Audacity Forum

I know this is possible to an extent, examples below.

http://www.youtube.com/watch?v=W13JAwQ7Odg

http://www.youtube.com/watch?v=JPJhFhCYjbY

http://www.youtube.com/watch?v=V1lyHhAB3ts

Notice, the background music IS there, but it has been suppressed so much it gives the illusion of only the voice being heard, when singing.

It should also be noted that these songs are in Japanese, when singles are released in Japan, they also come with the instrumental version of the song, if that has anything to do with it. Whether or not the two versions of the songs are the same, I have no idea. (By same I mean frequencies lined up perfectly.)

I tried contacting the video uploader, but the person said someone else did it for them. When I tried contacting that person, I found their account was deleted.

I've tried so much. Messing with high pass, low pass, inverting. Getting the insturmental version of one song, zooming in to the start of the song and matching the frequencies to the best of my ability, and downloading the Kn0ck0ut plugin.

Google searches return results with people in forums saying its immposible, but those three examples prove it can be done. Some better than others.

Any ideas? Thoughts? I would really love to listen to some of my favorite songs with just vocals.

not easy if it could be done at all

they started with multiple tracks and mixed it that way

now if you could find a good karaoke program that removes the voice and makes a second file with music only
then you could take the music only file and subtract it from the original to leave only the voice
wont be perfect but might do what you want

dont you hate it when they bury the voice under the music
never can figure out why some folks do it that way

This has been discussed many times - it's not easy and requires special tools - see here: http://forum.audacityteam.org/viewtopic ... 580#p58782

whomper wrote:now if you could find a good karaoke program that removes the voice and makes a second file with music only
then you could take the music only file and subtract it from the original to leave only the voice

You appear to be suggesting this technique that does not work: http://forum.audacityteam.org/viewtopic ... 580#p58796

well then, how was it done in the two tracks posted above? japanese singles all come with a karaoke track without the voice, but ive read everywhere that it doesnt work if you try to subtract the voice and voiceless songs

<<<japanese singles all come with a karaoke track without the voice>>>

Because they were made that way in the studio? Karaoke is or was a big deal in Japan and I'm not shocked to find that karaoke versions of the song are issued right along with the regular performance as a standard product.

Right up at the top of each discussion of this vocal management thing is the admonition that it's a snap if you have all the studio capture tracks and mix-downs. They don't take a CD or, worse yet a download and try to rip the vocal out. They load the original song from the studio backup drives -- or tapes -- and turn track 47 off which has the vocal on it and possible two or more tracks as well which have the stereo vocal effects and "sweetening." Instant karaoke mix -- and it's in perfect fidelity and stereo.

If you can do that, all things are possible, perfectly executed and in any combination. Doing any of this in post production -- in either direction -- fails many more times than it succeeds. It can be said that it fails most of the time.

There was one posting a long time ago that didn't believe us and said he was going to keep pounding away at it until he found a way. We never heard back. He's probably still pounding.

Koz

Johnnymushio wrote:japanese singles all come with a karaoke track without the voice, but ive read everywhere that it doesnt work if you try to subtract the voice and voiceless songs

IF you have the full song on CD, (1)
AND the full song without vocals, also on CD, (2)
AND the backing mix on both recordings are from the same digital master, (3)
AND the mix on both CD recordings is identical with the exception of the vocal (which is absent from one) (4)...

then YES, you can isolate the vocal, and you can probably do it to quite a high degree, and even in stereo. It's then just a matter of lining up the two recordings accurately and inverting one of them.

but... you need 1,2,3 and 4.

If you only have the original full recording, or worse still, a badly compressed version of the full recording, then no chance.

Typically it is asked "I have this song that I recorded from YouTube - how do I remove the backing to leave just the vocal?" The short answer is that you can't. The long answer is that if you spend a lot of time and effort with the specialised tools mentioned previously, you may be able to achieve partial separation, though the result of hours of detailed work is not likely to be impressive.

Of course, if you have access to the original multi-track recording, then as Koz says, it's a snip, just mute the channels that you don't want.

incorrect. i have 1-4. it doesnt work. if that did work, you could remove the vocals from the songs with audacity, since that is very easy to do with inverting. then take the vocalless track that was made on audacity and the original track, both being the same length, then invert that to get just the vocals. but alas, that doesnt work.

my three examples above are proof that IT IS possible to remove music and keep vocals only. but no one can give a simple tutorial on how to do so. contacting the video uploader on youtube, they say their friend did it. not much help there

Johnnymushio wrote:my three examples above are proof that IT IS possible to remove music and keep vocals only. but no one can give a simple tutorial on how to do so

See here: http://www.elevayta.net/product13.htm