How To Fix Suno Ai Vocals

Рет қаралды 34,543

Chris Unlocks AI

Күн бұрын

Пікірлер

@ChrisUnlocksAI 2 ай бұрын

Download EASEus Here. bit.ly/4cC7Dot

@NoahtheGameplayer Ай бұрын

Nah, use Lalal AI

@coinagnostic Ай бұрын

I'm extremely disappointed with the AI EASus separator. The results were subpar, with muddy vocals and faint, residual instruments.The main issue with Suno is that the audio quality is decidedly lo-fi, making it impossible for the separated tracks to sound good unless they're in hi-fi. Unfortunately, this means that the AI-generated music will always be easily identifiable. What's even more frustrating is that Suno's stems are worse than those of EASus, which is unacceptable considering we're paying for the service. To improve, Suno needs to provide the option to record in hi-fi and export in WAV format. Until then, the quality will remain trash.

@coinagnostic Ай бұрын

@@NoahtheGameplayer They're both trash.

@IanHollis 2 ай бұрын

I notice in your music style you have "jazz playful". Try adding things like "clear, confident vocals" or "Sultry female vocals" (or velvety, or whatever you're after). I find it helps A LOT in improving Suno's voices.

@KarstenJohansson 24 күн бұрын

Now that version 4 is here, the vocals are way the heck better (as are the instrument sounds), and you can also split your track into stems. Also, if you've left your 3.5 generations in the Library, you can remaster them to make them sound more polished and less "sizzly."

@chillwalker 2 ай бұрын

You can "unnoise" the Suno sometimes noisy, scratchy vocals (they seem to mistake "airy" for noizy) when you are working with a Spectral Editor like Spaectralayers. Split the vocals into its 3 Components "Tonal, Transient, Noise" The quick way is turning the Level of the "Noise" Channel a bit down, but that works only semi. The better way is to work this Channel with an EQ. You definitely dont need much of the Lover End noise that is just rumbling. And you dont need the Frequencies that are causing the scratchy Sound... It takes a bit of time and practice but you will find out what Bands you need for good sounding vocals and what not. Works in 90% of the time. Till now I did not see that anyone had this Idea before so: youre Welcome! ;-) (Please don't ask me any more about this, if you totally dont understand what I m talking about, sorry, i am not a Teacher of your junior High, ok?)

@HappyHumanist369 2 ай бұрын

I take it just as it is. Nice description!

@DJKMonkeyMusic Ай бұрын

I can't do this because these frequencies also stay on the instrumental. Unless I have to make the instrumental somewhere other than SpectraLayers

@baldman5k 11 күн бұрын

@@DJKMonkeyMusic separate the vocals from the instrumental, and EQ both tracks separately. Then slap them babies back together. Also, if you remove some of the transient frequencies from the instrumental, that the vocal track is using, and then do the same with the vocal track, and remove some of the frequencies that the instrumental is using, it’ll mix back together fairly cleanly, with minimal competing frequencies. It’ll fit well together better. A lil headroom freed up for the vocals to sit in nicely.

@mysticCelt 11 күн бұрын

@@baldman5k Definitely helps. My process is like yours, EQ the stems separately for clarity (clean up the low end and smooth out the highs). Then add just a smidge of reverb to the vocals to make 'em sparkle, get rid of the dullness. Then bring the two channels back together via a bus, and use another 'verb to widen the stereo image just a little more.

@StanRain Ай бұрын

Im surprised that no one has developed an AI driven "repair vocals" service yet. Imagine being able to feed AI vocals into audimee, feed it text lyrics, and the the AI automatically smoothing, filling, and fixing the lyrics, correcting unclear artifacts and replacing it with clear sound with new vocals. I'd gladly pay for that!

@MentalBlue Ай бұрын

I am surprised how many people are producers and songwriters bit no musician anywhere. I'm sure they will produce the magic button for youv, so you don't have to do anything else than press the button. Next to the button will be a fader. If you push it up it will generate your social media popularity algorithm ... Or even better fck those buttons and faders. Just put an electrode to your forehead and think about how popular you want to be and you don't even have to make a song! 🤦‍♂️🤦‍♂️

@StanRain Ай бұрын

@MentalBlue Yes! See? This guy gets it! You're really reading my mind!

