you'll need to use the mobile browser website to download atm)
If you have the lyrics for the song how do you denote which part to be sang by female and male?
The answer is not a solution (prevention) but an understanding. Presently, Replace under 4.0 doesn't do well with "attaching and blending" the placemark or right edge created by your highlighting segment -- IF , IF it is longer than your original selection. If it's the same, it's great and I use all the time now.
The results, if samples too long, are that if your sampled outputs are not exactly the same length as what you highlighted for replacement, then that sample, if chosen, will, until this is fixed, simply bleed over entirely its length on top of the remaining song -- or -- Suno makes an attempt to move and attach the remaining bit without good results, creating a same bleed-over or overlap.
The "solution" is reviewing your samples and do 1 of two things:
- make sure the "Make Same Length as Selection" [MSL S] is toggled on. Then most of your samples will be the same time length plus 2 seconds blend time on both sides. Those, in theory and practice, should work. I've had no issues if I do that. If longer, avoid, forfeit the credits and keep trying as likely the longer ones will overlap and disappoint you.
Or 2) whether the "MSL as S" toggle is on or off, review the samples first before confirming. Some may be the exact length, some may be seconds more and some 2 minutes long. Listen and decide if it's usable and worth knowing that you may wish to use the sample anyway and start again to extend from, now, somewhere before the bad blend of the new "Full song" you confirmed to create.
Number 2 can be fine if Using Replace mid-journey by those that create by extending songs every 1 minute during their song creation.
But yes, some final song sound perfect already, but then we're having to use Replace anyway, because Remaster of the song created a beautiful and perfect song except for that one mispronounced word now in the second verse. So, that's when we stick with #1 and just make sure we only use exact time replacements.
Whew. Glad Woody didn't block me. 🙂↕️
https://www.suno.wiki/faq/style-and-lyrics/duets/
Duets are tricky to achieve with consistency -- while they call it "continue from time", this article echoes my thoughts about using Extensions to control the style of each section you generate
Hi,
I just noticed this today—only the Custom section isn’t working for me. I can still create songs using the Suno auto-generation section, but I really hope the team can fix this for me soon. This issue is preventing me from delivering my projects to my customers. Thanks!
Can you elaborate on what isn't working? Are you receiving error messages? What do you see vs. what do you expect to see?
The songs aren't complete, and the covers look like this:
All the covers are the same and sometimes it shows with chains.
https://suno.com/song/650a8c7e-6a13-40b8-afac-b4f5ff7931ed
Energetic pop-rock, driving guitar riffs, orchestral elements, and female vocals, with a cinematic flair. song. Listen and make your own with Suno.
the covers look like this
the covers are the same
Just to check, are you referring to the artwork attached to the song, or are you referring to the "Cover" feature?
No, I mean, when I generate the song using custom lyrics and music style, Suno generates the same cover, sometimes with chains on all songs, which are all incomplete.
Please check this one: https://suno.com/song/890fcd31-5285-45b1-8998-4676aefc116a it's incomplete.
Synthpop, atmospheric synths, driving beats, female vocals song. Listen and make your own with Suno.
re: artwork / chains; Suno uses the style of music and the lyrics to determine the artwork. The first song you shared mentions Stormy Skies & Lightning, the second song mentions Chains, so it's possible to get artwork with these images.
re: incomplete songs; the first song you shared is 2m47 and ends naturally -- the second song is cut-off, but this has always been a possibility with generative AI -- "up to 4m" doesn't mean 4m every time -- if you enjoy a song and it ends abruptly, you can use the Extend feature to continue from that point, this is a common and very powerful method 💪 🙏
Thanks for clarifying, but I had the issue a few days ago, and something went wrong with my account. I don't know what it was, but everything seems normal now. (Still, I need to create more songs first to make sure it's working again for me.)🤝
Are you using anything other than the official Suno website at https://suno.com or their official iOS/Android apps to make music?
The example you shared a few days ago had a "SUNO" watermark over the image, indicating an account issue. As I'd mentioned, this can be due to things like billing details being incorrect. But, if you are using something other than the official Suno apps/website this can also cause problems.
I'm only using this website: https://suno.com I just installed ios app on my phone and created some songs using the app.
Ok that should be fine. Just making sure because there are fake apps and websites. The official app is published by "Suno, Inc." in case you want to check.
I don't know why, but the songs I generated now became very short again, but this time there was no watermark or anything like that.
why is SUNO repeatedly pumping out 2 minute songs and not using the full lyrics and adding words before MY lyrics... stop hijacking my song. I paid for full songs. give me full songs. also add a contact me section to your webpage so I dont gotta post publicly on discord... like wtf
re: contacting support -- the help page, https://help.suno.com/, is linked in the same corner where the Discord logo is, which would have lead you to support@suno.com, or billing@suno.com is listed on the suno.com/account page.
Best of luck 🙏
Yeah the main help page just has the Discord link at the bottom.
I clicked through Troubleshooting and got to the "something isn't working" page.
Unfortunately, the same problem still occurs now, it cannot finish singing the lyrics
But my account was not flagged.
The black-background image I linked to is the main website, showing where the help page is.
The page I linked to doesn't have an outright "contact" link at the bottom, but it provides a search bar for you to type "contact", so it can show you to appropriate link.
Seems like they don't want to be contacted by going that route
This is stated on the account page, and they broadcast support@suno.com on a good number of articles
A simple short typed note to them when reporting songs would almost remove any need to contact them
They might be contactable on their web page... but that ppl end up here asking how to contact them is a design indicator that a problem exists and needs fixing
At a guess, they probably receive a lot of emails asking how to do things, which the community can help with. At the moment they seem to have a massive backlog of emails so fielding "how do I..." queries is best done by the community.
As a side note, as far as I know the bug reports are aggregated, and aren't necessarily reviewed by humans -- it's great to provide feedback (there's a feedback link in the same corner section as the help article from suno.com! 🙏), but it might not be necessary to reach out to support for each bug report.
Again, it'd be great if the #known-bugs channel was available for everyone (it's currently locked and hidden!), because this way people could check that page and upvote any bugs they're experiencing -- it's absolutely fair to want to be heard when you've paid for an exciting product and the experience is marred with some kind of quirk or flaw, and it's human nature to want to be heard.
I tried to contact them, but they didn't respond.
How long ago did you reach out to them? They've been sending out automated responses to let people know they're dealing with heavy traffic, but they're working to get caught up.
I genuinely believe having maintained #known-bugs and #known-suggestions channels would cut back on a number of support tickets.
To an extent. Or even better, a help bot to ask questions to.
It wouldn't be drastic, but giving users a "semi-official" place to see their frequently reported suggestions, issues, and bugs would go SO far in terms of making people feel heard
do we at least get the credits we paid for back when it produces something that isnt remotely usable?
I wrote to them before Christmas last month
But later I used my own account to buy a membership and the problem was solved. The songs I made no longer had watermarks. The account was not marked.
By default, the system refunds tracks that are under 8 seconds long.
While it's not impossible that Suno can generate a track with "zero usable content", if you're running into the issue that frequently, it might indicate something that needs to be looked into one way or the other, because it's not a universal issue.
