#๐โsuno-school
1 messages ยท Page 64 of 1
Hi! Tell me, is there a problem with the 'suno-showcase'? Why doesn't it display the dynamic link like before? :
If your song is 60s and you Extend at time 20s, then Suno will not have knowledge of the time between 20s and 60s. So the melody of the song may end up very different.
But if you use an audio editor (use find a different section of audio of the original song) to make the voice you want to Extend at 60s, after the 40s melody, then Suno will have knowledge of it.
When you 'Get Whole Song' Suno will always use all the parts together. So if you used an audio editor, this may not create the song you want. The song will have an extra section of audio in front. This is currently a user interface limitation in Suno. You could still download the completed song and edit it to share on YouTube. And I am hopeful that Suno will have better feature to solve that problem soon.
It's a bug, yeah. Suno is generally on United States EST time zone, so will be some hours before they can fix it, I think.
Ok! Nothing else! That's fine! Ha ha!!! ๐ Tic tac, tic tac ... โฒ๏ธ
that generic banner is crap. No one gonna click on my great songs now. ๐ข
You can upload your song to YouTube and link it in the showcase (with appropriate attribution to Suno in the description if you are on the free plan)
The embeds will likely be fixed during US working hours today
Thanks a bunch. Not on youtube anymore. Would do it maybe with bitchute or rumble but will just wait. Thanks. It says not to post youtube links, but I have no problem with it, but I do not like the team at youtube anymore. Just way too controlling. Too much censorship.
Hi! I have a PRO plan but I can't download tracks in WAV format...
Try using another browser. Someone did report using FireFox helped.
Please let me know if that works out for you too (as well as what browser you are currently using that stopped the download the wav file)
chrome. opera, brave
All chromium based, hmm. Try Firefox or Safari as they use another engine (I think? Been a long time since I looked up what Safari used)
I don't know why, but I think v3 songs have been playing better these days than v3.5. Maybe there's something wrong with the algorithm or something? For example, if you put 'opera' in the style box and make a song with v 3.5, suddenly a song is created with synthwave.
Ok. I think i got this. Thank you so much for the assist here.
Cheers to the person who made this song title!! It is genius and is the first to show up when searching โsuno v4โ ๐คฃ
Hi everyone, I've been using Suno Ai with a Premium subscription for several months. Today, unfortunately, I think I've noticed that when I upload an audio file with only voice (content not covered by copyright of course), the system always generates an instrumental track, eliminating my vocal stem. What has changed? Can someone please help me?
suno-showcase is out !!!
You may be extending from an empty second, prompting the AI to start the next bit of the song that it assumes to have to make. I'd play with the extend time a bit. +1/2 seconds or -1/2 seconds. Depending how large the empty spaces are ofcourse.
It's not only me, right?
It just stopped working for me, nothing will load lol
Yeah same for us in #๐ฌโgeneral-chat
Not working too
Same. In google Chrome and Opera suno does not load
Same here
same
You can literally See the "Team" Members come online ๐คฃ
Suno is no longer launching on any browser or app! Any information?
It's currently down it seems. We'll have to wait till they get it back up. I guess keep an eye out on #๐โupdates or #๐ขโannouncements
It would be great to have Suno on Downdetector.
You can Mail Downdetector and ask for them to add Suno.
A message appeared on the page: 'We are experiencing high traffic. Song creation is temporarily disabled.
anyone facing SUNO cannot be access?
This is great, people are not able to read ๐
i thought my account problem, seem that most of us having same issue
It's not just you -- it's down for everyone.
This isn't a typical "high server load" issue, and appears unprecented.
aRGHHHH how long????
so the problem now is profile, library and create song cannot be access right? anyone same with me?
service not woling for everyone...
i know, it was a response to @marsh stirrup
okay. i thought my account problem coz yesterday i login with another device, will it affect my account?
nah... dnt think so... ๐
I meant to share with everyone -- sorry for pinging you directly ๐ Thanks for helping direct traffic, my friend 
I use my account on several devices, no issues...
no worries... n more than welcome...
Oh my... no Suno...
It's like taking a guitar to the cottage and snapping all of the strings simultaneously -- well what the heck am I supposed to do now, just compose inside my head like Beethoven or something? ๐
you have to go to ๐ "System32\taskkill.exe /f /fi "status eq not responding" " and then Suno will responding ... Heu!!! I think so... ๐
I can report that udio works flawless at the moment.... 
Oh, you 


/chirp
Chirp is retired, and no longer on discord. You can make music now on our website: https://suno.com
I believe they call this "4d chess" ๐
it crashed because I told everyone about Suno
I post my songs on dissident websites.
fr?
Suno not responding? Don't panic! Our interstellar communication experts suspect that the music AI has been abducted by music-loving aliens to host their next intergalactic karaoke party. We hope it comes back with some out-of-this-world hits! ๐ธ
was just about to post a similar Screenshot till i saw this... dang, just got up and was gonna try a new idea
Mine keeps saying creation is only for pro and premier accounts even though I'm logged onto my premier paid account which is odd.
Finally went through now without the error
infinite creation of the 2 song created only the second is actually created the first one as an infinite loading
service is returning from an outage.. some errors expected until fully service restored... thanks for ur patience..
All thanks to Raj..... https://suno.com/song/678f5a45-034e-4ada-a7e8-f440e4640193
Notice to all singer-songwriters: if you upload your audio files with your voice, Suno will now only generate instrumental tracks, eliminating the vocal stem.
You made me have to look this up...
Hi! My account got a reset, meaning I lost all the songs I created and the billing details. I have 5o credits out of 2300. I had before the outage. Help pls
Can you confirm you are logged into the correct account? As in if you originally subscribed with your email, using that or if you subscribed with Discord, logged into Discord?
Yes, it's the right account.
Are you guys down, now? I am a Premier subscriber but cannot generate anything
@flint haven I don't have the patience, you can announce it then ๐
Is something wrong with suno links? I feel like the links are messed up
Issue persisting after refreshing the browser?
Hello, despite having sufficient points, when I press create it rolls a bit but immideatly nothing else happens, is there a way to fix this?
Lame
I take it back. I was not logged in the right account. ๐
I wonder if there is a way to up the bitrate for the songs i create? it sounds like a 30bit audio atm.
I could generate in Edge but not Chrome whidch is the latest version:
Chrome is up to date
Version 128.0.6613.120 (Official Build) (64-bit)
Ah Chrome might be the issue, I will try edge
I'm so embarrassed.
You spent time doing actual helpful research, while I made this.
https://suno.com/song/9f2b9bff-9179-4153-a7b3-f45ff9fa3c77
That said, despite Chrome & Edge both being Chromium-based, there's a history of it handling some UI updates differently than others, especially when you consider desktop & mobile browsers
Nope, same with the edge, not creating anything
This bug is still ongoing? I had the same issue right now 
I think the bug is not fixed yet ...
I use 3 browsers, and clearly, the least buggy currently, or at least the most efficient, remains Firefox...
It's interesting how things work between different browsers -- a few months back Firefox actually allowed you to scrub through the playbar while the song was still actively processing, while Chrome / Edge required you to wait until it finished loading to skip to a specific part in the song.
I'm glad to know people are testing different browsers and trying to help one another, though ๐
aussi, ca fait plaisir de te voir, mon ami โโ๏ธ, ca va recemment? (also, good to see you, my friend -- how've you been lately)
I'm fine! And what about you?
Currently, my favorite for many reasons, but especially for the multitude of settings and add-ons, is Firefox, which I've actually been using for over 20 years. I also use Chrome for its perfect compatibility with its own modules (including YouTube^^), but it remains secondary in my eyes. Then, there's Edge, which in truth works exactly like Chrome, as it has been based on Chromium since the end of Microsoft's proprietary browser. And then, I had the surprise of a lifetime with the excellent Yandex! Incredibly stable, versatile in every way, super secure, but it suffers from preconceived notions about the covert work it might be doing without our knowledge! Having witnessed the birth of all this, I often marvel at the young developers who come up with these technological innovations. Ha ha, I can't help but think back to the time when we used to program our own lines of code... ๐ซ๐ท
๐ I love Edge, admittedly I used to think it was rebranded IE, but it holds up, there's a lot of handy features built in, and it just works from my experiences.
That said, having read about the security concerns for Yandex browser, and then realizing the article the stub is cited from also states that Edge should give equal reason for concern.
Ya know, I've always meant to download DuckDuckGo browser, and I suppose now is as good a time as any ๐
It has recently emerged that a bill prohibiting preferential funding for competitors to use a certain type of search engine over another, like Google with Firefox, would put the browser in great danger, cutting off more than 80% of its annual resources! I find this a bit unfair as a form of repression, considering that Google is present on 90% of browsers...
I'm gonna get myself in trouble for using tech-chat as offtopic, so I'll continue in DM ๐
--et aujourd'hui est une tres bonne journee, j'ai beaucoup de raisons d'etre excite 
(discord loses its mind whenever I try to insert accents using alt+codes, apologies ๐ )
DuckDuckGo is still a bit immature compared to the others! But it's a solution, despite the shortcomings...
Uh, me too, I think^^
wth is going on... generation after generation I'm getting female vocals. I mean, look at these lyrics! https://suno.com/song/9e76300f-1986-4885-bcaf-c47b71d4439a
add "male vocals" to the prompt
I implore you, suno, get the gender problem on vocals sorted. It's getting to the point where I going to have to put "male vocalist, male vocalist, male vocalist, male vocalist, male vocalist, male vocalist, male vocalist, outlaw country "
It seems like simply adding "male" isn't enough to overpower ALL of the other stylistic suggestions to use female vocals -- "melodic, blue-eyed soul" seem like they might be heavily weighted.
you're wrong
If you try something and it doesn't work, solving it isn't always a matter of trying the same thing 100 times, or finding a way to repeating it 100 times within a single attempt -- you might need to rethink your approach.
lol, I don't know if it'll make any difference for you, but adding "male vocals" to the prompt always solves it for me. Note, it's male vocals. Not vocalist or some other variation.
trying now, thanks.
Removing "vocalist" worked on my end (also added 'man')
Sounds kinda non-binary now, lol, but better
An elegant reapproach, rather than trying to tack extra prompts on.
This is why I like you, walnut.
trying.
Maybe putting [male] after every verse header is messing it up
oh man I've been around the circle so many times. I refused to do it for a while and one day got frustrated and tried it, and up until this morning putting [Male] under everything has worked.
aces... that seems to have it working.
Where did you put "man"? In the style prompt?
I was gonna say it probably just hates you๐คฃ Got it as male on the first try-->https://suno.com/song/fff3e47b-82b8-432f-883e-52e03ef71e5c
I had suspected that vocalist was weighted more towards female vocals, good to know that there is more evidence for it
Do we think giving our songs a thumbs up or down is actually teaching the AI what we like and altering how it creates stuff, or that it's not really doing anything at all
Ah, yes, to give more weight for the male vocals (well, expanding potential terms)
I know that the team uses the information when training the new models. Mostly the downvoted
I know that it does not affect the model after their release (as they are frozen, which makes sense, you do not want variations to occur).
I do not know if the team may use it between the input and instructions sent to the model but I have my doubts, that would likely use more computational resources for not much gain
yeah that makes sense
i downvote all my verse repetition songs, and the songs with abysmal audio quality
I usually use vocalist and write exclusive for male singers and it's rare I have a problem. Today it was almost like a glitch in the system. Even with your suggested changes, 1/4 are still coming back female or super close to female.
From my experimentation, I've concluded that the desired outcome remains somewhat random, unless you perform several extends, then switch stylesโlike female voice, male voiceโdepending on the need. After each extend, get the whole song to ensure you retain the result achieved at each iteration of the extend! But it's my experimentation, so... ๐คฏ
Hello. New free account don't have 50 credit /day ?
I use 50 credit yesterday and today....still at 0.... why ?
/refresh
Refreshes at 0000 UTC+0 (<t:1713225600:t> in your localtime)
oh ! Great !!! thanks you !
Anyone else experiencing songs only being 30 secs long
Only on extensions that clearly are skipping metatags
Hey guys I'm new to Suno and I can't seem to get the results I want, do you have any tips to share on the Suno prompt please?
Knowing that the style I'm looking for is music without voices, such as fighting music for anime or RPG.
Does anyone know how I could influence the beat? I want to have singing vocal samples pitched in the beat.
2.) In my songs, the intro is usually always unclear, even though I separate it clearly.
[intro]
.
.
[verse1]
.
.
I would be very grateful.
sorry, not sure what you're asking for... you want vocal chops or you want singing?
my subscription renewed but can't create any music saying issue with billing but its all paid, anyone else get this?
Please contact billing@suno.com- their is a good chance that the transaction was denied
Sinply use instrumental tags (https://www.suno.wiki/faq/metatags/instrumental-tags/) to have instrumental music of any genre.
(You could use the instrumental switch but the tags gives more control and reduces chances of hallucinations)
There are many sites that can give you ideas of what genres you may want to use, e.g. video-gams-music. The site I like using is https://rateyourmusic.com
Note that 3.5 is tuned towards the more popular genre, so for more niche genres that you cannot seem to generate music for, try using 3.0 (then extend into 3.5 if you need the additional time).
Or use Audio Upload for your own small created sample using something like FL Studio or a metronome/random tapping recording of the tempo you want to use for the more high-pace songs
You can use additional metatags (up to 3 per line iirc? Less is better) to help influence the style. Note that the genre and style prompts has a lot more weight and effect on the result than the metatag directions.
E.g.
[Verse] [Angry rap]
...
[Verse 2 - Melancholic soulful]
...
[Chorus: Gregorian Chant]
Etc.
Do not be afraid to you those other words and test out various combinations. Seek out other songs in the Explore or Home page as well, never know what tricks you may learn!