@Leviathan3000-m8b Ай бұрын

@@MentalBlue Yeah man! But wait, thats exactly like DJs who think they're musicians when they're just mixing together other people's creations, tapping on tech stuff and puzzling pre-made sounds... Ok, maybe instrument makers also used to think that way? At least, one has to have a certain musical sense to create new inventions for musical expression; or even a higher one... Certainly both are an art onto themself.. I take back my impulsive agreement, because according to your sarcasm, people with good vocal chords could also be criticised and natural talents who never had to practice to sing grandiosely; according to your logic, one would also have to condemn people who play sports on a console instead of sweating themselves outside - then also people who watch other people on TV; instead of themselves living actively in that moment - this will go on and on - better let's clarify the question of whether and to what extent your desire to be famous prevents you from becoming famous?

@MentalBlue Ай бұрын

@@Leviathan3000-m8b you're completely missing the point

@Prestige_Sokrat Ай бұрын

@@MentalBlue oohh ok... you convinced me 🤣🤌

@GrumpyhubMusic 22 күн бұрын

I create my music on Suno, I try to fill up the whole music style section as much as possible by adding vocal styles. Plus adding tempo, what instruments or synths, Etc Heres an example of one I have recently used dark pop with electronic elements vibe, haunting, ethereal synth pad fades in with a subtle low pass filter, Soft Piano Melody, echoes, 90-100 BPM, reverb, delay, modern VSTs Serum, female voice.

@twilightbiscuit 2 күн бұрын

super cool! thank you! 🔥💪💪

@criteria9886 14 күн бұрын

Great video, mate! I had to subscribe, like, and comment. The step-by-step tutorial was incredibly helpful. I'm learning how to fix audio, and this has been a big help. Cheers!

@mysticCelt 11 күн бұрын

Thanks for the video! Suno v4 really improved a lot. That said, my SOP is always having the [Clear Audio Production] metatag in use. When the song's done, I take the stems into my DAW to add the pixie dust. Results have been pretty good.

@TheShadowWraithCodex 14 күн бұрын

great video. Thank you for this. I have been looking for a tool like this. You rock :)

@BrandonDennis1 2 ай бұрын

I use a similar workflow, but often with heavier genres the mixed audio causes loss of syllables. If you use a separate stem splitter, for the vocal copy I recommend boosting eq in the vocal range about 200-600hz or so depending on the male or female vocals. Then high pass filter below 100hz to cut sub frequencies. Also you can layer multiple tracks with a limiter if you have a studio setup to bring out some clarity. I’ve had better luck not performing spectral cleanup since it sometimes drops off some frequencies that audimee or kits can pick up.

@AlanMaynard-y6m Ай бұрын

way better?? Its a game changer for me. Thank you so much!!

@Shayzare 16 күн бұрын

The input sound quality was purposely replay at 8:40 to make you 🫵 think that the out put was way better than the original sound quality. That was a horrible vocal separation of the vocal steam. The end results was excellent though. Great explanation of your process. You've earned a subscription. 😍

@Bartetmedia Ай бұрын

I'm new to SUNO but not Audimee, Audimee was my next step to repair my extracted vocals. Thanks for the example, now I know for sure it will work. Cheers!

@coinagnostic Ай бұрын

I'm extremely disappointed with the AI separator I tried. The results were subpar, with muddy vocals and faint, residual instruments.The main issue with Suno is that the audio quality is decidedly lo-fi, making it impossible for the separated tracks to sound good unless they're in hi-fi. Unfortunately, this means that the AI-generated music will always be easily identifiable. What's even more frustrating is that Suno's stems are worse than those of EASus, which is unacceptable considering we're paying for the service. To improve, Suno needs to provide the option to record in hi-fi and export in WAV format. Until then, the quality will remain trash.

@RubenAlfons 29 күн бұрын

@@coinagnosticI totally agree with you

@DJElectraFry Ай бұрын

Thank you so much for showing me how to do this! I didn't even know these programs existed and I make a lot of AI music. Well done sir!

@Bartetmedia Ай бұрын

If you didn't know about these programs, what are you using apps from 5 years ago? LOL!

@DJElectraFry Ай бұрын

@ I knew about Suno, but not about the other programs he talks about. I’m old Ok, it’s not easy keeping up.

