[MPlayer-dev-eng] Re: PATCH 4/5: substract surround from front in pl_surround
Tobias Diedrich
td at informatik.uni-hannover.de
Tue Dec 11 13:13:23 CET 2001
Stephen Davies wrote:
> Thanks for your effort and thought on this.
No prob. ^_^
As I said I am using pl_surround for testing my ao_multi output
and was a bit surprised at how it handles things.
And I read some of the docs on www.dolby.com yesterday.
> > Now this would mean we have downgraded front to mono :-(
> > But remembering that Surround is supposed to be low-pass filtered,
> > this is not really true, as mostly the high frequency carry
> > directional information, whereas you can't locate very low
> > frequencies.
>
> Surround should be filtered to 7kHz. That's hardly "non directional very
> low frequencies"...
ACK
I just meant that after lowpassing my attempt does not take away all
directional information, in fact I found the results to be not bad,
but I didn't do much tests and I don't have decent equipment attached
at the moment...
(I've got a probably 30Years old Pioneer SX-770 Receiver as main
Amplifier and two rather small noname speakers for front and
surround streamed over tcp to my notebook where I attached a portable
radio ^_^;)
> My code drives the two rear channels in anti-phase deliberately. In this
> I followed the Dolby Surround documents from Dolby, which recommend this
> to create a "diffuse, non-localizable" sound at the rear.
OK, I probably missed that point in the Dolby documents...
> Remember - this is not a Pro Logic style active decoder. In the Pro-Logic
> case you would drive the speakers in-phase. The adaptive matrix would
> steer the sound to attempt to localize it.
>
> If you drive the rear in-phase with a passive decoder, rear sounds
> localize always in the rear centre. For ambient sounds, reverb and such
> typical surround sounds you want the diffuse effect.
Hmm, at least my Dolby Prologic Receiver can't steer anything on the
rear speakers because they are electrically in series...
So surround is always in the rear centre there. I'd have to check if
they are in phase or in anti-phase.
--
Tobias PGP: 0x9AC7E0BC
Hannover Fantreffen ML: mailto:fantreffen-request at mantrha.de?subject=subscribe
Manga & Anime Treff Hannover: http://www.mantrha.de/
More information about the MPlayer-dev-eng
mailing list