Question about M/S mic technique decoding regarding digital or analog domains.

Brento

Brento

New member
This question springs from a digression off topic on another thread, so I thought I'd start up new...

I have heard from a few people that the M/S stereo recording technique does not encode/decode "properly" in the digital domain. On the face of it, this could make sense as the technique was developed prior to the proliferation of digital recoding technology. Most people are happy to record an M/S pair, then duplicate the side mic in post, flipping one channel of this digitally. However, I know people who say you have to do it with three channels of an analog desk, or similar.
There seems to be a fair amount of misinformation about this and it is difficult to know who to believe. Can anyone answer this question with some level of authority borne from experience?
Cheers,
Brento
 
miroslav said:
OK....then ask those people what they are talking about exactly, and then share it with the rest of us. :)

Who are they.....and what did that specifically say about it?
Click to expand...

(This is more or less the same reply I gave on the "Analog only" forum, Miroslav... BTW, I am just asking a harmless question about whether anyone has heard of this. I'm not sure it should matter from who or where I heard it. You're right though, I should just look up the bloke and ask him what he was on about, which I will endeavour to do. I realise I am a bit short on details.)

Hi,

Just to clarify, I have used M/S many times in the digital domain (ie. duplicating the side mic track and flipping it with a real time plug-in, or via off-line process). I achieved results that varied from excellent to messy, but I think this was due to application rather than technique.

I asked the question because the subject came up in another thread, off topic. My confusion stems from an off-hand comment by a bloke who ran a live sound company I worked for over a 15 year period from around the mid-nineties. We were talking M/S as we were busy preparing a system for a gig the next day, and when the subject of flipping the phase of the side channel came up he made an off-hand comment that went something like this: "you know you can't do it digitally, right?" - I said "really, what do you mean?". Then he must have gone off to do something else like answer his phone and the question was never resolved. Never came up again. We were pretty busy. Thing is, this guy (I am not going to name him without his permission) has about 45 years experience in audio (he's semi-retired now). I am no longer in regular contact with him, but I think I'm going to look him up after the silly season and try and find out just what he was talking about.

Sorry to be so vague about this. I was basically just asking if anyone had ever heard anything like this about decoding M/S before.

Cheers,
Brento

Cheers,
Brent
 
Well...you need some logic/math to do it in either analog or digital...though I think I know what you are saying....


I bet this guy has probably been simply avoiding the digital jump...which is nothing unusual for guys who find an SOP that works for them, and after x-years, simply don't bother to try anything new. His view is probably just one of not-knowing rather than simply bashing digital.
I've been doing M/S for awhile...and the first time I thought about how to do it in the DAW....I had to pause for a bit and think about it because it's wasn't identical to the analog steps, but then in both analog and digital, there was a logical way to do it.

Anyway...I thought maybe this guy had some deeper insight, and it was going to get into some sort of artifacts/conversion thing....but it looks more to me like he just never tried to do it in a DAW. :D
 
Last edited:
Nah, Tim (there you go, that's his first name) was always keen to embrace new technology, more so than most, and this included digital. He was one of the first guys I worked for that bought digital desks, although I would say that he likes his old His Masters Voice gramaphone a little too much...
As I said, I'll go and visit him after my January hols, drink a few beers and ask him what it was he was getting at. I can only speculate that it has a) something to do with artifacts or aliasing in the very high frequencies (a bi-product of the conversion process as you mentioned in your last post, Miroslav) or b) digital summing. I'll get back to you...

Happy New Year - as usual, I'm working a gig :(
Brento
 
Brento said:
I can only speculate that it has a) something to do with artifacts or aliasing in the very high frequencies (a bi-product of the conversion process as you mentioned in your last post, Miroslav) or b) digital summing. I'll get back to you...
Click to expand...

I'm curious to hear what he has to say...and looking forward to it, as we haven't had a good analog VS digital debate around here in like at least a month already.

