The No.1 Website for Pro Audio
All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
4 weeks ago
#1
Lives for gear

Correlation vs. Phase?

Although worked several years as a semi-pro (or advanced hobbyist) in producing, playing, and writing music, time to time I notice, that there are basics in audio technology and audio as physical phenomenon, which I haven´t internalized, which I have to re-study.

One of these areas is a very basic one, namely concepts of MID/SIDE vs. MONO/STEREO and IN PHASE/OUT OF PHASE vs. CORRELATION.

Lets put some some basic definitions here first:

Mono = Mid (M) channel only
Mid (M) = L + R
Side (S) = L - R
M + S = Left channel in stereo
M - S = Right channle in stereo

"Polarity" is the positive or negative orientation of a waveform, relative to another.

"Phase" is about the position of a waveform in time, relative to another.

"Correlation" is about polarity alignment between two channels.

Now the question: Why the S (side signal) isolated from the stereo signal is always 100 % out of phase (-1), measured by correlation meter?

There is obviously something I have not completely understood?

Can someone explain me in mathematical terms, how the waveform in time (phase) is related to polarity alignment between two channels in stereo.

"Phasing issues" and "out of phase" have negative connotation in producing - but the S-component (Side), on the other hand, is very essential part in producing music, and creating stereo picture (3D-picture).
4 weeks ago
#2
Lives for gear

Quote:
Originally Posted by Harry_O
Now the question: Why the S (side signal) isolated from the stereo signal is always 100 % out of phase (-1), measured by correlation meter?
Because of L - R. This cancels out everything L and R have in common (everything that's equal and in phase). The difference between both.

Worth noting this covers only the very short term (i.e. memoryless correlation). The true, perceived stereo correlation is more complicated and longer term (e.g. delayed reflections fuse with their main "cause" under certain circumstances).
4 weeks ago
#3
Lives for gear

Thank you.
What my common sense still quite doesn't get, is, that according to the above, the
stereo picture always means, that those elements, which create the stereo (Side S), are
by definition always 100 % out of phase.

This means, that Mid M, or mono, which by definition is always 100 % in phase, cuts out
something of the stereo picture. i.e., which we can't hear in mono.

On the other hand, Mid (M) = L + R.

There is still some missing part, so far (in my mind).
4 weeks ago
#4
Lives for gear

Verified Member
Quote:
Originally Posted by Harry_O
Thank you.
What my common sense still quite doesn't get, is, that according to the above, the
stereo picture always means, that those elements, which create the stereo (Side S), are
by definition always 100 % out of phase.

This means, that Mid M, or mono, which by definition is always 100 % in phase, cuts out
something of the stereo picture. i.e., which we can't hear in mono.

On the other hand, Mid (M) = L + R.

There is still some missing part, so far (in my mind).
Look to where you use the word "phase" when the word "polarity" may be more suited...

Whenever you combine 2 signals (L and R in this case) that contain polarity inverted signals, those signals will of course cancel out. I see no conflict between the statements in your post. Yes, some of the sound cancels out in L+R... because it is L+R.
4 weeks ago
#5
Lives for gear

Quote:
Originally Posted by Trakworx
Look to where you use the word "phase" when the word "polarity" may be more suited...

Whenever you combine 2 signals (L and R in this case) that contain polarity inverted signals, those signals will of course cancel out. I see no conflict between the statements in your post. Yes, some of the sound cancels out in L+R... because it is L+R.
Concerning "phase" and "polarity" : Side-component - 1 correlation, proved by using correlation meters, and the user guides talks about "phase", not "polarity", see e.g.
https://www.voxengo.com/product/correlometer/
4 weeks ago
#6
Lives for gear

Verified Member
Quote:
Originally Posted by Harry_O
Concerning "phase" and "polarity" : Side-component - 1 correlation, proved by using correlation meters, and the user guides talks about "phase", not "polarity", see e.g.
https://www.voxengo.com/product/correlometer/
Yes, that's a correlation meter. It can detect phase issues if they are present. It can also detect polarity issues if they are present. The two words often are conflated even by highly educated people. It's actually hard not to do it. For example, nearly everyone wrongly calls a preamp's polarity switch a phase switch. A vibrato bar is wrongly called a tremolo bar. Etc. It's just jargon. But none of that has anything to do with what's in the side signal.

See your own definition of "phase" in the OP. Do you think that being panned in the stereo field has an effect in the time domain? Is a panned signal delayed or advanced in time relative to a centered signal or a signal panned opposite? If no, then I think "phase" is the wrong word for it.

