Does it do that anymore? I've only seen where suno creates the song from prompts, it will throw in what's genres it's running. And maybe only in 3.0. I liked when it would analyze itself and give the styles it was going to use, then I would try it that way on a next song. I haven't seen that lately. It seems to just copy back whatever I write in styles box.
#šāsuno-school
1 messages Ā· Page 78 of 1
Beside the final weights (which are a bit a general average for the whole song, it will also tell you the genres at any point in the song, so yes, it changes constantly. This song, for instance had one area that "religious" was ranked highest.
At one spot, hardcore was topping the list.
Extend not working. I get 2, 10, 18 seconds. WTF?
Any info about this from SUNO?
Look at this, none of these are usable. Garbage. Fix it please.
I wasn't even refunded for any of those credits.
Bullshit.
Refund are for anything 8 seconds or under. So extends can run anywhere from 8 seconds to 2 minutes.
If you get short extensions, it may be from a slightly silent point or completely void point in your song. It's suggested to find a spot a second or two back from where you were trying to extend. Listen to the music and choose a reasonable spot after a "section" but not at a break where there may be a drop in sound, like a bass-drop for instance.
Only run one or two extends, if all short. Then that's a sign that the timemark should be moved slightly and try somewhere else.
I moved the time 10 seconds before the end of the track when its loud and busy.
Its 100% not a quiet part.
I've been having this error with every one of my tracks.
10% of the time it actually works and give an extention that is about 1 minute.
Most are just straight to the trash with only 2 or 8 seconds.
Seems you're trying everything you know to do.
I'm just looking to see if other have had similar issues. If anyone know a solution, if SUNO has a fix on the way.
Because this is 100% broken.
Is that an instrumental track? Those don't seem to work good via Extend from my experience as well...
Completely unreliable. Yes, instrumental.
So one of the main features of the service, they are working diligently on a solution right?
Whatās the length of the unextended instrumental?
The bast track is 1:10 seconds, others have 1:30, 1:50. All don't work.
You can try to put in some Lyrics / don't put it to instrumental and put stuff like [Instrumental Interlude 1] / [Instrumental Interlude 2] and such in the track. That kinda worked for me at some point.
No real lyrics at all ofc
Are you using a blank style?
Hmm, interesting experiment. I'll try that.
I've started out using the original style, and it didn't work, so I tried blank style same random results.
free Palestine
Can you make one of the instrumentals public?
Short extentions often are suno thinking there is an ending in sight. Are th extentions garbage or are they ending the song? This might hint to what suno is trying to do.
@polar spade Hmm, I feel I'm choosing a part of the song that maintains the general vibe of the main section, but that it something to experiment with.
If one of the "bad" extensions is still a viable sound, then try one extension off of the shortest extention that sounds decent. We've called it a pivot point when we throw in an extension and then Extend off of that at 00:01 . But ive extended into 00:02 or 00:06 for instance to keep the style consistent. Its a thought.
It eats up 5 credits, but sometimes a quick fix to continue on. These pivot extensions can though send the instrumental in another direction altogether as it doesn't find anything to feed off of when split up like that.
Sometimes the change up is cool, sometimes odd.
Thanks I'll give that a try.
Much appreciate your attention.
Posting one up may help.
Oh yea. I'll do that now. Got distracted over here.
80s synthwave funk, Streets of Rage, 16-bit video game, high tempo, upbeat rock song. Listen and make your own with Suno.
As always, throw a line to support@suno.com as well to get things started in case we can't find a workaround in here as just members.
Ah great, I'll be sure to mail them.
So, it DID possible think as if the song was ending.
Not much I can analyze from that, but good to know. I feel the close it gets to the 2 minute mark, the more often (30%) I'll get a short "end this now" extension. Some short ones sound fine enough to use further -- from there, I'll extend off the short extension to hope to to fool it to go on.
That's about it for me. Let us know if things work out. Otherwise, start another song and get back to this one later perhaps of there was a hiccup in the system that put you in extension limbo.
soryy to bother mate i didint use suno for few days ,there came out a tag in my song ,what happen .i subscibled?
Can you give a detail of what you mean " a tag" in your song.
before the lyric there came out like "the song is made by suno .com for free"
Have a question
I spent a fortune making my songs.. to get the right vocals
Now ppl can make personas based on my vocals⦠is that not considered stealing ⦠I spent hundreds of dollars getting those vocals
I see. I heard of issues and confusions over using iOS apps regarding this. But there was comments last night, so I don't know personally what was done.
You've been a Suno member in discord since November. Have you been using the browser version since then?
i did use chrome on my pc
And I'm close to say you've now tried an app on your iPhone? How does a new song result on chrome if you go back to that?
This has been happening to a lot of people. It's related to people using websites/apps that have unauthorized access to Suno's API. Since it's happening to you when using the official website/app, you should report it in #šŖ²ābugs-web
Suno is basically cracking down on those unauthorized sites, but it sounds like their approach, while entertaining to most of us, may be a wee bit buggy.
There's actually a conversation about this very thing happening in the bugs channel right now.
yeah i found my ac was login on an unknown ip address 3 days ago
You can keep the personas and songs private.
Uhhg! I'd report the song bug to the bugs channel and contact support@suno.com about the potential hack.
okk thanks mate really appreciate!
Ohh.. my. Well as @naive quail mentioned head to https://discord.com/channels/1069381916492562582/1082775596767117312 and see if there is further info to help get this resolved. Sorry you have this happen to you.
Suno.com is a great site and it's good to see others are finding the REAL ....SUNO. perhaps use your chrome browser version of your suno account you've had for a month, until you get more information regarding the iPhone app issue.
Or if you've logged into a browser site which was no longer Suno.... of course.
Dude that means that nothing you make can be sharedā¦
I own the content of my song,
You would have to make a persona voice of your song, THEN you'd have to make it available, public, for other to use. Otherwise they'd be never know.
I can can reuse your prompts, but may not get your exact singer vocals that you got. So, I can't exactly steal your singer straight from your song.
I can only use your prompts. Technically, your lyrics are there as well, but I don't want your lyrics, so I like your style only, so I write my own song, and likely will get a different vocal, maybe male or female and so on. But YOU can make a persona and keep it private if you wish.
You donāt understand
Some one took an old song of mine and just made a persona out of my song.. no extensions no work
Is this intended ?
Now the song states that itās made out of a some one elses persona
I could be totally wrong as to how it works then.
I get you, brah if that's the case.
Look at the picture and the time stamps
Certain styles might get the same vocals though. That much is true. If they also used your lyrics, then that's dissing you.
Thereās guidance near the bottom of https://www.suno.com/terms on how to issue a takedown notice, as far as I remember
We are building a future where anyone can make great music. No instrument needed, just imagination. From your mind to music.
Or at least direction on who to contact
Look at the picture
Itās my song that the persona is based on⦠no own content
Iām not able to effect change about that, and Iām not sure anyone on the discord would be able to.
Sumoās Support/Legal teams would, though
I am confused is this not part of suno ?
Yes, but there isnāt an official support presence at the moment.
The Support Lead has been stepping in increasingly, but outside of community help that we can provide, the only official way to get help is through their official support channels.
If someone has infringed your content, Suno is willing to help you make the situation right, but approaching the matter through the appropriate channels is still important.
I see it's an issue. As @charred merlin notes, it's more of an intrusion thing which would be over at Sunos support. (See above).
As for answer whether someone can do this always, that's an interesting š¤ position. I'd be concerned as well. I'll look at my own and experiment a bit. Thanks for bringing it up.
Head to support and let us know what you hear. Hang back in here and help others, if time, to work this out. We're all just members in discord trying to lend a hand.
Hopefully this can be helpful šŖ
Actually, can we make Personas out of other people's songs? I don't see that option available
My fear, on the other hand, is that if I make my Persona based on my song, someone using it might get almost exactly the same melody as my original song. Who then has the rights to it? š¤·āāļø
Look at the picture
Itās my song that is made before personas
And the persona it self saying itās based on my song.. but sits on a 0 content acc
Persona's picture can be changed
I can also put your picture into my Persona, but that doesn't mean it's created from your song
The announcement doesn't explicitly state that you can only make personas of your own content, but if someone uses your content and you don't like it, there's ways of handling the matter that are more actionable than others.
Can you? I can only make Personas out of my own songs
Sounds like "only of your own content" is enforced through UI.
Again though, if someone makes a persona out of your song, there isn't anything the discord members (or feasibly even the Support Lead, who DOES hang out on Discord lately) is able to do -- they've set up dmca@suno.com for a reason.
A long time ago I extended someone else's song and from my extended version I can already create a Persona. š¤ It's a loophole hell š
Look again who made the song
Yup -- new tech is definitely the "wild west", but drawing the line between moderation and freedom-of-use can't be as easy to pen as we'd think
But does it sound like your song that it is supposedly based on? Especially the voice.
these new Personas, can I use it to fix lyrics in a song and keep the same feel as the original?
I've tried using Covers but for the most part the original lyrics get passed through instead of the changes.
I would bet that you have a better chance of doing this if you use the Cover function. I understand Person functionality as a way of keeping the voice between songs.
Dude
I canāt even make a persona on my own song since some one else took itā¦
Again who made the song that the persona is based on⦠you c it in the picture
Itās my song
Can you post a link to the persona here?
Can I may I ask: Regardless of how the persona is listed in styles, does the voice of their song sound the same as the voice in your song
Personas is a GREAT FEATURE. BUT:
When I create a new song, choosing a Persona make only small changes in the lyrics, a complete identical song is created (all new lyrics will be conpleItly ignored) can't do minor changes on the lyrics. The conplete lyrics must be conpletly different to get the Persona feature working. Im hoping its only a bug for now.
The answer could possibly be, yes, of course
I'm at work, so I can't listen to any of the two songs to compare just to let you know
If I write a country song and the only style genre, the prompt is country and someone reuses my prompt as well as my lyrics. The song's not going to be the same, but does it pass the persona and the actual audio of the persona's voice over FROM my song?That's my question
Unless it's been changed from the alpha release, there's no restrictions on how many personas can be made from a song.
š« appears on many songs that have restrictions (which some people have found ways around).
I've tried to explain appropriate ways to address the matter, but could you please explain what you would like to happen now?
I still have some doubts as to whether the persona he's talking about is based on his song, but we need a link. I can't find it just from the picture.
Let's assume that it IS his song -- what then? If the other user named the persona "Electrician's stolen song" -- what then?
We're still not able to do anything about it from Discord, and they'd be directed to contact the appropriate team (support / dmca / š¤·āāļø) .
Then we can't help, it needs to be solved by Suno's staff
If it is NOT his song, it's just a persona with his picture. People impersonate and imitate lots of stuff.
I'm not a horse, it's just my avatar picture š
I'm sure this is going to be a talking point moving forward, so I'm not meaning to say it's not worth discussing -- I'm interested in seeing how all of this plays out -- there's "acceptable community rules" regarding using someone's song as an extension, so I imagine Personas will follow
How the hell do you explain this photo I found of you, then?
I like styles that tend to get the same voice though. But that's rare.
I did a persona because of the falsetto voice all the way through the song. I thought I had something. Nope. Just a regular pop voice the next few songs I used with that persona. So it's keeps the general nuance, and if the style prompt of the song are similar, likely the persona, is my guess, will sound close as well. Again, I'll test opera and death metal and see what difference it makes.
I spent half of Persona credits on creating 2nd song š 99% of the samples had the same voice, but I got the same exact melody from the root song 3 times which worries me. I like the feature already. Even more than covers.