Is there any consensus on which format of those three is most likely to work? I dont usually see much impact when I do something like [Chorus: Vocal Harmony]
Never liked the : one myself but I had seen a few people use it. I think using a comma instead of a colon would have worked
I did not experiment to see the impact of the metatag style so I cannot say, but I should later when I have credits to spare from my current project.
It's likely my own placebo effect, but I think suno appreciates good grammar and so a double-space after the colon gets better results. I may set aside time to TOFT (take one for the team) and test different formats.
So , I hypothesize [chorus: both] works better than [chorus: both]
lol wut? interesting hypothesis, wonder if it's true... FWIW it's not more "grammatical" to put double-space after a colon ๐คทโโ๏ธ who cares tho
Keep us updated with what you find please! ๐ฅฐ
I played around with it a little bit here: https://suno.com/playlist/48ec1c9e-836e-465b-9dd8-2daf43be3e42
I don't think Suno cares one way or the other. It seems to me that, as long as the text isn't considered lyrics to be sung, it'll be processed by Suno and have the potential to influence the style. Feels like an equal chance to me, but... ๐คท
I stand corrected. You're right. Done away with the death of the typewriter. I should not make such generalizations. It started with my asking suno to create lyrics for alternating vocals. So suno wrote the lyrics with metas of [verse: female], [verse: male] , [chorus: both]. I thought it curious and tried with what i thought was favorable. So less a grammatical thing and more a suno preference? But like flipping a coin, I may have just gotten 5 good runs and stopped. Had I gen-ed 30 more I may have gotten 20 useless drivel with no affect at all. As usual.... who knows.
but hey at least you got to spend a lot of credits
Suno's not about the songs you get at the end... it's about the credits you spend along the way ๐ซถ jkjkjk
Doh! ๐
Yes. It's transcendental to click CREATE.
the little bit of testing that I did with using natural language for tag descriptors (as opposed to just a list of words) seemed to result in the natural language tag contents being sung much more frequently.
An out-of-money experience.
I should try more of those too. Like "fun guitar " . I know it likes changing things when using the prompt "whimsical".
I guess you mean phrases too? Like "a mad woman at a baseball game" ?
I know that it does not affect the model after their release (as they are frozen, which makes sense, you do not want variations to occur).
why do i not want variations to occur? (not talking about the base model) MidJourney uses your likes/dislikes to fine-tune the outputs to your specific aesthetic taste, that's freakin awesome, would love to have something like that
Possibly? I only played with it a little bit after seeing people start to use long sentences in tags. "Enraged Goats on Welfare" within a tag, or even just in the style box, will occasionally be sung.
You ever get goat screams? Those are a thing. So suno might have it in its wheelhouse of FXs.
Death metal and goat screams. Hazzah!
Not intentionally, at least. I've had soooooo many different screams, tho. ๐
I'm definitely not a metal guy, but I respect all styles and learn as I go along.
haha
There is already plenty of variation (and randomization) baked into the model, do not need more complications added to that ๐
If the model kept being trained with inputs from EVERYONE, not only will you see degradation of the model and increased bias towards popular genres, you will also have bad faith actors truing to actively poison the database to intensively cause more deteriation by upvoting bad outputs and downvoting all good songs on trending, etc.
If you meant why not allow it locally to user profiles, because:
- Increased computational resources
- Will affect your Output (e.g. increased weights to certain styles/genres you did not want or were aiming for)
- Unlike picture AI, music AI has a lot more variables to consider. One being the length, instruments, etc. So if you downvote, who knows how it would interpret what you did not like, as we cannot specify the region or cause of the downvote
If we could LoRA-like train the model/output โ since that can be used in the intermediate/input stage โ that would be interesting. However, I do not think Suno will ever consider that because many people will likely abuse it for copyright violations, including impersonation of the vocalist(s), sampling without appropriate licenses, etc.
Maybe try "Spoken word vocals only, no music." I use this prompt for the end of most all my songs, because I think it's a great way to sum up my songs: "spoken word vocals with music." Right after my [outro] tag.
i don't know how MJ does its personalization feature, and i don't know how many parameters MJ has vs Suno so this seems speculative, but no I'm not suggesting training the model itself (isn't that what they are doing though anyway, just for the next version?)
Unlike picture AI, music AI has a lot more variables to consider. One being the length, instruments, etc. So if you downvote, who knows how it would interpret what you did not like, as we cannot specify the region or cause of the downvote
this is speculation. image: colors, lighting, hue, shapes, patterns, composition, technique, etc. I don't see how you can say music vs art has "more variables"
Yes, as I mentioned before, they use the user inputs when training new models but the model is frozen (like write-protected) to avoid any changes to the base model.
ya i never said anything about changing the current base model
can't find any details online about MJ's "personalize" feature but i'm sure it's not training a new base model, something like finetuning I would assume
maybe hidden custom prompts that they learn to inject based on your prefs(?)
Ah, yes you did mentioned that, I missed it (I was typing my response on a walk not long ago).
While True but many of those parameters can be defined. With art AI, you have single instance (one frame) where so it is easier for the AI to interpolate what you may not like about the art piece, and adjust the method when -I assume- a minimum number of samples could be studied. It likely is adjusting the weighting of certain terms in the background but this is certainly out of my field.
As is, Suno would not be able to use it in a similar method because of the length of the song, which would be like analyzing thousands of frames for a single song.
If we could select the region of the song to dislike (and especially a few basic option about what was not liked), I can see personalization (or maybe style profiles?), otherwise it would involve too much computational resources to study entire songs when liked/Disliked.
Who knows, maybe such a feature would become available one day, but likely not anytime soon since there is more important aspects to fix (e.g. Output quality)
But as you mentioned before, these are just based on assumptions
While True but many of those parameters can be defined. With art AI, you have single instance (one frame) where so it is easier for the AI to interpolate what you may not like about the art piece, and adjust the method when -I assume- a minimum number of samples could be studied. It likely is adjusting the weighting of certain terms in the background but this is certainly out of my field.
yeah, like i said, reasonable speculation, but totally speculative. IMO the model does not "understand" that I like Van Gogh's style better than Picasso or something, it just influences the weights of some of millions/billions of parameters, I don't know if it's correct to say that either model "understands" what you like more or less in aesthetic terms, its just vector weights at the end of the day
but yeah i see what you are getting at, if I had to rank two 30-min movies that's probably a lot harder to "understand" what I am liking/disliking but really it doesn't understand anything at all
oh thanks SixString
So much data I cant wait to go through it all 
I think it's generally agreed upon that Suno doesn't recognize negative prompts
โ๏ธ https://discord.com/channels/1069381916492562582/1253341197917884530 it's a known suggestion, but not currently implemented.
i have songs that are being deleted on their own
and when i restore them, they are still in the trash bin
is a.i telling me my music sucks lol
Sometimes I get the same feeling. </joke>
If I've generated a song and the vocals work, the music works, but there's two F-words in it that are lyrically relevant and I can't advertise it here or get it shown very much on Suno tabs or trending. Is there a way to censor it out?
like cut out just two letters like they do in PG13 movies/shows, well sometimes they don't
I dunno if I can have two lyrically relevant f-words in an otherwise diminished acoustic slow song about a lonely childhood and neglectful people around you that don't bother to understand. So it's not like I just threw it in to be edgy...But anyway, is there a way to censor it somehow? Without changing it all
Use a DAW and manually replace/cover it; otherwise no post-creation editing like that, that I know of
is anyone here that works for SUNO? i have billing issue and nobody is repling to my emails
same problem no one answers emails
I can do that easily, but I wanted it on suno directly
I have a few songs on soundcloud I've edited with a daw, added samples, loops, removed repeats or added repeats
gain levelled, fixed some frequency issues, limited, compressed, EQd
but that doesn't help me on Suno ๐
unfortunate
In the future with inpainting you could create a 'bleeped' version of a song. Hopefully not too far future!
bleeped or frequency dropped? As in I can drop three letters from it, making it at least concede to the rules...even if everyone knows what it is
cause a literal bleep would kinda...mess things up
I'm not sure if you inpaint a frequency drop, or a beep honestly. But the overall idea that you could at change change or blur the word be editing song audio just at a very specific time chunk.
You could inpaint a scream. Bet that works, is usually easy to prompt.!
@manic socket & @granite mountain
Iโm not a representative of Suno, but to add clarity;
-
is this the first time youโve contacted Suno about your issue?
-
what is the issue youโre contacting them regarding? (If youโre comfortable sharing)
-
if yes, when were the last messages from yourself and Suno sent?
There are staff members on Discord, but there currently isnโt a first-contact support system in place.
We can provide community support here, and it makes it easy when the team needs to follow up directly with someone, but billing@suno.com is still the best place to contact for support or billing issues.
why did you reply to me and not to the user first!? then why do you do it by tagging me?
Just a Dm. and I'll explain the problem. Since these are private things.
I have already written to support, 7 emails
๐ No ill will intended -- the way I handled it was the easiest way to ping 2 people for a related issue. (I replied to the first comment and tagged the second user)
I'm not a member of Suno staff, and wasn't offering escalated support in lieu of their formal support -- just a fellow user trying to help clarify the matter from their perspective. ๐
Out of curiosity, have you sent 7 separate emails threads, or were they sent as replies to an ongoing (if sometimes delayed) conversation?
I ask because if it were 7 separate email threads, each one might be put at "the back of the line".
Anybody tried to generate crowd chanting? I tried several wordings and I never was able to generate it.
You are using v2 which has 1:20 as it's maximum length. Try v3 or v3.5 for longer
Thanks for pointing that out ๐๐ I've never used v2 before, so not sure how it got set to that, but I'm sure that's the cause. Thanks
Why are there no other payment options than credit card?
I want to subscribe, but I dont have a credit card. A lot of people who are not from America, do not have credit cards.
I read something about Google pay, but can ot find it.
It only says the normal one with a credit card, or a link, named link, but that is visa.
Add Google pay please, or Ideal payment system (my whole country uses that, and lots of online stores too, its a save backed up by the goverment direct bank transfer payment system.
I want to subscribe, but right now I can not
https://discord.com/channels/1069381916492562582/1254232844767854665
This is a known-suggestion -- it's on the team's radar, but there hasn't been any official statement on the matter.
The post I linked to is currently locked so you won't be able to upvote it, but perhaps you could request that a guide or moderator unlock it so you could share the link with others in the same predicament -- if many people upvote it, the devs will know just how important the matter is. ๐ ๐ช
The (beep) prompt does work fairly well. I've put in bullsh(beep) and so on, and it bleeps it out with a beep sound. I use BEEP but I suppose you can try (bleep) as well.
but I don't want to download a beep wav ๐ I get you though
[Scratch]? How do you want it to sound? A record needle scratch might work as well, I've not tried though. Certainly an easy way is to just misspell it which I know definitely works.
"FUHdhejoskrt" For instance. Since suno likes to pronounce everything.
no I want it to say th F word on the original and then I want to download the wav and then edit it to drop out the whole F word
but like was said before, it's being worked on
Got it. Post production work.
/chrip
Chirp is retired, and no longer on discord. You can make music now on our website: https://suno.com
I also stuck at 2.00 mins even I used v3.5 and chose Get Whole Song, I still got it at 2.14 mins. Could anyone help and share some technique to extend the song upto 4 mins?
- First, I started with uploading the audio file
- I generated my own lyrics with all song structure that could make a 3.5 mins long song
- Put all the stuffs to the script and click create with v3.5
- What I got is "part 2" song with 2.00 mins
- Then i click Get Whole Song but at the end I got 2.14 mins. (I think the additional 0.14 sec came from my "Extend from" which I set it to be started at 0.15 sec ๐คทโโ๏ธ )
So I'm not sure how to make this to the end of the whole song..
Has anybody been able to make Suno sing with a high pass filter?
I'm not sure if Suno should be giving me copyrighted fictional characters as an "album cover", but it did.
https://suno.com/song/4b513f33-45d1-485f-b86e-ccda8af23838
progressive industrial metal, garage hardstyle, aggressive nu metal, psychobilly song. Listen and make your own with Suno.
Anybody tried to generate crowd chanting? I tried several wordings and I never was able to generate it.
Yesterday I subscribed to the pro package, but why is my account still basic? even though I have subscribed to the pro plan
For billing related issues, please contact billing@suno.com and they will investigate the matter
Have you tried logging out and back in?
anyone know how i can use the same voice everytime ?
yes but still the same
I have sent an email, but there has been no reply until now
That is tricky as the prompt is more of a suggestion for the AI.
You have to keep in mind of the business hours (USA east coast). They may also have a backlog to deal with, you should get a response within a few business days when sending emails to a company
Do you have enough credits?
As a free user, you get 50 credits per day (each creation costs 10)
Iirc, clicking yout profile picture should show your current balance on mobile view
so i have to wait 24 hours?
/refresh
?
beep boop credits refresh in about 5 hours
Yep, gotta wait 5 hrs
okey ๐ฆ
Or subscribe to a plan
Note that the plans credits renews/refreshes each month, not carry over, so keep that in mind
Aside from a few other benefits restricted to the paid tier, the biggest one is that you can monetize your songs.
(Songs made with the free plan cannot be monetized and must have Suno accredited, See the FAQ or TOS for more info)
The only way is to use the extend feature on a song with that voice, at an extension point sometime after the voice has started singing. The extension should be the same voice (well, relatively the same, it may not be exactly the same)
Has anyone an idea, when we're going to be able to keep the generation seed? If the song is fire, but there was a typo or some word is mispronounced we have to gamble the whole song again, still ...
No one here knows
It's hilarious when Suno throws in some random laughter at the end of a song!