You also wrote - "Correlation" is about polarity alignment between two channels.

https://www.justmastering.com/articl...ined-part1.php

"Polarity allows you to encode stereo into Mid/Side or to decode Mid/Side back to Stereo, and the difference between the left and right channels determines what you hear in the mid (sum or L+R) and side (difference or L-R) channels."

So, as to M/S processing I think "polarity" is the more relevant word. But phase can also have a role in stereo reproduction when more than one mic is used on a single source, then panned in the mix. I think that's less common than just using the pan pot though, and has nothing to do with how M/S works.

Last edited by Trakworx; 4 weeks ago at 06:44 PM..
4 weeks ago
#7
Lives for gear

Thank you for commenting. For me this needs some more thinking.
3 weeks ago
#8
Lives for gear

Verified Member
Quote:
Originally Posted by Harry_O
Thank you.
What my common sense still quite doesn't get, is, that according to the above, the
stereo picture always means, that those elements, which create the stereo (Side S), are
by definition always 100 % out of phase.

This means, that Mid M, or mono, which by definition is always 100 % in phase, cuts out
something of the stereo picture. i.e., which we can't hear in mono.

On the other hand, Mid (M) = L + R.

There is still some missing part, so far (in my mind).
phase (deviation) usually gets measured in degrees, not by a percentage - and even with a signal showing a phase shift over the frequency soectrum, it may not be much of an issue as long as signals correlate, so when looking at phase, it doesn't have to be phase vs correlation...

maybe forget about m/s for a moment and have a look at a normal' stereo l/r signal which you feed into a goniometer (which i find to be quite revealing and mesmerizing): the relationship between the two signals does vary in phase (with music) but you'll hardly see them to be fully out of phase for a continued period of time.

now m/s is just another way of representing a stereo signal; phase relationship doesn't change upon matrixing.

Last edited by deedeeyeah; 3 weeks ago at 09:00 AM.. Reason: edited for clarification
3 weeks ago
#9
Lives for gear
in mid/side, the setup is actually kind of simple. Mid is left plus right. It is the sum of the left/right signals. If the left and right signals are exactly the same (mono in a stereo buss) the result will not add or subtract anything. In a “normal” stereo signal, anything not centered will be lower in level, extremely panned signals will be way down in level, and completely out of polarity signals (often misidentified as “out of phase” signals) will disappear from the sum signal.
The side signal is the difference signal. It is left minus right or right minus left. When we say “left minus right”, we actually flip the polarity of either the left or right channel and then add the two signals together. Mono/centered information cancels In the case above, where the left and right signals are identical (mono in a stereo buss), the side signal is nothing because the two identical mono signals cancel completely.
The extraction and manipulation of the side signal requires either a dedicated mid/side device (hardware or software) or some engineering trickery that most audio engineers don’t understand.
3 weeks ago
#10
Lives for gear

Quote:
Originally Posted by Bushman
in mid/side, the setup is actually kind of simple. Mid is left plus right. It is the sum of the left/right signals. If the left and right signals are exactly the same (mono in a stereo buss) the result will not add or subtract anything. In a “normal” stereo signal, anything not centered will be lower in level, extremely panned signals will be way down in level, and completely out of polarity signals (often misidentified as “out of phase” signals) will disappear from the sum signal.
The side signal is the difference signal. It is left minus right or right minus left. When we say “left minus right”, we actually flip the polarity of either the left or right channel and then add the two signals together. Mono/centered information cancels In the case above, where the left and right signals are identical (mono in a stereo buss), the side signal is nothing because the two identical mono signals cancel completely.
The extraction and manipulation of the side signal requires either a dedicated mid/side device (hardware or software) or some engineering trickery that most audio engineers don’t understand.
One reason, among others, why elaborating, and trying to internalize these things, is important in my opinion, is, that manipulating not only L/R, but M/S, and even "out of phase 3D", by adding delay, reverb and modulation, has become very common thing with the new, sophisticated plugins.

At the same time, we tend have less understanding, what these plugins really do to the signal.
The plugins are "too clever" for us.

Someone could argue to the above, that "what does it matter, if it sounds good, it is good".
On my behalf, I think, that we can get better results by understanding things, and the physical laws, under with the "sound" exists. . And If not for other reason, its intellectually rewarding.
3 weeks ago
#11
Lives for gear

Verified Member
Quote:
Originally Posted by Harry_O
Someone could argue to the above, that "what does it matter, if it sounds good, it is good".
Yes, but only if it translates well out of the studio. The answer tips towards "no" if we are making decisions based on a listening environment that doesn't translate well.
Topic:

## Welcome to the Gearslutz Pro Audio Community!

###### Registration benefits include:
• The ability to reply to and create new discussions
• Interact with VIP industry experts in our guest Q&As
• Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD \$20/year
• Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.