[mythtv] NuppelVideo(Recorder/Player)

[Craig Longman]

i have been playing around with the transcoder because i want to keep 
some of the kids shows i've recorded so i can play them for my child 
whenever she wants to watch them, but the file sizes for the hardware 
encoded files are a little too big, and she is not very discerning when 
it comes to resolutions... ;-)

i have a few questions/problems though.
1) it appears that the NuppleVideoPlayer doesn't initialize the 
NuppleVideoRecorder properely, the commDetect member is not initialized 
(afaict) in the ReencodeFile() method.  there are two ways to fix this i 
think (but the right way might depend on the next question).  just 
adding a NuppleVideoRecorder::InitCommDetect() would work (and it seems 
to work fine), or perhaps a check at the end of 
NuppleVideoRecorder::WriteVideo() to see if commDetect is NULL before 
using it, in addition to checking if we're hardware encoding.

2) i noticed that the ReencodeFile() doesn't honour the edit lists. 
 would it be acceptable if it did, as in, if worked on that as a patch, 
would it be accepted?  perhaps with a flag passed into it, defaulting to 
false.  in my case, i would like to flag off the parts i don't want, and 
re-encode it into a smaller file, with the cruft removed.

3) i also have some very strange stuff happening on the audio, the pitch 
is way, way down, but still appears to be in sync.  almost like it was 
recorded at a higher sampling rate, then not properely resampled.  both 
the default and the transcoding recording profile are set to 32000 
though, so i'm confused.   setting the transcoding sampling rate to 
48000 causes the following error:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 16384 (LWP 9660)]
0x4024e8a5 in AnalyzeSamples () from /usr/local/lib/libmp3lame.so.0
(gdb) bt full
#0  0x4024e8a5 in AnalyzeSamples () from /usr/local/lib/libmp3lame.so.0
No symbol table info available.
#1  0x4028cc18 in __DTOR_END__ () from /usr/local/lib/libmp3lame.so.0
No symbol table info available.
#2  0x44df2000 in ?? ()
No symbol table info available.
Cannot access memory at address 0xc492e000

setting the output format to uncompressed though, produces good sound, 
albeit a much larger file.  i've never used libmp3lame, so i don't know 
if this is expected (i seem to recall hearing something about non-32000 
causing problems though).  i'm much less certain about what to do to 
resolve the problem with mp3 encoded 32000 output though, sounding all 
fuzzy like that.  maybe something to do with the fact that i'm running a 
hardware encoder.

cheers,

    CraigL->Thx();