[FFmpeg-devel] [RFC] special "broken DV" demuxer

[Reimar Döffinger]

On Wed, Mar 11, 2009 at 08:47:42AM -0700, Roman V Shaposhnik wrote:
> > proper autodetection is simple and reliable, needs reading only a few
> > bytes and does not needs seeking nor relevant amount of buffering.
> > When you allow anything non-critical to be missing, allow reserved
> > values to have any value etc. autodetection will require reading and
> > analyzing huge amounts of data 
> 
> I don't see how reading one frame (even a DVCPRO HD one) and cycling
> through all the know locations for the redundant copies of the
> dv_sect_header would be "analyzing huge amounts of data".

Uh, the current code already searches for a dv_sect_header _anywhere_
(in the probe data as far as available, arbitrary amounts if the demuxer
is forced), that is exactly what causes the issue, pond.dv does not
contain even a _single_ one.
The only half-way reliable thing it contains is dv_sect_subcode, which
my other patch adds a check for in addition, but with only 25 fixed bit
values there is a chance of misdetection (> 1 per 128MB of random data
on average, compared to < 1 per 2GB of random data on average for the
dv_sect_header check), and there is no particular reason why a DV
file would need to contain that, since again most players would not care if
that one was missing either...

Greetings,
Reimar D?ffinger