I'm looking for a service (preferably free) which will cleanly isolate audio when I have a separate audio file that has everything I want to remove. More specifically: I have audio of someone talking and a song in the background. I also have audio of just the song. I'm looking for a program that will give me just the person talking.I found vocalremover.org which usually works really well but it (understandably) fails when the song also has vocals. I also tried doing it manually with audacity but the result was kind of scuffed.Any help appreciated.
>>1457686You could always become a feminist and fight against FemSec
>>1457689um okay
>>1457686https://github.com/Anjok07/ultimatevocalremovergui/releases
>>1457686
examplehttps://youtu.be/fNFzfwLM72c
>>1457712vocal extract
>>1457712instruments only
>>1457697>>1457705>>1457712i appreciate the help but i think you missed a part of my post. the audio has both a person talking AND a song with vocals, and i'm trying to get just the person talking. this seems like it would isolate both the person talking and the song vocals
have you checked out the links here>>>/wsg/5515011
>>1457796i did now. it looks like mostly music generation and vocal isolation like what the other anon suggested, but all of this is way more complex than what i need and also doesnt accomplish what i need >>1457762thank you for the link though. didnt know about this thread
bump
>>1457686Check out RVC v2. It's used to make AI voice covers but it comes built-in with multiple options for removing vocal shit from songs. https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI
>>1457910Also allows you to isolate audio, forgot to mention that.
>>1457910having a tough time understanding what *exactly* it does because so much of it is in chinese. you mentioned removing vocals but can it specifically do this >>1457762 ? all i see is an input file and an output file and no way for me to give it an entire track to remove from another track
still bumping
still looking for a software that will let me remove a specific song from an audio file, NOT just isolating vocals
if you have a copy of the song alone, AND that exact same copy of the song is playing under someone talking (no EQ differences etc.) you can line up the tracks so that they sync exactly, invert the audio, and thus isolate just the person talking.https://manual.audacityteam.org/man/invert.html
>>1458614i've tried doing this and it "sort of" works but i guess the audio isnt exact enough because it basically just makes the song quieter and lower quality. i've also tried using audacity's "noise reduction" tool and it's better at cutting out the song but it the remaining audio is scuffed. was hoping there'd be some AI-supported tool because eyeballing this is pretty difficult but it seems like every tool is just doing vocal isolation
>>1458617what you're asking to do via AI is basically impossible. no model is trained specifically to distinguish a song vocal from ordinary speech. the manual approach via Audacity is the best you can do.you can also try tools like izotope to zero in on certain frequencies and scrub them out. but you're effectively asking to unbake a cake, it can't be done.
>>1458682if i can almost do it manually then it's 100% possible to finish the process with AI, particularly using it to clean up the remaining audio. it doesnt need to be all manual or all AI
