[mythtv] Question/request: mythtranscode resampled output

[Joseph A. Caputo]

Geoffrey Hausheer wrote:

> On Sat, 18 Oct 2003 08:21:01 -0400, "Joseph A. Caputo
> jcaputo1-at-comcast.net |mythtv/1.0-Allow|" <s9v7hefgnx0t at sneakemail.com>
> said:
> 
>>Hi,
>>
>>I've been poking around in the code trying to figure out how to do this, 
>>but I'm just not familiar with all of the Nuppel/Lame/libavcodec stuff, 
>>and I thought maybe someone more familiar might say 'oh, yeah, that's 
>>easy...'
>>
>>What I'd like to do is add the capability for mythtranscode to resample 
>>the audio (I can add the support code; i.e., command line switches, etc; 
>>it's just the actual resampling that I'm uncertain of).  I'm primarily 
>>interested in having the output resampled when using 'raw' output from 
>>mythtranscode, but I suppose some might find it useful (??) in a 
>>transcoding profile as well.  Anyway, cursory examination makes me think 
>>that currently the AudioOutput/AudioBuffer classes kind of 'no-op' the 
>>output sample rate stuff and just assume it's the same as the input 
>>sample rate.  Is this correct?  I'm kind of lost as to where to hook 
>>stuff up.  If someone familiar with this code is willing to give it a 
>>shot, I'll happily 'trade' patch efforts for some UI-type work.
>>
> 
> I'm not sure of a good way to do that.  Once we have an algorithm that
> resamples the audio, it should be added to the AudioReencodeBuffer class.
>  But I think such resampling needs to be done in the frequency domain,
> and I don't remember any of my Laplace transform stuff.  Maybe it can be
> done easily in the time domain (it's been 10 years since I've done audio
> processing, so i really don't know).  If you like, I'll take a look at
> it.  I wouldn't expect it to take much coding to implement.

I believe Isaac mentioned libavcodec has a function to do it; it's more 
a matter of hooking it into the AudioReencodeBuffer class, as you said.

> I guess the
> real question is why you need it?  Most tools that can read the fifo
> output data should be capable of resampling the data.

True; I guess it'd just seem 'cleaner' if Myth could do it natively 
(well, that and the fact that toolame can't resample AFAIK).  Anyway, I 
suppose I could try piping the raw audio through sox to resample.

-JAC