(not meaning the brush hands and walk away from the issue, just meaning to highlight that there's a lot of users who are able to use the platform without issue, so there's clearly something to investigate further, which many here are happy to help with)
I am very new to this, emm...
Anyone know the best way to get a vibrating electro magnetic female voice?
okey, i try
The short answer is no.
The long answer is you can accidentally create them with effort :p
I'd love to see a focus on processing metadata in V4.x - calling for voices, tagging them with names so they can be called up later in the song by name, more control over the instruments used in a particular interlude
I wouldn't go that far, it is easier to get duets these days than before. You just need to have tags with Male/Female or Names to increase your odds (but, half that the time, the AI may switch it around).
E.g. (different variations)
[Verse 1 - Male]
[Verse 2] [Female singer]
[Verse 3] [Donald]
[Verse 4] [Minny: Female]
[Chorus] [Everyone, All]
Though using multiple extensions and specifying gender through the style prompt is easier (and likely better quality wise)
That article was written around the launch of v3, and while I agreed that v4 is a lot more receptive to certain suggestions, I still like to position the trickiness to prevent setting expectations that any failures in adhering the metatags is an outright bug.
If people know how notoriously tricky it has been, and they get a near-miss, they'll hopefully focus on what they're doing right in order to get more successes, opposed to focusing on "what Suno got wrong" and failing to iterate on their methods, just asking "when is Suno going to fix duets?".
V4 is more receptive to some, and quite a bit less receptive to others.
That's kinda cool actually
I can't never get it to do a duet. Sadly
Most times, my getting the right duet is just setting up the right parameters, and then clicking create about 8 times before Suno rotates through the combinations and gives me the right one. It's almost that simple. The focus becomes "burning through credits" and people get frustrated something doesn't happen on the 1st or 5th click and think they must need to change tags or prompts-- which may be true -- but those magic tags and prompts change every time you start a new song.
It's like a lock automatically changing the password with each song, and either you attempt to find it by changing prompts and tags, or you just click a few times until suno unlocks it for you.
Main issue i'm having when i'm using the Persona to generate a new song 90% of the time It won't finish the song. It always stops early around 3:32 long and cuts the song off or it won't follow lyrics
Personas seem to reserve a chunk of time from the song they're based on to help give style context for how it should continue -- it sounds like that particular persona uses a 28-second long clip of the original -- the length will vary from persona to persona, but you can extend from the clips to wrap them up, provided 3m32s isn't enough to conclude the song.
Similarly with the lyrics, if you find that it's jumbling the lyrics up, it could be helpful to generate one section of the song and extend on top of that -- this way your songs are composed of multiple sections that you manually selected as "the best" and discard of any segments that don't live up to your expectations.
It definitely don't ever work out in my favor.. thanks for the tip
What app is best for using putting 2 sections together?
Without the notice
In the a whole song, on the second part being added?
Hi, is it possible to download a MP3 including the image I added to the song?
wow I feel like a dope @charred merlin I thought personas just gave the song you put it on that vocalists "style" this explains why i'm getting pieces of the songs I pulled them from in the new material. that'...is a bit dissapointing.
Just incase you're not aware, you can click ... > create > get whole song on the part 2 extension to get the Full Song added into your library for free 👀
I didn't think it will add after 4:00 minutes
Individual generations cap at 4:00 but anything that combines tracks aren't
The .mp3 files are available from ... > Download on the website, or through the iOS app. The images can be downloaded from the website by right-clicking/long-pressing on the image and saving the file.
I've clicked on create whole song, and it never does.
@charred merlin I just sent you a DM. when you have time, no rush
Initial tracks from v3.5 / v4 can be up to 4m long *
Extensions with v3.5 / v4 can be up to 2m long *
If you take a 3m song, and extend a 2m extension, then even another 2m extension -- clicking ... > Create > Get Whole Song will add the full 7m track to your library for free 👀
* "up to" == full length not guaranteed
You might need to refresh your page -- also, having the "liked" toggle enabled, or anything in the library search bar can prevent it from being shown -- but typically it gets added as the most recent track in your library.
I can't find (Generate) section and (Chirp-beta) section here! Please Help!
Chirp has been spun-off into https://suno.com -- you're not able to generate through the discord anymore
If there are old Chirp songs you're trying to get, I think the moderators might be able to assist?
Some personas are better than others, I've noticed. There's some that used to constantly regurgitate sections from the original song, but there's some that ABSOLUTELY understood the assignment.
I feel like a dumb butt. Wow it worked. I have used so so many credits trying to get my song to fit in 4 minutes. Wow thank yall

@charred merlin @hollow hare and @unique stone are really knocking them out of the park today! Thanks for the patience with others and myself in finding good solutions and getting to the bottom of some frustrations or just lobbing some excellent hints and pro-tips. You all are wonderful!
It's my first time here and I want to generate songs and music and music song videos, I am looking for "Generate: ---> Chirp-beta" , I can't find it here!
Hi @opaque plover -- I responded just a few messages above -- you cannot make songs from the Discord anymore 🙏
The "chirp" bot was shut down 😱
All is good though, the songs are made on https://suno.com now 🙂
.... and the day is not over... 😄
ummm... what?
Probably means everyone stopped making songs 😉
(couldn't help myself) -- likely just a temporary hiccup, but it'd be neat if this was a sign that the front page is getting a shakeup 😄
Hi! Has anyone encountered that Pro-Plan has limitations and all songs started to be created in poorer quality even on version 4, and there is also a limit of no more than 2 minutes of a finished song? What to do about it?t it?
Not on Pro, so that counts me out. Besides the intentional hallucinations, 90% of the last days worth of first run songs have been excellent on 4.0. It's later when I cover or remaster, things get dented up a bit.
Across all tiers there's usually been some variance between audio quality between generations, but if this is happening on a regular basis, if you'd like to share a recent result, we can put our heads together & try to figure out what might be impacting them?
Just showing some of the lengths of my recent instrumental songs, to highlight that there isn't a system-wide issue getting songs beyond 2m
for example, two minutes and no more. Regardless of the length of the text
Someone asked a question about poem song reading how do I reply to them since I’m not familiar with it. Do I just forward their question in here?
Doesn't look like he's under custom
Did Suno come with an official answer recently about the daily API questions?
Hi, I have a question for you all, can SUNO on my beat make a song and sing it?
Are these extensions?
Initial generations have a 4m limit, but extensions go up to 2m, but you can "get whole song" to stitch the whole track together, as well as extend multiple times in a row. But the 2m limit only comes in after extending from your initial song (which can be up to 4m)
If they plan on making one available or not, and if yes, maybe then can give an indication of a timeframe
@dim cloud
I have a question about “get stems” function. Extend parts and merged whole songs are not support to get stems. Is it normal?
There's some quirks to the UI, but for most of these restrictions you can use the Crop section, highlight the entire song, then "crop" it to create a duplicate that you should be able to get the stems of
If you upload a track, Suno can extend from it and try to continue the beat you've provided, if that's what you're interested in 🤘
I had SUNO create a song then I copied the lyrics, made a few changes, and pasted the lyrics into the lyric box to create a new song with the same persona. The new song did not follow the edited lyrics exactly. Did I miss a step?
