Any proper way to mix quad down to stereo?

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I'm not sure why you want to do this. As Fizzy said most mixes come with a stereo version. You can use an audio editing program like Audition or Audacity to do a mixdown or easier still just use the "Matrix Mixer" in Foobar2000.

You have many mixdown options like a direct fold down (combine Lf with Lb and Rf with Rb). Or you could pan the rears toward the centre. You could encode the four channels to RM/QS or anything else.

You could also get the Involve encoder to encode (QS compatible), it's said to make encodes that still sound like stereo.
 
Encode with QS. It can be used anywhere.

The Scheiber matrix system is very similar to Sansui QS but without the 90 degree phase shifts, it might be easier to create a Scheiber encode in software since it only uses phase reversals (probably can be somewhat decoded with a QS decoder and a Dolby Pro-Logic 2 music decoder, among others).


Kirk Bayne
 
The Scheiber matrix system is very similar to Sansui QS but without the 90 degree phase shifts, it might be easier to create a Scheiber encode in software since it only uses phase reversals (probably can be somewhat decoded with a QS decoder and a Dolby Pro-Logic 2 music decoder, among others).


Kirk Bayne
We usually refer to that as RM, although technically RM is a term that includes QS and other similar systems. The advantage of QS is that everything is mixed together in phase. Pans to intermediate positions will not disappear as they can with other matrix systems.

You just need to mix the two stereo pairs by 7.7dB. Then shift the phase +90° on the left and -90° on the right of the rear pair only. Then mix the front and back pairs together. Not hard to do with an audio editing program like Audition.

For RM (Scheiber) you would just do the rear mixing "out of phase", as you say. Other variations are EV-4 or Dyna which better preserve left to right separation of the mixdown, but still suffer from encode cancellation issues.
 
The hardest thing to do is matrix an existing discrete 4-channel tape, because the phases in different channels were not recorded with matrix in mind.

Often the trick is to reverse the phase of one back channel so that the front and back stereo pairs are in phase. One side pair is backward in phase, but it is not heard without turning the head either way.
 
Anyone have (and could post) the sequence of steps needed for Audacity to make a quad to (encoded) stereo conversion?


Kirk Bayne
 
Most of the quad albums I know of have a dedicated stereo mix as well.
I'm not sure why you want to do this. As Fizzy said most mixes come with a stereo version.
There are those rare cases where a quad mix differs significantly in actual content, from simple things like level differences to the use of a completely different take or recording, such that there is no existing stereo equivalent. In those cases, I think a non-matrixed stereo downmix could be useful. Some have even been released officially from the '90s onward.

To somewhat maintain the relative levels and panning of the four channels, what I would do is this:
  1. Split the four-channel file into four individual mono tracks, routed to a stereo master track
  2. Pan tracks as follows:
    • Ls to 100% left
    • Lf to 50% left
    • Rf to 50% right
    • Rs to 100% right
  3. If the stereo master track clips with the gain at unity, reduce the gain of all mono tracks equally until clipping is eliminated
    • Leave stereo master track at unity gain
    • Leave 1-2dB of headroom on the master track (i.e. do not exceed -1.0dBFS)
    • Best practice: if possible, use true peak level metering and do not exceed -1.0dBTP on the master track
  4. Bounce to new stereo file (or print to new stereo track)
The resulting stereo downmix collapses the front stereo image to 50% of its original width, while placing the respective rear channels at the far edges of the new stereo image. If that makes the former front image too narrow, experiment with wider Lf/Rf pan values.
 
There are those rare cases where a quad mix differs significantly in actual content, from simple things like level differences to the use of a completely different take or recording, such that there is no existing stereo equivalent. In those cases, I think a non-matrixed stereo downmix could be useful. Some have even been released officially from the '90s onward.

To somewhat maintain the relative levels and panning of the four channels, what I would do is this:
  1. Split the four-channel file into four individual mono tracks, routed to a stereo master track
  2. Pan tracks as follows:
    • Ls to 100% left
    • Lf to 50% left
    • Rf to 50% right
    • Rs to 100% right
  3. If the stereo master track clips with the gain at unity, reduce the gain of all mono tracks equally until clipping is eliminated
    • Leave stereo master track at unity gain
    • Leave 1-2dB of headroom on the master track (i.e. do not exceed -1.0dBFS)
    • Best practice: if possible, use true peak level metering and do not exceed -1.0dBTP on the master track
  4. Bounce to new stereo file (or print to new stereo track)
The resulting stereo downmix collapses the front stereo image to 50% of its original width, while placing the respective rear channels at the far edges of the new stereo image. If that makes the former front image too narrow, experiment with wider Lf/Rf pan values.
Why would this not be ideal in 100% in original perspective Lf and Rf thus keeping the solid and exciting separation?
 
Because then there is no separation between front and rear.
But if the front and back are collapsed to regular stereo, there is not going to be any real separation left between front and rear anyway, other than some imaginary spatial imaging presence left over from the recording & overdubbing sessions.

Taking away some of the separation between left and right, does not add separation between front to back (once it's folded to stereo) or does it really?

I can see where this might still be a nice mix and a good listen. But I don't see where this more narrowing of the fronts creates more depth. But I might try it.
 
But if the front and back are collapsed to regular stereo, there is not going to be any real separation left between front and rear anyway, other than some imaginary spatial imaging presence left over from the recording & overdubbing sessions.

Taking away some of the separation between left and right, does not add separation between front to back (once it's folded to stereo) or does it really?

I can see where this might still be a nice mix and a good listen. But I don't see where this more narrowing of the fronts creates more depth. But I might try it.
It would make more sense to me to narrow the back channels than the front.
 
Could use the old Electro-Voice Stereo-4 matrix and encode the surround sound into stereo that way (AFAIK, could be done with a DAW).


Kirk Bayne
 
Back
Top