[mythtv] Splitting out audio output code in preparation for native ALSA support

James Knight foom at fuhm.net
Sat May 10 00:12:13 EDT 2003

Hi, I'm working on adding native ALSA support to Myth to avoid needing 
to use the OSS emulation layer.

My plan for doing this is:
1) Split out audio output code from NuppelVideoPlayer and make a 
generic base class for audio output
2) Add the ALSA output class.
3) Split out audio record code in NuppelVideoRecorder and make generic 
base class for audio record
4) Add the ALSA input class.
5) Do the same sort of thing for the mixer control.
6) Update the configuration interface to support this, somehow.

I have done step 1 and started step 2. All the code in 
NuppelVideoPlayer having to do with OSS DSP audio output is now in the 
class AudioOutputDSP, which is a subclass of the pure virtual class 
AudioOutput. AudioOutputALSA is half-implemented, but I have not 
included that here.

I have attached the diffs to libmythtv.pro and 
NuppelVideoPlayer.(cpp|h) as well as the new files AudioOutput.(cpp|h), 
and AudioOutputDSP.(cpp|h).

I do have one question I posed in AudioOutputDSP::OutputAudioLoop about 
its use of video_frame_rate to determine how many bytes should buffer 
before sending the buffer to the sound hardware. As far as I can tell, 
this is not necessary, and perhaps even detrimental with low framrates, 
so I hardcoded it to 30.

I consider this code a first draft, and comments are certainly welcome. 
This is based on current CVS code, and it works as far as I have 
tested, except that, for the moment, I have completely commented out 
the "NuppelVideoPlayer::ReencodeFile" function, as it was doing evil 
things with the audio buffers that I haven't bothered to fix yet.


-------------- next part --------------
z'µìmjÛZržžÜ²Ç+¹¶ÞtÖ¦y«ŠŠ.¶›­¶ÓÊ—š¦™bq«b¢ñƒ8©J,Þמöå¼­zÀÞ±É赫ڊV›•å+—úb¥êæj)jÛZržžÛ?ÛM7ӝ=ÿÞùs^9sö®v*(ºÚnµ¸§

More information about the mythtv-dev mailing list