Side as Stereo

Official support for: meldaproduction.com
Post Reply New Topic
RELATED
PRODUCTS

Post

I'm back here again, tried to ask about a year ago but didn't get anywhere.
So in a Melda plugin there is a button labelled "Side as stereo" with the following description: "Side as stereo switch makes the Side button audition the side (aka stereo) signal in a stereo way as it originally is."

How can I do this in MXXX?

It really annoys me that every time I think I fully understand Mid/Side there is another thing that makes my brain hurt! ha ha.

So I will try to explain what I don't understand, hopefully someone could explain it to me? Here goes:

As I understand it Stereo is 2 channels one for Right and one for Left. Some of the content of these 2 channels are the same (kick drum for example) meaning that these frequencies leave the 2 speakers at exactly the same time and are perceived to be in the center. Some of the content of the 2 channels is different (ping pong delay for example) where the sound from one speaker happens before the other or sometimes not at all.
So it is these differences between the two channels that creates width.
So we can change L and R channels into Mid/Side.

To get Mid we add together the Left and Right. This leaves behind everything that was the same in both channels.

To get Side we minus Left and Right. This leaves behind the differences of both channels.


So here is the part that confuses me:
Why is the Side signal a single channel?
Surely if it comprises of the differences between left and right we should have 3 channels now:
Left Side, Mid and Right Side.
Melda Production & United Plugins
Surface Studio = i7, 32gb, SSD.
Windows 11. Bitwig, Reaper, Live. MTotal.
Audiofuse, Adam Audio monitors + sub, iLoud MTM.
Polybrute, Summit, Pro 3, Tempest, Syntakt, AH2.
Ableton Push 2, Roli Seaboard Block.

Post

jmg8 wrote:So here is the part that confuses me:
Why is the Side signal a single channel?
Surely if it comprises of the differences between left and right we should have 3 channels now:
Left Side, Mid and Right Side.
You have answered your own question. Mid is the sum of L and R. That's all there is. I guess that's why IIRC the C in a panner is also called the phantom center.

Of course you could use 3 channels like the front speakers in a surround mix, but that's not how it is done. Tradition? Actually I wonder if it made a difference if a complete mix was done with 3 channels.

Masi

Post

OK, we have L and R at the start. These may be totally independent audio signals from the standpoint of this discussion.

Now, we may arbitrarily juggle that data and retain two independent encodings:

M = (L + R) - 3dB
S = (L - R) - 3dB

From these values we can precisely reconstruct the original:

L = (M + S) - 3dB
R = (M - S) - 3dB

If working with pencil and paper, rather than subtracting 3dB, divide by the square root of 2. Doing it that way makes it trivial to prove that we can convert from one form to the other and back again with exactness.

Now, forget for a moment that M and S are mid and side. They are just another encoding of some data. And it's clear from the simple math that either version of the data is complete and can be used to un ambiguously go to the other and back again. So, two values are all you need, period.

Now, one of the values is clearly an average (times a constant), and that makes perfect sense as the expression of the middle of the L and R signals. The other, the difference, isn't nearly as easy to wrap one's head around. But the math makes it clear that it gets the job done fully and completely. Long story short, just accept it and things work out.

All that said, I still have no clue as to what this means: "Side as stereo switch makes the Side button audition the side (aka stereo) signal in a stereo way as it originally is." Huh?????

Post

Lets say we have a simple song with a kick in the center, a hihat panned hard left and a shaker panned hard right.

The mid channel would only contain the kick right?

The side would contain both the hihat and shaker?

So if the side signal is just a single channel, then it has mixed the hihat and shaker into one channel.

How then, when we convert back to LR does it know to put the hihat to the left and shaker to the right?
Melda Production & United Plugins
Surface Studio = i7, 32gb, SSD.
Windows 11. Bitwig, Reaper, Live. MTotal.
Audiofuse, Adam Audio monitors + sub, iLoud MTM.
Polybrute, Summit, Pro 3, Tempest, Syntakt, AH2.
Ableton Push 2, Roli Seaboard Block.