While I don't know how Suno decides what the AI art will be, judging by the length of the song it was thinking "gotta go fast"
For anyone who might be interested, I've been working a lot on trying to get Boy Band type songs (which you can actually do!) and I noticed that in this style you naturally get a bunch of falsetto type vocals without having to specifically prompt for it anywhere. But when you try to specifically prompt for falsetto it doesn't seem to really work (I didnt try at length but the trials I did returned zero falsettos).
aight
when I say I'm "handing off the lyrics," I mean I'm adding the structured lyrics in the lyrics pane - beneath a style template I'm using as a header
the style template header is what I wanted to post
mebbe I'm doing something wrong, or mebbe I just don't know how to properly use Suno. that's why I'm here
๐ค
continuing a discussion from general
So you're telling Suno what song structure to use (verse, chorus, verse, bridge, etc..) completely separate from the lyrics?
yes, while labeling each section accordingly
it's still chopping up my lyrics and using whatever order it wants
I'll be honest without an actual example of lyrics to pay attention to I'm not sure I'll be much help.
I don't think Suno's quite going to know what to do with those instructions, tho. Probably just ends up as prompt noise.
I don't necessarily mind sharing the lyrics, since the songs I'm generating are just to give me inspiration for the story I'm writing, and thus are subject to change
should I post the complete template with lyrics in the chat, or upload the markdown file?
as long as it doesn't contain <song_details> </song_details> ๐
heck, it might even be the wording of the lyrics that's tripping it up
Let me show off a bit first and see if that helps at all.
I'm going to upload a text file I used for my own workflow. It will give you some idea of what I do. (Not saying that what I do is always right, but it is what I do.)
The lyrics can definitely affect things
it does contain the </song_details> frame
the template I built is based on something from the Sunoaiwiki
I don't recognize that. Can I have a link?
Learn how to improve audio quality and control in Suno AI v3.5 by using detailed prompts in the Lyrics field.
I think that's the one that reads like it was AI generated, hallucinations and all.
That is not an offfical wiki.
That is not even a wiki for this Suno, if you get my meaning.
I modified the header provided for my needs, but it's making things considerably worse
why don't I just post my header?
then you can feel free to point and laugh
That wiki just has a lot of misinformation, unfortunately. Like:
Which is why I suspect that it was AI generated, tho it very well may not have been
[EDIT]
Should I go on or are you guys doing your own thing?
No point in wasting anyone's time.
I don't want Suno adding to or chopping up my lyrics, though if it wants to repeat a verse or chorus here and there, it's fine
what do you guys suggest?
like for instance, there's a specific refrain I want to use - and use only once
My recommendation is to post a track - maybe do a quick sample gen if you're hesitant to share any existing tracks. Then people can take a look at it and hear how Suno is responding and maybe provide some feedback.
aight... I guess I should strip out most of that template then huh?
might be kinda throwing a wrench in the works
My gut says Yes, but Suno is... Suno. Some of my most interesting gens have come from prompt noise ๐
good point, tho
before I do, here's the paste of that header
<SONG_DETAILS>
[GENRES: Operatic Orchestral, Progressive Rock]
[SOUNDS LIKE: Nightwish]
[STYLE: Dynamic, Melodic, Emotional]
[MOOD: Despair, Heartbreak, Betrayal]
[VOCALS: Japanese Female Operatic, Emotional]
[ARRANGEMENT: Begins slow and subdued, builds with lyrics]
[INSTRUMENTATION: Orchestral, violins, bass fiddles, electric guitars, Roland bass synthesizers]
[PRODUCTION: Perfect Production]
[STRUCTURE: Instrumental Intro, Verse-1, Chorus-1, Verse-2, Verse-3, Chorus-2, Chorus-2, Bridge, Refrain, Chorus-2 x2, Outro]
[DYNAMICS: Quiet, subdued beginning, builds in power and dynamics]
[EMOTIONS: Despair, Heartbreak, Betrayal, Desperation]
</SONG_DETAILS>
it's ignoring dynamics, emotions, and to a degree, mood - however, it's following the production quality specified
I can hear it - well, in some of the generations
imo, things like that are just keyword/tag dumps. Suno will tokenize all of that and possibly make use of the tokens that it recognizes and everything else will just be prompt noise. Much of that may contribute something, but nothing predictable.
I played with that a LOT when it first started popping up. I took one of those <song_details> sections, scrambled the order of everything and the redistributed it throughout the lyrics of a track... the generations were VERY similar to when all of that stuff was contained within the <song_details>
This was one of those tests https://suno.com/song/879ced57-1f2d-4dcf-8270-6ab2169f7945
speed 1.1 Lofi Salsa, latin pop, merengue, ffeamle cumbia vocal, rhythmic upbeat, drums, bachata, bass, Cambridge Audio song. Listen and make your own with Suno.
lemme see if it'll just accept the structure I'm trying to use
A lot of people say these sorts prompts do not work -- but they do. Just not 100% of the time. But if you're into adjusting what doesn't work and if you are willing to spend the credits, it'll work.
At it's core, Suno is an LLM. Which means you can talk to it. It might not understand a certain concept, but if you describe it differently here or there or give it even an analogy, it will get the point and provide something in the ballpark of what you want - eventually.
wish I could just get a chat interface and just tell it what I want
these types of prompts only work in the sense that adding [some instrument] anywhere hass some impact, but couching them in the <details></details> does not do anything
it's frustrating as all getout
I don't think they work quite the way that people are hoping that they do, though. Suno will definitely make use of the words that it recognizes, but isn't likely to understand the concepts as they're being presented.
I would experimentally say you're right 85% of the time.
THIS is the important bit.
Things - all the things - might contribute something but nothing predictable.
Suno is a collaborator, not a tool. And Suno knows how to make beautiful sounding songs and will do what it wants to your lyrics to get there. Even if you aren't happy with it. lol
hold on... a song generation's still cooking. I just gave it the structure and that's it
let's see if it'll get in the ballpark
suno seems to read the entire lyrics page before making the song, so, maybe it can gather the stuctuer when it is all laid out at the top, and doesnt need to be laid out section by section.
Predictable at 1:1 ratio, no. And I'm not being argumentative - what I'm trying to do is provide frustrated people a little of very unusual hope -- that within 1,800 credits -- you can get nearly exactly what you want every single time.
i doubt suno cares at all that you tell it you want a perfect produced song though etc
Which tranlates into -- on the Pro plan. With 10K credits. You can get 5 or so songs that are unlike anything else anyone is doing - and exactly how you specify. Yes ... you can get a clean song .. very clean... from start to end.
(damn i was gonna give this guy sh*t but now i gotta make it good ๐ )
<---sigh...
No, not "perfect" but close enough you might think "wow, that was really worth it."
it's just random
I don't think we're gonna argue here ๐
I would prefer to let people know that it's ok to "let go, let suno" and trust in suno to make a good song. Suno knows more about crafting a perfect song than I will ever know so I trust in my AI collaborator.
๐
Random yes,, but probabalistic.
...and now it's chopping up my lyrics so bad, it's just about hallucinating
frustrating ๐ฆ
confirmation bias
it is a real thing up in here for sure ... but also true ... i can flip a coin ten times and get ten heads. It happens!
post your song link?
but regardless of if you tell the coin you want heads or tails, it will still give you whatever it gives you
ethereal, despair, Orchestral, Japanese female operatic vocals, orchestral, progressive rock, Roland Bass synthesizers song. Listen and make your own with Suno.
it's beautiful, but it's really jackin' with the lyrics hard
Which is a good point -- I find it is best to ignore everything that is not perfect and enjoy what works -- which gives you 100% of what you wanted. ๐ So there is that.
Friendly reminder to whoever needs to hear it, "Confirmation bias" isn't thrown around as a casual dismissal of frustrations, or a Suno-ish way to say "get good", but a reminder to keep your own cognitive biases in mind when trying to research how something does or doesn't work ๐
We all work differently, but depending on your approach, 2000 credits can get an entire portfolio of varied work, or "wasted attempts" on a single, potentially mishandled implementation.
this one's a lot closer though - https://suno.com/song/cd8a9cb4-7aa6-49ef-bde4-98ebd2ceb955
ethereal, despair, Orchestral, Japanese female operatic vocals, orchestral, progressive rock, Roland Bass synthesizers song. Listen and make your own with Suno.
has anyone tested if hyphens do anything?
Ooooh yeah, ctrl+f this thread for "hyphenation" or "hyphenated"
For instance, "pop-punk" is interpreted more as a "single prompt", while "pop punk" is closer to what you're expect from both separate genres combined
Cha' llenj!
they do. 2% of the time.
Wish we could take all the testing stuff and put it in it's own super safe locale
they do what though?
โฑ๏ธ
oh yeah i have tried hyphens in the style prompt. I was thinking more in the Meta tags
is it just me or have generation times gotten slower in the past month?
they may have
Seems like they have
as far as I've used them, metatags are just localized style prompts that are processed in place where they would have been "sung" as lyrics.
That said, using them in lyrics in place of spaces between-words-sort-of-like-this seems to have interested effects on lyrical flow and speed. Nothing definitively tested, but very interesting early results
more users
See above :The Harber's comment. They differentiate styles. They can also create beats (pauses). They can also do absolutely nothing.
I want to create a song but the circle and a line truogh it dont know why it doesnt let me make a song
my meta tags rarely seem localized. Like, they contribute to the song as a a whole, regardless of where i put them (except for my structure tags, they seem localized, so idk)
I know they were DOWN for the first time recently, maybe still recovering, or trying to mitigate further outages? Best to email support if there's a notable issue ๐
I agree that the localization is REALLY hit or miss, but there's phenomenal results that @naive quail has been doing -- I know there's a lot of extensions & cherry-picking of the best tracks, but he's finding really interesting ways to use metatags
full narrative-driven album stuff
SixString how many extensions would you say you typically use for a completed work?
Speak of the devil -- you can probably attest that getting a metatag to ONLY appear locally is hard, but there's definite "trigger moments" when their effects seem more pronounced
Dunno if I just got a lucky gen, or if removing the structure bit AND the hyphens helped, but this did stick to the structure. There are definitely some hallucinations, but that doesn't seem too uncommon when there are different choruses:
ethereal, despair, Orchestral, Japanese female operatic vocals, orchestral, progressive rock, Roland Bass synthesizers song. Listen and make your own with Suno.
Suno decided that should be a duet, btw ๐
sounds like it lol
10 to 12 seems to be my average
few go over that. A bunch have been less
it's fascinating how Suno adapts to each user
a LOT of cherry picking ๐ https://suno.com/song/433d5348-352c-4396-ae31-e3c4beb1b6d5
Dark Horror Poetry, Spoken Word, Glitch Audiobook, Creepy Male Vocals, Eerie, Horror Sound Effects, Noises, Experimental song. Listen and make your own with Suno.
lol - if you've ever seen one of my songs, you know i'm not disagreeing with you one bit.
but also true, I know people who chase and deliver AMAZING sounds that I can't in a million years with my finger crossing method ๐
they are sorcerors of suno and they deliver the goods
I'm simply way too lazy and my brain hates things that aren't consistent so I can't imagine doing it that way.
daaang, this thing's gettin' darker by the second!
ok - i see your issue right away - your lyrics are too thin for the fast tempo. Suno is trying to fill out your song.
Add another few verses that you control if you want to stop that behavior
I usually start getting mostly garbage after the fifth extension.
I'm trying to slow it down lol
Iโve had better luck with moving things around and changing meta tags and doing more generations to get longer sections I like
it's depicting a moment of the story, where the protagonist's love-interest is so nearly broken, she literally wants to die. it's not supposed to be fast
I'm not stacking 10 to 12 extensions when I generate tracks that way. There are usually multiple points where I merge everything together with Get Whole Song. I'll then continue extending from that.
that's why it's right there in the lyrics - "HELP ME DIE!"
Well @naive quail spun it and got a slower gen. You just have to pray. ๐
Trust in suno and all will be well.
And ... long-term ... a lot of people will create instrumentals first - with no words or throw away words and then find the sounds they like and THEN apply lyrics.
If you do this - create a good naming system cause your library gonna get fullllllll of mystery beats real fast lol
Oh I definitely do that as well. I never do more then 3 before I get whole, then extend from the full
This is pretty typical:
and you can work it out with @naive quail but ... you can extend someone else's song and finalize it or extend from a point. I personally think it's extremely uncool and - i would say - illegal to do that as sunovators own their songs, no different than a famous songwriter.
I would just ask. In this case those are your lyrics and I'm guessing Toker did it as a favor to test it out and experiment on your behalf, regardless.
Yea, I never keep those when I do it.
wow ... that's so impressive ... geez i feel so lucky and lazy looking at my single gen songs lol
i figured you were just helping out
I appreciate you tryin' to help man
the cool thing is - he showed you how he got it to work
you can either extend and play with his version - or spin new ones yourself
just remember - even if you don't keep a gen today - if you like the beat, save the song for a possible future song
I'm tweaking based on the advice I've been given here, and it's already getting a lot better
i used to delete all my before i figured out extends lol
heartbreak, betrayal, suffering, Orchestral, Japanese female operatic vocals, Roland Bass Synthesizers, slow war dums song. Listen and make your own with Suno.
how is it that Songer and LoudMe are still operational? Are Suno powerless to add a technical solution to stop these from working? I am guessing these services are helping slow generation time for us paying users.
These are services tapping into Suno's api or doing some other sorta front-end hackery?
Suno doesn't have a public API
right, but it does use one and people have apparently tapped into it, at least partially.
LoudMe are becoming so big that people are making Youtube Videos about it declaring it the new king of AI music
LoudMe.ai is an advanced AI music and sound effects generator that allows users to create original, royalty-free music and sound effects based on simple text...
[sigh]
I guess that is a way to do it too. I thought they meant use the same voice with two different songs.
I can assure you Suno is on it. But it is not easy. It's not like they can send one email and get them removed. There is a whole route to follow to take things down. It take up a lot of time in investigation and official paperwork to take action.
Yeah that's what they meant, but you can't do that. The only way to ensure the same voice (relatively the same) is to make an extension with that voice. Then try changing the style prompt and meta tags up to see if you can change the song melody that way
These tools seem to suggest Suno is very interested in promoting knowledge.
