How To Make AI Cover Songs (Free Method)

Learn to make AI covers songs easily in a few minutes. Many ai cover songs have gone viral using voices like Frank Sinatra, Spongebob, Squidward, Drake, Plankton, and many more. It's surprisingly simple and puts a fun twist on your favorite songs.
--------------------------------------------------------------------------------
More from Futurepedia:
⚒️ Get recommendations on the best AI tools for your work: www.futurepedia.io/
✉️ Become the office AI-expert, 5min/week: futurepedia.beehiiv.com/
🐦 Follow on Twitter: / futurepedia_io
🖥️ Follow on Linkedin: / futurepedia
--------------------------------------------------------------------------------
Ultimate Vocal Remover: ultimatevocalremover.com/
Google Colab: colab.research.google.com/dri...
AIHub Discord server: / discord
Finished AI Cover song: • Shake It Off by Iron m...
--------------------------------------------------------------------------------
Chapters
0:00 Intro
0:21 Overview and demo
1:07 Isolate the vocals
2:29 Remove reverb and harmonies
3:35 Step 2 overview
3:55 Google Colab
4:13 Find voice model
4:48 Final Colab step to open interface
5:15 Vocal swap interface
6:18 Merge tracks
6:45 Demo

[@SandraKim]

Thank you so much ! Other "free" options online sound horrifying but your method is incredible and it sounds so good thank you !

[@sa1nt363]

yo what the fuck? same profile picture is crazy bro stop meatriding me dont copy me ever again

[@Official_Proxy]

I tried pasting links from the RVC model list but it just gives me an error "JSONDecodeError". After the session expired, I tried again, and the second time, it wrote "ModuleNotFoundError". Am I doing something wrong or is this method no longer working?

[@Akio3000ify]

I could just say thank you for pointing me in the direction of UVR alone! Thanks!

[@moppupaws]

this works perfectly! my only complaint and warning to others is that this process is actually INCREDIBLY difficult if you have slow wifi. the loading times for everything are drastically increased, and with the colab basically having a very short time limit on your link due to the code "not being allowed" on it, you can lose it very easily and have to start over simply because something didn't load in time. it's a pain and very tedious but it does work!

[@deadeyedingo9874]

do you know how to fix it my internet is terrible and i can never finish my thing

[@akaka6377]

@@deadeyedingo9874 if you're talking about being disconnected from the google colab, you can't really avoid it unless you buy the subscription because google doesn't want you doing that for free lmao. i bought the 10 euro one and made my own voice model no problem.

[@tslkieran]

Great video mate. Well explained, straight to the point and a high quality method. Cheers!

[@xxintixx4638]

It doesn’t work

[@Sadako2602]

@@xxintixx4638 Yes it does

[@XLost_LordX]

@@xxintixx4638 it works really well

[@mald34]

you probably did something wrong@@xxintixx4638

[@AmandeepSingh-kt2dc]

You keep getting better man. Keep it up.

[@umario251]

Thank you so much! So simple to follow and was way easier than other methods! Appreciate it!!!

[@yutyrannushuali3967]

Question, could you instead remove harmonies first and then reverb second? I ask this because I notice it has options for instrumental or vocals only and you select the ladder, would this save a step or does it work better when you isolate vocals, then reverb, then harmonies?

[@Whiskey_Rowan]

Thank you so much!! I had been wanting to try this for a while and was intimidated by all the other websites and stuff. So this really helped me. Have a great day! :D

[@GeanImation]

That's both amazing and horrifying. Possibly the last nail in the coffin for great new, genuine artists to emerge. The information technology age has been killing all incentive to dedicate your life to perfect songwriting and this may be it now. Nice find on Bruce singing Rush's Tom Sawyer

[@itsonlyatree9185]

idk I think a lot of people share the same sentiment, but that means there will still be a market for great authentic songwriting. this could even cause more people to seek out genuine artists rather than generated content

[@PartyTime432]

Dawg, we just wanna joke around and make random people sing songs. No need to get depressing

[@jennyfisher3765]

Homeboy people are making spongebob characters sing billie jean it’s not that serious

[@SaphyOW]

@@PartyTime432FR

[@MukeshKumar65387]

People are making covers with Hitler singing peaches I don't think artists are gonna be hurt that much

[@TheSoundYouNeedAlong]

Amazing and fantastic! Thank for sharing!

[@Tygermite]

Had me worried at first but solid tutorial. Thanks!