Everything we provide to Suno is a suggestion, even the lyrics 🙏
Yes I read that and some other things, but I didn't see any statement from the actual team. Maybe I missed something?
I'd recommend following #📢┃announcements and #🚀┃updates -- if there's ANY mention of an API, guaranteed they'll announce it there.
I just have a beat, and do not have the opportunity to sing now, I wanted to try to SUNO to sing, on my track and actually I wrote the text, it's a pity that it is unrealizable
I do that multiple times a week, figured it would be better to just ask someone from the team. Also with people asking about it every day it would be good to have some statement in one of these topics
Suno doesn't have the ability to generatively layer new content over existing content, but there's a lot of creative ways you can approach songwriting (or song recreation with lyrics included), it just depends on how involved of an audio engineer you'd prepared to be 😛
That said, generative overlay seems to be a common request, I'd love to see something like that soon 🙂
Yeah, true, but even when making a new song in custom mode, it rarely gets more than 2 minutes, regardless of the lyrics. At some point it just stops and that's it, no ending. This all started at the end of December and still is.
@steady kelp -- are you able to chime in? As far as I've understood it, while things may change some day, at this juncture there are no plans to release an API, correct?
The only references to an API I've been able to sus out was from the notion.site from back in the day that more or less equated to "hey, we're considering releasing an API, sign up here if you'd be interested"
But having a simple "There's no current plans to release the API" would help clear the air
hey! no publicly available API at the moment, and no sign of it happening in the near term.
i'm not really the authority on the topic, but i can tell you with certainty that i'd personally want to clean up the issues with bad actors misusing Suno before having a Suno-authorized api available to the public.
That would help clear the air, but we rather hear when it might come 😄
Thanks for clearing that up, dude 🙏
As reasonable as the assumptions were, they technically were assumptions based on the timing of events.
Hi, do you know how to make a "persona" keep the voice in the following song?
I’ve seen people coming in General asking about the APIs. Isn’t there a faq page that specifically says there isn’t going to be one and the reason why.
hah, at the risk of sounding grumpy, and this isnt a reflection on the company's thoughts, but i generally think "maybe if the bad people didnt do bad things, we'd have a different stance on API usage at this point" 🫠
If there is an API available for a price you're happy with, there cant be bad actors, problem solved 😉
I don't know if there was an outright statement that the company had shifted away from it -- a while back they were considering the idea, so I understand why people are curious.
It could definitely rock the boat, but I currently see it as a seller's market -- why would Coca Cola license out their recipe if people still want to buy Coca Cola?
It just seems repetitive for the mods and guides having to answer when people are asking about it.
Why would Coca Cola forbid shops to sell Coca Cola 😉
I don't think we're arguing the same aspects -- the stores don't get to rebrand the coca cola, you still see the original company's brand.
An API would mean they accept money to sit back and let other people sell music generation services, which would directly compete for the customers that they're catering toward. It's not quite the same as other APIs, as commonly referenced, as ChatGPT is a fun front-end that grabs low-hanging dollars while OpenAI sells their intended product, the API.
If the only reason Suno has to release the API is another revenue stream at the cost of some control over their market share, I wouldn't say it's a fair trade at this point, regardless how much it would help up-and-coming developers.
Me too xd
Appreciate the replies. Keep up the good work in here 🤟Have a good day too
is it just me or the website for some reason is super non responsive while using chrome browser (might be something on my side of course) 🙂 as with edge looks ... a bit better
Has anyone found a way to prompt Suno to perform Duets more effectively?
again, my personal opinion only here (not that of the company), but as long as there are copycat apps called sunoaiai or sonuai or some silly shell app saying theyre using "v5", attempting to use Suno behind the scenes while scamming people and charging more money than the service itself... i don't see how offering an official api could be a win for creators as a whole. i wouldnt want music making to devolve into a ticket reseller mentality where somebody is paying 100 bucks for a ticket that was 60 bucks directly from the venue because they dont know any better.
@chrome cypress
Text
[Person B]
Text ...```
Have you tested that ?
Not like that I haven’t , I have done Something like:
Rapper:
text
soft vocalist:
text
does sun do better when it is Person A ,b cause it knows that should indicate a different vocalist for this sections?
У меня в целом он медленный стал в последнее время, во всех браузерах
«I've had it slow in general lately, in all browsers»
Tested it - Works!
Text
[Chorus: Male Voice]
Text
[Verse 2: Male Voice]
sings everything as female - 🙂 the other way around ... same stuff sings as female - doesn't matter
oh other way around worked - and it's a duet even
Text
[Chorus: Female Voice]
Text
[Verse 2: Female Voice]
sorry, i'm not Russian, but think you said something about all day 🙂 haven't spoken/read/write Russian for quite while 😄
Here, made in custom mode from scratch, a lot of text, and still the song interrupts at some point, more than two minutes does not give to do, as if some restriction on the account appeared. How to check this?
Sorry, I forgot to translate the text. I wrote that lately it's been slowing down in all browsers.
It's not that is slow for some reason on desktop versions it uses mobile version for me - while using chrome, will probably clear cookies and it might get fixed who knows 🙂 Edge is ok-ish ... but still the design can be improved but not my thing 😄
I agree with your statement about the wrongdoers. Literally re-branding is a terms of service thing, but I agree it's a troublesome thing to handle. No matter if there is an official API or not, they will try it anyway.
@charred merlin example of ChatGPT/OpenAI is pretty good. I don't see anything wrong with that, but of course it depends on what is more important to the Suno team. The choice seems to be between serving a bigger audience and wanting to keep users solely on their own apps, or maybe it's just technical stuff that's not resolved yet.
About users paying more on a different platform, that's just normal, like Cola is more expensive in one place then another. You can't go under the API price if you want to be sustainable (simply speaking). Offering an API allows users to be free to use their preferred app/platform and results in more users and integrations.
It doesn't seem uncommon that songs get cut off abruptly without resolving, but I would recommend emailing support@suno.com if every song you make is 2m or under -- doesn't necessarily mean there's been a restriction, but the support team might be better positioned to advise on whether or not there's a reason these songs come out shorter 🙏
is there an issue with extensions today? I'm getting nothing but garbage.
Not unreasonable points being made, and of course it would be a "nice-to-have", especially if you're looking to launch an app or business based on it, but understandably there's a LOT of moving parts that might not make it feasible.
Maybe a slippery analogy, not 1:1, but I'll still try; a restaurant makes the best food in town -- people love it and beg & plead for delivery so they can finetune their optimal experience (eating while still in bed watching netflix, for example), and subjectively for them, that might be the best experience. However, if the restaurant benefits from having customers dine in-house (like by offering dessert, upsells, daily specials, people walking by and seeing the restaurant full & busy), it wouldn't be in their best interest to make those changes.
Regardless, the team is likely well aware of what positive & negative implications releasing an API would have, and they seem to be not in favour of it atm.
Is your text just a text or is the prompt complicated/code style one?
