Just wondering what program I could use and how would I be able to remove voice from a song. My friends and I want to try and make songs like the Grey Album, except we won't be selling it, we just want to do it for our own entertainment. Thanks for the help.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
How do I remove voice from a song?
- Thread starter nsknike
- Start date
- Sort by reaction score
Basically it is impossible to remove the vocals and leave everything else sounding the same. You can try EQ'ing the track, removing the frequencies that the vocal range is in, but this will also remove the frequencies of any other instrument that lies within the range. Other than that, you really cant do much unless you get a version sans lyrics to begin with, find a MIDI version, or spend mucho dinero on Pro audio apps that still will not do the job.
Many rap artists release songs in instrumental and a capella (vocals only) versions so that people can remix them. That's how the Gray Album was made, I would guess.
faintember said:
yup, it is pretty messy to do but not necessarily difficult (nor do you need Pro audio apps). are you in school? take a signal processing class, go to your school's lab, load up matlab. write your song to .wav, load song, fft away... but like faintember said, you'll null out frequencies with instrumentals.
removing vocals from the song requires lots of skill in pro audio, bundles of expensive equipment/software, and loads of hours skimming away frequencies on waveforms. there is currently no software that can automate this, and if someone were to make one then they'd be absolutely minted
the fact is, huge music companies spend loads of money getting people to do this, and it takes a long time. people are still trying to separate tracks off old beatles albums because they used a 6-track in the 60's so they would have to bounce tracks down. people are still trying to do this so that they can digitally remaster them
the fact is, huge music companies spend loads of money getting people to do this, and it takes a long time. people are still trying to separate tracks off old beatles albums because they used a 6-track in the 60's so they would have to bounce tracks down. people are still trying to do this so that they can digitally remaster them
Hi you could try this http://audacity.sourceforge.net/
Read this as well http://audacity.sourceforge.net/help/faq?s=editing&i=remove-vocals
Have fun
Read this as well http://audacity.sourceforge.net/help/faq?s=editing&i=remove-vocals
Have fun
With some stereo recordings you can extract some parts by inverting the phase of one channel, google for "oops effect". This works best on older recordings from when stereo was still somewhat of a novelty, and there was a tendency to try for an exaggerated 3D effect.
Well, it is all but impossible to do without altering the original (which is what i was referring to). And some Pro apps do this job better than, oh Matlab, but for the average user the difference is negligible. Either way, there is no good way to remove vocals and leave everything else as it was before the vocal removal.hhlee said:
nsknike said:
i very much doubt soundtrack pro is possible of doing it. no existing program can automate this process - it has to be done manually
Strip Vocals from Track
In order to remove something from an audio track you use something called common mode rejection. The same principal used in balanced audio cables. Basically you flip the phase of part of the signal 90 degrees and then flip it again 90 degrees when it reaches the end of the chain. Any noise or signal picked up is cancelled out, i.e. revese the left channel 90 degrees and anything common to them both should cancell out.
To rip vocals out of the song this way would also take out some of the frequencies of other parts of the mix.
However, for about $80,000 Cambrige(sp?)/CEDER makes Audio Forensic and restoration hardware based on some very crazy algorythems. Check it out.
http://www.cedar-audio.com/
In order to remove something from an audio track you use something called common mode rejection. The same principal used in balanced audio cables. Basically you flip the phase of part of the signal 90 degrees and then flip it again 90 degrees when it reaches the end of the chain. Any noise or signal picked up is cancelled out, i.e. revese the left channel 90 degrees and anything common to them both should cancell out.
To rip vocals out of the song this way would also take out some of the frequencies of other parts of the mix.
However, for about $80,000 Cambrige(sp?)/CEDER makes Audio Forensic and restoration hardware based on some very crazy algorythems. Check it out.
http://www.cedar-audio.com/
Even though this thread is already kind of dead, I thought I'd add my two cents.
I have tried both Audacity and Audion, and I've realized both give you the exact same quality. However, I prefer Audion simply because it's easier to use. Audacity requires you to split the stereo tracks, subtract one, and make them both mono tracks. Audion does that all for you.
As for the quality, it depends on how lucky you are. Some songs come out virtually untouched while others sound like nails on a blackboard. It ultimately depends on how the song was mixed in the studio. My brother is a recording artist, and he and his band made a song, but decided to add some extra vocals and retooled keyboarding into the song, etc. several months after the original was recorded. When you take the newer song and squeeze it through one of these apps, you can really hear what they've added in, which is pretty much taken out of the song (except for the extra background vocals.) It seems like just about every song I put in these, I can somewhat hear the original singer's voice, but it is very quiet and I doubt the average listener will notice. I think I'm a bit of an audiophile.
I have tried both Audacity and Audion, and I've realized both give you the exact same quality. However, I prefer Audion simply because it's easier to use. Audacity requires you to split the stereo tracks, subtract one, and make them both mono tracks. Audion does that all for you.
As for the quality, it depends on how lucky you are. Some songs come out virtually untouched while others sound like nails on a blackboard. It ultimately depends on how the song was mixed in the studio. My brother is a recording artist, and he and his band made a song, but decided to add some extra vocals and retooled keyboarding into the song, etc. several months after the original was recorded. When you take the newer song and squeeze it through one of these apps, you can really hear what they've added in, which is pretty much taken out of the song (except for the extra background vocals.) It seems like just about every song I put in these, I can somewhat hear the original singer's voice, but it is very quiet and I doubt the average listener will notice. I think I'm a bit of an audiophile.