3 out of 100? That's not bad.
I changed the style a bit comparing to original. I don't think you need to keep styles the same.
As everyone else, I've been itching to create albums by the same "artist". The consistency will benefit when posting on media sites.
My intent is to also get into children songs, and KidsPlace Live. That's a special door to get through, that distribution sites don't cover, but requires a marketable , promotable album and not just one song. I'm not sure it can be done with a fake artists though. But many "bands" over the years have been one or two people with a big "band" name.
Hello, good morning, afternoon, or evening. I've been testing out persona and I'm noticing that some of my generations have some random 'Ducking / Dips' in volume in some of the generations.
I've did a quick search but I might have used words that don't match to what someone else has said. Would this be a technical issue, I didn't think it was a 'browser bug' per se.
It only happens every few generations or so. So, if it's a bug for now then oh well. It's not overly bothersome, just odd.
If any help, in 6 months of using, I've only just this week got a song (no persona) that oddly lowered its volume halfway through (not fade out completely) , and on an extension, raised it back to normal.
I noticed this as well, but only when I tried to replace a section on a persona song š¤
Are either of you using an 'odd' addition in the style settings? I saw someone suggest using 'vocals -3 to -6 db, music -6 to -8 db,' to help with quality issues. I've only notice it do this volume dipping with persona though and out of initial haste, I think I deleted the song's that didn't have that style addition.
No, but if you have only some sections with that you can actually try to replace that part and regenerate
That's a good point. I've yet to use in painting as I find there isn't enough of the generation I want to keep so I scrap it.
Man, playback on the site is terrible today. I Can only seem to listen to 5-10 seconds at a time
hey i never heard of Suno before, i downloaded an app call AI Song (for which I paid a subscription too) and one song generation came up with a "This app is made using the a free Suno account, don't use this app its a fake, use the real app Suno" is that true? I imagine this song was generated because of this. Glad I found Suno though if that is really the case that app has some bad interface issues.
Welcome home 
I was using it to put music to my lyrics for demos to figure out which sound I want, then using that to get the song produced. this seems much better
I am having the opposite effect, so far of the 20 persona generations I have done the generated songs are coming out extremely hot, like really loud.
Ive been getting this also
I find Suno just does that normally, especially with some bass intense song's like some forms of electronic music. I haven't tried using the 'vocal/music -db' trick on any of those genre's yet. I tried with hardstyle but Suno and hardstyle just doesn't work I have come to see.
I didn't know there was a trick for volume, is that the prompt for it in your last reply? I'll have to give it a go
I saw it in a youtube comment a few days ago and I haven't tested it out all the much. So, I don't actually know if it works. For me, I found the best method for keeping the quality of songs clear, despite it's over use of compression and limiting, is to limit the amount of prompts and to structure your lyrics according to the style and genre of song.
As an example for my current set of musical trends, I try to keep it to 2 genre type's that are generally some-what overlapping metalcore, post-hardcore, and then add the 3rd being the 'off / infused' industrial, doom, some form of edm, the same goes for the style of song, aggressive melodic and the 3rd being the 'off / fusion' soaring, roaring.
Also to Note. Metalcore has a variation, Melodic Metalcore, so I can remove melodic from the style and try out others. In this case as well, Roaring, would sort of fit the 'aggressive' style also, but it's a bit of a variation so including the 2 together gives the AI a bit more definitive guidance.
The next would be song structure.
Dozens of genre's use similar song structure, but Suno doesn't always follow the structure, and we as the user don't either. So, it's up to use to figure out the basics of the structure and help guide Suno to, well, 'perform' the desired structure.
In most case's Suno goes with:
Verse > Verse 2 > Chorus > Verse 3 > Bridge > Chorus
With the bridge / verse being interchangeable to some degree's.
For me, metal, and in this case, metalcore specifically tends to have a different structure.
Verse > Verse 2 > Chorus > Bridge > Breakdown > Chorus
Where as in some instance's, you may wish or find that;
Verse > Chorus > Verse 2 > Chorus > Bridge/Breakdown > Chorus
Might better suite the combination of style / genre you're trying to get out of the song. And, to better balance the loudness of instrumentation and vocals. However, Suno cranks the output and is playing the loudness war's like everyone else is and has been.
Pardon the story time.
Is there a way to keep the generation from wandering from the custom lyrics, it seems to take some liberties and readjust the location of phrases, reuse verses?
this actually might be my answer - if I labe those sections in the lyrics and stick with a progression close to that it will stay on point?
Almost anything you provide to Suno is a suggestion, and isn't guaranteed to be obeyed as "hard commands" -- using structure tags can help suggest to Suno that the upcoming block of lyrics should be interpreted differently, the same as adding anything as a metatag, but it's still generative AI that's actively being developed, and there's a large degree of variation and reasonable interpretation that can arise from even the best presented & carefully crafted prompts.
Even the lyrics themselves are suggestions, though certainly "more strongly suggested" than metatags or complex style prompts
Thank you Dylan that is very useful information and thank you for taking the time to type that all out š
Soory 4 late reply
I donāt have a problem with the whole using other ppl personas⦠the problem is that l canāt make a persona of songs that l made if ppl beat me to itā¦. So al the old content on suno is up for grabs . Thatās my problem. I payed a lot of money to get the perfect vocals and now other ppl can just block my use of ifā¦. And yes l can extend and make a new one.
But this is songs that trended and are popular outside of suno, now it looks like someone else made the vocals for meā¦. And othersā¦.
Hope that made things a little clearer.
Making a persona of a song doesn't claim it so others can't use it -- unless you're explicitly receiving an error message that states this, in which case, please share a picture
Hi, is it possible to explicitly tell suno to "speak" a certain part of the refrains instead of singing ? For example a spoken intro, before the actual song really starts?
Yes thank you. I have found that syllable and line count has a lot to do with the way it handles the pattern. It is obviously trying to pick the right things to go where it thinks it should be placed to hit right in the beat. Good part, match syllables and have an even line count or a (...) at the end of one that is odd. Still this only works sometimes. The Bad part is that you can't specify or predict the actual BPM to match up the syllables. However, I have found that if you don't give it space to choose that helps. Changing Tempo in between sections helps. Definitely using [Fade-out] [End] or [Instrumental Finish] [End] Helps
Yes [Spoken Word]
Also using (...) in verses can prompt it to speak that part.
You normally need to use the tag though
This doesn't demonstrate what I asked?
That isn't how persona works at the moment
It is fairly close to just a quick prompt. Any changes to the prompt after the fact makes more weird crap than anything else. Mostly very slow songs.
I had to pull a song out and bump up the speed over and over then push it back
I think the persona is supposed to allow for descriptive vocals without using it in style.
@charred merlin Clear ?
However, that is not how it is working right now
I was able to get some powerful vocals with the persona beta. It took some time to scrape it together, but it came out fairly well. Check it out.
https://suno.com/song/6eb4b0b6-7752-4d20-981a-3afe6c05df60
Epic Orchestra Alternative Rock Soundtrack, piano, duduk, sitar, shamisen, dulcimer, distorted guitar, Battle Drums song. Listen and make your own with Suno.
That was the best I could get going. It sounds good the persona made the Vocals strong and clear most of the time. Very powerful compared to before. That I dumbed down too.
All it actually does is save the prompt you used to a picture and layer the song you used beneath it. Making it play the persona song many times and others it will mess up the tempo.
"An error occurred" isn't the same thing as "someone else has already made a persona using this"
Have you attempted any troubleshooting methods? Attempting with a different name? Trying on a different browser/device?
Thatās the message u get
Al other songs no problem
I understand that's the error message you got -- I'm just saying it doesn't state what you're conveying it does.
Also if you try to make a persona of a song that you already (that you own ) have a persona on you get the same error
Suno likes consistency and unity. If your lyric's don't have a flow Suno will inevitable toy with your lyrics so the flow is maintained, I.e, you get hallucinations or artifacts.
There are time's where lyrics can be a little more obtuse with the flow or pattern and that is generally in the 'Breakdown' or 'Bridge' sections of music. Meaning, they don't have to follow a consistent rhyming scheme or pattern.
But, if your Verse sections don't follow a consistent rhyme or pattern. Suno doesn't like it.
Are you using the app?
Replied to me but I meant @mighty lagoon
i didn't hate what it did, i did have my lyrics labeled, and it appropriately repeated a section that still fit
Good catch -- this is a change between the alpha & beta releases. Though the UI still restricts personas to songs that you've created only.
While I'm glad that we've discussing the details behind this, I still want to stress that the only people able to take action would be the support teams @ support@suno.com
In the meantime, crop out the last second of your song & you'll have a brand new track you should be able to make a persona of -- that's the extent of what community support is able to suggest, really š¤·āāļø
They my friend donāt care š¤·āāļø
The only workaround this will be to make a persona every time you release a song and to take down all you content and only share the most current one⦠many of the ones that are almost always trending are already taking there music downā¦.
I think the official support team cares more than we do 
I donāt
Label the lyrics and make changes to the Tempo, the changes to the Tempo helps stop the AI drift that can happen. Keep your syllables as even as possible, apps like Lyric notepad will keep a syllable count.
Computers like numbers...lol
What is the difference between "reuse prompt" and persona ?
At the moment not much besides it just filling in the style
It is closer to cover in the way it works though
Persona uses AI to analyze the vocals and/or instrumentals to affect a similar sound in later generations (work in progress)
Reuse Prompt just enters the same lyrics/style of music to save you from typing it out again, but is an entirely new generation, not influenced by that particular generation, just using the same text-input
I think I am going to keep a note to copy paste about Persona...heh. The only Beta feature that I consider actually Beta on Suno is the negative prompt. That works because almost all generative AI uses negative prompts. Besides that, cover and persona I would consider in Alpha and Persona more like Pre-Alpha phase
In the end I believe the point of Persona should be:
- the ability to use a very descriptive vocal prompt to get the vocals you want from the 120 characters
- allowing for the 120 characters on the song using the persona to be used to describe the style of music.
Ok. So the known limitations are not documented in the release note.
In this regard, if you mean syllable and line count to match BPM. You can (sort of) determine the BPM. The more lines in a verse, the faster Suno will try to get through the lyrics.
1st work around: Use parenthesis to dictate your flow ' ( ) ', causing a separation from the 'Main', vocals into 'Backing' vocals. So the flow is consistent and it won't try to rush your vocals, if you have a BPM in mind
Example It's the most wonderful time of year (To go out for beer)
2nd work around: Shifting your lines into shorter phrases rather than a long phrase also has an impact. But, be careful when using 'joining' words like, ' but, and, if, when. Suno will connect phrases into longer sentences when you do.
Bad Example
Look in the mirror
And what can you see?
Good Example
Look in the mirror
What can you see?
3rd work around: Shorten your generations. instead of doing a full song run-through, chop it into 2 sections or more.
Example: 1st iteration: Intro > Verse > Verse 2/Chorus > Chorus/Verse 3
Extension: Chorus/Verse 3 > Bridge/Chorus > Chorus/Outro/Finish/Etc
Perhaps I miss understood what you meant. Apologies.
Yes, the length is huge. Since it would sound stupid to opera sing a very long verse quickly, it will break into a rap and many times because of the majority of rap it will attempt to use male vocals although that I consistently circumvent by tags at the top of sections
Alright, I read again. Yes, by using the structures I showed, it should help dictate to Suno where and when you lyrics should be used.
