[wdmaudiodev] Re: GFX/LFX fullduplex support.

  • From: pete c <prchan_0779@xxxxxxxxx>
  • To: wdmaudiodev@xxxxxxxxxxxxx
  • Date: Thu, 4 Sep 2008 15:06:16 -0700 (PDT)

i appreciate the info,
Im assuming you ran into the same starvation capture/render situation that we 
previously, when the LFX APO render side falls behind the WDM driver 
filter(capture side); did the AEC handle this situation or did you prevent the 
WDM driver filter(capture side) from being starved of LFX APO packets from the 
render side.  
Do feel there is any benefit of the WDM Filter Driver(capture side) / LFX APO 
(render side) implementation over a GFX/LFX implementation?
Im trying to determine what the best approach is here.

--- On Thu, 9/4/08, Tim Roberts <timr@xxxxxxxxx> wrote:

From: Tim Roberts <timr@xxxxxxxxx>
Subject: [wdmaudiodev] Re: GFX/LFX fullduplex support.
To: wdmaudiodev@xxxxxxxxxxxxx
Date: Thursday, September 4, 2008, 4:28 PM

pete c wrote: 

So for your Vista implementation you were able to use a LFX APO for the render 
side that feeds render data to the AEC which lives in the WDM filter driver on 
capture side.

Did you have a problems with render/capture synchronization as we've been 
discusing with the GFX/LFX approach?
Yes.  Note, however, that I only did the driver work.  The algorithm work was 
all done by the client.  I was working with a group of audio processing gurus 
who have more letters after their names than I have.  They knew the algorithms, 
but they didn't know drivers.  That's why they hired me.  I have very little 
visibility into the processing library itself.  I just feed them packets.

Did you make any attempts to pursue a GFX/LFX framework for AEC as we've been 
No.  We had to have a WDM filter for XP anyway, and since the WDM filter was 
going to work in Vista (because it's not a WaveRT driver), it didn't make sense 
to pursue a capture APO.
Tim Roberts, timr@xxxxxxxxx
Providenza & Boekelheide, Inc.


Other related posts: