From: Jens Arnold <arnold-j_at_t-online.de>
Date: Mon, 25 Sep 2006 08:48:27 +0200

On 24.09.2006, Jacob Rau wrote:

> Also, why are we wasting CPU time bit-swapping? We can have
> the software do this--put something in Talkbox that asks "Are
> you using this for an iPod" (or any other platform that uses a
> software codec) and have it spit out a properly formatted
> file.

The bitswap was meant as an intermediate solution. The proper
solution will be extending the voice file header with a format
and target indicator, and building target specific voice files.
This does not only mean to avoid the bitswap, but also to
exclude voice clips not needed for a target, and to be able to
use a dedicated voice codec (i.e. speex) instead of mp3 for
swcodec targets. Unfortunately we still don't have a speex
codec, in spite of a lot of talking about it...

But for now, I don't think the bitswap is a big problem. It's
done only once when the voice file is loaded.

Regards, Jens
Received on 2006-09-25