@PeterBecker-m6c 2 ай бұрын

This method has a big disadvantage: The original vocal stem from Suno contains a lot of artifacts and sound material that belong to the instrumental track - if you generate new vocals without these artifacts, they will be missing in the new vocal track, thus in the new mix and you can hear it clearly. The vocals sound better, but the instrumental that is laying under the vocals sounds worse because this sound material is missing. By the way, who needs an online Stem separator when you use Logic as a DAW? You can easily use Logic for extracting stems into 4 tracks ant it works very well.

@Artsonica Ай бұрын

Studio One Pro 7 has this stem separator too. And Cubase Pro via built-in Spectralayer

@jeffbull8781 Ай бұрын

Or use Udio which has great vocals out of the gate. You can even generate a song in Suno then just remix it in Udio if you prefer Suno arrangments

@nf_alr Ай бұрын

Thank you so much. Million thumbs up!

@AutoHeadbanger 9 күн бұрын

I do this with coding written into the custom lyrics box to guide the AI. It does a hell of a job to clean up vocals and make them crisp and clean. Then I run my songs Through Diktatorial to master them

@laboonmusic 4 күн бұрын

can you make a tutotial

@GeoTactics Ай бұрын

A couple of observations: the original vocal track has to be clear and loaud otherwise Audimee will not produce a good vocal line with a lot of drop outs. I tried it on a song that had very lush reverb effects, regardless of whether I removed the effects or not, or what I did to the track the Audimee track had a very dry vocal performance with drop outs every few words making it not usable. Still, it could work on other tracks which I want to try out later. The chord app only works on Macs o/s. It would be awesome if they ported that to PC.

@kolermusic5483 Ай бұрын

Maybe filtering and compress voice previous import in Audimee?

@AntonSoundcheck Ай бұрын

Mind blown!

@HarryDeeProductions Ай бұрын

bro.. your an absolute legend,

@SupraCx 2 ай бұрын

this is what i needed ty for this , i use fadr/ stems for stems Separation it can separate up 14 instruments and finds key song and chord progression , also u can download them as midi , been using it for 4 months

@azankyzan 2 ай бұрын

Great job man! Thank's a lot!

@VoltaSlime Ай бұрын

Thanks for this! But just a few hours ago, audimee put my favourite voice behind a pay wall, which is an absolute bummer, but I'll adjust to others. Thank you again!

@ChrisUnlocksAI Ай бұрын

That is a bummer. You can always subscribe for a month then cancel after you have created an album. Good luck. Thanks for watching.

@DesiRockRevival Ай бұрын

Great info

@Level-Sound 4 күн бұрын

Hello, I’m using a translator and I’m based in Germany. My question is this: most of the songs I play are always 4 minutes long, and after 4 minutes the song abruptly cuts off. Could you maybe make a video or show/explain how I can shorten a 4-minute song in Sono so that it has a clean ending? Is there a way to edit it in Soner? That would be amazing. I always watch your videos and really enjoy them, but I have to rely on subtitles. Still, the way you explain things is awesome, so thank you for that!

@user-cg3tx8zv1h 2 ай бұрын

Good stuff... FYI Chord AI link is missing...

@ChrisUnlocksAI 2 ай бұрын

Sorry about that, Here you go apps.apple.com/app/id1446177109 Thanks for watching.

@RubenAlfons Ай бұрын

Do the voices generated by suno sound weird or is it my ear playing tricks on me?

@user-cg3tx8zv1h Ай бұрын

Hi, I would really appreciate if you could please share the "Chord AI" link. I tried searching myself, but couldn't find the one you're using in the video.

@ChrisUnlocksAI Ай бұрын

sorry, here you go apps.apple.com/app/id1446177109 thanks for watching.

@billkeithchannel Ай бұрын

The link is in the description. It looks to be an Apple thing.

@user-cg3tx8zv1h Ай бұрын

@@billkeithchannel Appreciate it... That's really a little nifty thing...

@KirkNugent Ай бұрын

WOW! Mind Blown. Thanks for sharing this. Tremendous value.

@johnsydneymcnair7077 2 ай бұрын

Great video

@420SAVAGE 18 күн бұрын

Problem is with this method everyone's song is gonna have the same 5-6 female vocals from Audimee in their songs.