There is also the Pre-Chorus option. Which, when/if you do Verse > Verse 2 > Chorus
Suno has a tendency to take the ending lyrics of verse 2 and create a Pre-Chorus to flow into as a 'bridge' into the Chorus. Or, if the music doesn't flow well from Verse 2 and into the Chorus. It will repeat Verse 2.
As well, it's the same for Chorus to Verse 3 / Bridge. If it doesn't flow musically into the next part, Suno will repeat a section in a different style of the original part it is copying, to flow into the next part of the song.
If you don't mind and it works, then cool. but, I hope this let's you know why it is doing that.
Amazing thank you, I didn't know about Lyric notepad.. learning so much here.
this makes sense.. I should probably do some learning on song structure no musical background other than band in school ( and listening to a lot of music) making my own songs though has been pretty rewarding
It has ups and downs in suno lyrics, it is normally used for just lyrics so when you put your structure in it counts it and tries to highlight rhymes, Rhymer's Block also works well for helping you find rhymes in real time, but I don't think it has a syllables counter
The structure isn't really difficult, i have seen things very simple work
Like Intro Verse Chorus verse Chorus Bridge Outro
Rhymezone (website) is what I use for help with figuring out rhyming words and it has a section for words that have multiple syllables. But, it's not like that notepad you mentioned.
Yea, Eminem really changed the way many think about rhyming schemes...lol. I love to rhyme whole multi-syllable words with multiple words. Like in my Only the Crown song.
Faster and Faster,
Passed him and Passed her.
I think there is something wrong with my account, since yesterday all the generated songs I have had "SUNO" printed on the image cover, and all of them are being abruptly cut-off and the song stops at around 2 minutes, the last 20+ generations have been like this since yesterday!
any ideas what could be the reason?
Has your subscription ended, or do you have one?
Are you using Suno directly? I ask because many people have come here from other sites that used Suno and it has added protections now.
Am using SUNO directly, and my subscription is still active
Website or App?
website
Can you drop a link to a song that is doing this? We're documenting false positives for followup.
There is more than one, this is one of them: https://suno.com/song/c8b07a75-ead4-4044-b7de-379209d7b80c
Futuristic serene atmosphere, Spacey Lofi, xylophone, synthetic sounds, and space ambiance, 70 BPM song. Listen and make your own with Suno.
It's just images? Not audio in any way?
what do you mean?
I see you using custom are you using lyrics or at least a song structure with tags?
Many instrumentals get cut early
There are some users getting audio messages telling the listeners to visit Suno.com instead of using a copycat generator piggybacking on unauthorized API access.
This would be a "false positive" since you are a Suno user and accessing Suno directly.
@plucky gale The 2 minute generation problem seems like you are generating song's with the 'Extend' option.
Then again, instrumental songs don't usually go over the 2 minute mark either. only sometimes.
I'm not using extension of another song
The track had a nice sound. (I listened to check.)
I haven't made enough instrumentals to be sure, but I know that they get cut if you use custom without a structure.
Even if I don't use custom tags, it still getting cut-off like this https://suno.com/song/bd538200-96a2-4ea4-a8b9-ae5f48ae012c
Futuristic serene atmosphere, Spacey Lofi, xylophone, synthetic sounds, and space ambiance, 70 BPM song. Listen and make your own with Suno.
Even when I use normal lyrics it's still is the same
It can vary depending on your lyrics/style prompt, but the songs don't always generate the full length -- do you find that the tracks are consistently short between generations, even when changing the style prompts?
Yes, I even tried using the same lyrics and style of songs that were being fully generated before, they are having this issue now
Forgive me if this is a silly question. Have you tried logging out and back in.
Yes, this issue is happening since yesterday
I think my account is being flagged or something
This is just a quick example of things I do with custom instrumentals, ignore the tags in the middle.
[Intro]
[Soft Intro]
#&##&&##,
#&##&&##,
#&##&&##,
[Distorted Guitar]
#&##&&##,
#&##&&##,
#&##&&##,
[Verse 1]
[Synth Arpeggio]
#&##&&##,
#&##&&##,
#&##&&##,
[Bass Drop]
#&##&&##,
#&##&&##,
#&##&&##,
#&##&&##,
#&##&&##,
#&##&&##,
[Chorus]
[Drum Machine]
#&##&&##,
#&##&&##,
#&##&&##,
[Pulsing Synth]
#&##&&##,
#&##&&##,
#&##&&##,
[Verse 2]
[Percussion Break]
#&##&&##,
#&##&&##,
#&##&&##,
[Synth Arpeggio]
#&##&&##,
#&##&&##,
#&##&&##,
[Bass Solo]
[Break]
#&##&&##,
#&##&&##,
#&##&&##,
[Chorus]
[Heavy Bass Drop]
#&##&&##,
#&##&&##,
#&##&&##,
#&##&&##,
#&##&&##,
#&##&&##,
[Outro]
[Soft Outro]
#&##&&##,
#&##&&##,
#&##&&##
Trying to extend, slice, replace, etc, and it's just SO off. The lyrics it thinks are there, aren't even in the song, the part I highlighted in yellow is said in that selected section.
(which it's repeating, already happened)
That part is normally off. Use your own timer. And just know when thing happen by listening
Im keeping them all private.
In the "back end" so to speak, are the lyrics not on a separate audio track in sync, or is it all one merged blob?
Blob.
Uploading your own loops lately makes nothing but noisy songs. it's so odd, there is so much audio noise now it's not worth doing anymore.
Oof, that's significantly harder to work with.
I have never had an unsuccessful replacement when ignoring all the lyrics parts
(but not surprised)
It really isn't hard. Just look at the wave pattern
I was failing to replace a few days ago, but it was the LAST 30 seconds of the song, thus the instrumental extension (as adviced here).
Oh use extend
I've yet to execute now that I've etended and it's not the last 30.
There's a minute of instrumental now, so hopefully it works.
I say 'Blob', but it's hard to tell. If you ask for the stem's, the instrumental and vocals aren't fully separated.
When I have a bad ending I'll extend with the short end I want using [End] to tell it to stop and then crop any extra, after cropping then I will do "Get Whole Song"
And I do it all on an android tab on a browser...lol, I have to triple tap things just to get the red highlight working...haha.
Can't say I blame you -- if I want to use the song externally, I rarely even share the suno.com/song link
Change your (Chorus) and other's to [Chorus] The parenthesis generally tell Suno to put what's inside the '( )' in the song. So it may not be registering your [Bridge] and [Breakdown] Properly.
Are you saying to change from parenthesis to brackets for those?
So no parenthesis, always brackets, for headers?
Yes, parenthesis are used for backing vocals and other items pertainging to the song. brackets dictate what's to happen.
I wonder if I would have better luck just covering (this is a cover) the original, while fixing that.
It might change your song, then again it might not. Do you mean to make a cover of the song you are trying to fix?
The cover was close enough to the original that I was happy with it, it kept the primary wins.
It was quite close to the oriignal.
(all covers were landing in this territory)
Interesting, it isnt showing up in the Persona 
https://suno.com/persona/7dfd7288-1279-45a9-b5e1-24cbe8266a50
https://suno.com/song/33edf0fa-f38c-401f-8a2a-6097a4bb4e92
South-African English Female Vocals, Anti-Hyperpop, Deepstep, 404-error-808, Grime+Tearout, Happy, Anti-Pop song. Listen and make your own with Suno.
If it's a case of trying to tinker with a section from a cover song, that is the issue.
Have you tried making a persona from the original, and using the persona to make a cover, rather than use the cover function?
Nope, made this all pre-persona, just saw that.
More or less, I got a prompt that nailed the singing, ups and downs, style of saying the words, it just didn't get all of my lyrics. The entire goal has been to preserve that specific rendition, while giving it all of my lyrics.
Don't care how I get there.
I tried to do cover's to do just that, and I can't be bothered anymore. it doesn't work well enough and I don't have the effort in me to thumb up and down all the generations that try to become a cover.
I'll just take the lyrics and dish out a new song in a different style and hope for the best.
Generally, that's exactly what I'd do too, but I started seeing the new tools (replace, cover, etc), and wanted to see what I could do.
I get the chaos involved here, not complaining by any means, just trying to learn what I can / can't do.
Also, this one version is a certified banger
It cannot be lost to time and or technical frustration.
Tinker time 𤣠that's fair, that's when I got burnt out and lead me to my reponse.
And, you got the bug. now it has to be finished, lol.
Story of my life
I've spent over an hour on the phone with Netflix's higher up support talking technical bugs trying to solve a years old bug with them, with video, code, etc š
One edible, one problem, one company - eternity of problem solving
To clarify NOT because I didn't have a workaround or solution, it just "had" to be solved haha
excuse me i don't know if i should put this here or in the bug section, but am i the only one who's cover generations gets shorter and shorter each day? like, last time i checked it was always stoping at 3:43 or so, now it seems to stop at 3:13
I would reccomend my changes with the () to [], and using the extend function. But, set it to go back by 1 or 2 seconds more, or forward. It might help. Also, the brackets might give Suno better direction.
also, i haven't tried personas yet but i don't really understand how they are supposed to work, especially how i have to write the description. does it have an impact on metatags or is it just for aesthetic only?
Does any know if it takes a while for a new "persona" song to appear on the persona page?
Just changing to brackets on a fresh cover (of the OG that worked) seems to yield pretty good results, I got my lyrics at least.
I'll have to listen closer, but I think I heard a few "blips" where it almost misses 1/16th or 1/32 beat or something.
Both in music and lyrics, might be a Cover bug
Keep in mind that there are no real hard rules when it comes to prompting for Suno... https://suno.com/song/77e41229-81bd-4317-ba7e-e8871b31678f
Crooner, Exaggerated, Dramatic, Flamenco Guitar, Cheese Sticks song. Listen and make your own with Suno.
You've just proven to me that I'm using "certified banger" all wrong.
There is a new bar.
You mailed it. I release this stuff so I don't want it out there. I make public the songs I've passed on. they aren't bad they just aren't what I was looking for.
A cover won't always be the same length. Different interpretations will have an effect. That's why when a musician does a cover, it doesn't have the same length.
In the case of Suno, if my generations we always the same length, I would start to think the 'Cover' feature was just re-creating the original over and over again.
As for Persona, The idea (I feel) is to keep the same general style and genre of a song you created.
Example You make a song and really like it, you create a Persona or a Band/Artist from it.
Now you have a persona for that style and genre and will generally sound similar so you could (theoretically) create an album and it would share a theme and feel.
Where as if you just make songs and throw them into an album, it would be more like a compilation album, rather than an Artist's or band's studio album.
well i assure you that the cover generations always stop at the same timestamp, litteraly midsong. i can use the extend feature of course but that's not what i'm looking for right now. i'm taking advantage of my last free tryals to experiment a bit
A person just mentioned a little while ago that their song's would abruptly stop. You may be experiencing this issue as well.
ok but what do i write in the escription? do i write the genre like i do in metatags with comas and stuff or can i go with something more free in style? does it matter?
The description doesn't really matter. It's a stickynote for you or others to use to ssay what the style of the persona came from.
Has anybody made a "persona" song and got it to appear on the persona page? It looks like it only lists the "root" song, even if you make a second one.
It's because it's showing the song you created your persona from. You would have to make a new persona from a different song to get a new Persona on the page.
You misunderstand my question
Fair.
we need super-likes for shit like these.
ok that's what i needed to know, thanks :3
I made a persona from an existing song. Then I used the persona to make a new song. The persona page is still only showing the original "root" song and not the new song.
Right, because it's a new song. Not a new Persona.