Post

Sorry for the stupid question... but:
So in a Melda plugin there is a button labelled "Side as stereo"
Where is this button? I dont see it in the main channel selector modes...

Post

vectorwarrior wrote:Sorry for the stupid question... but:
So in a Melda plugin there is a button labelled "Side as stereo"
Where is this button? I dont see it in the main channel selector modes...
I was hoping not to answer that question as it is a plugin that has been in beta for a while, some of us already know it but not sure if im supposed to talk about it?

Just look up at the top of the plugin.
Melda Production & United Plugins
Surface Studio = i7, 32gb, SSD.
Windows 11. Bitwig, Reaper, Live. MTotal.
Audiofuse, Adam Audio monitors + sub, iLoud MTM.
Polybrute, Summit, Pro 3, Tempest, Syntakt, AH2.
Ableton Push 2, Roli Seaboard Block.

Post

jmg8 wrote:Lets say we have a simple song with a kick in the center, a hihat panned hard left and a shaker panned hard right.

The mid channel would only contain the kick right?
Ah, there's the source of your confusion. The mid signal would contain all three components. The hihat in the mid would be the exact same thing as the left signal. The shaker in the mid would be the exact same thing as the right signal, but the shaker in the side signal would be phase inverted. This side signal would have the sum of the hihat and inverted shaker, but no kick.

Run this through the simple equations and you'll see how things sort out.

Disclaimer: I may have not quite gotten the description of what's phase-inverted quite right, but this is the basic explanation that's not too far off base.

Post

dmbaer wrote:
jmg8 wrote:Lets say we have a simple song with a kick in the center, a hihat panned hard left and a shaker panned hard right.

The mid channel would only contain the kick right?
Ah, there's the source of your confusion. The mid signal would contain all three components. The hihat in the mid would be the exact same thing as the left signal. The shaker in the mid would be the exact same thing as the right signal, but the shaker in the side signal would be phase inverted. This side signal would have the sum of the hihat and inverted shaker, but no kick.

Run this through the simple equations and you'll see how things sort out.

Disclaimer: I may have not quite gotten the description of what's phase-inverted quite right, but this is the basic explanation that's not too far off base.
OK, so I will give this some thought and come back to you when my brain accepts it. Cheers.
Melda Production & United Plugins
Surface Studio = i7, 32gb, SSD.
Windows 11. Bitwig, Reaper, Live. MTotal.
Audiofuse, Adam Audio monitors + sub, iLoud MTM.
Polybrute, Summit, Pro 3, Tempest, Syntakt, AH2.
Ableton Push 2, Roli Seaboard Block.

Post

jmg8 wrote:OK, so I will give this some thought and come back to you when my brain accepts it. Cheers.
I couldn't sleep last night and while tossing and turning, thought of this example as explanation. But first, I'm going to switch conversion formulas and use something other than the one involving the square root of 2. I like that version because you can you the same formula to convert either way.

More commonly, we have these:

M = (L + R) * 0.5
S = (L - R) * 0.5

L = M + S
R = M - S

The advantage here if doing this repeatedly in a plug-in is that we have half the multiplications required.

Now back to the example. We have HH hard left, K in the exact middle, and S hard right. K in the exact middle means that there will be exactly the same K content in both the L and R channels. Let's call that PK for partial kick.

So, starting off, we have this:

L = HH + PK
R = S + PK

Run this through the conversion:

M = (HH + PK + PK + S) * 0.5 = (HH + S + 2 * PK) * 0.5
S = (HH + PK - S = PK) * 0.5 = (HH - S) * 0.5

Given that the M signal is a stereo mixdown to mono, the M equation is just what you'd expect it to be. The S signal doesn't track to anything as concrete as the notion of a mono-mixdown of a stereo signal. Anyway, converting back we have:

L = M + S = ((HH + S + 2 * PK) + (HH - S)) * 0.5 = ((2 * PK) + (2 * HH)) * 0.5
= HH + PK

I leave the other formula as an exercise for the reader. :D

Post Reply

Return to “MeldaProduction”