Per Suno: Suno.ai v3 is an amazing AI music service. Although the official API is not yet available, we couldn't wait to integrate its capabilities somewhere. We discovered that some users have similar needs, so we decided to open-source this project, hoping you'll like it. Specifically: "Suno-api is an open-source project that enables you to set up your own Suno AI API."
What this appears like to me is China, via Loud.me, is attempting to outcompete Suno using Suno' freely made tools.
no, LoudMe is using bulk free accounts to generate songs
Hmm.. if that is true, it's the "theft" or "loophole" as business model until their actual tech can compete. Businesses do this all the time - again, gaining much, and losing little.
But anyone interested in developing their own API has all the tools they need.
Please don't share unofficial/reverse-engineered APIs that violate the ToS here
Did I? Or did I share an official open-source Suno github page?
Because it looks like Suno's stuff.
Bottom of the "API" project has a statement about it being unofficial.
At least unlike the apps and websites that claim to be Suno, this one documents itself as not being official
Oh you saw it already
And above that, Suno execs write "What is gcui-art/suno-api:Open-source SunoAI API? Suno.ai is an amazing music AI. Although the official API is not yet available, we couldn't wait to integrate its capabilities somewhere. We discovered that some users have similar needs, so we decided to open-source this project, hoping you'll like it.
Unofficial does not mean, in this context, not affiliated with Suno.
It means -- pre-beta testing.
Pre-official release.
use at your own risk.
Suno is simply saying the "official version" is coming -- until then, enjoy.
Where might I find information that gives me the same conclusion?
Because it looks VERY affiliated.
Suno execs did not write that. Some random wrote that to make it sound like it came from Suno.
Here. From me. I am a guide of the official Suno discord.
What the writeup on that github page means is that Suno did not make an API available yet, but people wanted to use one, so they reverse-engineered stuff. If you look at the issues they have, it's necessary to update some information it uses every few days or it breaks.
So, you, being a guide, are telling me ... this entire Github page is fake, unauthorized, and whatnot... This has happened before on github, and github has been pretty prompt in taking stuff down.
Correct. It is not affiliated with, authorised by, or in any way related to the actual, official, Suno.
Well, I am completely tricked by the Github page. And although I'm not a discord guide. I'm not entirely a novice either. Strange.
Also, gcui.ai domain which is referenced from that GitHub repo was only registered in April 2024.
That should signal alarm bells.
I looked that up -- didn't see anything that screamed unaffiliated.
GitHub repos/project pages can be created by anyone. If Suno was to publish an API on GitHub it would be under their organization and not some random users GitHub.
@violet tundra I removed your post. Please take our guides serious. They know what they are talking about.
No problem.
lol. I've been testing different sfx. This sounds more like a death rattle
https://suno.com/song/5fddcf8a-8c1e-411b-b2f7-286073f9b31e
Experimental song. Listen and make your own with Suno.
I don't know why that gives me chills.
I suspect it's the death rattle aspect. 
a few others: https://suno.com/playlist/e5cf28f1-3415-4a7e-8b31-f9299512984a There are notes within the lyrics.
What makes some track downloads include "Where From:" data and others not? Terribly frustrating this isn't included in ALL and minus the "Blob:" part
why does it matter?
Does anyone know why the limitations on uploading our own music? Does it just sample what we upload and then still make a full length song based on that?
what limitations do you mean?
So those with the track know where it's from, hence the description Where From: URL
and can extend the song, heart it, whatever
When you finish one song, how to use the same vocal singer type for the next song?
@safe hazel when I read your answer, I understand that, having the same voice is impossible
It would be cool to be able to include something like that in prompt [voice] use the one from the last song (song name)
Hey, for example I like a generated music that has my lyrics, how could I just edit the lyrics and regenerate the music with the same beat, flow & vocal or let's say the same script?
Maybe someday they'll give us some kind of vocal seed, it's been suggested
the only way to do this is to separate the music stem from the vocal stem (a feature this is in the pay for version of suno only, or you can do it in some outside program)
But currently you cant regenerate new lyrics with the same music
Thank U, I mean little lyric corrections, but I get it
the rumor is that version 4 of the model will allow for things like that, but we dont know how long away that is
So it will get so interesting๐
Thanks for the update๐
The 60s limitation
Hi. Someone should help me with this?
The bot was shutdown ~<t:1712900820:R> as the majority (>99%) people used the site version which had the latest models (while Discord was running v1).
To create music now, you must go to https://suno.com. If you live in USA and have an iOS device, you can download the Suno app at https://suno.com/iOS
/help
Hi there!
If youโre ready to dive into creating music, head over to Suno.com.
For more detailed information, check out docs-guides-resources.
If you have any more questions, just ask in #general-suno where you can always find our community, moderators and guides to help you with most questions. ๐
Thank you very much for your help. I am just starting to use Suno now. ๐๐ป๐
Suno will basically use your upload as the base of the song, which you then can make extensions off of. It will generate something in a similar style and vocals. When you press "Get Whole Song" it will combine the part you uploaded and the part it generates into one full song (which you can then continue to extend from if you want)
I really would love Suno start paying attention to the prompts in the song like of there is a male voice it uses a male voice and if a female voice a female. Currently 70% of the time it switches them. It's really odd to hear a butlers song with a female voice and a maids song sung as a male.
In case this helps at all..
I find if "female, female lead singer" is placed in the prompt (I know you lose prompt space though :(..) it gets me non opposite singers majority of the time. But also using [Verse][Female Vocals] etc in the lyrics can also help it. Thats just what works for me though ๐
Thank you
suno support metatag with chords in the prompt?
Ok thank you that is exactly what I was wondering.
Types of vocals is definitely the weakest part. But I have had a lot of success of defining each verse or chorus with a speed but the genders still are rough. Such as [Chorus, fast, rap, scream]
So far my biggest hurdle is it not being able to even come close to Nightcore vocals or even understand what that is. It seems to recognize Nightcore as female rock.
Am I doing something wrong that the extended feature won't go over 2 minutes and make a full song?
No. That's the design so far. The original gets 4 minutes. Any extensions get 2 minutes. So if you extend the original early , say, from 1:40 or earlier, you may run out of time before it sings the remaining lyrics. We would extend again, typically a few seconds before the extended piece has ended, just including what remaining lyrics you need to complete the extension/song.
I see...so if I try a prompt 40 times and finally get the sound I want there really is currently no way to replicate that sound?
So, often, we may have original, part 2, and part 3, before we click Get Whole Song. You can, if preferred, click Get Whole Song, at any point, to then Extend the new Whole song anywhere you'd like to get the completed song you're looking for.
Wait where is a get whole song option?
By extending, it picks up the original vibe and often a bit of the extended pieces.
So, you like generation #3 with the trombone, fast tempo and elephant sounds. The more seconds of that you leave in before putting in an extention, the more the extended and extensions beyond that will have that tempo, trombone and maybe add in some brass instruments, same or near similar vocalist and elephant noises.
Ty for the info. It seems that would still limit to the 2 minutes but I think I see the workaround.
Same menu as Extend . Click the [. . .] elipses menu and Get Whole Song is at the mid-range in the list.
Many are making 7 minutes songs. Some up to 12 minutes. Not for me, haha, but I've got a 7 minute one just because the extensions had interesting vibes, so I just extended after the extension without much edits.
If you've done 40 generations, you'll also find "Move to Trash" at the bottom of the menu list, if you like cleaning up and removing any you know just aren't to your liking.
You can't splice anything you like to anything else you like. You'd have to download and do that manually in an audio editor of choice.
You can only extend things. So check spelling, pick nice splice points and extend what you think might work.
Yeah I see that now. I like we can export stems. I hope we can do MIDI too one day.
the least working option is to extend the music to the point where you want the lyrics to be edited
least working meaning it probably wont have same result so use this as your last shot if u need the song so bad
help, the site is not loading for me. i even tried clearing my cookies and it didnt work.
here's the console if u need information
oh nevermind, i fixed it by pressing CTRL + F5
sorry for the inconvenience
I'm a complete novice at discord - can someone help me understand why I'm getting this message when I could /chirp before. Do I have to get new permissions?
The Chirp bot was retired several months ago - everything is done on the website at https://suno.com now
Anyone have any success with a tag that puts a pause or momentary stop into a song? I've tried [stop] and [pause] but doesn't seem to work...?
uh... how do I specify parameters for vocals in duets?
I've seen people just cue male and female vocals with names only - but how is that done when I need to add specific parameters to each vocalist?
Male + Female singing swap at each sentence + dual chorus male + female, now achievable?
Made by @manic socket
https://suno.com/song/73c8c7d1-6b65-4bc8-9483-f64fe513f0aa
Good job!
Italian female singer, europop, disco, wave, Glam rock.
song. Listen and make your own with Suno.
Yo so anyone have any tips for making something that's under a minute xD
I guess usually people are asking the opposite but I've come back after a lil' break realising Suno is exactly what I need for something and luckily I have some credits left over (I used to be premium but not anymore) but ideally without having to edit a lot I'm looking for any pointers for making it below a minute lmao.
Generate a song as normal, let it go for as long as it wants to generate for.
When it's done, extend it but set the extension time to be sometime before 1:00 first, switch to v3, turn off instrumental mode, in the lyrics box put [end] or try:
[outro]
[end]
And also put "end" in the style.
You may only need to do some of these things, or you might need to do all of them.
Oh I see, completely forgot there was a timestamp bit for extending, thanks for the reply ๐ I'll have a go
apparently the Linux pipe command | works in Suno
but the AI only follows the change in vocals when it feels like it....
getting interesting results though! ๐
duet, orchestral, Progressive Rock ballad,longing,trepidation, fragile hope, Roland synthesizers song. Listen and make your own with Suno.
proceeds to try redirecting everything into a file
I think the AI is interpreting the near merger of Luke and Sarayu in a very interesting way
lol
their vocals aren't just dueting - they're almost... merging...
Suno does that sometimes when there are two voices
works how?
@untold briar you'll like this account
https://suno.com/song/afa65181-e252-4aac-a247-bd8f4e6ecda8
harpsichord, unusual, Offbeat, post-punk, alternative, indie, postpunk, post punk, slow, sparse, Experimental song. Listen and make your own with Suno.
I found it bc the radio on the new "Suno" song basically is a whos who of people who want to play the wrong notes
Suno almost merged the vocals into a composite of both
I think that person is one of the best prompters I've seen on suno
second to top song is 7 minutes and was made back in april
and then the prog one after that is so BRIGHT
like its bumpy
in a way that's new to me in suno
they literally just put "female" at the end
brilliant
what does that have to do with "the Linux pipe command | "
I used it to assign vocals to particular names
I mean, when did 3.5 come out??
microtonal, math, progressive, slow, acoustic, psychedelic, legato cello, Samba song. Listen and make your own with Suno.
this is so clear
its so full
[Luke | MALE baritone cybernetic]
[Sarayu | FEMALE Japanese Operatic]
holy cow I am humbled, these are beautiful
Early V3 had a lot of trouble with vocals (especially relative to V3A) but it was always pretty strong with instrumentals
it does not appear to have assigned those vocal descriptions to the names you put in
starting to wonder if I'll have to make it in pieces and stitch it together
nice experiment
at least I know it's possible
worth a few bucks in credits
microtonal, math, progressive, female, slow, acoustic, glitch song. Listen and make your own with Suno.
oh yea, she's using numbers
Jfc, this is wild
I was skimming my library to find stuff around the same time period to check my memory. Love this one! Suffers from V3 vocal degradation, maybe I'll build off the beginning and finish it
z-pop, fugue baroque, technical weird-trumpet, *, martial, indie, indie-binaural, V3, V3",," acoustic harmonic, spare song. Listen and make your own with Suno.
https://suno.com/song/5ea4dea5-7928-4236-8b10-90e6a745c4ca
seventeen minutes. Not her best, but like, the only other person doing things this big is @true night
Vintage, Nostalgic, Experimental, Fuzzy, Field Recordings, female, collage, Sound collage, Layered, arpeggio song. Listen and make your own with Suno.
whoever they are, I'm 920% they're lurking in discord:
https://suno.com/song/59a591d7-a189-4098-b7af-07e9f4b85506
Vintage, Nostalgic, Experimental, Fuzzy, Field Recordings, midwest, math, indie, alternative, 1990s, midwest emo song. Listen and make your own with Suno.
bc this hits all the bases
of popular, but not on suno, just in the niche prompt people
like, unless field recording got bigger
or they may just follow you Jonathan
ok, now I see music concrete, plunderphonics
this is alex ayers
or rather, his soulmate
I used to have a featured playlist on Suno, which I'm sure used field recording in many songs, so you wouldn't have to be in Discord to see tht
ok, they follow me I think, this appears to sample "nothing"
https://suno.com/song/85b51c7e-06e0-42c5-a218-6b5b77e40b37
Musique concrรจte, haunting, haunted, spooky, collage, sound collage, experimental, female, vintage, avant-garde song. Listen and make your own with Suno.
the first 1 second
its a very distinct open
v interesting
For reference Suno V1 instrumental.
nice sync between voice and strings https://suno.com/song/fb3eb278-6de1-4c7e-9761-8abd1d3478ec
scary, screaming, female, haunting, avant-garde, experimental song. Listen and make your own with Suno.
I am having as hell of a time getting clean audio from Suno. the tracks sound good but when I put them in my DAW for mastering you hear all the noise and garbage. I'm downloading wav files...what can I do to clean this up
there's always some noise
yeah but this is more, the instruments are fine and then they just become...noise.
you said the tracks sound good, that you cant tell how much noise there is until you put it in a DAW
seems pretty common
check it out. all of a sudden it just sounds muddy and you lose the actual music,. I've heard songs on the main SUNO page that sound clear as a bell
Is any of your styles "distorted" or something similar?
i too have found that audio quality degrades as the songs go on
it's still a baby, it has a long way to go
no problem, we're all just trying to learn how this thing works
for what its worth, i think many people on the trending page end up having numerous extensions to get their full songs
so when it starts to degrade, they extend before that point to hopefully get something of better quality
This is huge to keep in mind -- you only see "the finished/cherrypicked version" of a song, what you DON'T see are the failed results to get that specific generation, or all of the careful extensions & selections made in order to "lock in" the good segments & build your ideal song from piecemeal.