You misunderstand my question
From what I understand, the persona page is showing the song from which it was generated
So the persona page will only ever show exactly 1 song?
hey guys, keep getting the please wait a few seconds before generating more lyrics error, for 2 days now, how do i fix this?
But ignoring the highlight lyrics part works right?
create a playlist for your character
https://discord.com/channels/1069381916492562582/1259902708765167709 š
This happens when you click the "write about" button with more than 1248 characters -- you can fit 3000 characters into the finished lyrics and click "generate", but you can only use up to ~1248 characters when clicking "write about"
I assumed the persona page would show every public song made by anybody using that persona.
What's the purpose of the persona page listing the number of songs, if only one song will ever appear there?
pretty sure Suno team's cooking some serious cake behind curtains, wait and see
Ok, so it's not done yet then
The purpose of the persona page, in your library. Is to show the Persona's you have created. Not the song's created by it.
My "Persona" Has the voice i want... but no matter what i change, lyrics or settings.. instead of changing those it just keeps making the exact same crap and it pisses me off that SUNO takes credits away for bullshit copy when you clearly change things.. it does not even follow the damn original lyrics.. randomly jumps and picks randome parts of the text
Hey man -- do you mind if I shoot you a DM? š»
This still doesn't explain why a number of songs is listed. This feature wouldn't be here if it could only hold 1 song.
I don't see a list of song's. Just the image of the song that I created the persona from.
Can you post a picture, please.
If someone creates a song using one of your personas and makes it public, it should show up on your persona's page.
Right now i think Persona is garbage and it clearly does not work at all.
Why would it count the number of songs if only one can be there?
It isn't the best, but you can get some really cool clear belting vocals out of it then extend from early on to get the pace back up
here's an example of a Persona page listing songs created by other users: https://suno.com/persona/f25fdacb-6e1a-446e-a2e1-348e17a55004
Cover and Persona both need a lot of work, but you can see the possibilities
Ah, then I think it's what SixStringtoker has said. If your persona is publicly available. If someone creates a song withit and it is publicly available, it will show up in that page
So it only adds more songs if somebody else makes them?
AND sets them to Public
But you can only make one song per user?
I used the persona from that link to do a test song, but it's not public, so it won't show up.
I'd be surprised if there's a limit
I just want to save the voices so i can use the same ones in other songs, and the style i get so that mabye i can use the same for an album.. but it wont do any of that.. right now it only makes copies and not a single change no matter what you do
It's only showing one song for me, even though I made a second one.
I know what you are talking about but it can change. I can show you. I'll drop two links here and remove the embed so it doesn't take up space. One sec
Ok, I think I figured out the issue with the persona page. If you do additional edits to the song, such as extend, crop, or replace section, it will no longer appear on the persona page.
Only original generations directly from the persona will appear.
You use persona as a cover option?
Yea, I saw someone report that yesterday
No. But they work the same way kind of. The persona definitely tries to poke through. So you have to cut it and extend the new vocal style you get from it that you would probably not get otherwise. Not as clear at least
Honestly I don't think it should be public yet. If it were me I would have fixed cover first.
Thanks. Hopefully it's something they fix.
Let's be honest though, Suno is amazing...lol. Yea new stuff ain't working the best, but what is happening is magical.
I've never heard so many global voices saying what they need to say.
I think the cover option should be something we can do with other people's song, which may have been the intention. I don't often look at other peoples music. And with persona, we are making a 'band' or 'artist' to generate cohesive song's to formulate an album. As for the public option, for others to try and do the same with the Persona we make public.
I agree, there would need to be a remix tool in place to fill the gap left by cover on your own songs.
And that there, might be the problem. A cover is not a remix.
Exactly, but it is used for that
And this is why there is a fuss about it š¤£
Yep...haha
I think the issue is that "remix" is used too liberally atm, being used to describe any alteration
Similar to how "AI" simultaneously means a program trained on a hyper-specific task like image generation, but also means self-reasoning computers like AGI
Oh, don't get me rolling on that...haha. AI has become the name for chatbots...lol
AI has been around longer than everyone in this chat
ok i'm back after testing covers on other songs and i can confirm that they are getting shorter and shorter. i'm starting to thing that the longer the original song is, the shorter is the cover generation, cutting midsong. and it's sad since there are pretty good renders i have to toss
Not at all, I was out walking zeh dog.
Of course by self reasoning you mean self correcting or adapting though right...lol
tbf, I think some of the musical terminology needs to be reapproached in the context of AI-gen -- by the logic that a band can't cover it's own song, are we considering the points of separation to be each individual song, each user, each style or "theme" that artists use, or all of suno as a whole? Can Suno cover Suno?
I need a drink or two for this.
Yea that is all we were saying about it. Naming.
self-reasoning == AGI, like what OpenAI's endgoal is. Not currently feasible, but still described as "artificial intelligence", which also technically describes the logic behind the opponents in atari 2600 pong
The overuse of the term AI muddies things bad.
Yes, that is what I mean
I find it to be more, 'Algorithmic Interpretation'. Not artificial intelligence.
AI is tossed out by news and everything else as if it is new
I diverted into that tangent, but we were relating it to how "remix" is similarly muddied -- people consider reusing eachother's style prompts to be "remixing"
As a glorified garbage man (1800GotJunk, part time job), I literally corrected a client yesterday in her own home about the news spitting AI.
Haha
She genuinely loved the clarification that we weren't about to have self-thinking cray cray bots.
The news loves to make stories.
And to that, I liked the lyrics of the song you were playing with.
LLM and the like are SO far from that concept, it's not even funny. The only reason it gets used, is because the END PRODUCT is deviously close to uniformed, which is to no fault of the end user, but spreading the concept that they're remotely related is just hype.
Is it normal to post a WIP anywhere?
"There's rudimentary AI controlling the enemies in video games, but they don't hold grudges and think about you after you beat them -- don't worry."
I'd showcase it, but was trying to finish - "finish" is turning out to be harder than I thought haha
This is the 3rd time I've ever spoken in the discord. no idea.
It's my second š
As a game dev on the other side of the trash man coin, I appreciate this.
@neat nexus Sorry to bring this up again, but just to confirm, I have reported the issue to support@suno.com, do I need to do something else now or I just need to wait now?
I have to explain all the time down here in the dirty south. "No it is not a self-aware or sentient being, and you have AI in everything you use, all the way down to a power adapter that tells the charger to quit when it is full. What they are talking about are chatbots that use predictive algorithms to predict what word is the most probable to come next. That is all. As of now"
If you emailed them then yeah, waiting is the way to go now. Bear in mind that it's coming on the weekend though.
Even then, "AI" in video games is SUPER simple. You assign objective(s), order them with importance (can be altered in real time based on conditions), you tell it what it can do with its world (advanced ones can "figure" this out based on underlying capabilities), and they just go do that. Jumping over a ledge in a 2D platformer is as simple as "Can I walk here?" "If not, what happens if I don't?" "How can I overcome this with my array of abilities?" "do"
i just answer but to link my own issue, since i seem to be the only one with it should i report it to suno? since nobody seems to answer me here...
Does anyone still have the bug? I change the lyrics a lot, but the āoriginal songā is sung instead of the new one
Using persona?
Yes
Yup
I have not had the Persona use anything from my original song at all.
Excecpt maybe a guitar riff, and even then it was different from the original.
It tends to happen when you leave the description it puts in there or one similar to the one from the persona
Not sure what your issue is, but if it's a false flag like me I think it's better to report it aswell, it can only be fixed from their side!
That is why I keep saying it needs to be reworked. Not that I can do it myself, but as it stands, it seems like it sits underneath a new song influencing it so much that it pokes through some times. And definitely is jacking with the tempo.
That's because the Persona is carrying over the original style. if you change the style it will have issues.
When I don't that is when the original can pop up. I'm surprised it hasn't happened to you if you aren't changing the style.
i don't think so, my issue is that every song i cover seems to get shorter and shorter, cutting abruptly midsong. i tryed on defferent songs but the generation as always shorter. like the longer the original song is, the shorter is the cover which makes no sense for me at all. and i'm talking about static timestamps which don't change at all. it was stuck at 3:13 earlier, now it's stuck at 3:08
I didn't use Cover songs much, but I had this issue aswell, I think it's a generic issue in cover songs
This is getting confusing. I keep a few of the original prompts because i want a selection of songs to sound like they came from the same 'Artist' and only changing a few so the new song's I generate are different song's, but share the same genre.
Such as; my Persona was made for metalcore and metal. So I don't try to make electro or house. it doesn't fit the persona I am using.
Ahh I see. Either I read it wrong or they said it wrong, but I believe the description was vocal style. Let me read again real quick
If I use completely different lyrics, there are no problems. But if the song is too similar to the original, it ignores the new lyrics
Weāre excited to introduce Personas! Personas let you save the essence of a song - vocals, style, vibe - and reimagine it across your creations. Want to bring your favorite vocals or style into a new song?
The commas to me split it up, why not just say bring your songs style
Yea, in the thing I just posted it said the same. I misinterpreted the comma delimited reasoning
Regardless, when I do that is when the original song from the persona will poke through or sometimes be completely there as if nothing changed at all
Still I haven't done it yet, but I will try doing it with a stem later if you can do that. Pull just the Vocals. If not a stem then I'll make an vocal only song. I just want the vocals not the track. Like in life same singer different songs.
Gotta run. Be back
Which is the odd part. Back when I started and it was V3. I did about 3 or 4 song's with the same genre and near identical styles and the singer sounded the same. yet, in V3.5 I don't get the same singer's voice when I do that. Although, in many case's. The female voice is close at times, but you can hear a difference.
What's the address for legal? Is it legal@suno.com. I think its more than that.
Are covred credits replenished monthly just like normal credits?
Also, why are they a separate pool?
If there's reading on this, link me.
Nevermind. That's right. I found it.
As in the number of covers & personas songs you can make?
Seemingly so, yeah. It just says free credits remaining, I missed the word "Free", so I assume that's just for everyone, normal paid credits can do it after?
There's a set number of free generations everyone gets, then it costs credits after those run out
Huh, I just noticed that some replaced lyrics in a cover (original -> cover) used the old lyrics, even though every other spot updated.
Like, a single 5 words
I also have the same problem with personas
I need to remind myself that the lyrics aren't magically a separate track.
Logically I want a text change to just happen, it's not that simple, especially with a cover.
Definitely a lot of "feeling out" in terms of how these features act consistently
I got pretty good at wrangling MJ into doing pixel art in my desired style for me, don't I know it.
I've seen a few people do this, how do I go about a few verses being more "said" or chanted"
Almost like the singer does slam poetry mid-song (usually with a minor radio voice filter), for just a verse or so.
Side note - I should definitely spend more time looking at other people's creations, I'm flying in the dark here.
Suno Wiki
Suno AI Wiki (Suno Wiki) is designed for beginners to easily learn and get started with Suno AI, discover how it generates realistic music with vocals and instrumentation or purely instrumental tracks based on text prompts.
šŖ Even though the devs have emphasized the differences from how other AI services work, I'd say that a general familiarity with how to make the most out of other AI platforms is 100% a transferrable skill to Suno.
"in my desired style" -- not sure if it was intended as such, but that really emphasizes something that I feel about Suno -- affecting any particular sound isn't just a matter of "using the right prompts", because the weighting of all prompts used is connected, so you really need to familiarize yourself within a specific style to get the most consistent results -- but even achieving "mastery" in a style doesn't mean that you implicitly understand every style and how to create it
There's the gold
ftr -- if that's the site I'm familiar with, a lot of sections appear to be hallucinated by ChatGPT and don't actually hold water.
