On Mon, Feb 20, 2006 at 11:46:30PM +0100, michel Xhaard wrote:
> Luc
> I did not really understand what you mean :( 
>  0x2600frame.txt is get from pixart0x2600.log.tar.gz file
> data.jpg is the binary image of 0x2600frame.txt
> the frame header  ff ff 00 ff 96  is know for a lot of manufacturers like
> Pixart, Mars-Semi, with small change Sonix ff ff 00 c4 c4 96 ..ect
> The PAC7311 chips brief speak about a full jpeg stream, that is why i suspect 
> a jpeg, but who know ..
> the end of frame seem ff d9 
> end of frame or start of frame are not always synchro with a new iso packet 
> buffer :( if you look at the Pixart PAC207 or the sn9c120 from Sonix  
> (spca5xx) you will see i have to split the iso packet and ask a new buffer in 
> the middle of a iso desc,  In interrupt context fun :)
> regards
 
Ok thanks for the explanation, now i understand better the format. I think
it's very stupid to not begin a new frame at the begin of an iso transaction
as the pwc driver do. So hardware designer is not always a sane software
designer :). If i found some times, i'll try to decode the stream, but
perhaps it's compressed.

Luc
_______________________________________________
pwc mailing list
[email protected]
http://lists.saillard.org/mailman/listinfo/pwc

Reply via email to