Closed GoogleCodeExporter closed 8 years ago
CoreAVC only parses the video from the stream. If the audio is not working,
that
would seem to have something to do with the transport, but is not really in my
domain, nor is it likely to have anything to do with coreavc. On my system,
pass
through works fine with CoreAVC.
I really have no idea how to debug this further.
Original comment by alannis...@gmail.com
on 7 Jun 2008 at 2:32
Maybe you're right....maybe is not coreavc related...but is definitely related
to
h264 transport.
You are lucky Alan, coz you have none of above symptoms...looks like a second
poster
has exactly the same problems:
http://groups.google.com/group/coreavc-for-linux/browse_thread/thread/337b2f4ca0
6574e1?hl=en
Original comment by ppinc...@gmail.com
on 12 Jun 2008 at 12:19
fuser shows that dshowserver opens the snd card - there is therefore a race
(even
with dmix) as to whether mythfrontend or dshowserver gets it first.. Can we stop
dshowserver from doing this?
Original comment by foobum
on 4 Oct 2008 at 1:15
I've been seeing the same issue lately. I am looking into it now.
Original comment by alannis...@gmail.com
on 8 Oct 2008 at 6:26
The issue turns out to not be in dshowserver but in myth itself.
Basically the call to dsshowserver in avformatdecoder.cpp doesn't close open
filehandles. Myth internally opens the sound-card, calls dshowserver, closes
the
sound card then reopens it (which fails if the close happens after the
dshowserver call)
The fix is to close all open filehandles before calling dshowserver. I can't
test it
at the moment, but this one-line patch should resolve the issue:
{{{
Index: libs/libmythtv/avformatdecoder.cpp
===================================================================
--- libs/libmythtv/avformatdecoder.cpp.orig 2008-10-08 13:09:22.000000000 -0700
+++ libs/libmythtv/avformatdecoder.cpp 2008-10-08 13:05:42.000000000 -0700
@@ -331,7 +331,7 @@
//Create read/write semaphores in locked state
ds_mpi->sem_wr = sem_open(sem1, O_CREAT, 0644, 0);
ds_mpi->sem_rd = sem_open(sem2, O_CREAT, 0644, 0);
- system(cmd);
+ myth_system(cmd);
ret = sem_twait(ds_mpi->sem_rd, 10);
shm_unlink(shm);
sem_unlink(sem1);
}}}
Original comment by alannis...@gmail.com
on 8 Oct 2008 at 8:10
The change appears to work for me. I've committed it now.
Original comment by alannis...@gmail.com
on 9 Oct 2008 at 1:47
Original issue reported on code.google.com by
ppinc...@gmail.com
on 6 Jun 2008 at 8:41