@TerenceKearns 27 күн бұрын

Chord ai has the same monthly cost as netflix here in Australia which is about AU$15/month. So over the whole subscription thing. Would have been happy to pay a decent up front cost.

@ChrisUnlocksAI 27 күн бұрын

When I went to subscribe it gave me an option to buy it out right for $100.00 USD. I'm not sure if they still offer that or not. And I am with you. I'm so sick of the subscribe model. Especially when most of the video ai generators are coming is at $30-$90 per month. Thanks for watching mate, I've got a real good friend from down under. Definitely will be visiting it one day.

@jerryjoys4561 Ай бұрын

doing the vocal conversion every now and then I notice that the new singer goes out of tune or croaks during playback, why does this happen?

@barneycasa 29 күн бұрын

Seems the Voice changer function is now only available as a local app for Windows / PC =( Or am I missing something? Too bad, this looked really good. Any other (online / mac) alternatives that people know of?

@Chihiroryuu 18 күн бұрын

if i send it to music aggregator is it safe?

@PraiseBeVinyl Ай бұрын

I'm still getting mechanical synth sounding vocals. But totally appreciate this process. I'll play with it a bit more and get better results I'm sure.

@EnhanceYourJourney Ай бұрын

Could not find the link for the vocal replacements like you mentioned

@ChrisUnlocksAI Ай бұрын

Here you go. It was further down in the description. Thanks for watching audimee.com/?ref=Chrisunlocksai

@billkeithchannel Ай бұрын

Link is in the description.

@sanjibsinger2010 Ай бұрын

Is Suno Accept Own Voice. For Duet Can We Include Both Female & male Vocal on Suno. Also Can We upload Our Composition and Suno Can make the Music Arrange According to Our Music Arrangement & Tune 9Composition).

@ChrisUnlocksAI Ай бұрын

Hello, yes you can input your own music and it will take it and use it to create a new song.

@HappyHumanist369 2 ай бұрын

I don't understand. When I chose a singer to convert, the one on the right didn't sound like original singer with cleaned up vocals. How did you get it to match?

@EduardKraft 2 ай бұрын

Chord AI for Mac - Link please 🙂

@ChrisUnlocksAI 2 ай бұрын

Here you go. Let me know if this doesn't help. apps.apple.com/app/id1446177109 Thanks for watching.

@EduardKraft 2 ай бұрын

@@ChrisUnlocksAI Thanks 👍

@Peter-gu9ph Ай бұрын

@@ChrisUnlocksAI Hi Chris thanks but this link is for an iPad version - you are showing a Mac desktop version and I cannot find where you got that - please help!

@HubsByG Ай бұрын

Sorry for the questions, beginner here. When you used ChordAI to detect the key of the song it said Dm 195bpm. When you imported and mastered it in Logic it said Cmaj 120bpm. My questions are why? Why did logic not detect it the same as ChordAI? The mastered output, was that in Dm or Cmaj? Thanks.

@ChrisUnlocksAI Ай бұрын

Hello, logic defaults to Cmaj 120 I didn't manually change that as I didn't need to for mastering. Chord al was the correct key. Hope that helps.

@ftaylor613 Ай бұрын

What about those terrible instrumental SUNO stems? They have the vocals still present in them. Can the instrumental be cleaned up also?

@ChrisUnlocksAI Ай бұрын

you can clean it up with RiP.x pro. it breaks the stems down into midi files. hope that helps.

@TargetPlusPoint Ай бұрын

Talk about persona selected option ,in suno

@KrushnaMohite-q3f 19 күн бұрын

Can I upload songs created using Suno AI V4 on KZbin and earn money from them? Will my KZbin channel get monetized if I do this?

@ChrisUnlocksAI 18 күн бұрын

Hello, yes as long as you are a paid sumo subscriber. And use the songs you create. I use them all the time and I am monetized.

@AI_Expedition Ай бұрын

And then introduce V4 released this week :)

@philharmonicAI Ай бұрын

What version is LogicX you are using ?

@ChrisUnlocksAI Ай бұрын

The latest version. Thanks for watching.

@DisabledTestedTravel Ай бұрын

What I think is funny is most of it just sounds like auto tune to me. My friend (who has a few LPs out) has been accused of using AI but she just used auto tune in some parts (high notes) to clean it up.