@naive quail is this the one?
I have translated quite a bit from udio to here.
The other thing to note is that Suno was or is trained on and from 'industry standard'. So, if you veer from industry standard, things get interesting.
https://suno.com/song/ebbcf4a4-c56d-48a9-a573-5766288b92da with enough luck and the right stuff, you can get stuff like this (im still surprised i got this vocal)
Ambient Nordic Folk, female vocals song. Listen and make your own with Suno.
Just wanna throw this out there... Personas are awesome and something I've wanted for a long time!
it's been pretty good so far with what im doing
Yeah!
Yea, that's the one.
That definitely lands where you wanted it to, given the prompts!
Impressive
i made 5 personas but only used one so far and its looking good.
š thanks man -- I mistakenly called out the chirp v1 documentation a couple weeks back, so I was trigger-shy š
for sure, i was so surprised when i heard it the first time
I can only imagine how thrilled you were, given the absolute specific nature of the wintery / aethereal vibes
I just posted my first piece in #š£āsuno-showcase - feedback appreciated āŖā«
Any way to make good EDM remixes of ballads that don't sound very different, just change the instruments? Do I need to use covers, personas or what? I want to do something like David guetta does pretty often, transform a ballad I created with Suno into an EDM club banger, not changing too much.
You could try covering, as it can work but sometimes gives you completely different results. But that might be the best option atm
But persona could also work since it would use the similar singer to the other song
Is it just me or the sound quality and the sounds of the instruments got better after the persona update? Less muddy quality generations and more realistic sounding instruments instead of the previous MIDI sounding unrealistic ones? I'm purely enjoying this so far
Things move FAST in there -- can you link to your post?
is it normal to have another song in my song
I've had that happen when I try doing a cover and change the lyrics.
like it changed fully its style lmao
It's the one about hating working for pennies, being a millenial who has programmer skills, but works for a glorified trash company, you know, that one.
handwritten lyrics, it feels like, based on the flow?
talking, am i, like yoda?
I just ran a Persona song, and it... changed all the lyrics, and I think some is in a foreign language??
Hybrid - I use chatGPT for structure / syllable count, manually change a ton, fix flow, etc. I also give very very non-generic prompts with a lot of detail, as I hate the common rhymes and crap it spits out.
I bet that goes to the AI algorithm, if a certain voice is learned from a specific singer, for instance, who sings in a non-English language, and you force the song to use that voice...
i have no persona i'm a basic user, thats worst
Like ordering a rough-formed bust-shape for a sculptor to work with -- still 100% their finished work, but the scaffolding / heavy lifting is offloaded to ChatGPT -- a good collaborative process
I use tools and get results, I've never cared how I get there š
My first ever game was "programmed" in Flash IDE using ActionScript 1, and I "copied" the game X times, where X was the number of items, and altered the game ever so slightly to allow the next progresson (point and click). I had no idea how to program, but I knew how to tell the timeline to "gootoAndPlay(FrameNumber); - which was enough to get my entire idea done with that terrible hack.
Just like I use gpt for lyrics, BUT, I constantly have it revise and tweak it so it works better (in my vision)
It's a pretty good workflow, TBH.
Way better than me trying to write lyrics myself, since I suck...
Sounds like a hallucinated song. Suna throws those in now and then. They can be quite interesting sometimes.
Yeah, it sounds good, but singing total gibberish.
Haha, yeah...
Some of us rewrite over that .melody with real lyrics, rerun with the instrumental stem, refresh with a cover and get a decent full song out of it. Now I suppose we can add a preferred persona if we want to take out the accent.
How might one go about finding a public Persona with a specific sound? Like, something that is akin to a specific band or singer.
I'm having an issue where I'm getting verses repeated. I kind of have a feeling it's down to I don't have enough words in the song and it's trying to feel the Gap to make it 4 minutes but I'm not totally sure.
Oh, interesting.
Yeah, I get that on occasion too. For that I just do "Extend" from right before the repeated lyrics, then [end song] at the point I want nothing else. Sometimes that works. ^_^
Usually it's the first verse but the rest of it is great so I would hate to do an extend and lose all that good stuff
You might find a fun sound out of all you've done, or someone might share a sound. If you get a sound that sounds exactly like a celebrity artist, there might be issues of publicity rights. I'm personally enjoying some "nice" vocals but try to scratch the ones that sound too exact to an artist.
Yeah, I get that, though I have no intention of making any of this music public in any way. Just trying to capture the feel of a specific artist for a song.
If I make it public to the app and someone scales it for their own use, it traces back to me, I would think. I just don't want to take the time to defend myself. Lol
Thatās is true, I really like the persona feature, but someone can still use it from the extend feature
I may be totally wrong, just prefer to Err on the side of caution
Yeah, fair enough.
If I had a nickel for every time someone said, I think too much
haha, better safe than sorry!
I need to remember that just because I keep my stuff private, anything someone else makes public could still be an issue.
But at the same time... https://help.suno.com/en/articles/2410177
Yes! Well... let's talk about it. If you made your songs while subscribed to a Pro or Premier plan, those songs are covered by a commercial use li...
I would have my own game dev studio at this point, and not the kind I had at 23 where I shared a room that was residential and business haha
As a fellow everything you said, who has also lost in the stock market following (proven) companies I believe in, make sure not to put all your eggs in one basked.
Unfortunately proven track records and passion don't always translate to success on the market.
Sadly.
Disclaimer: This IS financial advice š
I made a persona from a duet and the result is pretty great. Never been able to make back and forth vocals and two singers singing in unison so easy
https://suno.com/persona/ac1447ce-b7fd-4e9d-8564-129389b4e88b
That's really cool. I tried for the longest time to get a song with female verses and a male chorus and it sort of worked out...
How do I delete a persona so I can empty out my trash list?
There's not currently an option for that, unfortunately.
Okay thankies š«°
Big difference between someone taking a public Suno song and being able to extend from it than it is for someone to take a song they don't own & upload it onto suno
Another good addition for personas was one of the things i was looking forward to appearing at some point, however im noticing a issue with it being unable to actually give endings atm, have not had a single one pop a actual end so far with testing it, just leaving this here, since it seems like a technical thing then just bad luck at this point.
I do get ending's with the use of my Persona.
What I have noticed with testing stuff out is after it finishes a chorus and goes into the instrumental breakdown. After the breakdown, it's a white noised washed out background of the song made from the persona.
If there is nothing after it
Strange never got one of those, am testing a second persona now, but still not getting any endings, im trying to change the speed tags im using, to see if its just the persona i used first was too slow already, or something, but most of my songs are ending before the lyrics are done, and just cut off prematurely.
Although for your find, I havent gotten one thats actually gotten to the end of the lyrics yet, so, might be why im not seeing it
I also noticed if you don't set its vocals it still switches between male and female randomly, even though the male vocals sound weird (at least for the persona i setup with female singer) when its in use.
I haven't tried changing the vocal's. I'm gonna try that right now.
Ahh FINALLY got one with a end using one of my smaller lyrics songs, and it didnt do anything from the previous song, try changing the tags a bit, see if it removes any of the white noising
This one had a pretty normal end on it, that id expect from the normal generations
although its the only one so far, and ive used a chunk of the free rolls so far (At 167 out of the 200 at this point of posting)
I'm at 154 of my 200 and that was the first I've noticed the issue.
As for the female lyrics, It tried once and just wouldn't after š¤£
I'm guessing if stray too far from the original song style that you created the Persona of. Suno just says no.
I've noticed that Suno creates songs with Persona pretty different than it does if you just add lyrics and make a song the normal way. I'm getting cooler song structures and rhythm but still getting all that damn noise in the background.
like...the first 3 songs, wonderful, but dirty with noise.
This. yes.
Think your right on that, been messing with changing the tags and its hit or miss if it keeps vocal style, and other bits, but theres sometimes it does what im hoping it will do so far.
If they could fix all that noise man....I'd be so happy these songs are great but they are all unusable.
And yes, I am seeing this too, the structure is better utilizing aspects from the song genres then it does normally
It's like Suno is trying to get a stereo sound but is getting stuck in a mono quality. It's kinda funny but too bad.
I think I have one song out of 7 so far that isn't dirty or have other noises in the background that ruin the song.
Can u privately share the songs so i can see what you mean by dirty
cause i dont seem to be getting this atm or maybe im not hearing it
just curious if you dont want to then thats fine
The songs will have noise in the background, sounds like wind or distortion.
here second
what you think is synth, is noise, it only happens in this spot in the song and no where else.
Sounds like it's repeating the vocal but it's been thrown * to the wind* (š¤£)
Looking back through the server chatlogs,
it seems that people sometimes complain about Suno generation blocking bad language.
This is quite a surprise to me,
since one of the tracks I made last month has six F-bombs in it. š
Is this just very good luck?
i've never had that issue
Not for me. I have plenty of swearing.
Ah, then it's a case of those people being unlucky
Noice
The censorship from ChatGPT was much stricter a few months ago, the devs did tone it down
one of my songs has the F word stretched out in the begining for like 4 seconds and its very easy to understand
Thanks for the info!
Ahh I see, i sent one of the songs, and some text in DM, however what song did you use to make the persona?
what song? this was from scratch didn't use cover
Oh what did you use to make the persona that your using though?
OHHH you mean that. I used the voice I had from Luna Flare, my Kpop band.
Ahh, just the vocal mic in then?
no it was a song. are you saying that the "persona" actually takes the song as well and not just the vocal style?
Yes
cause that would kinda suck
Style and voice.
doh.
I was going to also ask if the mic in also had any overwhelming overhead
like a feed in sound or something
lemmie try that.
but if u added the song in then its probly something layerd in the song used
What do you mean by 'mic in'?
I think they mean "vocal" only
so you can create stems as we all know. maybe upload that and then pick it>
The persona is especially focused on a vocal style, but you may notice that it also affects the music itself. It can sometimes work like a specific artist with all the members and not just one singer.
When you record vocals depending on the quality of the mic it can have outerlayered sound, usually you can edit it out, but the layers still in the audio file usually just been toned down, i was assuming this was a real band he/she was talking about, so mic in ment the mic recorded audio file,
hmm...I wonder though. see I didn't fully read the announcement I just jumped in and started playing. I was hoping it just took her vocal "sound" meaning her voice sound and took that. Like if Beyonce was singing a dance track and then did a metal track, its still Beyonce's style and voice. but if its adding from the music as well, well, that's not really a persona now is it.
That's what I was thinking you meant. But, I was hopeful there was some new settings I missed xD
or rather the feed in audio to the audio file, honestly when im thinking on how to say stuff it doesnt really go through well, my english is shit
that's just taking the vocals and applying it to different tracks with the music. its just remixing it in simple terms
Your english is great, don't knock yourself.
Well maybe grammar was a better word, or even how i present what im trying to say, either way.