On a related note, I wanted to see if I could max out an instrumental track in v4 by using @naive quail's /\/\/\/\/ "guitar bends", and both generations came out bugged as hell
https://suno.com/song/4f8f09f0-7752-4e02-a293-6a9f86db9d36 -- bugged audio @ 0:32
https://suno.com/song/9744fe5a-7b1d-4a81-81cb-e2a202172d81 -- the EQ "puckers" @ 0:15 and bugs out the same as above @ 0:31
Hello everyone, I'm back to raise the same problem that I raised several days ago and to which I received no response... I use Suno version 4 and I noticed that on each piece that I create, the sound gradually degrades by going higher or lower in the treble and with reverberation!?! Whatever the style of music, the problem always comes back. Am I the only one who has noticed this? 🤔
No, a regular promt with lyrics and verse/chorus division
Yes, it is. The voice is usually always better in the first versions, and the quality gets worse with each next attempt
At first, the voice, the beat, everything is fine and then the more we advance the more it goes up in the treble with more and more reverb present.
ah! So it's a problem
i mean, a real problem 😄
Because I spend a lot of credits before I get to something that I like and that is a little out of the ordinary! The AI has to push itself to its limits so that it comes out with some slightly original melodies
On average I try 40 times a song!
Either because the melody is not good, or because I am entitled to a man's voice when I ask for a woman's voice...
It's quite consistent, too. I thought it might be a result of all of the slashes being on one line, so I added spaces and line-breaks - still get the bugged audio.
A question! Does anyone have a trick to make the AI understand when we want the song to end?
I tried different prompt but don't work
the question is for me?
🤣
ok
Is anyone else unable to Crop at the moment?
🙏 They're currently spinning indefinitely before erroring out on my end.
Okay, I'm feeling confident calling this a system-wide issue.
v3 seemed more willing to wrap the song up if you extended & used end or something similar as the sole style prompt -- doing this brings the "get whole song" quality back down to v3, but you can remaster to bring it back up afterward
I just cropped a song. It deleted the part I wanted to keep and kept the rest..
there's no way back and I can't recreate it :/
Cropping isn't destructive, it leaves the original and creates a duplicated clip of just the content you selected.
But when you say "just cropped", do you mean within the last 30m?
No more shimmer? Haven't heard it all day
well I am missing the short piece and the long piece is missing the piece I selected
Hello , how to get suno premium students plan?
going back to double check
It's been greatly improved upon, though some genres are still more prone to it -- we're happy to help shine a light on prompting methods if we can help 🙂
Sign up for a google account using your school's email, sign up for an account using that email address, then you should be able to.
its not there
Can we distribute music on Spotify after getting premium plans?
Yes
but only the songs you create with the premium plans
@cinder gorge thanks for sharing your experience getting signed up with the student plan 🙏
If you create the song with a free plan and later upgrade, you cant monetize that song
Can't get it back can I?
Thanks
But you CAN email support@suno.com to request the rights to those songs if you wanted to to distribute them 🧠👈
Thanks
wait fo real? good to know. I didnt see that info on the website.
so I went back to the editing instance on the browser and that same clip shows the beginning crop I made, I'll try again. Wish we could duplicate and save a copy
Not quite sure what to say, as none of the features are destructive -- both crop & replace section create new entries.
Typically you can "duplicate" a song by "cropping out" the entirety of it, but cropping still seems broken on my end -- the newly added tracks never finish loading, and just disappear.
I tested crop about 30 mins ago: #🪲┃bugs-web message ... it's STILL spinning. lol
Let me try it
100% system-wide issue
Fixed it for you! https://suno.com/song/134e0caa-4f40-48dd-8856-45fd87d2a5ac 
metal song. Listen and make your own with Suno.
...tho it didn't max out the time. 
Yup, and when you refresh the page, that crop file is gone :).
My Suno app does not show my playlists anymore. It did right up until a few hours ago. I’m logged in, can see some of my library.
Extend is BROKEN
Ahhh, maybe that’s what the other user meant — I thought they meant the crop deleted the original song
Seem like V4 might be able to do End song soon. My extended songs sometimes have shorter versions now. Like under 1 min and it ending pretty good.
And it has nice ending.
APP version - not showing my playlists anymore - was fine until this morning? Even the DISCOVER section is jacked - shows Greek Text as the tiles
Greek text? Sounds like something broke temporarily - OPA!!!
anyone else having issues croppping songs?
Can someone explain how Suno is trained? No-copyright music or copyrighted music from streaming sites for example
@spiral bane yeah, it's system-wide
ok lol thought it was, needed a confrim thanks you are a champion sir
Does it give an error message?
You just tricked me into following @teal raft as I went to test it, didn't you?
😛 Just my way of saying that I wanted to test for myself if following was broken, so I followed KakerMix (it stuck)



App says you have no likes or plays, sorry you had to find out this way 
(that was a joke-y attempt at saying things seem buggy ATM, I'm sure they're busy behind the scenes -- all will be good in time)
Does anyone else have trouble cropping tracks? For some reason, it’s not working for me. ⁉️
3%, you aren't wrong.
Cropping is broken atm, which is why I'm thinking @grim wolf shouldn't worry too much about the follow thing being down for him atm
...the team doing some work at the moment?
Hi all! Is anyone having trouble creating music? Every song I make has an annoying background noise that gives me a headache... for a few weeks now
A high-pitched fuzz that takes over as it goes? If so, thats just audio degradation
Yeah, and i don't know how to solve the problem
you can't. Not without either a ton of credits on a workaround, or just waiting
Yhea..i can't make songs actually and i'm still paying 
and I have some hope, because I've been genning Remasters of my latest song in v4 every couple days to see if they've licked the issues... shimmer seems to be greatly diminished, but the degradation is still as bad as ever
I'm still trying to remaster, but the degradation/high treble is still there.
Yeah, i hate it.
...but I'm wondering how much of that is caused by degradation in the original song
The biggest determing factor on Suno's output is what we use as input.
Everyone is using the same models (those who are seemingly unable to make even a single song, as well as those making new songs daily) -- when we've worked together, we've made large strides to figuring out some of the issues as a community.
If you'd like to share what you're using as lyrics/metatags/style prompts, you might find someone willing to help figure things out.
If so, means that ALL my original songs are degradated.
Even the new generated in V4
Generated by Suno, yes. They will have some degradation, based on length
some songs and styles are less affected than others
This is generated in V4. No extension, no replacementhttps://suno.com/song/414bb503-0943-44f6-8099-b891d14801f1
And it's a sh_t
my main question... is if I Persona either the v3.5 or v4 version of this song, how much of that degradation will be preserved in the Persona?
Always used this one because i like how songs comes out. V4
Now...i don't know what happened
I make songs only for mr rp character so basically I fix the lyrics that inspire me and create the song, nothing complicated
see, this is why we, as users, get so absolutely heated when it comes to the audio quality issue... these songs that we love, even through the degradation... we want to be able to rescue them and make them sound like they should
They've trained it on publicly available music. They probably scrubbed various streaming services, youtube, etc... It was definitely trained on copyrighted data.
I swear... the next time someone says 'just make a new track lol', I'm gonna make a pet replacement analogy...
If all the songs I create give me a headache it's pretty difficult "just make a new track"
right
but even then, without a way to rescue what we already have, fixing the quality for everything going forward will still be rough on people with plenty of music they love already
I dont know who mentioned using 3.5 instead of 4.0 with Suno for
more abstract songs but.. thank you..
