[alsa-devel] basic questions

[Lee Revell]

On Fri, Mar 14, 2008 at 6:57 PM, Eliot Blennerhassett
<linux at audioscience.com> wrote:
> On Saturday 15 March 2008 11:35:20 Lee Revell wrote:
>  > On Fri, Mar 14, 2008 at 12:35 AM, Harsha priya gupta
>  >
>  > <harshapriya at gmail.com> wrote:
>  > > Hi,
>  > >
>  > >  I am ramping up on ALSA architecture and framework. I have a couple of
>  > > basic questions. I apologize if these questions are very basic to be
>  > > raised in this mailing list. If anyone can help me find answers, it would
>  > > be great. Thanks in advance.
>  > >
>  > >  1. My assumption is that ALSA handles only PCM streams today. Does it
>  > > handle encoded streams from application to sound card (Which can do
>  > > hardware decoding)?
>  >
>  > Someone else could probably give a better answer, but I believe that
>  > currently ALSA has no API for cards that do HW decoding.
>
>  Not AFAIK, and certainly the architecture assumes a constant data rate. I.e X
>  bytes consumed implies elapsed time of (X * constant)  = the _period_
>
>  Alsa headers do define some format constants for compressed audio.
>
>  We would like to support compressed audio for our cards, which can do hardware
>  decoding. I did try to implement it early on, but got lost in the quagmire of
>  driver midlevel code and alsa-lib.
>
>  A first step would be to allow CBR compressed formats to pass through to the
>  driver unaltered. For these there is a defined bits per sample, though it can
>  be non-integer which might be a problem.

I would look at how ALSA implements AC3 passthrough.

Lee