[MPlayer-users] Re: divx 6

[Michael Niedermayer]

Hi

On Sat, May 06, 2006 at 11:44:58PM +0200, Alexander Noe' wrote:
> Michael Niedermayer schrieb:
> 
> 
> >>They are not correct files since they are unplayable without the index
> >>and thus hopelessly fragile (and also unplayable before download
> >>completes). This is idiotic and broken and regardless of whether it's
> >>within the letter of the spec it's absolutely wrong.
> >
> >
> >does lavf as demuxer play them? and yes i agree these files arent valid
> >(assuming i guess correctly what we are talking about ...)
> 
> You obviously not knowing the AVI specs doesn't make this claim any 

uhm, i do know the avi spec pretty well, though i am only guessing what
these low overhead odml files are, we had enough flames so please stop
telling others what they know or dont know

do these low overhead files have a odml index? obviously yes ...
do they have a old style index? no? (doesnt matter)
do they have 2 or more frames with differnt sizes in a chunk? rich said
they are unplayable without index, not only unseekable, so my guess is yes
and thats where you break the avi spec


to quote the spec:
-----
dwSampleSize 

Specifies the size of a single sample of data. This is set to zero if the samples can vary in size. If this number is nonzero, then multiple samples of data can be grouped into a single chunk within the file. If it is zero, each sample of data (such as a video frame) must be in a separate chunk. For video streams, this number is typically zero, although it can be nonzero if all video frames are the same size. For audio streams, this number should be the same as the nBlockAlign member of the WAVEFORMATEX structure describing the audio.
-----

compressed video frames can vary in size -> dwSampleSize = 0
"If it is zero, each sample of data (such as a video frame) must be in a separate chunk."


> more valid. A 'valid file' is a file that meets the specs. There would 
> even be valid OGM files if anyone had properly written down OGM specs.

certainly ...


> 
> The AVI header has a flag (AVIF_MUSTUSEINDEX) of which, as the name 
> says already, the purpose is to indicate that the index is necessary, 

yes, mplayer, lavf and probably everyone else supports this


> and the OpenDML index points to the beginning of the data, not to a 
> chunk header, 

yes


> and the specs do not mention that there would have to be 
> a chunk header preceding the point an index entry points to.

that is where you are wrong, see above

also note that the odml spec doesnt seem to change the definition of 
dwSampleSize, so this is still true for odml-avis

these files are not valid avi, and they are not valid odml-avi

[...]
-- 
Michael

In the past you could go to a library and read, borrow or copy any book
Today you'd get arrested for mere telling someone where the library is