[FFmpeg-devel] Bug in the mpeg-ts writer code

[Baptiste Coudurier]

On 02/17/2010 04:21 PM, Yann Coupin wrote:
> On Wed, Feb 17, 2010 at 11:25 PM, Baptiste Coudurier<
> baptiste.coudurier at gmail.com>  wrote:
>
>>
>> The patch is good, however the first cc will be 1, I'd prefer if it would
>> be 0.
>>
>>
> I realized that but neither VLC (not sure I should consider VLC as an
> analyzer though) nor this other analyser[1] complained about it. And while I
> didn't do it myself, I got positive reports about files generated after the
> patch was applied while files generated before both this fix and the one
> about CBR padding had errors according to the cisco analyzer. On top of
> that, and I can totally be wrong here, I thought that in a real life
> broadcasting scenario, you?d be lucky to tune to a given stream and get a
> packet with a cc equal to 0 so I don't think it's really problematic. If you
> think it can cause problems, it's easy to correct by initializing the cc to
> 15 instead of 0 and it will wrap and be equal to 0 after the first
> incrementation. Or do think it's a hack and it should be done differently?

I should have mentioned that it is not problematic at all. I said that 
I'd prefer if it would start at 0 because in the perfect scenario of a 
complete file, it seems natural for the cc to start at 0.

-- 
Baptiste COUDURIER
Key fingerprint                 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA
FFmpeg maintainer                                  http://www.ffmpeg.org