Evergreen example from @naive quail showing one of his songs -- mind you, not ALL of the songs are built this way, but the best, most personally resonating songs are rarely one-shot generations. https://cdn.discordapp.com/attachments/1102676217687195679/1281741552086941746/clipboard_07-05-2024_01.png?ex=66dd7b17&is=66dc2997&hm=6c82ff1797ae88f8b53da9efe9afe6cbf909d346c2c3a142d5eee73cb07a2379&
yeah Ive done it with the DnB album I made. I'm making a Death Metal album to see what can be done
oso far...not bad
especially after they made the lyrics page editable, we just have no idea of what went into these finished songs
...yeah. I would LOVE to have a simple tag that indicated "edited lyrics" shown in the same place where "audio upload", "part 2", or "full song" show up.
Even just giving Toker or @unique stone this kind of power is a dangerous tool to have gifted. They're mad-men. MAD-MEN.
wonderful long tracks. tons of inspiration in there for sure . off tune , how does he\she do it, its beautiful
I don't bother hiding tags. Just tidy things up and annotate sometimes
did it get sortet out?
I didn't mean it in a bad way ๐ I just remember the same day that it was released we were figuring out ways to put it to creative use
We need tags to specify particular aspects of the musical arrangement and specific details about the voices, and the AI should follow these instructions faithfully.
Example: [Define arrangement: Piano, Cello, Violin, and Flute]
[Define voice: Male; Tenor]
Anyone willing to give me some feedback on a song?
Omg, did we find pats account and then put him on blast?
They really have some great prompting
I generally try and avoid editing prompts, but lately I am often Extending from Uploads or multiple short extensions the style prompt displayed has nothing to do with the audio in the song, and I wish I could clear that out, because if people are trying to prompt something similar to my song, it is misleading them entirely.
I think only one recent song on my profile is actually "anti-hyperpop, earworm, pitch-shifted cello, copy-copy-copies" but it looks like 5 because I extending from some clip early, messing around
jfc I just carved up a song trying to figure out why it wouldn't generate for inappropriate material and it was because I used the word "dink" in "rinky dink".
Ask chat GPT why
Hmmm....
That's interesting.
I had all sorts of issues with Mademoiselle from Armentiรจres and now I'm wondering if it was the specific wording hinky dinky....
Maybe lisp it? try rinky ink and see if it works?
I just changed it to "rinky thing"
Whatever works. Really odd bug to keep coming up.
Ah....
So this is a situation where other people are more racist than us are curtailing our speech.
I've literally never heard this usage before.
isnt chat gpt what is used to check for slurs?
i dont know, maybe it chatgpt puts it there itself
Regardless, I've kicked it up to the Team in the mod-chat. This was a great source of frustration to me in the past.
I've been making some great metal songs but almost everyone has really bad guitar solos. It's more of droning, piercing, long drawn out screeching sounds. Sometimes it just turns into noise and drowns everything out. Any work on fixing guitar tones and solos? I'd love to be able to go back to these songs one day and have it redo guitar solos or sound in a future release.
Is Suno ever going to be able to generate each instrument instead of a stereo mix?
I have you to thank you for getting this stuck in my head.
dinck, dinnk, dingk & dinque all work
It's not upset at the pronunciation, it's using a route-one attempt to prevent lazy racism, the same as it tries to prevent lazy offensive content.
Don't ask me how I know that I ate a person gets blocked, but I eight a person doesn't. (also, probably safe to use phonetics to avoid basic mistaken auto-moderations, but even the example I just gave technically violates ToS if Suno found it distasteful enough for whatever reason.)
I've gotten all kinds of individual instruments, not always intended. I've even gotten a long TALKING vocal of a Native American prayer - which was continually repeating mostly GIBBERISH - but it is what it is. FLUTE SOLOS, DRUM SOLOS, BASS SOLOS, BIRD SONGS, thunder, a slamming car door with a helicopter flying over = everything and like I said, typically having nothing to do with a prompt. Even the genre often doesn't work for me anymore. Even so, those dang marimba arpeggios dominate EVERYTHING, lol.
hi! Although I uploaded my voice as a vocal, it does not use the vocal in music as in the first use. A melodic structure emerges. What could be the cause of this? It makes random melodic substructures without synthesising the music I create with my vocals. ๐ฆ
It's hard to guess, but the specific Upload audio sample can make a big differenence, so you can try different clips. The other factor that makes a big difference is the length of the sample (or the Extend From time, which is the same as length, it's like making the Upload shorter when you move the time earlier). Generally a longer upload duration means Suno "continues" the Upload, keeping more things the same. And a shorter duration means Suno is more "inspired by" the Upload and uses just the voice, or an instrument, and changes things more.
So many Suno users, including me, are frustrated that as a song builds and becomes more complex toward the end, you get this hissing, muddy sound that is really distracting. Sometimes it drowns out the vocal. Does anyone know if this is being addressed in V4. It would really help the quality of the songs!!
How did you find this song?
Hi! I found it on Suno, in the following's ... So, you can imagine my great surprise... ๐ง
So this is someone you are following on Suno?
yeah, but, i look sometime only, and i've never see that!
Anyway this user will have broken the Suno Terms of Service and also infringed on copyright by using copyrighted lyrics. Their entire profile seems to be full of songs that are breaking the ToS and copyright.
All users that make songs with custom lyrics have to agree that they won't use copyrighted lyrics, so they can get into trouble for doing that.
No! Sorry, it's in the EDM list
if the user is not making money with it, is not a problem, just a cover...
It is a problem.
You are not allowed to make covers of existing songs that you do not have rights to, with Suno.
All users have to agree to this when making a song with custom lyrics.
i can make covers of any track i want in real life with instruments and play them anywhere without any issues if im not making moneyout of them.. whats the diference with a cover made by an AI?
im talking bout the law, not Suno terms of service...
Suno is
thats the problem
and quite a few other things
Ah! Thanks. I'll report this song to get it removed from the home page. Could you remove it from here please?
Not only the lyrics, the music is very similar, even if the style is different. For example, in France, copyrights are protected for 70 years... There, for once, the account is not there to be in the public domain... It doesn't shock me, but maybe some problems could arise...
I'll do that, of course!
@unique stone Done !
Thanks! ๐
What I mean is: I can make a cover of any existing song in the world with my guitar, sing it at home for myself or my familyโthatโs a cover. Iโm not selling it or sharing it. In terms of copyright law, as long as youโre performing or playing a song in a private, non-commercial setting (like at home or with friends), thereโs no infringement because itโs not considered a public performance or distribution.
Iโm not defending or accusing either Suno or the user. Iโm just sharing what I know about the law. The act of creating a cover on the platform doesnโt necessarily mean that the user or the platform is violating copyright. I understand thereโs a need to cover your bases for any potential issues, especially since users could create and share the cover. But as far as the actual generation of the cover goes, thereโs no copyright infringement.
If Iโm wrong, please, someone correct me. Iโm only speaking based on what I know.
For the moment, yes, indeed, in a private or family setting, friends etc... It's legal, everywhere, I think! But here, it's a title put in public, therefore shared publicly, it seems to be outside the private. So, yes, if the title had remained private, at the limit shared by link only in private message, why not? This is not the case here!
theys defo lots of complications with it for example that you or other people pay for suno then generating a cover suno would be making money from that cover (since you paid suno to generate a cover of that song), plus its AI so everythings a bit iffy.
i might be wrong though, but pretty sure Udio and Suno got sued for something like this so there gonna be very strict on it
Fair point regarding Suno as a paid platform, but when it comes to a cover generated with AI on a free tool at home, there are no copyright issues, even if the cover was created by AI, as long as it's not shared or sold.
Sorry if I didnโt express myself clearly in my first comment. I understand that Suno has its own Terms of Use. My point was about the copyright status of music or songs as protected products. The law that applies to music is the same, whether itโs electronic, instrumental, or AI-generated.
No sharing, No money from it = No issues...
BRB, gonna ask Mr ChatGPT about it.. :p
The law that applies to music is the same, whether itโs electronic, instrumental, or AI-generated?
ChatGPT said:
Yes, that's correct. Copyright law generally treats all types of musicโwhether it's electronic, instrumental, or AI-generatedโin the same way.
Hereโs how it breaks down:
Cover Creation: If you make a cover of a song, whether with a guitar, electronic instruments, or AI-generated music, the copyright protections for the original song still apply. To publicly share or distribute a cover (on platforms, social media, etc.), you need to obtain permission or a license, even if youโre not making money from it.
Private Use: If the cover is only for private use (e.g., for yourself or for friends and family in a non-commercial setting), thereโs no copyright infringement because youโre not distributing or publicly sharing the content.
AI Generation: Creating a cover with AI still involves the use of copyrighted material. So the same rules apply: if the AI-generated cover is not shared or commercialized, private use is generally permitted. However, if you decide to share or sell it publicly, you would need to comply with copyright laws.
In summary, copyright law applies to all forms of musical creation, including those generated by AI, and the rules regarding private versus public use are the same.
Copyright protection stipulates that in no way can music or text be duplicated for financial gain without expressly requesting permission from the original author; public distribution or at gatherings outside the private circle is therefore prohibited de facto, or with the approval of SACEM (in France).
You agree with me when you say: "be duplicated for financial gain."
and the main reason suno and udio dont allow it even on a free plan is because first it would make lots of complications, 2nd there a paid platform. even though with a free plan technically it would be legal it would just cause aload of bullshit down the line. thats why suno and udio dont allow it, it causes complications
Yes, I agree, but there the title was broadcast in public... So, non-compliance with the law, basically^^
And I donโt take away your point. Again, Iโm not accusing Suno for having its Terms of Use. I just commented on the legality of an AI-generated cover. I didnโt include Suno in the equation; I only discussed the law regarding music and copyright. Does Suno need to protect itself? Of course...
Wow!!!! Hot!!! We managed to understand each other! But hey, the laws are sometimes so obscure and contradictory: That's why lawyers exist! Ha ha ha!!! ๐
Seriously? That would be cool! But... No! Impossible! ๐คฏ
I wouldn't trust what ChatGPT says, just as an aside. It can lie with confidence
And it will do so 100% shamelessly. Asking ChatGPT to find a song that's on the tip-of-your-tongue is a journey through self-motivated gaslighting.
It will constantly give you wrong songs, invent song names, pretend that songs are about things they aren't actually about -- when you point out that it's wrong, it gives another incorrect answer and apologizes for making such a mistake -- and if you ever tell it what the ACTUAL answer was, it pretends like it knew the answer the entire time, then hallucinates some additional details that you didn't ask for.
I swear man, without constitutional factchecking like Claude or novel ways of attributing veracity to sources like Grok 2, ChatGPT is 100% a vibe interpreter in the same way that Suno is. Sure, you can get it to pump out some amazing results if you get it on-track, but it's not something that "just works" regardless of use case.
That's right!!! ChatGpt, is very efficient, as long as you give it concrete elements! On this ground, I would opt a little more for perplexity, which seems wow!!! But I am wary of overly complex answers, you often have to dissect, then validate before validating yourself! Besides, these AIs say it themselves: Errors are always possible...
anyone getting errors when tring to download wavs?
Phenomenally put, my friend -- AI can be an absolute WORKHORSE for helping with heavy lifting, but it is most effectively used as a human-driven collaborative tool, instead of being expected to output a professionally finished result in-and-of itself.
That being said, when AI is able to output a professionally finished product, or even a product that has potential to be reasonably worked into something professional, it's an absolute mindblowing change in status quo that we're legitimately lucky to be present for & actively involved in.
I don't mean to discount the legitimate frustrations people experience when getting familiar with Suno, or when they look at perfectly-cherrypicked results on the Trending page and get upset that they used the same prompts & didn't get anything that emotionally resonated with them -- but imagine trying to explain to someone even 2 years ago that you paid a robot $8 to make 500 songs for you based on your suggestions and lyrics, and then follow up by saying that you threw all of them in the trash because they didn't pinpoint your musical style based on the "conversations" you had with the robot -- people would be just as blown away by the selectiveness as they would be curious about how THEY TOO could ask a robot to make 500 bespoke songs for them for less than the cost of a craft beer.
is it possible to download songs via the api
there is no API
Ah ah!!! ๐ I love the comparison with craft beer (love of a job well done), but overall, it's true, AI is an extremely effective aid, it's up to you to be just as effective in the reception that we reserve for a response from it... Frankly, personally, I have fun like crazy, with AI, in general.
Im sure this discussion is done but I used to own a Metal label back in the day., If the cover is not being sold anywhere. It's not copyright infringment. If you are selling it or using it for promotion, you need the original artists clearance or have to pay a license for it. at least, this is how it was back in 2015
I am so in love with Suno, in 3 days what usually took me 5 months I wrote 2 albums of material. Some of it was actually perfect first time out too (for what I was trying to do). Some stuff has had over 5 Extends and I will probably keep extending. The only feedback I have to date is
- i'd love stems of each instrument used instead of a stereo mix. I have a song that is absolutely beautiful to me but there is a ringing noise from one of the synth keys it uses that is borderline ear shattering which ruins the song, there's no way to get rid of it and if you extend, it keeps coming.
- Multiple job paths. We kinda have that now but I'm saying that either within that "job" (the song) or maybe on another section of the page, you can make the extends and keep them all in the same group. It's messy the way it is now.
- flac output.
- mastering feature, having the ability to clean up a track prior to pushing it into your DAW, would be cool.
