Can you effectively record three channels with two if you use a M/S decoder?

[mjfe2]

Am wondering about making a live recording of a funk band by sending the mixing desk outs to a four-channel interface. Would it be possible to combine a stereo submix of the horns and a mono guitar signal and then use a M/S decoder to process the signals separately when mixing afterwards?

[The Elf]

Short answer - no.

[mjfe2]

Short answer - no.

There are four horns. What if I panned two of them hard left and two hard right -- surely then I could end up with the guitar as the mid signal and the horns as a sides signal? Then I was thinking I could narrow the width of the horns track to mitigate against the extreme panning. Would this really not work? I realise it's a bit ad hoc!

[The Elf]

No matter how much you squirm you only have two channels - no matter what you do to them. That means two mono signals - not one mono and one stereo.

[Tim Gillett]

As I understand it M/S only works when the M/S signals have a specific relationship to one another such as with the cardioid mic for Mid and the figure 8 for the Sides, both mics being physically close to one another and picking up the same sound stage.

But your question is an interesting one.
I've resorted to electrical subtraction with some live 4 track recordings of a band I played in years ago. At the gig there obviously werent enough tracks for me to record all instruments or vocals separately so I had recorded it in this way: 1: FOH mix. 2: guitar, 3: bass, 4: accordian.

At some points the accordian was too loud in the FOH mix. By inverting phase on that track and mixing in back into the FOH mix I was able to lower the accordian's volume. The same could be done for the other two instruments recorded separately. Their level in the mix could not only be increased but now decreased, re the live FOH mix.

It would be possible to create a new FOH track where the guitar, bass and accordian have all been stripped out. Then you can add each of them afresh, panning if desired.

Of course you can only do this if the signals to be cancelled are identical or very similar.

Ah the joys of algebra and simultaneous equations!

Tim

[John Willett]

M/S = Mid / Side *NOT* Mono / Stereo

[mjfe2]

M/S = Mid / Side *NOT* Mono / Stereo

I know that much, thank you! I just wondered if it was possible to remove mono sound from a stereo signal (like those plugins do for amateur karaoke backing tracks).

[Hugh Robjohns]

Would it be possible to combine a stereo submix of the horns and a mono guitar signal and then use a M/S decoder to process the signals separately when mixing afterwards?

As others have said: no! You can record a stereo submix of the horns with the guitar panned centre if you like, but the separation between them will be limited. You could use MS processing to adjust the relative balance, EQ, or dynamic compression of the Mid (guitar) and Side (horns) signals with some independence, but that's about as far as it goes.

Hugh

[Hugh Robjohns]

Short answer - no.

There are four horns. What if I panned two of them hard left and two hard right -- surely then I could end up with the guitar as the mid signal and the horns as a sides signal? Then I was thinking I could narrow the width of the horns track to mitigate against the extreme panning. Would this really not work?

It depends what you're trying to achieve. The MS signal is created and extracted by a sum/difference matrix.

Mid = left + right Side = left - right

And for the decode:

Left = Mid + Side right = Mid - Side

So the mid track will contain not only the guitar, but also all four horns. If you don't want horns on the Mid channel, then you have to mix the horns into a mono signal, and then split that in two, one part panned hard left and the other hard right, but with a polarity inversion.

Only in that way will you have the exclusive separation you require with guitar only in the Mid, and horns only in the Sides.

If you place some horns left and some right, as you propose, there is no exclusivity, and no separation, because they are different signals in different places.

As the Elf says, MS is a two-channel medium, not three-channel.

Of you want separation between the guitar and horns, your only options are conventional stereo with horns panned hard left and guitar panned hard right, or MS with guitar centre and all horns summed and split as described above. Two channels.

H

[James Perrett]

A friend of mine used to mess around with putting signals into the rear channels of a Hafler style surround system with surprising success. This involved inserting a difference signal between the 2 stereo channels. In theory this would also appear in the front channels but the level in the front channels was much lower than I would have expected. So, while you won't get total isolation, it might be worth a go.

One other thing to look at would be the Center Channel Extractor in Adobe Audition - this can isolate signals in a given position in the stereo image although the sound quality isn't always great so I treat it as a 'use when absolutely necessary' tool.

James.

[Hugh Robjohns]

I just wondered if it was possible to remove mono sound from a stereo signal (like those plugins do for amateur karaoke backing tracks).

Yes, most easily through phase cancellation... but then your stereo horns become mono too (and possibly with some phase issues in the process). There are other more sophisticated means -- the Adobe Audition Centre Channel extractor that James mentions, or Roland's R-Mix -- but at the end of the day, you're trying to get more out of the system than it is capable of.

H

[mjfe2]

Okay, thanks Hugh and everyone else. Looks like I'll have to give up this pipe dream! Will be recording horns to one channel, guitar to the second, keys to the third and drums/bass to the fourth -- just like in the '60s!