                      User's Manual for Vocal Eliminators

                               by Zhang Zhichao
                            (zzcworld@hotmail.com)
================================================================================
* Preface

  This document is not written for a special vocal eliminator, with this you can
use most of vocal elimination software and get a good result for your songs to
be devocalized.

  I'm not good in English, so please do not complain about my English. If you
understand this document, that's OK.
================================================================================
* Introduction

  Have you heard of "Vocal Eliminators"? Suppose one day you hear a song on the
street, and you like it very much. You want to song it for your family, but when
you search on the Internet, you cannot find the background music of this song.
What should you do? A vocal eliminator can help you.
================================================================================
* The Best! Vocal Elimination Software

  I have heard of many vocal eliminators and I tried some. in my opinion, the
best vocal eliminator is "Center Channel Extracter" Filter in Adobe Audition. It
uses phase analysis to catch the center channel(in fact you can remove a channel
at any position), and remove it. If the vocal track is panned in the center, the
result is almost the same as other vocal eliminators. But if not, this filter
works much better any other vocal eliminators. As a commercial audio workstation
software, Audition is very expensive, we all cannot afford it.

  There seems to be a famous vocal elimination software called "Vogone". I tried
the demo version of Vogone, and I find it very ordinary. Vogone uses phase
substraction (L-R), an old method which can only produce a mono result. New
version of Vogone can process mp3 and wma files, but the price is also very
expensive. I won't by it if Vogone's algorithm is not to be improved.

  I also tried a VST Plugin called "ExtraBoy Pro". It's not free, but a demo is
available to download. It's so "advanced" that until now I still cannot use it.
ExtraBoy Pro seems to be based on phase analysis, and you adjust parameters by
choosing a rectangle with frequency and phase. It's particular and difficult to
use, so if you have time and money, you can use ExtraBoy Pro. ExtraBoy is in the
same family with ExtraBoy Pro, free to use, but costs you a lot to download.
It's easier to use, but the result is not excellent, I think.

  If you use Winamp, dsp_centercut is a good plugin. Although it can only reduce
the center channel, dsp_centercut provides us with a good quality. And what's 
the most important, is the open-source proprety of dsp_centercut. I like
open-source software more, so I recommend it to you. But if you use professional
audio editors, you cannot use a Winamp plugin.
================================================================================
* How does it work?

  The vocal eliminator analyses the phase of a stereo song and reduces the vocal
track form the song. So the input song must be stereo. A low-pass filter and a
high-pass filter keep some part from vocal track in which voices are mixed with
instruments like bass, drums, etc.
================================================================================
* Parameters

  Pan: Pan for vocal track. If the vocal track is panned in the center, set the
       value to 0(0 degree). If the vocal track is panned in the left, set the
       value to -100(-90 degrees). If the vocal track is panned in the left, set
       the value to 100(90 degrees).

  Low: Cutoff frequency for low-pass filter. If the background music is piano
       solo or guitar solo, set the value to 0 or 50 Hz. If the background music
       contains bass or kick set the value to 200 Hz. Normally with a song by a
       male singer the recommended value is 150 ~ 250 Hz. With a song by a
       female singer the recommended value is 200 ~ 400 Hz.

  High: Cutoff frequency for high-pass filter. If the background music is piano
       solo or guitar solo, set the value to more than 16000 Hz. Normally with a
       song by a male singer the recommended value is 6000 ~ 8000 Hz. With a
       song by a female singer the recommended value is 8000 ~ 12000 Hz.

  Exception: Songs of Vitas.
================================================================================
* Troubleshooting

  * Vocal Eliminators can only process stereo songs. (Important!)

  * If you want to process songs in MP3 format, sometimes you will get a bad
    result. Please uses CDs as source to devocalize.

  * Vocal Elimination does not always work: sometimes it can not reduce the
    vocal track, sometimes it will destroy the background music.

  * Do not rely on vocal eliminators. A original background music is always
    better than the devocalized one.
================================================================================
* Copyright

  This document is copyleft(you can copy, modify, etc), but you must write
"Based on the article of Zhang Zhichao". Please respect my works. Thank you.
================================================================================
* References

  <<The Truth About Vocal Eliminators>>, Ethan Winer
================================================================================
END OF FILE