If your just using the vocal audio
it may work better?
since it will use only the vocal part
that's what I am thinking...maybe.
but i have noticed some very upsetting oddities with vocals so far
This is understandable and I get it. I assumed it was taking the persona of the song as a whole, which it kind of is. So, your creating a 'band' to some degree.
mainly if it switches between male if it was female vocal
If you wanted vocal only, you may have to create an acapella
If you just want vocal
yeah I think so. I was hoping for "her" style. I put her in quotes cause...she isn't real.
do the add in audio thing in your song
so the ai gets the base of it, then have it create a low instrumental music genre or just generate till you get one with low instrumental
I'm not noticing a lot of the music in the Persona tracks coming into the new songs. but I need to listen closer.
the lower it is the more it gives just a specific part thats higher then the rest
but from testing I havent gotten anything super amazing from low instrumental personas
have a few i trashed already cause it just sounds like normal generations where the others based on full songs sound way more influenced
so far though im still getting ending issues, I think its more so due to slow progression on the songs im using so Imma mess with fast tags and see if that can fix any of it
i sometimes get cut offs in the end and have to extend 3 or 4x before it finally ends the damn song.
yeah this song I am in absolute love with...super dirty..can't use it. which sucks
and if I do a cover, its going to change it all and possibly, probably keep the noise.
Ya i got a few songs that are good with no ends that ive kept for later to do some extends im just trying tto see if i can make it give me full songs someway on this specific persona
Are your tags different from the persona song?
yeah
since i wanted to try making some story elemented song progressions
slightly
both are drum and bass
but the original is Jungle, Drum and Bass, Dub, Dark, Groove Oriented
Try one with those same tags, Maybe add a new one or two from what your trying.
Try adding Clean Instrumental to tags for one generation
and see if it fixes, then piggy back off of it
Try both
if able
not saying waste your creds on it, but having it in both may do something on one of em, but may not so just give it a once over if nothing, then move on
meh it's fine i'm premiere I have a bunch left still
Well theres always a chance of regeneration over a multiple that it will give a usable result, but honestly its always a waste of credits even if your pro or more
specially just for one song
oooof as a cover its sooooo dirty
So no dice?
the one with clear tags is still cookin
I went through 7,000 credits on one song. š¤£
damn I hope it kicks ass
When i first started, I am unsure of how much i spent on single songs, but id spent the next day or 2 deleting my trash
Yeah is there a Empty Trash now where it all gets dumped at once or no>?
i've not checked
The song is decent. Wouldn't say it's anything over the top. I've never emptied my trash. I have everything since day 1 xD
Ive asked for it earlier but nothing popped so i figured its still same
trash from what i saw makes the ai remove variation for the regular generations, but thats speculation, I havent been able to fully confirm, however when i deleted my trash I always got better generations
Hmm
What was that? š¤£
It sounds like you want to remove the reverb
itrs just bad noisy synth that is just...horrible and noisy and flangy
maybe try No Reverb
But my ideas are limited on this
since suno is made or broke by autotune needs
and thats alot of reverb
That was definitely a reverb issue.
listen to it. that's ...ugh...so so b ad
Ya try any tagging you might think would remove reverb
or lessen it
but atm im not sure of any way to actually remove that entirely
That's the original song poking in
balls
Well i think it uses alot of autotune-esk stuff
so if you got tons of reverb in your song you wanna use
Acoustic Guitar, Dobro, Resonator, Enraged Goats on Welfare, A Very Grumpy Grandma. Not Dave, Or Kevin song. Listen and make your own with Suno.
it probly will mess your stuff up good
That what I get for breakdown sections.
I'm guessing if the vocals for the OG song are not in the same spot as your newly generated song. The OG song pokes through.
and ruin a perfectly good song.
Hmmm, ill try the fast test stuff now see if thats the case
this is when I with we had stems for each instrument. cause then I can just cut it out.
or at least see if my results are different
atm i find people dont tend to have the same results so its hard to say which is or isnt
They need to make stems fully remove parts first, when i was messing with it, most of them had parts left over
Also just for retro's issue I cant really test well, most of my persona's are reverb heavy
@verbal mortar Curious, if you haven't. Try, vocals -3 to -6 db, music -6 to -8 db,
See if that does anything.
where would I put that
I have noticed every now and again that the song's with that addition, are quieter and I thought my sound got turned down.
My credits renew tomorrow if you want me to spend a big chunk on a Prompt Experiment Playlist so people can use comparisons to see if they think it works?
But, it's also Suno. So š¤·āāļø
Sure
I've been trying it, but it's just me. I don't mind forking out the extra few dollars here and there. I gave up other habbits for suno xD
yeah same I don't care, i'm having fun and I can afford to be a bit stupid with credits if I have to.
ya changing speeds very hit or miss as well
If there's anything worth exploring, I'm happy to look into it so others don't feel like they need to spend their money to see examples of whether or not that prompt works
I notice if you use a slow persona and add speed up tags it will make each insert a bit faster but the actual song flow is whatever it was in the persona song
this one for instance on the persona im having the most issue with
has some speed ups but the flows still slow
female singer, minimal dubstep, fast paced song. Listen and make your own with Suno.
Yes, because if you use a fast paced persona. Suno thinks the song should be fast, and vice versa.
Iām not saying anything about songs themselves. I was just asking about public Personas. I wouldnāt want to steal someone elseās song, certainly.
Mine have been ending, but usually have to extend because my songs are always long anyway lol I am super feeling that 3:25 limit all my personas seem to be giving me. Lots of extra credits being spent ugh
Ya this persona thats giving me grief needs extends pretty much every song it makes sept one or 2 so far
I didn't mean to imply that you were doing anything untoward -- I just meant to imply that if something isn't shared explicitly within Suno it's a degree "more safe" -- but that raises a good question about whether using a persona is different than using an extension... legal complications abound, to be sure...
Oh man. I think my persona broke at the end of one of the generations š¤£
Ive had similar things I think, but i figured it was due to the songs ending prematurely, but it very well could be the ai took a early nap
[Clean Instrumental] doesn't work
Ahh unfortunate
This is a little ways into an [Instrumental Interlude] and is ending.
It sounds like the OG song is trying to poke through then it just has a bad day.
I see that
Have not gotten a single one like this yet
maybe its because of genres used
or maybe for the effects that the original song had
my current personas are popish
I can try some metal
and other heavier effect/sound genres
see if theres some difference
Same exact prompts, the difference is. This generation is Verse > Verse 2 > Chorus and not a full one shot song like the OG
I see, so is that the end of the song?
No, it's supposed to be going from a Chorus into, this part, the instrumental interlude. Then I would extend and do the next part. Verse 3 > Bridge > Chorus
Oh so thats just a internal part of the song, thats not great, cant even extend over that for a once off to complete it
š metal, -5 to -10 db gave me a bunch of AMAZING metal screams, but they're not exactly quiet.
metal, -5 to -10db song. Listen and make your own with Suno.
Although I am wondering if this is genre based issue, what genres are you using, I am curious on how shell shocked the AI gets if I introduce very unsimilar aspects
Exactly as that -5 to -10 db You didnt add the 'vocals' or 'music'?
@charred merlin Sorry, forgot to 'reply'
melodic metalcore, metal, aggressive, emo, shredding guitar, electronic, Are my style prompts.
I'm currently compiling some with "-5 to -10db vocals" that are just as loud, but not in ways that are as funny
https://suno.com/song/22e7f905-38cf-4fd6-a531-c3ecf076cddf
melodic metalcore, metal, aggressive, emo, shredding guitar, electronic , vocals -5 to -10db, music -5 to -10 db
melodic metalcore, metal,aggressive, emo, shredding guitar, electronic , vocals -5 to -10db, music -5 to -10 db song. Listen and make your own with Suno.
If we are gonna test some random stuff. Can we, or would you. Like to agree on the lriycs section and some prompts?
Can I grab the stuff off that and try some stuff?
Have at it.
Hmmm Aggressive seems to circumvent the slow persona stuff somewhat
I'm getting short generations, but there hasn't been a single of these instances I'd consider "quiet" in comparison to tracks without the decibel stuff. I have templates that I use for lyrics as demonstration, usually nothing too wild because I don't want the lyrics to impact the music style -- ultimately, if an idea is worth suggesting to others outright, it shouldn't be dependent on what specific lyrics they use
Out of curiosity, am I the only person that, when trying to generate an avatar for a new persona, it literally generates a random image that disregards the input?
I'm thinking the music and vocal prompt's are more for quality than loudness. I think it only balances vocals to the instrumentation. Maybe
I also tried this on full song's, not quick prompts
If you enjoy doing it, and you feel like it works for you -- it literally hurts nobody.
But if it's being suggested as a way to impact the audio levels of the stems, nothing from the initial tests would suggest that.
I am not seeing any difference in loudness.
@cloud shore alright i put the 2 made with that stuff in DM
I would agree, but I also wouldn't lambast someone for choosing to include it. Likewise, I have a prompt experiment playlist where I tested "dolby atmos" and didn't find anything conclusive, but it doesn't HURT, and arguably it does SOMETHING, just not what we might initially suspect.
Hell -- if that were included in the style prompts of a song I made that turned out well, I'd likely leave them in when using "reuse prompt" just for the sake of "not wanting to break what worked before" š 
So suprisingly going from soft calm pop persona to a heavy aggressive and harsher sound seems to work better then going from another calm to the other, but I feel that could be my ears personal choice on that, although I liked that metal stuff more then the popy sounding stuff it was giving me using similar styles
Is there a limit on personas?
If I use section replace, my song will lose its connection to my persona. Anyone can confirm?
None of these effected the loudness.
was there any documentation for this change? I didn't see anything mentioned about it....
Finally got Suno to successfully make me a Persona! Now I tried it out in making a cover of one of my earlier works, and it's just amazing so far. About to hop on a second attempt, this time with a completely different song, and see if it still holds up. This feature is so cool. Just wish I'd been able to get it working sooner! Thanks to the whole Suno community for all of the valuable advice, input, and stuff that they do to help make this the best piece of software to ever come down the pike. š
I won a prize in the Flosstradamus contest. When will I recieve the prize, anyone know? Cant find any info
First issue I noticed. I created a Persona, and successfully created a ne song with that same voice and feel to it. Success!. Except when I extended it to fix the ending of the song up, now the extended song does not list the Persona, and does not show up in the Persona's playlist
It happens with 'Extend' as well. I fixed up an ending and now it doesn't show up in that persona's playlist
You got me for a sec. I thought you selected 'Cover song' and then tried adding in persona. I don't see that and I thought that was a genius idea.
Not coincided with any major change, just a stylistic UI update, it seems
#š”āsuggestions-web or #š§āsuggestions-layout are routinely checked by staff if you wanted to share feedback about the change, though šŖ
Being able to do it that way WOULD be a genius idea. Maybe someday. š Before Personas became a thing, I had to use extensions from a common base to get even vaguely similar vocals... I'm so glad that those days seem to be receding. Though it's still a darn helpful tactic in a pinch.
I also won a prize in that competition, but have not received anything yet
weird stuff
hey i hope its the right channel but im from germany and i cant pay with paypal and dont have a credit card so is it possible to pay with normal bank card via VISA?
Actually, the extend itself doesn't have information about persona, but when you get the whole song, it does still have it.
Mine didn't
All of mine keep the persona on the whole song
What is this about decibel tricks?
For the speed I download the song and put it in FL Studio then raise the speed and lower the pitch and cut anything that sounds weird.
This is a simple one time Extend off of a Persona creation that ended at the 3:33 mark suddenly. Used 'Get Whole Song' after and no trace of persona https://suno.com/song/67cf4e2a-25be-4dd0-a6af-29ffbcb1ea01
Ethereal Synth, Balanced EQ, Crisp Clean Sound
CD Quality song. Listen and make your own with Suno.
If you click the link for the song, it mentions the persona... but click on the persona, and the song doesn't show up in the playlist
Oh, you mean whether it counts on the persona itself, nah I have no answer for that
Showing that it didn't seem to work as a prompting method
That's a pro-tip regarding adjusting the tempo or pitch externally then reuploading, though šŖ
Does adding the prompts of balanced EQ and CD quality help?