[@wajahatkhan9455]

Thank you for sharing with us. It's mesmerizing.

[@metaversenft]

Is therre a paid version that speeds this up?

[@simonsanchezkumrich8489]

Sounds pretty well, i might try it, thanks for the tutorial!

[@futurepedia_io]

It’s definitely fun

[@tiktoktrendsetters8187]

You are an absolute life saver I had a project in mind and was using the training model in colab and I just couldn't get it to work for me. This is awesome thanks again

[@yoursecret]

thank you mate , this helped a lot clear and precise.

[@HauntFreak13]

THANK YOU!!!!!!!! This was great and easy to understand.

[@cib2938]

Gained yourself a subscriber, thanks for this well explained video !

[@scoobydoo5390]

Amazing tutorial! Thank you!

[@MrUniversalTouch]

U my friend deservers a sub I like how you executed your video without saying anything else and just kept focusing on explaining instead of doing something completely unnecessary like some videos

[@alinenin4028]

Thank you so much!! This helped me a lot 💚💚

[@GaryStudio]

Whoaaa thanks man for using part from my cover, really appreciate that \m/ great job!

[@stevePurvis1]

Brilliant video pal, Thank you.

[@Tinkatonix]

Thank you for the information, easy to understand.

[@Nameless-vm9wy]

whenever i try to download the model it says theres an error

[@thewebstylist]

You ROCK all ways my bro!!

[@TriforceChad]

This is super cool, but trying my hand at it the colab has been breaking at step 2 with a "JSONDecodeError". Any ideas? I've tried several voice models from the discord and get the same error each time

[@aurodj]

Thank you so much, a very explained video !

[@LoudPxckk]

Gotta love a creator respondin to da people who aint acutally try dis and sayin oh great video but not respondin to da people who did try it and havin issues

[@D00m1no]

4:27 Kamil Steinbach, Matemaks, Mirek Stabiński Tego się nie spodziewałem

[@kareemyoussef2304]

so glad i refound this video, spent like 30 mins searching for it

[@combineelite6476]

Amazing guide, thanks! for those stuck with "Connection errored out", you need to be quick when uploading and converting your vocals because the code on google colab will stop running after some time (from what I understand) causing the error to appear

[@wetaughted4640]

Everytime I try the runtime just ends before I’m able to download the file.

[@ToneTechStudios]

me too@@wetaughted4640

[@manowarjr80]

me too, is there a solution?@@wetaughted4640

[@diegovalentino2083]

This video is just great man. Really well explained. Same as other in the commnents, it would be great if you explain how to train your own model. There are a lot of spanish speakers who would like to train a model for spanish singers. Most of the videos in youtube are for english speaking singers. In my case, I would like to train some models for Vicente Fernandez or Juan Gabriel.

[@futurepedia_io]

That makes sense. I'm not sure I'll be able to make one any time soon, but this is a google doc from the aihub community with all the steps docs.google.com/document/d/13ebnzmeEBc6uzYCMt-QVFQk-whVrK4zw8k7_Lw3Bv_A/edit#heading=h.bjzhhhcn3f69

[@jamesmartin2063]

@@futurepedia_iodo you remember DarthAnonymous how make all AI covers and other stuff on his youtube channel sadley his youtube channel Got Deleted

[@yellow0285]

@@futurepedia_io cant you do this with a instrumental song or does it need to be the vocals

[@user-lk8qn7cp8h]

Step 2 doesn't work for me, when I paste the link it gives me this error: JSONDecodeError: Expecting value: line 1 column 1 (char 0)

[@Eggybreadcool-dl8vj]

same here

[@BenVanLooy]

@@Eggybreadcool-dl8vj same!

[@3km143]

Same here

[@Itsreal85Clips]

@@3km143 have to pay. step 2 is not free.

[@ninjamonkey0497]

same

[@maryamnavabi140]

Thank you so much, It was amazing.

[@CurtisCameron]

Thanks man this is perfectly explained

[@8batz]

thanks! it worked like a charm

[@kuruma6969]

Thanks, sir. Because this free options, i was make some AI Cover song on anime characters especially on Indonesian Cover Songs. Such a great tutorial. Also can swap speaks voice. Thanks the tutorial, sir!!

[@acatisfinetoo3018]

Thanks. I am new to this and your video was very easy to follow and get started.😉

[@Five_pebbles_official]

Bro your pfp💀

[@NoahFarber]

I can't believe UVR is free!!! Thank you for sharing 😁🔥

[@kobold7466]

excellent video. not too much time wasted on unimportant info and my cover came out great!

