Ok, but I use this with other encoders too, it's only the embedded
platform where the parsing cost is so critical. I think it shouldn't be
a big deal to change it to check for both the three and four byte codes.
On 03/25/2015 09:50 PM, Gilles Chanteperdrix wrote:
On Wed, Mar 25, 2015 at 02:42:29PM +0100, Robert Smith wrote:
Thanks,
I'm not sure what sort of constraints I can put on the number of slices so I
probably can't get away with short circuiting the parsing as nicely as you
do, but from memory our code for finding the start codes steps byte by byte
so your example here would definitely be a significant improvement.
I would need to search for three byte start codes though, I haven't analysed
the output of the TI encoder but I have had other encoders outputting three
start byte start codes.
By default the DM8148 encoder outputs one slice per image. And it
uses 4 bytes start codes.
_______________________________________________
live-devel mailing list
[email protected]
http://lists.live555.com/mailman/listinfo/live-devel