Yeah, I think that has tp be a bug... a song on the playlist gets removed when you extend it, or mess with it in any way
Oh, dang
Personally, I only care that it shows one was used at the moment so people can see that whomever made the song probably went through a decent amount of trouble to make it...lol
What's the point of a Persona Playlist then?
Again personally, nothing, don't have one and don't really care much. I do care about the quality of Vocals that are produced comparatively
For whatever reason the Vocals produced by covers and now personas are insanely better
ive noticed the vocal patterns are better with Personas they change the style and how they accent certain words. it used to be it would just go thru them in the time signature the song was in but now, they pause, they break words or lines up. its pretty cool. now...as I mentioned earlier...if they could only get rid of the noise.
From what I know of generative AI, the noise comes from drift when making long runs. Hence why it gets worse the longer the song goes on. However, I do find that changing the Tempo at points makes it check itself and restart its timing.
how does one change the tempo at points? that's something I was even aware we could do here
[Tempo Change] is a real tag, oddly the tags along with it are not actually about Tempo
[Syncopated Rhythm] [Polyrhythmic Percussion]
Reliably? Create an extension point at the point that you want to affect a notable change
so for Drum and Bass would I add a [Tempo Change] and then a number or just that as it is
Any reason persona-generated songs seem to all be the same length?
Yea, the difference is definitely noticeable considering before it sounded like a wave of noise was crashing on top or the end of my songs. With the Tempo Change even on one shot it helps
Is there a way to insert a key change?
Hello, can someone help me? How do I isolate a clip from a song produced and then when extending the song, get Suno to add that clip back in on a loop?
Ok with the keychange I would be giving you dated info, so sorry for the misguided. That was when putting the song details in the lyrics. Very hit or miss. I don't have any songs that use the key change
I don't think you can, but that would be great, if its not possible now, please make a post in suggestions.
Will it really follow it? That would be neat to make some super weird jazzy key progressions
Ok, thanks. Yeah I have been unsuccessful to add a key change on an ending chorus I have wanted. Bugging the hell out of me.
The tempo change really does help with the noise
Is there a way to get around that awful reverb eeeee noise from it not decaying properly in later extensions/generations? So many good extensions just get shat on by that sound
Anyone one else experiencing that your account dont revert to the free tier (to get the 50 credits per day) when your subscription runs out? or simply stop receiving the 50 free credits?
Where would you add tempo change? I dom't have much issue with noise per se, but I will also generate a song in 2 iterations.
Example
1st: Intro > Verse > Verse 2 > Chorus
2nd: Verse 3 > Bridge > Chorus > Outro
Unsure how to answer that. It is the not that is confusing...lol
Depending on the song but I normally go before a Pre-Chorus but in yours before the chorus
better?
[Tempo Change] [Syncopated Rhythm] Then after the chorus [Tempo Change] [Steady Beat] or another beat tag
Tier, not tire. And no, I have not ended my subscription back into free, sorry.
and this works with ALL types of music? Sycopated rhythm is making think of metal
I am not sure, I haven't tried to revert
and this would go in the lyrics yes not in the style.
I found out where the extend problem comes from. I noticed that when you go to crop or replace the section. The lyrics are way ahead. That's why sometimes, there's a difficulty to continue the song. It means that when you continue from a point sometimes, the lyrics don't follow and continue from the original end of the generation.
Then 1. The question was not for you 2. Tier or tire, speak some Swedish for me, see how well thats goes for you.
It seems like it takes a chunk of time to gather the context? It varies from persona-to-persona from what I've seen. Some of mine come out at 3:42
Honestly, just because the tag says that doesn't mean it will do that. However, what it will do is some type of change. If you are doing any electronic it is definitely a thing.
It reverts on its own if you dont resub. (when it works that is)
Thank you for this tip. Greatly appreciated
Oh I got ya. Sorry no idea on that.
The problem is that I don't know how to get around it.
is the suno site down for anyone else? š¦
No, is working fine on my end.
doesnt work in my firefox š¦ but safari works
thanks
ahh now it does! all it took was a discord message š
I found out where the extend problem comes from. I noticed that when you go to crop or replace the section. The lyrics are way ahead. That's why sometimes, there's a difficulty to continue the song. It means that when you continue from a point sometimes, the lyrics don't follow and continue from the original end of the generation.
The problem is that I don't know how to get around it.
Sorry I missed that, yes all tags in lyrics and not style
Since the lyrics tack on the bottom of the other lyrics instead of actually replacing them I just go through and delete the stuff that is doubled in displayed lyrics since that is what it follows when you extend.
My last song had like 3 stacks of itself after the Outro.
how do u do that ?
Just jam your song and hit edit displayed lyrics on the songs page or do it in details but it is easier on the songs page
I will go through and adjust lyrics as I hear them if Suno goes nuts but it happens to work...lol
Feels like I'm going to have an aneurism trying to change the last verse from female vocals to male vocals. Trying to get it to change by having the whole verse as it's own extension. Of like 20 generations only 1 had a male voice and it was.... meh. ._.
Anyone have any ideas of how to better improve the odds of having it swap to a male voice for the remainder of a song?
sometimes i feel we really need an option to "force" suno to accept a specific input regardless of what it thinks itll sound quality wise/what it thinks is "best". like in this case it's shown that it's able to swap over, it just doesnt regardless of what tags/prompts i use. 400 credits burnt so far and only 2 instances of it swapping from female to male
This is for extending?
just in general. like, as previously described, the song is sung with female vocals, but i want the last verse to be sung with male vocals, and to try and further encourage that i have the last verse on its own as an extension. ive spent 500 credits at this point trying to get it to change
What tags and style did you use?
I am able to get Vocals to change fairly consistently
im at the point im spamming things related to using male vocals in styles, and things related to female vocals in the exclude styles
Male Vocals should be in styles
they are
Both
I gave up on that. I just have a female vocal sing the whole thing, then split the stems in Logic and convert the vocal to a male voice in audimee. Works well.
@worn flicker ill give that a go, thanks
Emotional ballad, soft piano, acoustic guitar, strings, layered male vocals, female vocals, gentle build, duet finale. song. Listen and make your own with Suno.
with the 500 creds spent so far on just trying this ive gotten it to change only twice, and both times were kinda... meh
Best of luck. Never worked for me.
You have to walk the AI on to a Duet and use styles that use duets in the style for consistency
Even if your song isn't that style
Cover it after
Your welcome...lol
lol
I like duets
i want those 500 credits back demmet š
I feel special...š¤£
Or you could convert the female lyrics using audimee and splice the male vocals into the right places. Much easier to do duets that way with delay and pitch.
My mommy says I'm special. 
Dang, I do it with style and tags, I normally dont even need to cover it. I just keep things like broadway or whatever would use a Duet in lower case later in the style prompt
Here are my touched on responses in regards to duets:
In my experience, I get duets roughly 60% of the time when trying to achieve it, though half that time, Suno swaps the genders I wanted (which works out most of the time)
I had a much deeper coverage somewhere but cannot find it atm, but that gives you a gist
It's how I got Female duets
That is 100 percent accurate
Yuck. What a hassle.
You do have to write non gender lyrics, or just do extensions and not try to one shot it
I like my options
Audimee and conversion is so much better.
I hear ya, I use other remastering like BandLab and FL Studio, do you find it to be better than those? I am always down to learn a new thing.
For stem splitting?
For the easier way to do duets as a whole like you said. As I said I love duets, please let me know cause I would love to do it an easier way.
Well, to split out the vocals the key is to not have lots of noise. Apply some cleaning like a noise gate will help.
Iād try the stem splitting of a few DAWs (and Suno) to see which one is best.
I'm sorry I meant using Audimee. You said it is better...I was wondering what it offers that what I use does not.
I already use FL Studio
I don't do that unless I have to me that isn't easier than jacking with Suno a little. I like to try and get AI to do what I want, you know figure crap out.
Whatever, you get the point lol.
I made a persona from a song and tried to make one with different lyrics but it ignores the lyric and uses the exact same lyrics as the song i used as persona
I reverse a flanger
It is common. It will try poke through. I'd call it more of an alpha than a beta. However, you can get some great vocals from it they just may be slow and you will have to uptempo it. It not very consistent.
Audimee is an AI that changes vocals. They have dozens of male and female voices.
yea i using a slightly changed lyrics there i changed out the name from A to B but he sings the name A
Oh like the open source one R..something I have the github bookmarked has tons of vocal models. Like that?
US$30.00 payment to Suno was unsuccessful why
Yes, but much slicker.
We are just some friendly users in here. We don't work for Suno.
White names aren't even moderators.
Sweet I'm looking it up now, if it is iOS only I'm gonna be pissed...lol
Itās a service.
Also playback has been lagging for the past days when its generating
On Suno. I guess that means this technical discussion has gone off the rails if I have to ask...sorry
where need a suno protest
I will definitely give that a go to make some songs I have wanted to make that are telling a story. Thanks!
No sweat. I did a three act musical a while back. Relying on Suno for a male and female lead was just not workable.
Itās s little more consistent with Personas, but you need to be a paid sub for that
I am a paid āproā sub.
Duets in general are just hard to get
Which is why I use audimee.
Though Iāve been getting them more recently when not even trying to
Hello, I have a question about the Replace Section feature. As you can see in the attached image, when I select the section I want to modify, nothing appears in the Lyrics window. What could be the cause of this?
Best to add this to your exclude if your going to massage the vocals
mono, distortion, fuzz, hissing, clipping, overdrive, buzzing, artificial, peaking
The lyrics get highlighted on the other side, but I don't know what languages are supported.
This is a lyric created in Japanese, but other users have confirmed that it works well in Japanese.
That is fine. I just didn't want to give wrong info, then in that case the lyrics get highlighted on the right side.
You put what you want as the new lyrics on the left. It may populate but it isn't very accurate, I normally do it by sound and wave shape.
Yes, but why not detect it automatically?
Hey so I've been having a weird thing where every song is literally female singers how can I make it so that it's male I've tried with prompt version and custom excluding female voices and still getting all female voices
Put Male Vocals like that in style and add tags above as many sections as necessary in lyrics, [Male Vocals] sometimes you can just put it at the top, but nothing gets hurt by using the tags everywhere if you have the character space
If it gets real stubborn ill put Male Vocals in the front of the style prompt.
Bug? I Can't tell you other than see if they pop up by putting your time further along in the song, the timing tends to be off a bit
I had to draw a lot of that, not the main shape, but I forgot when I found it, I think it was just a thumbnail and I had to bump up the quality pre AI generation.
@worn flicker Oh wow cool!
You can see the exaggerated glow on the hood, that was when I was done...lol. I spent enough time on it...it was time to stop.
@worn flicker Still giving mostly female songs even with male vocals added to it
I don't get it
See #šāsuno-school message
Also see the song example of one method to "bruteforce" (per sƩ), well more like increase the chances of male vocals, by adding a male label in the lyrics as well
It seems like it can't detect Japanese very well. I tried combining it with English lyrics, but it seems like it can't detect the timing very well. I hope they upload a revised version soon.