[@Zombieirl]

This is the best tutorial out there thank you so much 😭🙏 would you consider also making a tutorial on how to train our own custom voices? Ai hub doesn’t have the specific model I need

[@schmdtta8106]

thats exactly my problem too

[@kuruma6969]

Um... sir, the 2nd step (swap vocals) isn't working again. Well it works at 1st to 2nd cell, but when gradio sequence (local/public URL), it won't working. Please fix it! Thanks

[@icardi187]

Bro I think thats the best tutorial i've ever seen

[@zilvinass965]

Amazing tutorial thanks! Quick question you mentioned that you do different channels for different content. What you advise do different channels on one account or better create separate accounts for each channel? As my content will be very different between channels and some will be experimental. And want to protect myself if one of the channels gets in trouble with KZbin, that won't affect all of them.

[@RoRo-OFF]

well explained! thank you so much!! +1 sub😊

[@griffinhealy200]

Thank you so much!!!

[@krwnik8419]

Doesn't work for me: ''JSONDecodeError: Expecting value: line 1 column 1 (char 0)'' Edit: It worked, but I used RVC.

[@jenko1017]

Same happens to me

[@krwnik8419]

@@jenko1017 I watched another video and it works.

[@jenko1017]

@@krwnik8419 Cool! Can you share the video or channel's name?

[@krwnik8419]

@@jenko1017 kzbin.info/www/bejne/pYClm6GqfpergZIsi=YlC5vCugzgua96is

[@krwnik8419]

@@jenko1017 name of channel is BidenWasTaken. Name of the program is RVC GUI.

[@KevinTesch]

Bro thanks you helped a lot

[@ornery8x559]

awesome video! have you done a video on how to train your own model?

[@flickdaboss]

thanks for this, i took a few liberties but this really helped :)

[@JesterOfLemmings]

Thanks, currently making a My Way cover.

[@salecroft1410]

This video was very helpful to me! Thanks for that I want to make an AI Cover with 2 voices, the two singers have their own part but also sing together at some points. Are their any tricks to know?

[@MasterBrain182]

Obscurious 🔥🔥🔥 🚀

[@Jswift77]

Thanks so much for the video. The one thing I can't get to work is the very last part. I always hit refresh and the audio I uploaded never pops up there. Any reason why?

[@thomasgreco391]

This is great. Is there another place to find more voice models?

[@emilysilverio6297]

I got stuck at downloading my link to the models list--no matter what link I try from the discord or spreadsheet, it says "error" in red. How can I fix this?

[@user-ib1qx9ip9j]

same

[@filipposk2612]

Same did you find a way to fix it if yes pls help

[@OneAnimeGuy]

bro same can anyone please explain how to solve this.

[@shehzadhaiderbaloch]

Awsome brother

[@Shaaashaaa_0]

thank you so much!

[@Horeffied]

How do you upload an audio file and have an Ai tool sing to it with lyrics you've rewritten?

[@feff33]

Ty. Do you know if there`s a way to change accent with ai? (i mean vocal tracks, not google translate stuff)

[@carlosauer2533]

Your video really helped me out a lot but is there maybe a way to install the interface as a programm on your computer? I've got a lot of files to AI-ify and always having to do everything all over again isn't that efficient.

[@user-te3mf6dp5g]

Hi, thanks for the tutorial! Is it possible to make a cover to music? Or for example, if there is a melody, what would the AI ​​come up with an accompaniment?

[@lavinchi6709]

Thank you so much🎉🎻🎸🚀

[@Miliano369]

Is google collab working for you? I think it stopped working. May download to run locally but not sure if it works on Mac

[@fatcatfinancialeducation9735]

Very good video. I have a question: could you make a cover of a song in which the singer's voice will not change, only the words will be changed?

[@Ysra-Otaku]

Thank you!

[@caniscanemeditbully214]

Amazing video, you won a like, I'll return to this frequently now!

[@theantil7]

FIrst off thanks for the tutorial, made a few great covers, but I can't help but wonder why when taken to the link given on step 3, that it gives me constant errors all the time. I have to go through steps 1-3 again and again, and only like 1/5 tries do I finally manage to get it to cough up the new vocals. If you could help me out lmk. Would installing locally improve my odds?

[@bobertbolero]

Fantastic video, I made my first AI cover!