A high-octane reimagining of “Vortex” by Calyx & Teebee, Faceless dives headfirst into the raw, electrifying energy of the original. With razor-sharp breaks, deep rolling basslines, and cinematic atmo
thats sick AF
like, I have one that starts out pretty distorted, and it doesn't degrade much, if any, over the 3:27/3:25 runtime (cut some mid-song silence in post)... but I dunno if that's because it had 'room to operate' on quality, or what
As a musician, who did about 2000 prompts and generations over the course of like 7 days, I can give you a few tips if you like?
Would be appreciated. Just wanna do a couple of songs for my OC qwq
okay here we go:
So the "style of music" prompt tells one part of the story, but it doesn't paint the whole picture.
https://suno.com/song/2b72cc10-bfc7-486a-9ed2-5b93410b69d8 , even without adding a comma between "female vocal", and "ispitational e musical teatrale", and without checking if it was supposed to be "inspirational", I generated a song using what you gave as an example -- as commonly noted, the audio degrades closer to the 2m mark, but the first 2m is a beautiful composition that could be extended from (extending from that section, not the end of the song), it could be cropped & covered, etc.
I'm glad Interkeukens is stepping in, but if you hit a brick wall, tech-discussions is here to get involved instead of getting angry 💪
/me angrily kicks @charred merlin in the knee
Thanks for your answer! Yeah, I think so too, but started to wonder when I saw that users will be held by law if you upload a song that is copyrighted. Do you understand the tech behind it, how the music gets generated and how big library of data/music they need?
there's 2 things to understand.
1 .make the song prompt contain keywords as much as possible for what the audible characteristic traits are. Such as all elements lets say:
- start off with genre, style, direction sound (raw / lush / heartfelt / pop / indie / rough / sweet whatever.
- add BPM although most people say it doenst work - combined with genre it actually does influence at times
- add a chord progression. this is very important. it will generate a much better output towards your derection becouse it understands the key you're trying to work in.
- mention elements you're imaginging. Drum type (dance drums / live drum tone / etc) banjo, congo's, shakers, sharp synths, neuro bass, whatever. From what i've experienced mentioning 'tone' or 'sound' behind certain important elements have helped me a lot.
- vocals mention their ethnicity, characteristisc tone of voice ( Adenoidal, Brittle Croaky and Gravelly, Guttural, High-pitched - whatever examples I can give). This helps subliminal.
- Most important I have not heard anyone off yet:
THE LYRIC PROMPT IS THE MOST IMPORTANT ONE (!) YOU CAN ACTUALLY USE THIS FOR SONG STRUCTURE, TIMELINE, WHERE TO HAVE ELEMENTS BE USED AND SO FORT! I WILL GIVE YOU EXAMPLES OF LYRIC PROMPTS THAT WHERE ONLY ** AND I ACTUALLY MEAN ONLY ** USED FOR INSTRUMENTALS (!)
THIS! 🤣
I don't know much about the underlying tech, but the Suno team has never denied training on copyrighted music. The court case will eventually get rolling and we'll see what the legal ramifications are, if any... it all depends on which side makes the best arguments.
Let me give a example here:
https://suno.com/song/111afa28-a05e-425b-bdcf-39f814ab3653
SONG PROMPT:
Drum and bass, nearofunk, raw, neuro bassline, morphing bassline, warping bassline, evolving bassline, octave gliding bassline, futuristic, sci-fi, synthesized elements, organic drum tone, psychedelic
(this version has audio attached - only a few seconds of audio - NO LYRIC PROMPT - so actually with audio you mostly likely sometimes are better off not using the lyirc prompt for instrumentals - export for more abstract work of audio such as D&B is also better off using 3.5 instead of 4.0)
Drum and bass, neurofunk, raw, neuro bassline, morphing bassline, warping bassline, evolving bassline, octave gliding bassline, futuristic, sci-fi, synthesized elements, organic drum tone, psychedelic song. Listen and make your own with Suno.
I will now find a example with a lyric prompt - hold on
Alright, that's cool. Want to understand the tech, suno and other AI music models currently feels like the "duplo" compared with lego. But you can utilize these systems in very interesting ways that only the model owner "suno" company currently can do. The copyright law might be dead soon, otherwise China will own the whole music market in the future
Example with lyric prompt that also has instrumental and song structure / arrangement:
Afro-house, Dreamy synth pads, light percussion (kick, hi-hats, claps), subtle bassline.
Vocals blend harmonies with reverb/delay, atmospheric depth. Minimal, balanced mix, smooth grooves, reflective song. Listen and make your own with Suno.
lyric prompt:
style of music - music prompt:
Afro-house, Dreamy synth pads, light percussion (kick, hi-hats, claps), subtle bassline.
Vocals blend harmonies with reverb/delay, atmospheric depth. Minimal, balanced mix, smooth grooves, reflective
do note: its a remake for a friends his demo track - so it has audio included. It could work similar but I had to make it a exact clone with vocal / tone and sound so I had no other option as it didnt generate anythuing similar. the lyric prompt does show all of what I'm asking it to do and you can tell it listened to it.
~~Do you have a song example of Suno adhering to all of these detailed metatags? ~~ I'm blind apparently.
I have sent it here
the last prompt is this
with audio attached combined with the prompt though - as explained.
but if you dont attach audio in the 'song prompt' you can mention the chord progression to get a better result towards where you're expecting it to go
The lyrics start ~0:33, "we come one, we set free", but that's not supposed to start until 1:46
example of song prompt with chord progression:
Party anthem, upbeat pop-electronic, 128 BPM, minor key, four-on-the-floor, synths, rhythmic clap,deep bass, layered textures, euphoric builds, dynamic transitions, nostalgic-modern. Chords: Bm-G-D-A
this is the problem with suno it sometimes mixes things up getting final results etc, when results are undesired you can 'replace song section'
Ive been to lazy to actually do this so far as I already am convinced this will work.
my friend is big in the industry (procuded and mixed chomatica album lady gaga, did "the middle" by zedd and grey, etc.. many big tunes) and he's not happy with Suno or AI in general. I'm understanding his point of view, although i'm enthousastic about AI and certain solutions, I do also see the downsides, as a musician and huge audio engeneer geek, I'm just seeing what AI can do and how it works. And its interesting for sure.
It's a fun song, but it seems like a lot of the tags aren't being respected, or they might be a bit subjective.
Bassline : becomes more noticeable, providing subtle bounce
I read that at the associated timestamp and thought "okay, the bass is pretty bouncy right now", but if I hopped back to 0:11 the bass was just as bouncy at that point when the tags say it's supposed to be soft, ethereal & dreamy.
It most likely will not listen to anything I mention, do not get me wrong, i
Really though, if you're having fun and finding successes, keep doing it 🙂
i'm saying it got me the best results
I find it WILL if you give simple style prompts and build from there -- there's a clear tipping point where you lose control of being able to predict where it's going
for all wondering:
I'm lazy AF - i had all these lyric promps written by GPT after learning it about what tags I want to implement. I had it study songs on youtube or with uploads and then make it generate similar song arrangements.
