Sampling question

[Cat Gas]

When sampling, how do you get rid of background beats etc?
For example when you're sampling a vocal from the middle of a song, how would you get rid of the percussion in the background?
So far putting a filter on it and EQing isn't really effective, because it seems to muffle the vocal

 

[thedjnifty]

When sampling, how do you get rid of background beats etc?
For example when you're sampling a vocal from the middle of a song, how would you get rid of the percussion in the background?
So far putting a filter on it and EQing isn't really effective, because it seems to muffle the vocal

Let me save you some time and let you know that it's virtually impossible

 

[Cat Gas]

Let me save you some time and let you know that it's virtually impossible

Alright, but is there anything I can do (apart from loads of EQing) to make the vocal stand out more than the background instruments?

 

[second opinion]

i remember something a while ago about phase inversion.
this is where you take a full tune put in wavelab with an instrumental version of said tune and you somehow in wavelab(this part i aint got a clue on lol)reverse the phase on the non vocal version and it merges with the original tune leaving just the vocal.(how successful this is i dont know but might be worth lookin into for what you need, or just get accapella of said tune and use non vocal bits of it to keep feel of what you doin)
hope this helps if not ill get me coat

 

[motion audio]

Let me save you some time and let you know that it's virtually impossible

This, once all the seperate tracks in the session for a tune have been summed to one then thats it, theyre just one audio file, theres not really a way to "seperate" them as anything you do to process instruments will also affect the vocals.

 

[thedjnifty]

i remember something a while ago about phase inversion.
this is where you take a full tune put in wavelab with an instrumental version of said tune and you somehow in wavelab(this part i aint got a clue on lol)reverse the phase on the non vocal version and it merges with the original tune leaving just the vocal.(how successful this is i dont know but might be worth lookin into for what you need, or just get accapella of said tune and use non vocal bits of it to keep feel of what you doin)
hope this helps if not ill get me coat

I can imagine this working to an extent actually, I've been reading up on phasing and how two identical signals can cancel eachother out somewhat

But if you can track down an acapella then that makes your life a milllion times easier from the word go

Trouble is, acapellas often don't exist for a lot of music

 

[Neomind]

This, once all the seperate tracks in the session for a tune have been summed to one then thats it, theyre just one audio file, theres not really a way to "seperate" them as anything you do to process instruments will also affect the vocals.

lies.

You can do the phase inversion thing... obviously the part which has the vocal will have an instrumental along with it, and then the vocal will shut up and the instrumental will go on the same, you can get that part put it in concordance with the vocal part and INVERSE IT'S PHASE, where you had peaks there will be "valleys".

http://www.youtube.com/watch?v=S0BLpoEYx4U

 

[motion audio]

lies.

You can do the phase inversion thing... obviously the part which has the vocal will have an instrumental along with it, and then the vocal will shut up and the instrumental will go on the same, you can get that part put it in concordance with the vocal part and INVERSE IT'S PHASE, where you had peaks there will be "valleys".

But what I'm saying is that it will still affect the vocals in one way or another, it will work to a certain extent, but just because your summing it with the reversed phased instrumental, thats not to say the vocals dont contain any of the frequencies in the same range as those of the rest of the track. When you play an reverse phase instrumental alongside a full track, if its aligned right and completely the opposite polarity then yes, it will cancell out all the identical frequencies in the original track, but what happens if some of the vocals happen to sit around the same frequencies? Theres no way to distinguish between the vocal part and the instruments, as theyre summed into one file.

Im not saying its not possible to seperate them to a usuable state, Im saying its not going to be perfect, as the vocal will be affected in some way.

 

[kama]

lies.

You can do the phase inversion thing... obviously the part which has the vocal will have an instrumental along with it, and then the vocal will shut up and the instrumental will go on the same, you can get that part put it in concordance with the vocal part and INVERSE IT'S PHASE, where you had peaks there will be "valleys".

http://www.youtube.com/watch?v=S0BLpoEYx4U

Its still 99% likely not to work because it would have to be EXACTLY the same audio without the vocal. When talking about played instruments, it's very rare to have exactly the same part (ie. copied loop of all instruments) without a vocal part. Even in loop-based electronic music if you manage to time the clips sample-accurate, there is still going to be things like delay-based effects, synthesizer randomisation, LFO's etc. which will be different from the other clip 90% of the time.

 

[kama]

But what I'm saying is that it will still affect the vocals in one way or another, it will work to a certain extent, but just because your summing it with the reversed phased instrumental, thats not to say the vocals dont contain any of the frequencies in the same range as those of the rest of the track. When you play an reverse phase instrumental alongside a full track, if its aligned right and completely the opposite polarity then yes, it will cancell out all the identical frequencies in the original track, but what happens if some of the vocals happen to sit around the same frequencies? Theres no way to distinguish between the vocal part and the instruments, as theyre summed into one file.

Im not saying its not possible to seperate them to a usuable state, Im saying its not going to be perfect, as the vocal will be affected in some way.

...but then again if you have the same track as an instrumental version, you will be left with the vocal, interlapping frequencies will not be affected. You can try this. Add 2 different voice samples on different tracks in your daw so that they play over each other. Use ones that have overlapping frequencies, like male voices or something, preferably the same speaker. Then duplicate either one of those tracks and when you reverse this new track's polarity, you are left with only the unduplicated sample.