We are due for one. ;)

:D
 
Well, I admit that even I love arguing about how wonderful good analog gear is -- But in this case, I can't even imagine the argument. M/S in digital would be *perfect* -- Whereas even analog might be wonderful and all, but still subject to the differences in the circuits from one channel to the other.

Okay, I imagine someone could say "that's where the magic is" -- But to say that it won't work in digital... :facepalm:
 
Massive Master said:
Okay, I imagine someone could say "that's where the magic is" -- But to say that it won't work in digital... :facepalm:
Click to expand...

It's all about the "magic"...;)


OK, so I had a REAL late night, and I'm just now having my first cup-o-Jo of the day, so excuse me if my thinking is a little off....
....but I think there is one aspect of M/S where it differs in the analog domain VS the digital domain...and of course, there are many studio rigs, so it may only apply in some instances, and possibly is what the OP's friend alluded to.

If you set up an M/S matrix with a console and you're tracking through it, you immediately can hear the L/R stereo mix of the M/S mic combination.
Doing it in a DAW...you have to first record the two mics and THEN create the M/S matrix, so you only hear it after the fact....right? Could it be that's what he meant by "you can't do this digitally"....or am I missing something here in my own thinking...?

When I record M/S mics via my analog rig...I'm recording the L/R mix of the M/S matrix, and that's what I am hearing in real time for my cue playback...NOT the individual signals of the Mid and Side mics.
With a DAW, I think you can ONLY record the individual M/S mics and THEN create the L/R mix...right?

Other than that...M/S works perfectly in the digital domain.
 
Pure speculation on my part, but since Tim was an early adopter of digital, he might have (had) a point. Since all digital signal processing introduces a certain amount of latency (& it was much worse in the earlier implementations), it's quite possible he attempted M/S in the days before automatic delay compensation, and the inverted S signal would have been slightly delayed relative to the "straight through" S signal. That would definitely mess with the M/S matrix.
 
bouldersoundguy said:
Copy/invert/pan doesn't cause latency in any software I've used in the last decade.
Click to expand...


Yeah...but even so, how can you copy/convert in *real-time AS you're tracking* the M/S signals in order to get a L/R cue mix...?

I think that might be the only difference with analog M/S....it's working in real time (assuming there's a matrix set up to encode during tracking rather than done after the fact.)
 
miroslav said:
Yeah...but even so, how can you copy/convert in *real-time AS you're tracking* the M/S signals in order to get a L/R cue mix...?
Click to expand...

I was countering the "all digital processing introduces a certain amount of latency" statement, not referring to real-time monitoring of M-S.

But why couldn't a M-S decoding plugin work on a live digital signal? Even if it causes latency it would still keep mid and side properly aligned. If not it would be just as bad with a recorded signal as a live one. It's only the manual copy/invert/pan method that can't be done in real-time.
 
bouldersoundguy said:
I was countering the "all digital processing introduces a certain amount of latency" statement, not referring to real-time monitoring of M-S.
Click to expand...

Yeah, I know....I was just following up with my own train of thought and questions.

bouldersoundguy said:
But why couldn't a M-S decoding plugin work on a live digital signal? .
Click to expand...

OK...now that might be a different animal....an M/S plug-in...though I don't know, as I've never tried them.
I was mainly going with the record, then copy, and manually set up an M/S stereo pair in a DAW, which is how most folks do it if they have no dedicated plug-in.

Again....this is all about trying to guesstimate what the OP's friend might have been alluding to (and not any analog VS digital thing for me)...and maybe the OP will come back with a definitive answer.
 
Actually, all digital processing does incur latency - it's just that in newer implementations it is minimal and the software compensates, whereas early digital (mid 80s - early 90s) was more problematic. I know, because I tried to be an early adopter, and was frustrated by the limitations at the time. I realize that things have changed, but understand how an "old school" engineer might be a bit behind in his thinking.
 
Latency doesn't really matter for non-real-time processes that have no time component, like copy, invert and pan. I don't think I've encountered software that did not handle these in the obvious logical way.
 
You must log in or register to reply here.
Back
Top