On Tue, 27 Nov 2007 22:34:50 -0500 (EST), Stéphane Doyon
>On Mon, 8 Oct 2007, Stéphane Doyon wrote:
>> On Sat, 6 Oct 2007, Marianne Arnold wrote:
>>> Having said this, I'm willing to test any voice patches on hwcodec, just
>>> can't test always and anytime but I was a tester for quite a few other
>>> features. Sometimes it needs some nagging and pointing to the patch entry
>>> in the tracker though. ;)
>> Wonderful. I'll call on you at some point then.
>I would like to request testers, in particular for HWCODEC (although
>others are welcome). The patch is bookmark-combined.diff in P#6240
>It improves voice feedback in the bookmark selection menu. Two twists are:
>enqueuing two thumbnails in the same utterance, and alternating between
>voice clips and thumbnails (which may be problematic on HWCODEC).
>Thanks for any feedback.
I have not tested your patch on my Archos V1 because I don't have an up
to date development environment, but would like to say something about
the following comment at your above link:-
>Bookmark.c had the following comment which I'm simply ignoring:
>/* disabled, because transition between talkbox and voice UI clip is not nice */
>So quite possibly there was a good reason for this, in which case I need
>to find appropriate fallback behavior for HWCODEC.
When I played around with this some time ago, I found that the MAS MP3
decoder would hiccup and loose one or two syllables when transitioning
between talkbox clips and Voice UI clips.
I found that this could be cured by ensuring that similar Lame
parameters were used for encoding both the .talk files and the Voice UI.
The parameters actually used were quite different, as the Voice UI
required that the clips were as short as possible so that the packed
module would fit into the available memory.
I hope that you find this information helpful.
Received on 2007-11-28