[MPlayer-users] Re: lavc-Options for *BEST* quality?

[Ketil Z. Malde]

D Richard Felker III <dalias at aerifal.cx> writes:

>>>   should be repeated a few times perhaps

>> Probably better to use an error correcting code, unless you
>> specifically want to add resilience against burst errors

> Of course we're talking burst errors. Think of a bad block on a hard
> drive (reading will return all 0's) or something. 

Well, but this will be generally caught by a checksum.  Just repeating
the packet headers are in general not going to be an effective (nor
efficient measure).

>>> codec specific header packet
>>>   64bit startcode
>>>   codec specific data

>> What's this for? [snip]

> Huh? I don't understand your question at all.

I'm just curious what kind of data would be retained.

>>> per frame header packet

>> Do you by 'frame' mean what I think you mean?

> Depends on what you think he means.

One image in a movie, e.g. 1/25th of a seconds worth, or similar.
Having a separate packet per frame sounds like a lot of overhead.

> Have you failed to read any of the past discussion whatsoever? This
> has already been covered many many times. The story goes like this.

Guilty, I suppose. I can't keep up with all the posts on
mplayer-users, so I only read it sporadically.  And I haven't use file
sharing, so I've naturally not encountered this problem.  Okay, so I
should probably have stayed out of the discussion. :-)

-kzm
-- 
If I haven't seen further, it is by standing in the footprints of giants