Re: Second monitor failure
- From: Geoff Turnbull <codistic@xxxxxxxxx>
- Date: Fri, 16 Nov 2007 17:59:57 -0800
The March Hare [MVP] wrote:
On Mon, 12 Nov 2007 16:28:20 -0800, Geoff Turnbull wrote:
I have discovered a variable that correlates perfectly with the misperformance: when the code is run in the debugger, whether debug or retail build, we load one pair of Audio and Video Decoders and the video plays properly; when run out of the debugger, we load another pair of Decoders and the video fails.
If I explicitly load (AddFilterByCLSID()) the decoders that are loaded under the debugger, it works properly both in the debugger and out.
What would account for these differences (different decoders loaded, different behavior of the two decoders)?
Some decoders may not load while you are in the debugger to prevent you
reverse engineering them.
You can manually add the ones that work to the graph in your app.
According to GSpot, there are two Video Renders on my system; both quartz.dll, one with merit 0x00800001, the other 0x00400000, so I suppose the former is always loaded, though it seems intuitively that the renderer would be responsible. How can I discover the CLSID of the renderer?
It's almost certainly the one with the higher merit (VMR7 if you are on XP
or Vista). You can examine the graph by adding it to the ROT and looking
at it in GraphEdit. There are examples of how to do this in the SDK
samples, search for REGISTER_FILTERGRAPH and it is covered in the docs.
By loading filters to the graph manually, I've tried every combination of the two decoders and the two renderers, and the behavior definitely follows the decoders, though it seems, intuitively, that this would be the province of the renderer.
.
- Follow-Ups:
- Re: Second monitor failure
- From: The March Hare [MVP]
- Re: Second monitor failure
- References:
- Second monitor failure
- From: Geoff Turnbull
- Re: Second monitor failure
- From: The March Hare [MVP]
- Re: Second monitor failure
- From: Geoff Turnbull
- Re: Second monitor failure
- From: The March Hare [MVP]
- Second monitor failure
- Prev by Date: Re: VMR9 transparency with 2 video sources
- Next by Date: Re: Render filter (or decoder filter) doesn't release sample buffer
- Previous by thread: Re: Second monitor failure
- Next by thread: Re: Second monitor failure
- Index(es):
Relevant Pages
|
