Effectively isolating the vocal from the rest of the track

rolly · Post by **rolly** » Mon Dec 12, 2005 12:46 pm

Hi,
I wondered if anyone has any tips using Ableton for isolating the vocal track within a song quickly and effectively so that the vocals can be used against a new backing for mash-ups and remixes. I've tried using the EQ 3 and 4 plug ins but I find it very difficult to isolate where the vocals are within the audio spectrum. Is there a nifty VST plug-in that works with Ableton that isolates male and female vocals as well as the doing the Karaoke reverse? On the understanding that most vocals are panned centre I've tried Adobe Audition's Central Channel Extractor filter but the results were not that great. Maybe there is no one plug-in that can do this, in that case, I'd appreciate other people's tips of what has worked for them in creating clean isolations of vocals. Thanks, rolly

hambone1 · Post by **hambone1** » Mon Dec 12, 2005 12:48 pm

Try a quick forum search. I think this topic has been discussed in detail.

The final conclusion was something like "taking the eggs out of the cake once it's baked..."

conny · Post by **conny** » Mon Dec 12, 2005 2:47 pm

Trevor Magnusson has just done a new plug, VoiceTrap.
Hes stuff has hight quality.
Have not checked this myself, though.

http://www.cloneensemble.com/

// C

spiderprod · Post by **spiderprod** » Mon Dec 12, 2005 3:21 pm

conny wrote:Trevor Magnusson has just done a new plug, VoiceTrap.
Hes stuff has hight quality.
Have not checked this myself, though.

http://www.cloneensemble.com/

// C

it works all right if the vox signal is in the middle & the instruments are panned like old rock recrdings , it's quite difficult with recent music .

rolly · Post by **rolly** » Tue Dec 13, 2005 10:57 am

Hi Thanks for the replies to date. rolly

ekitel · Post by **ekitel** » Tue Dec 13, 2005 11:08 pm

use a bandpass filter instead, this won't really isolate the vocal, but it helps. you don't need to have the vocal 100% isolated to make a cool mash-up, sometimes you just work with what you got and then see how it sounds, a little music from the original track might not hurt the final product

gomi · Post by **gomi** » Tue Dec 13, 2005 11:26 pm

hambone1 wrote:Try a quick forum search. I think this topic has been discussed in detail.

The final conclusion was something like "taking the eggs out of the cake once it's baked..."

i love the paint on my walls, but there is a bit too much red..

take it ou!

fabi · Post by **fabi** » Tue Dec 13, 2005 11:33 pm

Vocal Extraction

This technique is a scientific fact, it is humanly possible to lift a vocal from a track that has music behind it. Many people have tried to develop a plugin to do this procedure but have sofar failed. The process needs to be precise and has conditions for it to work. It is necessary to use both an instrumental and a vocal mix and the waveform must be identical. You can not use vinyl as the waveform will alter slightly as the vinyl spins. Also you cannot use MP3 because of the compression in the conversion. Finally, if the tracks you planned to use have been mastered by two different engineers, this process will not work because of the compression and eq techniques of both engineers would never be identical.

When you put two identical waveforms on top of each other the amplitude will double in size (which is about 6db).

Invert

When you run through this procedure later, this part of the process will make more sense. In order to get the vocal lifted away from the track you need to invert the phase. What this means is we need to flip one of the identical waveforms so there shape is oppersite to each other like a reflection in the mirror. When you put the inverted waveform together with the normal waveform. The musical element will cancel each other out leaving the vocal dry. The reason this happens is because the vocal version and instrumental version will have different characters to each other – that being the vocal. So if the two backing track’s are identical and cancel off each other, in effect you are left with the additional character’s of the waveform.

Vocal Extraction in Wave Lab

1. Import your digital tracks into Wavelab as audio files
2. Select the instrumental and drop down the Process Menu – select Invert Phaze and double click on the waveform.
3. Save this as an inverted file.
4. Create a new ‘Audio Montage’ from the file menu. File – New – Audio Montage and choose sample rate 44.1 kHz.
5. Create an additional audio track by dropping down the track number button
Insert stereo track (above or below).
6. Align the tracks together so they play at exactly the same time. Zoom in and try to spot visual landmarks like the beats or use a drop out section further in or the breakbown if the track has one with a sweet spot to work with.
7. Wave Lab snaps to sample level which means it opens to the highest level of the zoom detail. This improves your visibility when aligning the two tracks.
8. To render your accapella as an audio file, click on render on the master section panel,
9. Choose whole montage and tick create specific file box and choose a location on the hard drive for your accapella.