@RubenAlfons 29 күн бұрын

Well, the problem is that when you separate the voice from the music, the music becomes distorted and part of it goes into the voice... Do you understand?

@aaghardan2832 29 күн бұрын

You have option to download the stem version of your song in suno ai, which is seperated between music and vocal version of your song. But that feature only for paid version.

@RubenAlfons 29 күн бұрын

@aaghardan2832 ok, Thanks for the reply

@mac_erson9532 25 күн бұрын

@@aaghardan2832 how to do it in SUNO ?

@lonestaruk1561 Ай бұрын

Anyone know how to fix the instruments? And make it more natural?

@MrMoviemashup 2 ай бұрын

Link to Chord AI?

@ChrisUnlocksAI 2 ай бұрын

Sorry about that. Here you go. apps.apple.com/app/id1446177109 Thanks for watching.

@eccentricballad9039 2 ай бұрын

Hello Darlin' !

@TaylorLaurieAndTheAutumnLeaves Ай бұрын

Adobe Editor

@HappyHumanist369 2 ай бұрын

I don't like the fact that Suno doesn't give an option to save as flac or wav. Vocals sometimes do sound bad and using stems is too involving, time consuming and complicated for most people I would think.

@AbsurdImprobability Ай бұрын

downloading stems in suno gives you wavs. its a bit hit and miss on quality. Aside from the obvious conflict of interest with the sponsorship in the video there are alternatives. The method i've used that works great for me is to take the suno song and import it into udio, and then use udio's stem separation which gives you better separation than suno.

@C12omega Ай бұрын

cant even clone vocals perfect using suno

@rehdogg Ай бұрын

That’s just too much work. I subscribe annually to Suno, but when another AI generator comes along that has better vocal besides Udio I don’t like it, I will not be using Suno anymore I have emailed them numerous times letting them know how crappy the vocals are and I won’t be a repeated annual subscriber if it doesn’t improve. I refuse to do all the things you said here I shouldn’t have to.

@bleachedout805 Ай бұрын

I mean u wouldn't be an A.I. bro if you actually enjoyed putting in a bit of effort.

@rehdogg Ай бұрын

@ turns out in version 4 they actually listened and improved vocals

@Cr8tiveside 2 ай бұрын

You can separate vocals and instruments with Suno, Stems..

@BearBrave-wk3sy 2 ай бұрын

Go ahead and see what they sound like. He's already tried that.

@Clark-hj6vz 9 күн бұрын

Looks complicated. How bout I just give you the track to remaster for me?

@hyde.4737 Ай бұрын

suno Ai also messes with the beat:(

@ChrisUnlocksAI Ай бұрын

I've not had too many issues with that. Thanks for watching.

@Theofakustik Ай бұрын

Thx fyi , i want better to prod my song ❤

@HappyHumanist369 2 ай бұрын

I tried but it didn't work out. The voice doesn't sound like it's imbedded into the instruments. As if someone was narrating. No good.

@KuraiBeat Ай бұрын

ai voice .... use in lyricswindow [denoise] and at voice: [Ai Voice: -3Dba -lovely female voice, -pitched 2 tone up,] then what the voice setting and genre info be as example: [Ai Voice: -Screaming Ai Voice male, -young female voice, -hard riffs, -metalcore, -playing tone A G F C, -hard drops with trancevibe,] what you need more then suno? nerds are nerds no musicians! 🤓

@CzytamKsiążki 2 ай бұрын

Suno can separate stems too.

@pusabase 2 ай бұрын

I’d rather use my nutsack as a pincushion.

@timothyvass 2 ай бұрын

Suno can _pretend_ to separate stems 😅

@dreamer-1973 Ай бұрын

Suno is Russian roulette 😂

@nickerson1971 25 күн бұрын

I trash 9 out of 10. But it is still good

@generalawareness101 Ай бұрын

V4 did not help vocals either. :(

@AI_Expedition Ай бұрын

I beg to differ. I think it sounds a hell of a lot better. At least on most of my outputs.

@generalawareness101 Ай бұрын

@@AI_Expedition I remastered and overall quality improved but the vocals are still shite. Udio just does it better and Suno has the best interface, and features. Marry them, child is born, BOOM, the perfect system.