Mathematical transformation of stereo waveform to filter mono sounds?

Question

To filter out a mono component, It gives good results to subtract one one wave from the other, the resulting waveform will have any underlying sound that are identical in both channels muted.

Can you think of a way to filter out these mono sounds? (keeping only the sounds that repeat in both channels?)

to sumarise:
What I want is to keep the wave componts that are equal in both channels, and deleting the ones that differ.
This would be done by performing a mathematcial function transformation.

for example, the function that removes the identical sounds is:

NewWave(x)=Wave1(x)-Wave2(x)

where Wave#(x) is the value for the recorded waveform at the time x.

Answer 1

What your asking sounds similar to a technique used to cancel noise picked up along the length of a low impedance microphone cable. The cable has three wires: sig+, sig- and GND. The signal transduced by the microphone is sent down the sig+ wire as is but inverted before it is sent down sig-. Any noise picked up along the length of the cable will be in phase. i.e. will be identical on both sig- and sig+.

At the receiving amplifier sig- is inverted and summed with sig+. This puts the signal generated by the microphone in phase and puts the noise picked up along the cable 180 degrees out of phase. The noise signals cancel when sig+ is summed with the inverted sig-.

This wouldn't work for noise picked up on stereo wires because the R and L stereo signals aren't the same and you would get alot of undesired cancellation along with cancelling the noise.

Answer 2

Mono sound is the comparable sound out of each speaker (or in basic terms one) Stereo has distinctive sounds from the two audio device - in basic terms slightly distinctive, yet sufficient to offer a "stereo photograph" the place sound can seem to be on the left, the superb option or midsection.

Answer 3

If you do this in the digital domain, you can take the signal that represents the the difference in channels and subtract it from the original signal of either the left and right channels, then combine these.