[@RockBandPlayer84]

Your method produces better results than the expensive monthly ai subscriptions

[@TaraszewskiLabel]

thank you for your help

[@tylerczerniak9945]

Do you end up merging back in the reverb and melodies/retraining the melodies to the output model, or just leave them out?

[@chouchfroukh]

Incredible.

[@SimranKaur-qk8my]

thank you for a well explained video! How do you go about avoiding copyright issues with posting the actual ai cover songs?

[@ShidaSkunker]

I keep on getting JSON errors trying to use this method.

[@Umjion]

Same😢

[@zcasual4157]

same, guess its not working anymore

[@DavidxLinkan]

Anyone ells getting the error ModuleNotFoundError: No module named 'mega'???

[@Lvv3nd3r]

4:52- When I tried to download an AI voice from Discord, I got an error under it that said "ModuleNotFoundError" and it says Python Imput and other things under it and I don't know what to do, can you help me plz? Tysm :3

[@NyanTheFurry]

Is it possible to make a duet using this? I assume I’d probably have to split the audio when it goes to the next person, but I’m not sure what to do when they both sing at the same time?

[@Neroli107]

Thank you for your tutoriel. I have got an error at step 2 "JSONDecodeERROR" Could you help ?

[@grandmasoldpotroast642]

4:00 im getting the error "you may be executing code that is disallowed"

[@janyajones6975]

Me too and idk how to fix it

[@lostcause15]

just accept it

[@Zenbusebun]

Hey, Is it possible to modify a own created voicify model after it's trained?

[@DigitalVentures-jd8lq]

Do you know maybe an app (tool) which can separate an audio on instruments channels?

[@MidnightDystopia]

Thank you for this going to be doing this alot

[@futurepedia_io]

Awesome! I’ve been having a ton of fun with it

[@m9u7shuriken]

hey, what about if I want to use a voice from a video for narrating purposes, are the instrutions still the same?

[@AnthonyDev]

Excellent tutorial, now I don't need search more acapellas for remixing purposes. 😎

[@futurepedia_io]

It’s crazy how good UVR is. Especially since it’s 100% free

[@NoahFarber]

My mind is blown

[@vidnoz]

thx for sharing that

[@internetuser2291]

whenever i put any model link in the collab page,it says an error occured...

[@daniilstones418]

Amazing video man thank you, I'm having trouble downloading the audio from Audacity and opening it somewhere to put it to tiktok

[@ItsHayBale]

We need to get this video updated, we are all having problems on steps 2 and 3

[@ItsHayBale]

I am not sure what you’re talking about when you say comment out the codes. Shouldn’t each file have different code? I followed your instructions for step one but I still got an error in step 2 saying the client was not found. I deleted the code along with the line after it and it seemed to work. However I am still getting the “NameError: name ‘weeee’ is not defined” in Step 3 and I’m unsure how to solve it.

[@HijiyamasYakisobaPan]

@singingrobot316pls explain what commenting out means

[@Cheese_Doggy12]

​@singingrobot316 I haven't learn python I only have learned c# and java

[@tonihaltola6295]

@singingrobot316 you also need to comment out line 4

[@MisterE180]

What can you do if there isn’t a model of the voice you want to use?

[@jydenm95]

What model do i have to use if i want to remove the last vocals from the instrumental part? You can still slightly hear them on the instrumental part

[@codypolm7072]

If the celebrity you want isn't in the list, is it possible to extract their vocals from one of their tracks and upload it as their voice?

[@ARTICLERAUL]

Is there any tool which can convert image to anime or cartoon in bulk, having batch features - There are some including photoshop but just hangs while i try.

[@Goalkick_shorts]

gradio not working diverted to cloudfare and it works. However, connection errored out in RVC WEB while converting, last part awww. any help. thanks

[@CorinnaXcool]

How do people go about making songs that feature more than one character and timing the multiple different voices perfectly? Do they use Audacity? Are there any tutorials for that specifically? Thanks

[@mkall]

they make multiple tracks with different voices and then cut and mix.

[@AlexCaspian]

are there any voice AI models to make cover vocals free for commercial use and releasing on music streaming platforms?

[@joaobartolomeu4323]

Do you guys know why Mangio RVC it's always disconnecting while downloading? I cannot finish any download because of it

[@ampharos3106]

do you have a tutorial on how to make a voice model with this?

[@IntelliMindA.I.]

cool stuff maybe ill be big KZbinr like you guys !!!