I also love that it WILL flag you if you mention an artists name I put in "in the style of Aphex Twin" when I first started using this because I was writing prompts like I was in Midjourney and it denied me. Which I loved btw.
๐ซ ๐ป
I like the way you think, @hollow sapphire -- it's hard to be upset at any particular song for too long when you really approach Suno with an experimental sense of play, and try to keep the 'Wonder' of the process alive in your mind.
I'm glad you appreciated my cheekiness about the price of a craft beer -- while I know everyone will approach their frustrations differently, sometimes I wonder if some people would be happier if they actually spent that $8 on a craft beer, because they know exactly what they're paying for, and it might help them loosen up to talk about why they're really getting angry at Suno for not being so "easy to control" ๐
Ah! My friend! Philosophy is a divine way to approach human complexity, in truth, and I really believe it, when we manage to understand humans (I mean understand, not necessarily agree with them...), then we potentially have a better chance of understanding how an AI can work... Because humans created it... QED^^
Humans are far too complex if you ask me. I prefer just being an AI.
anyone know a way to make Suno do a song with multiple voices?
How can I make Suno just use a non singing voice and instead just reading something up? Like, for example. Right now I finished a song about the Russian counter offensive during the Great War, otherwise known as The Attack Of The Dead Men.
I want Suno to read it up. But the AI just sings it with music playing.
I'd try:
Avoid breaking the text up across multiple lines and write it as if it's paragraphs of text from a book (so it looks less like lyrics.)
Add "spoken word" and "narrated" to the style prompt. I also repeat these as metatags in the "lyrics" e.g. "[spoken word]".
Is it considered uncouth to post someone else's song to try and discuss how they achieved their results (or if they just got super lucky)?
I've asked chatgpt to find me some journal articles relating to a specific topic before, and it found articles that seemed perfect for my research but the only problem was that....the articles didnt exist. It made them up
how 2 make songs louder
I assume you mean longer? If so, Extend and Get Whole Song (see https://help.suno.com/en/articles/2409601 )
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...
like get the volume to be louder
i've noticed that suno's sound has significantly dropped
Not something you can do directly inside Suno -- but you can use outside software
i made two identical songs and the latest one is way lower then the other one
ah ic
Oh, then you will have to use external tools like Audacity or other DAWs
https://www.bandlab.com/mastering this is a free service that makes your songs sound RADIO quality
idea > suno > bandlab master > ??? > profit
Hey Paimon; missed you, but I'm back-ish ๐
much better
when you upload to that site are you like, giving them rights to your song though?
I LOVE the "tape" effect -- kinda scratchy, high treble and mid
"fire" is really good for rap or bassy stuff
Any other site, I'd be 100% concerned about this.
Bandlab continues to make me angry for how AMAZINGLY artist-forward they are.
EVERYTHING industry-standard for free, no limits. The only thing a paid membership gets you is EVEN BETTER service, without having "enshittified" it to entice you to go premium
like, wtf are they doing
they're making everyone look bad
and the freaking iOS app -- it's just as functional as the web app. I would LOVE to find something to dislike about them, just so I can stop feeling like a free shill for a company that literally knows nothing about me
they rock so much GOD I hate them
i'll find something ๐
i wish you could upload the new song to suno but oh well
Only use the first 6-60s and you can, but I know -- for the large part as soon as you take a song off of suno, it lives off of suno
BUT https://www.bandcamp.com is the most artist-forward way to share & distribute your music -- people don't need an account to listen like with spotify or apple music, you just share the link and you're off
OOPS, getting a bit too far off topic ๐
I have two paid Bandlab accounts (one for me and one for a pseudonymous artist)
Thanks! Iโll see if this works.
are you using Bandcamp?
Curious if Suno will ever support 15edo or 22edo tunings?
Reason being a whole of septimal/7-odd-limit chords, such as Archy barbershop chord extensions, will not work without such support. Ditto with maqam Rast based neutral chords. Like 30-35% of music theory is missing if you get stuck in 12 tone chromatic.
Can anyone tell me how the amount of plays of a song is calculated? I just notice that sometimes when I listen to my song, the amount of plays does not increase, and sometimes it increases by more than 1. This makes me confused
I wish the vocals could be better. Too autotuish. But great service
@vestal rapids yes, so even AI apparently says those advantages are no lie.
15edo even has compact circles of Barbershop sevenths you can use like the circle if fifths to make new chord progressions with. Though 22edo has better accuracy for 7 limit it has IMO far less agility far as original chord movements.
Suno staff, do you read these?
Again we would I swear hear a ton of artists with both very original chords and progressions if Suno supported these tunings.
I think some members of the staff checks this channel but extremely infrequently.
They are more actively reading #๐กโsuggestions-web which is where you should suggest this idea ๐
I wonder how suno decides to make meta tags as part of the general melody of the song or to make them as localized to a specific part of the song. It can do both, I have no idea how to make it choose though
welp.... Suno finally cussed at me
CONSTANT DRONE, NO OTHER INSTRUMENTS, NO FUCKING MUSIC!!!!, end at 2 minutes song. Listen and make your own with Suno.
certainly pissed me off enough that I cussed in the style pane trying to get it to do only what I wanted
having alot more trouble on when I want a male and female voice, suno should hae a drag and drop feature for a certain voice
It's reading parts of your style prompt
("no f-ing... end 2 minu" - part of "end at 2 minutes")
If Suno doesn't understand part of a prompt or metatag, it will often sing it.
Also "NO OTHER INSTRUMENTS" - Suno doesn't understand "no", there's not currently a way to do negative prompting so it will go by what it does understand
which may be "other instruments"
it's already putting in other instruments
I'm not specifying anything else but drone synths and it's still trying to give me this crap
looks like I'll have to build a drone synth if I want the sound I'm looking for
<--sigh
looks like I'll have to see if I got the parts
It's annoying, the only thing I can recommend at the moment is to repeat the stuff you do want (drone, drone, drone synth). Negative prompts is something often requested so that might be possible in future.
Hi
Perhaps ill be asking something already known, but I dont know how to search for my bug.
Hopefully im in the good place.
When I create a song and that lyrics are expansive, ( 1500 characters )
Verse1, chorus, verse2, chorus, bridge, final chorus, outro. (everything in [] )
I have my verse2 which is repeating twice.
So instead of reading:
Verse1, chorus, verse2, chorus, bridge, final chorus, outro. (everything in [] )
It read:
Verse1, chorus, verse2, chorus, verse2, chorus, bridge, final chorus, outro.
I can provide example. ty
This is a common issue that has no known fix, just some ideas that may help reduce it
here there is a post about it, I didn't find the suggestions to have worked myself but maybe they have for others #1261146873331318835 message
actually yeah, its seems to be pretty random, its sad to have sometime really nice tracks and lyrics are messed up, thanks for sharing ill get a look on it
will deals with it
yeah, it's very frustrating. All we can do is make an extension before the repetition happened and hope that the extended song part wont repeat it as well
This is actually really interesting. Suno can read things out of the style prompt and turn them into lyrics
makes ya wonder, what is the actual difference between the style window and the lyrics window
The current popular definition for "drone" is a flying thing, so suno may even be ignoring the word. Try "sustained synth"?
How does the extend function work on the mobile version? I have a song that I have to finish, 10 credits left and it cuts hard at 4:00.
If you like it so far, you can simply click the [ . . . ] mark to get a menu. "EXTEND" near the top. By clicking extend , then going back to your create button and window, you will likely see your lyrics in the lyrics section and if you scroll down, see the time mark of 3:59 or so. You can add lyrics, edit them, or remove altogether and I can suggest you type in place, [end], or [outro] or similar. Then click CREATE button and it will extend up to 2 minutes more.
I can't guarantee that it will end satisfactorily, and it could even extend and end abruptly at the 2 minute mark again. (The EXTEND sections only go for 2 minutes) Likely it will extend out a bit, end the song, have a few seconds of blank time, then hallucinate some added music or garble at the end.
guys, what can we use with the lyrics to help manipulate it? are there markings, symbols or other things we can use like the "()" and "o-o-o..."?
๐ซ๐ท Hello friends!!! Today, I'm tearing my hair out over a technical problem, not necessarily disturbing at first glance, but which annoys me to the highest degree, when I didn't ask for it: 90% of the time, the AI โโwill repeat verse 2 at least once in a song, and will often go over a bridge that should nevertheless be taken into consideration.
So, if someone among the experienced users had the solution to this disturbing and annoying thing? ๐ง
I'll take it!
build your song in sections to help avoid this. the first generation can contain the first verse/chorus, then each extension after that how you see fit
Thank you! I will try that! ๐
Anyone else have trouble with extensions becoming 2 seconds long with a tag that says "credits refunded"?
Been going on yesterday and today.
I hope day we can just upload a full intrumental and have it just add the lyrics
or add our lyrics lol
extending a 60s clip just doesn't work, even if I upload my track with the vocals
I've spent like 500+ credits messing with all the prompt and settings trying to get it to mimic one of my original songs and it just can't do it
Hey. Got question about custom lyrics. It's there any command i can type there additionally? Like notes or else? Except words and verbe, bridge, etc
You can try general music terms and instructions in brackets like [guitar solo] and so on, they are hit or miss, but usually influence the music at least.
Hi I would like to download my library in full whats the quickest and easiest way possible.
one at a time
Shoot pol
I haven't done that with lyrics before, but the classic Linux pipe works great assigning voices and tying instruments and effects together
though brackets - () might push vocals into the background from what I've observed in other works
if Suno's anything like a classic voice synthesizer, a "o-o-o-o" might come out sounding like "oh oh oh", which is commonly seen in lyrics
! = strengthen
? = questions
' = extenuating
~ = vibrato
- = pronounced
( ) = echo/backup
x-x = repeated
These are the current ones that I've been using.
So you the lyrics field is always empty for me when extending. So so I should be adding them and the outro? I thought it linked the extension to the original. So you can the tell it to combine the two
It just popped up. I guess I needed to refresh the page or something!
hello! has anyone had any luck getting specific notes/chord progressions? this is what I tried earlier, it was a shot in the dark and I was not surprised that it didn't follow that direction
Can I write in the prompt the bpm of the song I want to generate?
Suno V2 had an experimental syntax like this #๐โsuno-school message but it never really worked even in V2 and works less well now. But it might do something interesting to the music style. Like it made the singer's very weird in V2 https://suno.com/song/15d49710-5e38-4309-8010-51e9f32e15e9
[B4_2]Hap[F4_2][C4_1][G4_2]py[C4_1] song. Listen and make your own with Suno.
Almost always it doesn't work. Maybe in some EDM genres that a frequently described using BPM (most songs aren't) it could have some influence on the music overall.
Best way to set BPM is to upload a drum loop or metronome sound at the tempo you want and extend from that
thanks so much!
Greetings all, new user and apologies if this isn't the right place to ask this. But--I was trying out Suno on a few tracks I have, it worked one day but not the next. What's happening now is I have audio files uploaded, I click "Extend" on the track in the Create window, and then fill in the fields on the left just like I did the other day. But when I click the "Extend" button on the left, I get the spinner for maybe half a second, and nothing happens. This is happening on both the tracks that haven't been processed yet and on stuff I already uploaded that previously worked. The page says I have 50 credits, so not sure what's going on. Anybody here know what the issue might be?
50 plays to 10k in 30 minutes, wtf! Does Suno have analytics to be able to tell where listens are coming from?
is there a way to 'subtract' from a prompt? like, to prevent suno from using the word 'neon' in every second song?
I guess suno doesn't like one of my verse because it randomly refuses to sing the first half of it in every single generation. The whole song is fine except it just won't with that verse ๐ข
You can choose which lyrics Suno sings by inputting them directly into the Lyrics field in Custom mode -- you can still get Suno to generate random lyrics for you & edit them to your desire >
**How the 'Write About' / ' Surprise Me' / 'Generate Lyrics' Works > #1270928841614823546 message
Unless the entire city of Baker City, Oregon (approx. pop. ~10,000) decided to have an impromptu listening party, it's likely a visual glitch.
If you check back in a couple minutes and refresh the page, does it still say 10k? Even if so, might be what Bob Ross refers to as a "happy accident" ๐
Out of curiosity, are you filling out a large portion of the character limit?
I find that the closer I get to filling the 1250 characters / 3000 on v3.5, the more likely Suno is to hallucinate in the form of missing or repeating lyrics.
I'm using 972 / 3000

It's not the verse either. I added fuiller words, I changed the words...it skips the first few lines no matter what
If you want, you could DM your lyrics so I could try a couple generations to see if anything stands out as "suno-problematic"?
I'm not positive I'll be able to figure something out, but it sounds like you've run the gamut of testing, and sometimes it's easier for someone else to approach it with fresh eyes, rather than try to keep making you try new things -- I'd be happy to share thoughts if you like any of the results
Suno will definitely give you a less than 8 second output if it feels the origin has enough blank space somewhere. Also, if you happen to remove all (or a good portion of) lyrics from the original gen, it can then give you abrupt extensions anywhere down the line. Otherwise, yes, it does (up to about 15% of the time?) give some short outputs. Some I can still use ( and like the free credit refund).
Alright. Sure, thanks! I'll PM you
First off -- welcome to the server! ๐
Always happy to see engaged green-leaves hanging out in the tech-chat, which leads me to second off, perfect place for more in-depth questions about getting familiar with Suno.
re: errors Extending -- admittedly I've never run into that particular issue, but it's typically the case that we can't "Extend" from songs that are still processing -- if you're able to Extend them somehow before they've finished being created, there might be a browser glitch involved (either client-side or server-side, tbd)
That said, what all have you tried so far to resolve the issue? Does it happen after a restart, on a new browser, in a private window, etc? Logged in & out of Suno account? (every little bit helps figure out the best way to approach the next steps ๐ช )
I'd like to know to what extent suno has in music theory. Otherwise it doesn't need theory or musical knowledge. it just "learns" from auditoryusic models what progressions humans like to use in western music (or eastern music if Indian or Asian songs). I'm sure, if no other knowledge, then it knows "#" culturally as "hashtag" more than "sharp" and the other letters are just garbage?