Tags:
[deliverance][mood][[songsection][timeframe]
Yeah, same as me. The first two minutes are pretty cool. I'll try to extend the song then and let you know if that works
its more then that with the lyric prompt. you can really guide the direction more then you're aware of
I just had a thought -- are you a paid or free member? (idea works for both, just longer uploads are allowed for paid users)
I've had "big huge choir in the background" with a tag [backingvocals] actually apply it on lets say a [bridge][02:45-03:10] tag
Im paid yeah
Sorry, juggling conversations -- that was a question for Drecias
Paid!
4.0 works great for 'band' ish work and 'pop-EDM' but everything indie and niche is likely still best to use 3.5 so all good for unpaid users.
ah sorry 🙂
At first I just suggested putting the extension point before the degradation starts to kick in -- maybe 1:50 for good measure, or wherever "sounds right" in your song that you'd want to continue from, -- but since paid users can upload up to 2m anyway, you could download your song, throw it through a mastering pass (https://bandlab.com/mastering is a great & super robust free option, or I'd be happy to use my LANDR mastering plugin for one of your songs) -- then you could upload that back to Suno to extend from.
Might sound like a pain in the ass to go through the extra steps, but higher quality in == higher quality out
Most of my stuff is like... ska-indie-math and I think v4 handles the composition better than all of them.
I used to find that v3.5 drifted toward "generic" sounding content unless it was extending from existing content, since it was the "optimized" version of v3 -- which was still labelled as being "more broad & versatile" -- so I would start the songs off with v3 to get a good structure or unique "niche vibe" that wasn't pure guitar solo ecstasy, then I'd swap to v3.5 so it could continue the established style, but tighten everything up and give it better adhesion to a reasonable lyrical flow (where v3 got too weird and had "run-on verses").
it depends on the stuff I think most complex EDM more niche related stuff is better with 3.5
Instead of "better song structure", now v3.5 is labelled as "better arrangements and creativity", and v3 is "older with improvd sound" -- absolutely tarnishing how much of a powerhouse it is
haha
I prefer my post apocalyptic freestyle folk with a dash of v4, thank you!
whatever floats your boat
so far Drum & bass etc related stuff ive tried both versions and 3.5 still rules that
even remastering is a huge 'no go'
it changes warpy dnb bass to vocals - lol
v3 is the creative young brother who wants to play, and is into the crazy new stuff -- v3.5 is the older brother who can tolerate The Ramones, but still knows how to play American Standards and translate the chaos into something the folks can enjoy as well (or rock the hell out with his younger brother)
"older with improved sound", look what they did to my boy, they massacred him
also some great results
I didnt know it existed but I A/B tested and I can highly recommend another instead - more control too.
This one is far better:
https://emastered.com/
regarding cropping not working, system-wide reported at ~1:30EST
oh ok thx
Except there's no free usage on eMastered. You need to subscribe to use this product -- I suggest BandLab because they're available for all to use without payment, or overhead.
"Upload your track below and master it now for free"?
You can master it, but try to download
Has that got anything to do with one of my personas refusing to use new lyrics and just doing the same song over and over with a few random non-words?
band camp is also asking money for the best results lol
So.. exactly how much information about the song a Persona is made from does it 'remember'? Will it keep degradation as well?
No, that's likely just an issue with that particular persona
you can even upload reference audio to match your mastering with its awesome
BandLab offers an incredibly versatile free tier, unlimited use, they just offer paid extras like additional masterng settings -- the service you offered requires a paid membership. You could edit your original post to mention that
meh I didnt think the mastering options where any good tbh
i dont want to sound like a bum but if you actually good with google chrome development tools you can even donwload the good reasults for free - lol I just tried
my issue with BandLab is that just before I signed up, they offered 4 different profiles, with 3 levels of each, for a total of 12 options.
Not long after, they changed those levels to a slider, and locked free users into the 50% level, taking the choices down to 4
so, I said screw it, downloaded Audacity with MuseFX plugins, and started mastering myself
Oh? I was always under the impression that you couldn't actually use the far end of the intensity sliders without a paid membership.
Now they added a whole extra 4 mastering styles, so paid users still get 2x what the free users get. Plus they get Spatial 👀
You could use all 4 free types, and it had Low, Medium and High
I chose a paid subscription, but nothing changed. What is the difference between the paid version and the free one?
for Suno?
Amount of credits & access to extra features
plus, owning all rights to your music made during that time
Yes
Have you force closed & re-opened the app yet? (not just minimizing, but doing that thing where you see all the apps, then you flick the apps off the screen to start them over again)
I have a paid version for 10.99€
Yes,sure
Yes, I did everything. Now I'll switch off the power and turn on the phone. And I'll write here again.
🙏 This way we can be sure the app has restarted as well, good idea
...you never messed with Personas much, have you, theharber.?
nothing has changed
I'm trying to find out if it's worth saving the Persona for Smashing the Egg if the original is degraded so badly by the end
I was playing around with them more during v3.5 -- still having wild success using a v3.5 persona with v4, all of my best songs are coming out of it, with very few extensions required -- it just gets me.
I haven't made too many of them since v4 launched, let alone during the Degradation Days
Well, my idea was to use the Persona to do a re-take on the song once the degradation issue was fixed
Does anyone know how to contact the developers? I write to all the addresses listed, but there has been no response for the second day.
I'd really hate to have to throw out this version... there's features about it that are absolutely perfect, but they come both near the start and near the end
Okay, so we know it's likely not a UI issue at this point -- the next step would be to log into https://suno.com with the same login credentials you logged into the app with -- if you see 2500 credits, we know it's an issue with the app. If you see 50 credits, it's an issue with the account/billing, and contacting billing@suno.com would likely be the best bet.
Thnx
also... do you happen to know what the problem was that required removing the ability to Extend while using a Persona? If that worked like it logically should, I could probably just burn the credits to keep the degradation at bay
I know it sounds like a cop-out, but Personas still technically are in beta, so if you think of the basic song generations as "the main product", then instead of it being a guaranteed feature that's being restricted, it's an indication of where they're moving.
.....
Yes, the web shows that I don't have a subscription. But if you look at the list of paid subscriptions in Google Subscriptions, there is information about a paid subscription.
I was about to continue my thought and extrapolate upon it, but something tells me that line of thought will fall on deaf ears -- and I don't want to be accused of making explicit promises again.
Extending from a track generated by a persona will provide the same style as extending with a persona would give, either way. The real "unlocker" would be using unrelated personas on the same track.
So... what of the music we've produced already that we love and want to keep as-is musically, but simply repair in quality?
Like I mentioned before, attachment to our works is where the real emotion in this whole situation comes from
Happy to help, but bear in mind that Discord isn't one of the official support channels, so billing@suno.com is the authority on the matter -- but in terms of investigating (and having more information to provide to the support team, if needed), do you see the same songs/username on the mobile app and the website?
Well, that was kinda the goal. The whole idea behind my usage of Personas would be to fix up songs that have certain flaws to them by doing a re-take, either with edited lyrics, or, like now, to combat audio degradation
https://suno.com/song/2f5cc25f-577a-4da4-926a-a34c61e0afe3 - audio cuts off with 2 more paragraphs left from the song, is this a known issue?
sort of? It happens occasionally. You can Extend it with those two remaining paragraphs to finish the track.
thanks for the reply, i have another question. Can i train my own voice?
