THE FACTS - WHAT SONGS CAN BE VOCAL ELIMINATED

[admin]

[li]MUST NOT BE A MP3 FILE CONVERTED TO WAVE FILE - A file that has been compressed with MP3 will not vocal eliminate correctly because basic data in the stereo channels has been changed. After doing an MP3 to WAV conversion, the resulting WAV file is NOT identical to the master file that was compressed with MP3. DO NOT TRY TO VOCAL ELIMINATE MP3 COMPRESSED FILES!

[li]THE SONG MUST HAVE BEEN MASTERED AS TRUE STEREO -The wave file must be true stereo with two separate tracks. The original master song must have been mixed to separate stereo tracks in the recording studio when it was created. If the song is from a monaural file that has been copied to both left and right tracks, this is not stereo, and there can be a total elimination of all sounds. "Digitally Remastered" songs can be this way, if the original was recorded and mastered in monaural. Hard to believe, but there once was a time where stereo did not exist!

[li]ONLY CENTER PANNED AUDIO WILL BE REMOVED - All audio that is equally loud in both the left and right channels of a stereo file will be removed. Where each sound track is panned in the stereo mix is decided in the recording studio on the mix console. The monaural microphone signal for the lead vocal is adjusted so its pan-positon control is in the center (eqaual in left and right). If the vocal was recorded in a stereo microphone (not done usually), it will not be center panned unless the vocalist was precisely at the middle of the stereo microphone, and did not move their head while recording (very unlikely). If the audio is slightly louder in one channel than the other (i.e. not exactly center panned), some of the audio will remain after vocal elimination processing.

[li]ECHO/REVERB WILL NOT BE REMOVED - The reverb that is often applied to the vocal track is usually set so the echo is NOT center panned. That way, the ear "hears" and the brain "perceives" the echo to be from a larger room. Because the echo is not equally loud in both channels, it might be reduced, but some portion will remain in the resulting vocal eliminated file.

[li]OUT OF PHASE CHANNELS WILL NOT WORK - If the original master stereo mix has any delay at all between the left and right channels, the center panned audio will not be removed. In fact, in this case you will hear some frequencies of the audio amplified and others reduced (attenuated). A wave file that contains out of phase audio will sound really funky... and will not be usable for Karaoke. If you hear some frequencies in the music and/or vocals amplified and others attenuated, this indicates the original recording studio engineer mixed the background vocal track slightly delayed on one channel. This is probably due to running it through a reverberation processor that adds delay and creates lower amplitude echo signals that are also delayed by different amounts. If the channel delay (really a phase shift in time) is "just wrong", processing with a vocal elimator can amplify or attenuate random frequencies determined by the delay. When the signals are inverted and added, the phase shifted signal will not cancel the original signal, and can adversely affect any frequencies.

[admin]

Vogone I uses the same basic technique as the $1,200 Thompson Vocal Eliminator, but does it digitally and can therefore deliver higher quality.

To keep the discussion as simple as possible, the original wave file audio samples are duplicated during the processing. The duplicate stream is inverted (positive voltages become negative, and negative become positive) and the channels swapped (left becomes right, right becomes left). The original and duplicate inverted/swapped streams are then added together. Where a sound was equally loud in both channels, there now are identical but inverted voltages being added, and they sum to zero. If they are not equally loud, they sum to something non-zero; i.e. they are not cancelled. This is the basic process.

You can understand that if the audio in one channel is shifted (advanced or delayed) any from the other channel, the voltages will be offset and thus not identical. Based on what frequencies are present, the sounds will add or subtract, in definable but apparent random locations. Thus, the resulting sound can have louder sounds and reduced sounds, and even what sounds like echo or reverb (called flanging) as some frequencies amplify and others attenuate.

The Vocal Elimination algorithm that Vogone I uses either removes the vocal in a song or it doesn't. There are no controls to adjust, because there is nothing to change. In a true stereo wave file, Vogone I will remove all center panned vocals and instruments completely and cleanly without distortion.