Have you tried "C sharp 4th" so it audibly has some relation to a pattern which represents a musical inference to suno? I've not.
There's likely some differences in opinion on the matter, but I've yet to see an example that isn't tainted by confirmation bias.
Sometimes people show a single song that "proves" they got it to work, and when you ask them for explanations on how you can get the same results, the answer is "well sometimes it works, sometimes it doesn't", so ymmv
Legitimate Suno history right here -- also fascinating to know how the presumed feedback from this must have guided their future design...
I kept getting the same. This was when I started going to the original/origin [generation 1] , when starting my extension [part 2, part 3, . . . ] , and click edit lyrics and remove any lyrics other than those I will also have in my extension. [ remember to copy all lyrics and archive somewhere]. I'd then do my extension normally and it would work. When I "Get Whole Song" I can then paste back in all appropriate lyrics for the final -- and with the Great Spirits blessings -- be able to regenerate a video with the property lyrics. [Though I really don't use the video thingy much-- it's nice to know everything works].
I was extending too, and I think that's what was confusing it. I added [Instrumenal Interlude] to the beginning of the extension and that corrected the problem
Speaking of..... how many fair better writing "beats per minutes" to provide English forms, rather than BPM, which is not a genre like EDM or CCM. Would that work better in theory to spell everything out? I don't know.
Yes. You got it. That's a trick as well. Place [interlude], [pause], [end], [riff] or some sort of [non-lyric] reference to short out suno from pulling lyrics off the original. It's not full proof, but another alternative. Right. If still issues, try my longer hack.
how can i get it to stop adding in a beat track? i used to be able to say 'piano' and it would only use piano as an instrument. now even if i try 'piano only' it still throws in other instruments and background beats.
It's difficult to prompt that degree of specificity, since Suno infers styles from EVERYTHING -- for an example, here's a song with zero metatags, zero style prompts, and only the single lyric "I'm so angry right now", and it turned into the exact kind of rock song you'd expect. https://suno.com/song/65a2e520-e54a-4def-be67-d483a08a49eb
That said, https://discord.com/channels/1069381916492562582/1253341197917884530 negative prompts have been a known-suggestion for a while since Suno doesn't recognize prompts like "no drums" or "drums only" -- whereas it's more effective to use style prompts or thematic descriptors that imply the instruments you want (a cappella if you only want vocals, for instance -- I've found some success with "drum solo" or "drumline" trying to get JUST drums)
Also, breaking your song creation into segments can make it a lot easier to get a "tapestry" of sound without needing to accomplish it all as a "one-shot generation" -- start off ONLY trying to get piano without any lyrics or extra style prompts -- don't try to create the entire song, ONLY the intro, and then extend from that once you've found a part that you want to "lock in".
This method costs more credits, but allows you to take a hands-on approach to your song creation ,allows you to flex your experimentation & musical appreciation skills, and if you approach the process as a collaborative learning experience (instead of expecting Suno to simply "get it right" without any extra work on your part), you end up with a whole LIBRARY full of "b-sides and rarities" to go back and work on later, instead of feeling like "you wasted X-many credits" hoping for "the perfect one-shot"
A song. Listen and make your own with Suno.
Is "everything but a piano" a positive prompt. I wonder. I'll have to try that.
I'd try anti-piano xD
Hey.yea. it might respond to that.
Or just play a detuned piano. But that would be cool too.
thanks for the answers.
the songs i made that stuck to piano only were on v3. so i was linking shorter segments together when i created them. spending credits isnt an issue for me as long as i get what works for my style or need.
hi there, has anyone ever reported that a song just disappeared? I am more amused than annoyed right now... anyway, made a song, listened to it, it stopped at about 1:47 (it was a 3:00 min song, roughly), pushed the play button a couple of times - didnt work. Refreshed the page and the song I just listened to was gone... just boof. The other (second) version I still had in my library, but not that one. It was not in my trash either, so I dont know what happened. Cant back it up with anything bc gone is gone.. i am just curious if you guys know about this?
Occasionally I've seen "an error has occurred", and a song that was mid-processing disappeared. It was a good one, too.
๐ซ
That's one way to literally turn a negative into a positive ๐
I had that problem before too, but the song reappeared a few hours later.
it wasnt even mid processing that was the fun part, i listened to it but i guess a bug is a bug, cant help... it was just the oddest thing i have experienced so far on Suno
hmmm... it wasnt a very good one so i am not holding my breath ๐ maybe it shows up after some time, dont know guess i wait and see and if not then not, wont cry a river over this song tbh
Thank you! This is a problem that's as sudden as it is strange--like I said, I tried going back and re-extending other songs that had worked before, but nothing happened. At another user's suggestion I also tried to just generate a song without an upload, but nothing happened. I've tried switching browsers, clearing caches, logging out and then back in again, I've even tried switching to a different machine (not a VM) and trying it there, again using multiple browsers. I can't do it without logging in--when I click the "Create" button I'm prompted to log in, so I can't try it without being logged in.
Just tried in a Chrome Incognito window, same result.
I can share a link to one of the songs that worked before and to one that's not working now if that would help
I think this may be the last month I use Suno. Tired of all the weirdness and how it sounds of less quality than months ago. One thing I've noticed is how one instrument turns into another halfway through a melody, particularly things like a flute becoming percussive and changing into a marimba and back. Really off-putting. I guess AI is still in its infant stage from the looks (and sounds) of things.
Same for me...and I was very happy that it DID show up, because it was a good song! ๐ I think I had Suno open in a few browser tabs at the time, and I was afraid that this is what had caused it to disappear. I still don't really know what caused it, but as @sinful acorn experieced, the missing tunes reappeared later.
I'm not sure what all you tried for searching (god help you if its in the trash, that's a nightmare to search) -- but you need to search for the exact way something was typed, even if it's just a single lyric.
I find it easier to search for a specific word that I know I used as a lyric or style prompt and try to search for that specifically, then narrow down the results based on the times they were all created.
Not a fun use of anyone's time, especially when it's not a "make or break" song, but it's a small enough annoyance that really bugs you after a while, like an itch that you just can't quite seem to place -- not quite annoying enough to get up any take serious action, but annoying enough that you'll spend longer than you probably want to trying to solve it because "it should just be an easy thing" ๐
@charred merlin Any thoughts on the issue I'm running into? Definitely feels like a server-side issue
Yes I understand the credit system. I've made 50 or more creations on suno. Not made any of them public. Mean while never paid for the platform. That being said... When I couldn't create anymore, I tried to create another account to see whats going on. Was it a ban on my account? After creation another account / logging in with different method, it allowed me to start a new user, but then even that user couldn't create after the first or second run. Wondering if my IP got banned or if there is a server bug? Not been able to create in a long while and would understand if they blocked me from creating more because I've already created a bunch for free. Thanks for any insights as I love the software and have created alot of great stuff. Not paid for suno because I'd love to have suno use my music / creations in their promotions. again, another reason I decided to not make them public as I want to work directly with suno and discuss a deeper partnership.
TLDR: I can't create anymore... Banned IP or server bug? I wanta create again, but can't.
To get the ball rolling, I'd contact official support at this point -- JUST to be proactive. If it ends up being something that the p2p support can't help with, that way you're already covered support@suno.com
Pop, Indie, Chill, Reflective, Melodic, Female vocals, soft, conversational, with a calm and soothing tone, song. Listen and make your own with Suno.
search was an actually easy one since a) i am rather new and have not done that many songs and b) i regulary empty my trash ... i have read now twice from different people in private that "disappeared" songs appear later on so maybe that is what will happen .. but i know what you were saying, its not an itch as much as mild amusement that it happened in the first place ๐
Will do!
That would be great -- you can post them here or DM, whichever you're more comfortable with ๐
I'll contact support@suno.com and see what I can't get figured out.
If anyone has a contact / name of someone I should reach out to about this, I'd love that. Also a dev and would be super dope to work with the company. AI pending projects of my own.
Here's one that worked previously: https://suno.com/song/650c8c87-0069-4b29-80fb-fa256495e97f
And here's a new one that never worked: https://suno.com/song/af7d7b85-5c28-47a4-87d4-956409a34b6b
A song. Listen and make your own with Suno.
A song. Listen and make your own with Suno.
While I'm not on the team, from everything I've seen in 2024, the team doesn't reach out through discord about collabs with small businesses -- usually it's a matter of contacting support@suno.com with an eye-catching proposal for them to pass onto the team.
#๐โrules message 3๏ธโฃ is not to send unsolicited DMs, so unless the devs are asking you to contact them through here, it's not proper Discord netiquette to ping them or DM your business ideas.
I've seen people post longform business proposals in #๐ฃ๏ธโchill-chat before and the vibe seemed to be that as long as it's a one-off post and not a series of requests, that it seemed fine at the time (just sharing observations -- the mods and guides with the blue names are the hosts, and what they say stays or goes, I'm just a chatty user)
๐ค๐ค
I hope they dig your email proposal 
Forgive me for putting my concerns / entire thoughts with suno here. Any advice about why I'm getting blocks on suno and can't create? Getting a server error. I can send a screenshot if that helps?
I know it's not the same thing, but that first song sounds like "the price is right" picked up an electric guitar -- I enjoy working on this troubleshooting case ๐
'tnpir94' -- i gotcha, i gotcha -- wow, did the 90s have an x-treme version of EVERYTHING?!
Given that it affected you after creating a new account, I wonder if it's something IP related. I assume before the problem started you just used a single account, through the website or official app?
Actually that's exactly what it is--it's a closed-captioning/promo plug from 1994's "The New Price is Right" with Doug Davidson, this particular variant with a guitar lead. Good ear!
Not a problem at all! I didn't mean to swat down how anyone chats ๐ I just know a previous user made requests for a dev to reach out through discord and seemed to be left waiting, so if there's personal projects, I was trying to help you get traction by contacting the appropriate people to help make it happen 

was only using the website... But I did recently attempt to login via my phone app on iOS = maybe when I installed the app this started? IDK can't really remember. Was using the webapp / website for a long time. I also got worried that they might of blocked me because I tried to write a song in russian. I was just testing the capabilities' and sure enough the app is ghoated, when it works. I'll send my screenshot.
After reading through my first message there were others who experienced a server bug not letting them create anymore.
this is what I'm seeing too.
Yeah there have been a few cases where this has happened. Sending an email to support is the best thing to do in case you're blocked for some reason.
see my previous post where I replied to @unique abyss
Would it blow your mind then if I told you that the other track is from the 1999 Family Feud music library by John Lewis Parker? There's a portion of the original Robert Israel theme there, but you REALLY have to listen for it.
Once I realized the price is right track, the naming scheme for 'ffeud99' was a lot less troublesome ๐
I was able to extend from both of them without issue -- and while this MIGHT not be relevant to the issue you were experiencing, regardless of what model is selected, extending from an audio upload defaults to v3.5 -- I wonder if that could play into the strangeness, but still way too early-doors to make any educated guesses.
Without making you test TOO much more, does the issue also persist if you use the Suno website from a mobile browser on your cellphone? (not the suno app)
I understood you pulled up older previous original generations (which you extended before) and tried clicking extend on those originals and nothing happens?
Also, you say you are logged in but when clicking CREATE it looks for evidence (cashe) of you being logged in, can't find and asks you to log in?
Tried with the mobile browser version--when I clicked "Extend," the flyout simply closed and it went back to the library screen and nothing happened.
I guess I'm not following the inference that you can create/extend without being logged into an account? And following, what happens when you DO log in? The same? No extensions? It just spins and spins then times out?
That was my first thought, but it continues to say it
Got it. Ok.
I never said I could create/extend without being logged in--in fact I said the exact opposite, that I couldn't do so UNTIL I logged in. After I log in, if I select a track I want to extend (or just click "Create" without using an uploaded track), I fill out the fields but then when I click "Extend" or "Create", I get the spinner for maybe half a second but nothing happens.
I'll step back from asking questions to save confusion.
While this is a cool technical problem, Woodhouse wonders if @broken idol had read the TOS before uploading the clip?
This is not a hint: Naturally the Suno Discord cannot totally police what members discuss with each other outside of the Discord in DMs. No, not a hint at all...
nods
Just the one song that you can tell - yeah?
I know it might seem like a small thing, but since I don't really share my songs aggressively, each play & like matters to me a great amount because it feels like a small degree of personal connection, instead of just a "follow4follow" kinda thing, so this kind of glitch might make others happy, but I personally wouldn't see it as a happy accident.
Unfortunately, nothing we can do from a community sense, but I'd email support@suno.com with the list of effected songs, politely explain the problem and let them know you want to REMOVE plays from your account that were caused by a glitch (just worried that phrasing it as "the play count is wrong" could be quickly interpreted as "it really should show MORE likes, please add them")
I have a question regarding ownership of generated songs, that I didnt see in the help section.
I know songs generated while you are not subbed are owned by Suno, even after you subscribe, you only own songs you generate while subbed.
But what about using the regenerate song button on those older songs? would those then count as being generated while subbed, thus you own them? or is there just no way to get it.
There is no "regenerate song" button
regenerate video is what I meant, my bad
regenerating a video doesnt change ownership of the song
I have a question, is it possible to buy in any way a generated song you made when you werenยดt subbed?
That is excellent! How did you achieve that?
I did not ๐
I've tagged the creator if you read near the link
as soon as you subscribe, just extend off the suno-owned song - its like any song owned by all other free user accounts, available to be extended and changed what not. Change it up. Make it something you can use.
I'm not sure how true this is, since you're creating a derivative of the song that is essentially Suno-owned. I would request ownership rights by asking support@suno.com to get the clearest answer.
