## Topic: Pro Logic II Algorithms (Read 4429 times)previous topic - next topic

0 Members and 1 Guest are viewing this topic.
Pro Logic II Algorithms
##### 06 April, 2010, 12:09:38 AM
I'm looking for specifics on:

1. How to matrix from 7.1 discrete (Pro Logic IIx).
2. How to dematrix to 4.0 discrete (Pro Logic).
3. How to dematrix to 5.1 discrete (Pro Logic II).
4. How to dematrix to 7.1 discrete (Pro Logic IIx).

Pro Logic has only one dematrixing mode, but Pro Logic II and IIx seems to have three (movie, music, game).

• SebastianG
• Developer
Pro Logic II Algorithms
##### Reply #1 – 06 April, 2010, 03:09:28 PM
I only once found a document that explained how 4 channels (left, center, right, surround) are supposed to be mixed into 2 channels so that pro logic decoders can "recover most of it". But no words on how to "decode". I think this is more of a black art.

Here's how the 4-to-2 mixing should be done:
Let L, C, R, S be the 4 input channels (left, center, right, surround)
Let hilbert(.) be a Hilbert transformer (shifts the phase by 90°, I don't remember in which direction)
Out1 = 0.7 * (C + hilbert(S)) + L
Out2 = 0.7 * (C - hilbert(S)) + R

So, you have C and S attenuated by 3 dB and then mixed into both channels where S is additionally phase shifted by +90° and -90° depending on the channel.

I guess you have to make certain additional assumptions in order to be able to recover it. The decoding has to be some nonlinear operation exploiting these additional assumptions, because otherwise you'll have ambiguities. The only things you can observe in the end is the level/intensity of each of the channels Out1 and Out2 (in some local time/frequency region) and the phase differencee between those (in the same local time/frequency region). These are only 3 degrees of freedom for each time/frequency region which somehow might lead to a reconstructed 5.1 sound again ...

Cheers,
SG

• DVDdoug
Pro Logic II Algorithms
##### Reply #2 – 06 April, 2010, 04:13:17 PM
Yeah, the encoding information that Sebastian mentioned is on Wikipedia

There is some good decoding information on the Dolby website[/color], but if the nitty-gritty details (or algorithms) are there, I haven't found them.  (For example, when I searched previously I couldn't find anything about the channel-steering response time.)  I have a feeling these are trade secrets, and they might be revealed if you apply for a Dolby license.

Pro Logic II Algorithms
##### Reply #3 – 07 April, 2010, 01:25:28 AM

Anyway, I got my hands on "Principles Of" worksheets straight from Dolby.  The original Pro Logic specification calls for applying Dolby Type-B noise reduction to the surround channel.  Is this really necessary with PCM?

• knutinh
Pro Logic II Algorithms
##### Reply #4 – 07 April, 2010, 04:25:28 AM
Quote
Because of the limited nature of the original DPL, many consumer electronics manufacturers introduced their own processing circuitry, such as the "Jazz", "Hall", and "Stadium" modes found on most common home audio receivers. DPL II forgoes this type of processing and replaces it with simple servo (negative feedback) circuits used to derive five channels. In addition to five full range playback channels, Pro Logic II introduced a Music mode which includes optimized channel delays, and adds user controls to - for example - adjust apparent front sound stage width.

I dont understand this. Does anyone else?

-k

• krabapple
Pro Logic II Algorithms
##### Reply #5 – 07 April, 2010, 11:50:50 PM

Anyway, I got my hands on "Principles Of" worksheets straight from Dolby.  The original Pro Logic specification calls for applying Dolby Type-B noise reduction to the surround channel.  Is this really necessary with PCM?

Some years back, maybe in the early 2000s, I was able to glean from Dolby via emails that there was a low-pass shelf filter of high frequencies on the surround channels in DPLII Music mode...except when Panorama was engaged.  It was also turned off for Game mode and (I *think*) Movie mode.    I also seem to recall that it was sometjhing Dolby was thinking of eliminating or making optional for later iterations of DPLII.