Dolby Pro Logic II encoding test (90° vs 180° phase shift)

[raquete]

@ 3dsnar

So it is not that I disagree / agree with you
of course,i (we all) know that. ;)
I just wanted to clear things out. no need because you are always very clear in your posts.
remember,you are one "creator",i'm one single "listener".



@ Rockaria

c. difussion effect : (0, -180), (90,-90) > (0,+-90)...
difussion effect takes to listening effects when the audio is up or downmixed.(this is why i posted about speakers positions and respectives audio phases too)

all i can say about your (complete) last post is:
perfect (clever explanations,word by word) :cool: :cool:

edit: typos

[Rockaria]

The DIFFUSION effect looks like another expression(transformed energy) of the cancellation.
Like throwing another stone in the pond, the area between them will have some counter-wave effect making the power weaker.
If you change the phase from 0 - +-90 - +-180 between the two speakers, you will FEEL the center sound becoming weaker(scattered).

Not the entire cancellation(like digitally when mixing) but some type of transformed(to different type of energy : speaker) interference..
So I agree by including the digital phase altering mixing process(& delay,,,), the speaker diffusion/phantom center effects would make a full consideration of this environment.:)

[raquete]

ok.:goodpost: (one more)

one single test for "cancellations" :

1-create(generate tones) 100Hz -30 seconds stereo wave-same phase in L/R.
listen in one stereo amplifier(or in pc with good sound)where the speakers are in front of you(of course,left and right speakers)
you will listen one very clean sound.

2- now place one speaker in front the other.(face to face)
L --> <-- R
see(or better,listen) how much cancellation happen.

3- most important: i will that you understand what i mean :p

[Rockaria]

I see the sound becoming almost muted with extremely close faced speakers & identical stereo playback with a ch invert phased.;)

The speaker's transformed energy consists of several factors :
- direction : it has 360 degree but more directed to the front of the speaker
- carrier : air(thiner density of the matriels than water or circuit in digital/analog formats), it will be delayed, weakened by the distance.
- set up : 5.1(3 fronts and 2 rears facing each other but with some enough distance), angle & distance(&delay),,, fators are crucial
- room(ambience) : interferences, reflections(matriels), decay, echos...

Thanks for providing the excellent extreme example.:goodpost:

[edit]
The same DIFFUSION or cancellation effect will also happen when changing the one speaker polarity(swapping the red & black wires).

[Rockaria]

. for the invert-only versions(and these s/w dpl II decoders), the matrix3 is the way to go!
http://forum.doom9.org/showpost.php?p=844727&postcount=134
Based on the Dolby's and surecode's docs, the invert only versions should use the tebasuna's matrix3(in avs or similar scripts) :
- the DPL II is expected to have the out of phase signal in Rs : the matrix1 will have (-Ls, +Rs) encoding time
- the decoder movie mode inverts the Rs(as explained in the surecode doc) : the matrix1 will have (-Ls, -Rs) both inverted
- so the matrix3 will have (+Ls, -Rs) encoding time and (+Ls, +Rs) decoding time.
- but these models do not fully seperate the channels(unbalanced), although will never be agreed or admitted..

Because most 90 phase deg shift routines perform -90 deg shifts effectively, invert mixing the Ls(coeffs) will have (+Ls, -Rs) encoding time and (+Ls, +Rs) decoding time.

I have tried to find a plugin or ways to implement/utilize the 90 deg phase shift in avisynth with no success yet. Developing the plugin also does not have the proper environment(esp. time) for me. But as I found the softEncode surround 90 deg phase shift option performs correctly for the DD downmix encoding/play, I conclude it's the ac3 encoder's role to do the optional 90deg shift + downmix + (DD2.0 / WAV encode/STDOUT).

http://forum.doom9.org/showpost.php?p=850563&postcount=15
http://forum.doom9.org/showpost.php?p=853856&postcount=50
And the complete background of the DPL II downmixing process in DD encoding(rear 90deg phase shifts option) and decoding(auto downmix mode selection based on annex->dmixmod meta info) : http://forum.doom9.org/showpost.php?p=856250&postcount=98

I think the softEncode's batch mode will still be useful for the DPL II encoding automation if we don't care about the temporary WAV AC3 files.

As for the Dynamic DPL II encoding for the ds filter use(FFDShow...), there are some conditions I can think of for now :
. the freq-delay method must be applied to the all-pass(granular frequency bands) and aligned
. the buffer must be large enough to sync(align) the delays between the freq-bands as well as other channels.
. the accuracy of the phase shifts degree does not affect the quality that much
...
This technology is already fading even without being utilized of the full potentials and advantages.. that's too bad..

/.
[edit] added a reference in the same thread