Nope 😦
You can upload your voice, but it will prevent you from making songs Public on Suno
Well, sorta... You can extend from an upload of your voice.
and as Tiberious mentioned, you won't be able to make the full track public
Now I can't log into my account in the web version. I registered using a phone number.
thanks for the replies, i have some amazing voice samples that i want to use and eventually maybe make a song from them.
Can I cancel my subscription?and is this subscription for a month or minimum 12 months?
If you used your phone number, you would select the same option when signing in from the web -- as for the subscription, it would depend if you signed up for the monthly or yearly plan.
I'm going to guess that the lack of response to the question about rescuing the flawed-quality songs we like means that we're more than likely going to have to just dump them and hope to capture the same kind of magic again... :/
well.. at least there's Remaster, for what it does, as long as that doesn't capture degradation too
The new download button takes too long to download a file...
Stems is a disappointing feature.
I thought it would be pre-mixed export
But it uses the finished mix and filters the voice afterwards out and the instrumentals greatly suffer.
I dont need AI for that. That thing can do my audio editor^^
Thanks ^^
Do you know when this music stops playing because right after song ended it still continues playing the lyrics and music 💀 which idk how do I fix that.
Hey Folks anyone else is experimenting troubles with the creation?
I can't to reach 2 minutes long even with the v4
And every song create sounds glitchy and low
Do you have Suno's logo appearing in some of your thumbnails?
Hm, this might indicate your account is flagged for some sort of suspicious activity (can be a false positive).
There are several reasons why your account could be flagged. Fake apps, API usage, VPN, and payment issues are typically the common cause. Not that I am accusing you for any of it, there have been false positives in their detection system.
There are a lot of scam apps on the app/playstore. Ensure you have the official apps by going to
https://suno.com/Android
https://suno.com/iOS
(You will be redirected)
The official website is https://suno.com
To clear the flag from your account, please fill out the following form for someone on Suno's team to investigate the flags on your account: https://sunomusic.typeform.com/to/zGsJSzsf
Turn data collection into an experience with Typeform. Create beautiful online forms, surveys, quizzes, and so much more. Try it for FREE.
It's a copy-paste response since we did have a few users who did download fake apps before.
Just be sure to fill out the form and hopefully the investigation will be finished within a few business days
There's no pre-mix data to export unfortunately. Suno generates finished tracks
ok
Thx
I'd argue the main advantage to using stems is that it keeps things "within the Suno ecosystem" -- whereas you can only upload 2m clips, you can take a full-length song, separate it within Suno, then cover / remaster it 💪 (you might need to crop the stems to enable the other features)
Crop seems to work again 🙏
hmmmm.....
well, not like I'm gonna be making anything new for a bit... can FAFO and see if that helps with my song
maybe I can see where it starts degrading too
Should be noted that Suno AI most likely saves the generated audio in a vector array that is already mixed. So, the stem separation is generating the audio file off that mix, hence why the stems have artifacts.
but, a Remaster can clean them up pretty well
...mostly. Sometimes artifacts in the vocal stem can garble words
Are there no notifications or an overview page where I can see which new songs the people I follow did upload? I cant believe I have to manually click everyone to find out every time 😅
Anyone know how to make Suno just talk? I want it to just say a line of something at the end of a song but it keeps singing it and adding music.
[Spoken Word] and around 100 credits of rerolls usually does it.
eep 100 credits
where does [Spoken Word] go? in the music style or lyrics
oh and the shimmer is back...better than ever.
I usually put it in the lyrics if there's only a specific section I want it to do that with
like this
[Verse 2: Spoken Word]
The halls are paved in stygian gorse,
v3 seems to have the easiest time with it -- "spoken word, interview, speech" or similar style prompts that suggest recorded speech -- v3.5 and v4 can do it, but they both seem a lot more likely to drift toward music.
argument, debate, spoken, speech, spoken song. Listen and make your own with Suno.
Here's an example of v4 dialogue, but that's the longest result I got without music
(don't mind the bread-person)
dang that's impressive
Can I edit a song I created in a professional plan and distribute it as I want?
If all of the segments of the song were created on a paid plan, the track is yours to do as you please 💪
speaking of rights... what happens to a song if you made it on paid, but dropped back and did anything with it, then later tiered back up to paid?
or is it tainted the moment anything happens with it on a free plan?
Just following up. So can't add my vocals i.e. a demo or If I wanted to perform at open mic or showcase I would not be able to? as a PRO paid subscriber? trying to understand this more
Slight mix-up -- you can use your tracks however you'd like if you made all of the segments on a paid plan, but if you upload a clip to Suno and it detects vocals, you can't set the track as "public" within Suno, meaning it won't show up on your profile (but you can still share the link and place it in public playlists), a couple other restrictions for tracks that get restricted that way, but nothing that would limit your ability to perform 🤘
oh me was blind oopsi
I thought at home is only this Trending and new songs from anyone
But a bit scrolled down I see it now 😄
This will make my life a lot easier now
Now I understand how my followers notice about my new songs 😁
Yeah, he said it. The only things you wouldn't be able to do with it are:
Put it on your Suno profile, and
Create a Persona from it
the music gives me exacly what I did not want. You did it again Suno
https://suno.com/song/2c3df7fd-52eb-4a36-876e-5a3b29cae399
nature sounds; sea waves; beach; no instruments; no piano; ape sound song. Listen and make your own with Suno.
Suno does not know what "no" means, it is treating it as a style word, especially since it was not given in the Exclude Style section
I will try it again, but nature sounds music on Suno is difficult
I wonder if it might be easier to remove the nature aspect from the equation & try to focus on making the type of ambient music that would go well with it, then add some nature sounds afterward in a free audio editor like Audacity, or on the web at https://bandlab.com/studio
thank you
Yeah, Suno may not be the best for this. I remember https://elevenlabs.io/ having a text to sound effects thing, one of their examples had wave crashing iirc
babbling brook repeated on a low volume, occasional birdsong, slide whistle, cicada
I solved the problem. Just to let ya know..with Audacity!
My curiosity has not been this piqued all day -- do tell 
just did for the sake of the comment, but have to follow through with the other proper hints
https://suno.com/song/10c3947b-f724-47c0-a769-0e75e5bbe4a3
Your musical teatrales were degrading after the 2m mark, like a lot of people's generations do -- then we chatted about trying to salvage the first half of the song when it still sounded awesome and to extend from that, like we're bypassing the degradation;
Was that the right sort of approach, or how did you make out? 🙂
I tried to extend it, but unfortunately it didn't work, it still generated an unpleasant sound in some places, both in the voice and in the music. Audacity was recommended to me, so I downloaded it, followed a tutorial and finally managed to fix the song, removing that annoyance.
I suppose it could be a valid solution to the sudo "bug".
At least for me since I understand little about musical styles and I add an unclear description to the program
Does anyone know if it’s possible to generate car engine or revving sounds? I’ve been trying in a few prompts but haven’t been able to get it
I have a suno ai pro plan
Nd i want my song download Wav file but But I don't see the download option for wav file.