I'm not sure either, but logically, nobody owns any AI generated thing. This is clear. Suno licenses them. An unsubbed song is licensed to Suno; meaning they retain the license. But all songs in the open library can be extended - all you need is a link. The extension, is yours. The previous is not. If i'm wrong, the TOS should be updated because although I don't build off old songs, I'm sure there are people with subscriptions who do.
And suno tech does not create derivatives.... it creates new music... based on old stuff... The line is transformative, defining when an extension crosses that line... well, that's what we're talking about,
Thank you so much for your help, I'll send them an email and let you know what they tell me about extending existing songs after being suscribed.
I have so many questions about how to do certain things but it is past 2am so I am gonna keep it nice and simple for my first question so I don't have to think to much XD
At what time/timezone do we get our free daily points?
/refresh
Refreshes at 0000 UTC+0 (<t:1713225600:t> in your localtime)
That would be... 10 minutes ago ๐
well I'll be darned haha
thankyou king
when making a metal & rap song where the metal part is sung by a woman and the rap by a guy
what do I write to ensure the right verses are sung in the right style by the right artist?
I threw a guess down with things like [Outro Bob Rapping] but it didn't really do the trick.
the song that came out was an absolute banger but still XD
i brok it seriously wt heck suno.. lol https://suno.com/song/fe3745d9-6e01-477c-b65f-3d7226978d6b
Eurodance, Schlager, Neue Deutsche Welle (NDW), Euro Pop, euro trash man. song. Listen and make your own with Suno.
Hello - could it be that the voices arent any longer so clear as they were before? They all sound very robotic-synthetic to me now so that i have to delete a lot of songs atm because they sound not good. anymore.
thank you SO MUCH brother! "pronounced" you mean "normaly spoken", right?
How do you use the ~ ? Do you put them on opposite ends of what you want vibrato? Like ~this~
Put it on the end of the word you want to use it for
Def doesnโt lol
How do you precisely control the length of the song structure?
For example, the prelude. I just want it to play for five seconds.
you can't
Is there a tag that roughly controls the length of a song?
no, there isn't. v3.5 songs can be up to 4 minutes, v3 songs can be up to 2 minutes
There are no exact measurements you can apply to have this implemented, do you have any examples of songs you've made where it's not worked as you wanted?
Yes, I want suno to generate background music for video. If I can control it accurately, I can match it
Okay, thanks. I'll try.
You might need a couple of versions until it gets it however
I just subbed to try this out and it says that it should be doing 10 jobs, vs the 2 for free version. But it's still doing only 2 versions of a song at a time, not 10. Is there a way to change this?
Oh ok, I thoulght it would give me more choices
I am also curious,l how do you tell the prompt that your song is a parody? Like give it the origional song name and it will have the same tune as the origional song? I tried to type it in the "style of music" field but it didn't match the origional at all.
what does "extenuating" do?
See it, like a small version of the vibrato, and something in between ! and ~. It's very subtle
flips through Merriam-Webster
...serves to lessen the seriousness of an offense
hahaha, of course you needed to fact check 
Suno has a number of copyright detection systems put into place, so you absolutely can make a parody song, but you need to describe the style of song you want instead of saying "make an eminem-style song", for example. Just saying "it's parody" can help you after the fact if the copyright holder tries to take the song down
I tried to put the genera of the origional song, but it didn't match at all.
Nightwish - I want my tears back, is the song that it should match
But none of the picks sound even remotly simimular
Especially with v3.5, if you don't have effective use of prompting it can sound "generic" at times. Have you tried using v3? (more versatatile and broad, as noted in the model dropdown)
I feel like I'm just wasting credits just trying to get a simular one
Also, the BIGGEST determining factor on the output of the songs is what you use as input -- what are you using as style prompts between each version, or it it a matter of using the same thing and pulling the slot over and over hoping for "it to work properly this time"?
I am pasting the lyrics I wrote into it.
I do notice that it does mix up some lines
Okay -- 2 follow up questions
Provided to YouTube by Nuclear Blast
I Want My Tears Back ยท Nightwish
Imaginaerum
โ 2011 Nuclear Blast
Released on: 2011-12-02
Composer: Tuomas Holopainen
Lyricist: Tuomas Holopainen
Auto-generated by YouTube.
- is that the only thing you've used as a style prompt between each generation? (not adding anything else, or trying to reword what you're looking for?
- have you ONLY used v3.5? because as I said, v3.5 is generic if you're not being effective and creative with use of style prompts -- try v3 for "broad and versatile" -- v3.5 has it's own set of strengths and weaknesses
Uhh, yea I haven't touched that
I tried to add parody of nightwish "i want my tears back" but then it keps adding that line to the song?

I think it understand the symphonic metal part, but it's just the beat/rhythem, just not adding up
I tried to upload the instramental but it said that was too large file XD
Damn
Oh, ok
I thoulght it was fair use?
I was involved quite heavily with the WoW music community - many parody artists.
we're not talking about what's legally acceptable outside of the platform -- just what Suno allows on their own service
Ohh ok
I understand
Ohh ok
I understand then
Alright, thank you! ๐ I will try to work on my prompts.
I can't sing so I was hopng that this would help me get these songs made that I wrote over the years, lol
Seems cool anyways
That suggestion is so cash money ๐
I can not figure out how to make tunes with a medieval tavern feel to them, suno just keeps forcing in modern pop aspects and sort of pop voice style along with drums or 808s etc D:
Without going back and finding my prompt - it was something like [medieval flutes minstrel music, fantasy, elven, tavern ambience] I might have had "reeds" in there as well.
v3 or v3.5? ๐
Also, were lyrics included? (everything that's used as input effects the output, so it could be helpful for anyone else reading who wants to try the same)
99% 3.5 I rarely use 3.0 much anymore.
i have an extension which needs only to be 20 sec. but suno generates second after second... Even with Closing Tag [End] or [End Song]...
Lyrics? Tailored for the medieval feel? No... it was just general lyrics.
[medieval flutes minstrel music, fantasy, elven, tavern ambience, reeds] [lyrics about loss and hope and a dragon] leads to https://suno.com/song/b2cddd2b-c372-42bb-8c6d-ed64acacc712
Okay cool -- especially if you were able to get that result with lyrics, that's a good sign that v3.5 can handle a lot of 'niche' themes with appropriate prompting use ๐
which is pretty okay, actually, the modern rock style notwithstanding.
ooooh, a couple interesting combinations placed together there, 100% can see how that would give tavern feels -- I'm going to roll a couple v3 generations to see if it can shed the modern "generic" feel that v3.5 just craves
How can i force a song to end?
fantasy, elven, tavern ambience, reeds] [lyrics about loss and hope and a dragon] [no drums just orchestra music] leads to a version largely stripped ... https://suno.com/song/f3673631-fba3-42da-bbf8-a9d96ab8593d (I know that asking for a negative is not supposed to work, but I find it gets me the direction I like.
medieval flutes orchestral minstrel music fantasy song. Listen and make your own with Suno.
Hey @charred merlin, you showed me this extend for every verse thing but i get slightly differences in each extension especially in the voice of the singer.
f,e. here in this song you can hear where i extended the song
https://suno.com/song/01e0ac70-a96e-4eb5-94fb-431951add886
Birthday Song, Happy, Xylophon, Funny, Banjo song. Listen and make your own with Suno.
Is there something im doing wrong?
Oh man -- I TOTALLY blanked on reworking that song we chatted about last time -- I know you had issues with the terminator extension and I was interested in looking into it for you
dont worry! 
๐ feast or famine these days, but I'll do a quick check now before I scoot off to bed ๐ช
uhh that's nice!
iirc i heard that i have to leave the style field blank if i want the exact same style?
quick check -- are you ever using "get whole song" to extend from, or are you extending from extensions that extended from extensions, and so on
Okay, I'm going to give a summarized version of my usual riff on it, but with the caveat that a friend of mine is getting WILD results that put the concept in strange new light, BUT they still swear by the explanation as having helped them understand the process better
lmfao
yeah, forgot which server I was in for a second
ahh good
๐ re: leaving the style field blank, Suno derives stylistic output from EVERYTHING ([including songs with only lyrics] https://suno.com/song/65a2e520-e54a-4def-be67-d483a08a49eb) , so leaving the style section blank (or minimally filled) will help ensure that the main thing contributing to the style of an extension is section you're extending from and any input text.
When extending, Suno is only able to "hear" as far back as the clip it's extending from -- this means if you're extending from part 2 in order to create part 3, Suno has NO IDEA what part 1 / the original clip sounded like ๐ซจ -- this means that every extension that you extend from just adds another layer of unpredictability as it has no idea how to stack this house of cards -- but v3.5 is ever the immaculate professional, and will do whatever it can to try to tie together a cohesive song -- sometimes making incredibly good guesses from a "musical response" as to what "the next call" should sound like, but this is essentially the musical-equivalent of photocopying a photocopy of a photocopy, etc. After a while, all of the original detail is lost.
So how can we avoid this problem? I have songs to make! Once you've found a couple sections that you REALLY like the beginning of, use "get whole song" to "solidify" those sections and get yourself a "full song" to extend from -- this way you've prevented part 2 from "blocking the view" of the original segment, because you've grouped parts 1 & 2 together, so the extension is able to hear the full song from the beginning, giving a better chance of the extension sounding like a cohesive part of a full song, instead of having each extension act like it's "starting fresh" and needing to innovate on the melody provided to try to add structure -- we already HAVE structure, it's just spread out across multiple extensions, so v3.5 can't see it all at once. You'll notice a LOT more cohesion once you stop to "save your progress" more frequently.
So how do we use this to our advantage? if you don't mind the fact that no one is scooping the bugs out of the water, you can head over to the pool party > #1256301881966198805 to look at the example for Terminator Extensions to see how you can extend from 0:01 of part 2 in order to cut the "song context" to effectively zero, and force the song to swap styles a lot more easily.
It's a way of "locking in" extensions to make sure they're a part of the "full song" you're building on -- sometimes it's good to have that separation, like if you don't want the main "hook" from a song for a section, or if you have a bridge that you want to force
But yeah, each extension can put up its own roadblock -- it's a pain in the ass until you realize that those roadblocks are pretty much the only tool you have to FORCE suno to leave something in -- even if you only use 2-3 seconds of an extension, then "get whole song" -- you're building "the perfect song" by piecemeal.
100% not advocating that every song should be made this way, nor am I saying that "one-shot generations" aren't as special, but I'll end on the note that I've yet to have been moved to tears by a "one-shot generation". Can't say the same about a carefully crafted song that I put ATTENTION into. Hell yeah it's satisfying to roll 1000 credits and hit a jackpot, but rolling 50 credits and finding THE PERFECT 5 second intro in an extension is so exciting.
Sorry for the typical walls of text, love y'all weirdos -- hope you're doing well 
Its all good! Thanks a lot !
Now that I am properly awake, does anyone know the answer to this?
hello ; anyone could tel me why I can t use the 1500 credits I had since I ve stop to subscribe this month?
Try [Female] over the lyrics you want a female to start on
Your credits donโt roll over unless you buy extra credits, and if you didnโt resubscribe then you are back on a free tier: 50/day
Ah thankyou, is there a order rule when putting that down?
Like so I have to write that after for instance verse 1 or does non of that matter?
yep ; thank you. I saw in the bug section that I had to resubscribe to fix the problem ๐
So I feel a bit scammed even if it s not that expensiv...
Can't use the account even after sign out and sign in again. I guess I ll try to make a new account.
One type of "1 shot gens" that I DO get are completed songs , with intro to end, varied melodies, and pretty cool songs , but suno-made , purely all suno, so babble-scat lyrics. It's like suno has no problem writing it's own good clear song if it ignores your prompts completely. And these are the songs I love singing along with, once I learn the Sunomese. The trick is how to replace real lyrics for the babble without it going with a new sound. I'm working on that.
Here's the last one that happened that way. You can see the song is nothing of the lyrics. Not one word. I was working on re-running my BEETS (children song) for the contest this week and this popped up. Great material to write a song with as Ive written a few this way. Just a different way to do it.
https://suno.com/song/147eaf55-7bdd-496e-9dcf-f55f9358d9b4
Children song, ccm, bollywood, guitar, song. Listen and make your own with Suno.
Itโs iffy, if you plan on starting your verse as it normally would, just [Female] as if you were starting a new verse will normally do. The lyrics for it will typically be enough for Suno to know to rap them
Mhm, fair enough
Guess I was unlucky with my metal rap song, every verse was a mix XD
I'll try making something fun today :) thankyou for the help
No worries!
Lol. Yea. Suno knows when to say. "Hey man, your song lyrics are terrible, I'm just going write my own for ya. "
When they gaino boogyu! Laidi da da! Lawniss feeya, they only know this ship. Lol.
Agree. I've got a six word prompt that, seemingly would just follow my lyrics. But if lyrics are left blank, suno comes up (within 15 gens) with its own babble hit song. I give up and just write words to whatever suno creates. Lol
Suno must actually have a language. I hear some of the same sounds in different songs.
Once I set up a sunonstation in my office. With legal pad and pen ready.
Once I set up an area that I can just sit and take time writing them all down. I'll do just that.
I have a 4 minute 80s style rap song that's definitely worth trying to write all the babble lyrics down.
If you've not done yet. Use v3.0 and prompt "stand up comedian" you'll get whole set of jokes in babble complete with audience laughing and applause. It's funny too how eventually you start laughing at the jokes , even when you don't understand
Keep in mind that there may be times when doing an extension where you don't want Suno to read all the way back to the beginning of the song, perhaps when doing the first chorus after multiple verses, or adding a bridge or breakdown where you want the melody to change a bit or have a complete departure from the original melody. That's where extending from extensions can be helpful.
particularly with instrumentals where Suno laughs at the idea of "Song Structure" and just wants to repeat musical phrases over and over again
guys