Keep running into an issue on creating a persona, issue is only one one song 
and it's the one song I want to make a persona off
is it not possible to make songs with persona's that are longer than the original song the persona was created from?
ie, original song 3:37
all songs created with persona 3:21 - 3:28
all songs cut off before lyrics are done
most of them is just garbled noise anyway
It's not letting me upload cover pics
free user?
i managed to make a song that was longer than the original song, i've read a few users having issues with having nonsense coming out on the other end, i've personally not had any issues, although there was a song that had the original song mixed in with it but i could see how that could go really well with some songs but would be nice if we had control over our personas over-influencing the original song in the 'persona remix', i've had songs cut off outside of/before persona so i'm starting to think theres some rng scaling regarding weight of other song influences
I have the impression that when you change the prompting style different from the original, that the results become poorer
i stripped down the prompting in the lyrics to how i used them in the original, and now the song sounds normal. Still cuts off in the same range though.
honestly this might be the case, i didn't change the prompting style by removing anything, only added more prompts to it but since you mentioned that, i'm going to do some tests later brand new prompts
my original prompting was very basic, was on my first day using suno
since then i changed how i prompt quite a bit.
But when i now apply this newer prompting style, it's just garbage noise.
Funny thing is that one of my first songs, on first day, is still my fav š
No
Maybe it's too large, or an unsuported format?
hello chat
the Suno website just randomly stopped working for me. I tried switching my VPN on and off, restarting the browser (Yandex) and my PC but nothing changed - the songs just get stuck on 0 seconds, no matter what sound output device I use. I tried clearing the cookies too (if I actually did it right) - didn't work. The only thing that surprisingly worked is opening the website in a different browser - Microsoft Edge, just for testing. Does anyone have any ideas for why this is happening and how I can fix the issue? Should I just reinstall my browser?
oh wait what the.... huh
it works now?.... um, sorry for the falso alarm then (lol)
i just re-opened the page, instead of just reloading it. Maybe something was wrong with the URL I guess?...
Hello everyone,
Iām experiencing a serious issue with my Suno account. Since yesterday, every time I try to log in, it seems to create a brand new account instead of giving me access to my original one. As a result, I have lost all my credits, my active subscription, and all the songs I had generated.
I have already contacted customer support, but they seem to have misunderstood my problem as something related to audio watermarking. However, my issue is specifically about losing access to my original account and all my content.
If anyone here has experienced a similar issue or knows how I could get this resolved, I would greatly appreciate any help or advice.
Thank you in advance for your support!
same thing happend to me hours ago, now it's fine
Have you tried on wifi.?
I had the error too, but that shortly was followed by the website overloaded cloudflare
could it be that u are getting hacked š¤
Most of the time any connection errors are not your on your end if you are using the web version.
I wish I could help you, what you are talking about sounds like you possibly chose SSO login once and now it just goes in automatically when you already had a different account. Just a shot in the dark
š¢ Do you have any idea how I could resolve this and get back to my original account? Any help would be greatly appreciated!
omg this is ridiculous (though I can understand from a technical point)
In the original song, instead of using [verse] [chorus] and such, I used [spider] [pounce] [SCREAM] [snake]
I now used these tags again with the persona, they are the best versions I got out of it so far š¤¦āāļø
[spider] is now my personal favorite
10m at most. Apparently, we overload the backends. The Love Again is in full steam, plus a new cool feature with free credit for its testingā¦
Nonono, don't mangle your browser. I think it's a backend overload: #šāsuno-school message
Sooo just high load on the servers?
You use username/password or one of the four identity providers (Google, Microsoft &c.)Ā ? In the latter case, verify that (a) you use the right one and (b) you're logged on the identity provider as the same account you always use.
Uh-uh.
The server does two things. Firstly, it loads the whole page. If it's overloaded, you see this: #šāsuno-school message . Secondly, a web page already in the browser sends requests to the server when you click on buttons, or generally request it to do anything. If the server is overloaded, buttons appear dead: no playback of a different song, "..." menus show up but do nothing, &c. Just wait a few minutes. The song that was played will continue to play and seek, as it's already loaded whole into the page, but another won't load into the player strip.
Try using the resize slider to adjust the image after youāve selected it, but before pressing save.
I got it to work
Okay, I get it! Thanks for the explanation, lol. I hope they solve the issue soonā¢
Tu etes exceptionnellement eloquent et incroyablement serviable. š
so i noticed the UI is still bugged and does not display properly if an upload has vocals.
very worrisome as this already caused me and some others a waste of credits
hope it gets fixed
are there any tricks on reducing loss of quality when creating long (6min+) songs? I always hear a noticeable loss of quality when trying to extend songs even when I keep the tags consistent, and at some point it gets so bad I just can't continue anymore
Do you use "Get Whole Song" between extensions, or do you extend from extensions?
I'm extending from extensions, mostly
Extensions can only use the clip they're extending from as context -- so each segment might sound like the previous one, but without using "get whole song" to show it the earlier segments, you're going to iteratively build on whatever you give it
hm... that's good to know. I'll try extending from the whole song instead
do you know how long is the context btw?
Like photocopying a photocopy of a photocopy -- you lose detail and cohesiveness quickly. Costs zero credits to "get whole song", so it's a great step to take when you want to "lock in" the earlier segments, since that way you can use each "full song" as a stopping point to understand which song versions you want to work on -- "well... this full-song-path has that really cool guitar solo, but this one had the cool drums" becomes a lot easier when you're looking for "full song" instead of "part 6"
Not off the top of my head -- there's clearly some sort of cutoff point where if you have a particular sound at the VERY beginning of the song that it wouldn't be able to hear it, but unless you're getting LONG songs and expecting it to keep a LOT of stuff in context, it shouldn't be restrictive enough that the context should be a problem.
Yeah, I totally expect that there's a limit otherwise memory use would be crazy, it was mostly an "out of curiosity" thing. Might be helpful to know to try to get some theme replayed in a segment if I really wanted to keep it in the context
Wrapping up this ramble, I use a similar idea in the method called "Terminator Extensions" -- trying to capitalize on the fact that the extension is limited at viewing the clip it's extending from -- by only using a single second of the preceding clip, you've not only cut off the context of the clips before, but you've only given the current extension 1 second to process, so you can usually affect a greater style change without the song fighting to keep the current melodies (not perfect, but works a charm most times I use it š )
v3.5 can make clips up to 4m -- I think that'd be the upper limit in terms of context? Purely speculative based on conversation, not on observation
Hi, about persona: "I realized that when I choose the person, it also loads the person's style in the STYLE box, which I like, although sometimes when refreshing the page, stop loading that person's style.
An user has commented this and hapened same to me. Even if i set style before, the i select persona, the style changes to original song persona style, is it ok?
Speculatively speaking, I have a feeling that might be more close to 2 minutes - otherwise why would they limit uploads to 2 minutes if it could understand longer songs?
maybe 3.5 uses some sort of "sliding context" when creating 4 minute songs - that would also explain why sometimes it deviates from the original theme from the intro throughout songs even when they're not that long
Hello Suno Support Team,
According to the tutorial, the "Replace Section" function should allow me to click and drag over the section I want to modify, displaying the lyrics in red on the left side so I can easily see the portion Iām dragging over. However, no matter how many times I try, the lyrics do not appear on the left, nor do they turn red to indicate the selected section. Could you help me understand what might be causing this issue?
Thank you very much for your assistance.
Dear Fellows, I am a big fan of the Persona feature. I am experimenting, to mostly good results, with having my Personas sing lyrics I have used previously with random singers. Among those lyrics is a text by E. A. Poe, which is in the public domain. I have created dozens of versions with that text previously. But now it seems blocked to be used with my Persona, stating the creation would "violate Suno rules". How come all of a sudden?
I've noticed that issue if I try to use Replace Section after cropping a song, it somehow doesn't pick up the lyrics. I found I had to use the Replace Section on the full song first and then do the cropping afterwards.
Technically, public domain content isn't exclusively owned by yourself, so while it might have passed as a false positive previously, it's still against ToS.
I recognize this is for upload audio, but the wording is similar for the custom lyrics section
Hmm Hmm. This may be as it is, but does this imply any content which is under no copyright at all may not be used? Is there a way I can check my lyrics etc. before creating songs? I do not want to violate any terms, I quite like the service...
It's not a matter of whether or not the song hasn't been claimed by copyright, it's whether YOU have exclusive control over it.
Unless you're the only person in the world with access or control over content, it's frowned upon to use as input -- anything you originally create is fair game, but if someone else made it, you're only allowed to use it if they assign you exclusive rights to it, meaning nobody else is legally allowed to use it.
Outside of Suno, lots of stuff can be done with public domain, but private companies have a bit more leeway with what they want used on their platform.
In regards to accidental infringements, I wouldn't be too worried, as it's easy enough to use similar lyrics -- if you get an error message that the lyrics are protected, you perhaps can ask the "Write About" tool to reword the concept into a legally distinct set of lyrics šŖ
Thank you for making that clear!
a couple of weeks back, i used to generate songs with suno and the image that comes with it. Now i'm doing the same but the image that comes along with the song has the 'SUNO' logo on it. is this new update intentionally? How can i get the image without the 'SUNO' logo on it?
Shouldn't be the case if you're making the songs directly from suno.com or the official iOS app -- there's been some instances of that recently, so you'll need to contact support@suno.com so they can look into the matter for you š šŖ
Ah, just responded to this in general channel.
Still, valid advice above though.
If you post an example song which has this I can forward the info on.
I'm not in any backroom chats about how recent matters should be handled, so I've just been referring people to support until someone tugs my shirt sleeve
I use my Google account, and yes itās the right one, I can find on the invoice the same Google address š¦
The Discord server can provide community support, but it might be a good idea to get the ball rolling with an email to billing@suno.com -- afaik, they're active during eastern American business hours monday-friday.
All good. Support is an appropriate default place. Where we spot this we try to alert Mike B to it as well to untangle what's happened.
oh okay thank you, but i see it comes any Time?
I don't believe there's a set timeline -- just unavailable for the moment, with no public comment as of yet on when or if it will be made available.
alright thank you ^^
Good morning, my Abo will tomorrow start new. The Points they i have still will i Take with in the New Abo month or will i loose Them and today i must use them for Songs ?
If I'm reading this right. The credits, points you have. Will only roll over month-to-month if you bought extra. The credit's you get for one month, Say 10,000. Will reset the next time the subscription renews.
As far as I have seen. I may be wrong.
OK also must i use them today. Thankyou
I've been testing using Persona's as a tool for covers and it's Ideal.
I got a second one done with a Persona, it sure does a lot with Vocals. https://suno.com/song/8b69933d-61d2-445e-a005-169ab5d7bb26
Epic Orchestra Alternative Rock Soundtrack, piano, duduk, sitar, shamisen, dulcimer, distorted guitar, Battle Drums song. Listen and make your own with Suno.
For testing purpose's. May I try somethign with this song?
Dm direct?
Yea but I want to test your geekiness...lol
Oh? xD
You like sci-fi?
Sure
Lol...too long...dm me
For anyone else if you can tell me what show made the instrument famous at the beginning of that song then I tip my cap to you
Thanks, yes Iāve already sent a message, waiting for the solution š¤
Well the show famous using that instrument rather.
Keep your hat on for me I guess. I don't know, nor do i wish to google the answer xD
My favorite Composer on the planet at this moment. If he makes the score then the show is 100 times better
He has done the score for God of War since it went to PS5
Anyway, his name is Bear McCreary and the dude makes magic
That instrument was used a bit on Battlestar Galactica (not the old one)
Duduk
If anyone likes sci-fi and hasn't seen BSG do yourself a favor and watch a 4 season story unfold in an incredible way.
He did the score for Snowpiercer for the first 4 seasons I believe, when they changed composers, the shows feel changed.
I'll shut up now...sorry
Bear is pretty freaking epic, his reimaging of all along the watchtower was fire