Only shows normal download then download mp3 only
Pls help me how can i download Wab file
I have a pro plan already
Audacity is amazing 💪 I used to really rely on the OpenVino plugin (if you check out the "more plugins", or they might have an "ai plugins" section now), you can split songs into 4 stems right from audacity
Are you using the website or the iOS app?
iOS app is limited to just downloading .mp3s, Android app doesn't support downloads, but you can download .wavs and use the full range of features on the website. 💪
extremely hard, but go to search and look for songs through keywords.
I'm quite in denial, I admit, it took me a while to understand. Luckily I have a friend who plays guitar and understands more than me about music, so he was able to solve the problem in no time with Audacity!
Thanks for the help, anyway! Nice to know that there is a community willing to help
I did on my end without luck. I guess Suno does not know Motley Crue
I have one paid account + 5 free accounts, as I understand it, it is not appropriate to have so many free accounts, I want to delete my free accounts and stay with my paid account only, in such a scenario, can I face a problem in the future?
Sometimes it can be on Microsoft copilot extension
🙏
From everything I've witnessed, the fact that you've course-corrected upon learning the terms of service about multiple free accounts, you have "the optics" on your side
All I’ve managed to get so far is a guy saying vroom-vroom most the time, it’s pretty funny tho ngl
My buddy @naive quail has done a LOT of sound effect stuff -- even in his most serious songs he manages to sneak a [quack] or a [enraged goats on welfare] into the lyrics just to see how Suno will react.
https://suno.com/playlist/49a40042-a09a-421b-9df1-3f6d976221f9
Silliness aside, he's managed to get [lawnmower] working after putting his mind to it, so I'm sure there might be some tips about getting a car revving
haha
I was shooting for nature sounds and got the screeching tires of a car
probably someone dumping trash in the woods.
I did some more digging and found this by @PiosolverW https://suno.com/song/9a8de2a9-0f3f-4f8d-ae6a-d93eb5dda486
Sampling
Sound Fx
Gunshots
sound fx
Engine Roars
simulator
record
scratch
sampling
effects
car horn sound
birdcall song. Listen and make your own with Suno.
Another good thing to keep in mind is there's a good degree of luck -- sometimes a song that's advertised as "engine roars and gunshot" might have started off as "country death metal jamboree" and it slowly evolved into a track they made public
Gotcha thanks for the help guys
This appears to be a bang-on success -- I'm curious if this can be replicated easily
(the user's profile picture is hilariously terrifying, ai is so cursed 😂)
@naive quail, you didn't tell me you had multiple accounts
I only heard gunshots, tho 😄
Looking at the style prompts for it, it made me remember all of the debates into the differences between separating prompts with commas, newlines, or leaving them all together -- I haven't really tested a lot of that in v4, I always just use commas between
Ever since the v4 update, Suno tends to struggle to extend a clip without making the music skip at the point where the extension starts. Is there some way to fix this without an external editor? I've tried replacing the problematic section, but Suno doesn't manage to match it properly to the end, making things worse.
if i understand your problem correctly, then i suggest trying to listen to a point where you arrive at end of the beat and extend from there rather than just take a random point. Also start the extension with [break] of then mixes the parts better together.
and definetly avoid extending from middle of lyrics section
That's a good idea. I typically just extend from the end of the clip unless I need something specific earlier. Finding the beat when typing in the second to extend from would be tricky, though
just put the song on play and hit pause when you arrive such spot
It doesn't have to be perfect. You can go 1 second before the beat, the AI will take care of it. When you get the whole song it will bridge the 2 parts smoothly
and if the cap between sections is very tight or too tight then just roll back further there really is no other way, trying to use [break] in between or some other instrumental command to give it some space
The thing that sucks about this is that this problem never occurred before v4
Hi, can you help me with something.. Many times song ends without a proper end. just in the middle of the word for example.
and many times it puts music description into lyrics.. song starts with a description.
If you want to end the song. Try to end with the part that has less sound. V4 is hard to end the song. Try pick v3 where to you need to end, clear the style prompt and lyric section. put "End" in the style prompt, then remaster it.
For lyrics description, make sure you use [ ] instead of ( ). if the description is too long, it might also try to sing with the lyric too.
It'd also help if you provide an example of what you mean by "it puts music description into lyrics" (like a link to one of the tracks with the issue) - there can be a number of reasons for this, but without more information, all we can do is speculate.
Hi. Can you help me with suno fake issues, i already subscribe
natural language descriptors/styles/tags such as
Evolving synth textures with flowing arpeggios, soft choral echoes
...are MUCH more likely to be interpreted by Suno as lyrics and sung, regardless of whether they're contained within square brackets, paratheses, or bacon.
In the first case, the song and lyrics were still supposed to continue but ended in the middle of the chorus. Is it possible to extend the song with the same melody?
In the second case, I described the style of music as "funky rnb..." and the song began with the words: "Fubky rnb ...
Hey, don't you think that for free users they should give 400 credits instead of 50, plus that the credits accumulate and don't disappear the next day.
That doesn't seem feasible to me.
400 credits per day x 30 days = 12,000 credits
That's... more credits in the month than the top subscription package gives.
I suspect that all free users will be perfectly fine with that, tho... 
with the dots?
Well, then you can accumulate credits when you don't use it, as before
They're not offering the free accounts as a service to the world - it's a business model intended to generate subscriptions by giving users a taste of what the service is capable of but limiting their access to the full feature set.
Even as it is now, many users are perfectly content to use their free accounts indefinitely, despite the limitations. With access to 400 daily credits on free accounts, many Subscribed users would probably cancel their subs and switch to a free account.
It's just that I'm poor and in Cuba the restrictions don't let me buy anything
Maybe try to extend the song before the middle of the chorus part.
I get that, but Suno is still a business with employees and bills to pay. Free accounts for services like Suno cost businesses a LOT of money, both directly and indirectly due to the time spend combatting botters and those who'd abuse their free accounts.
Hm, every instrumental track I request still has an echo of lyrics and muffled audio during the places the lyrics were sung. Any chance it spits out a “normal” instrumental?
Is crop working this morning? I'm sorry I'm new to this server and I don't know where to find the current issues there's no status or anything like that I can see anyways
I know it was down yesterday
It was back up last night -- just checked & it's good now 🙂
Thank you!
I'm sneaky I use my phone for everything I'm on an iPhone. And normally I can get away with just using the browser version and the app but it will not work on the iPhone so now I have to go find a desktop lol
Yesterday I was using a desktop mind you I'm just not near it right now
It ends up loading music from other people when you try
I think this is one of the few services where not everything is available in both modes. But I don't mind bouncing back-and-forth I have a tendency to use my phone more on web access rather than using an app.
The on-site stem splitter isn' the most powerful stem splitter, but it's nice for the fact that the tracks stay "within the Suno ecosystem" -- whereas we can only upload 2m clips back into Suno if you split them elsewhere, you can keep the full-length stem in Suno if you split on-site -- while the tracks often have some "bleed" between them, you can remaster or cover the track to get a "clean" version. (if the cover/remaster options aren't there for the stems, you might need to "crop" the stems, selecting the entire song -- so that it makes a duplicate)