[XviD-devel] Re: XviD Mpeg4 Iso Interoperability bug!

Christoph Lampert chl at math.uni-bonn.de
Wed Feb 19 13:04:57 CET 2003


On Wed, 19 Feb 2003, Daniele Buffa wrote:

> Sorry if I contact you directly, I posted this problem on doom9 forum
> but I've received no response yet but I think it's anyway important...
>  
> I report what I wrote on doom9
>  
> <<
> Latest cvs sources and strange iso compatibility problems... 
> 
> Hi, 
> 
> Yesterday (16-02-03) I checked out last source of development three
> and compiled my "personal" version of XviD.
> 
> I have to compress some material....:/ 

If it's important material, don't use the development tree since that may
change from time to time (as you might have noticed). Normally, that is
when bugs are _removed_, no introduced, but who knows...

 
> First, the new feature, VHQ, cause my system to crash. However isn't
> this the problem, 'cause VHQ is very beta and unstable feature and at
> the moment I don't want to use it. 

Can you be more specific? I didn't have a single crash since I started
testing it. 

> The problem is on b-frame coding and mpeg4 iso interoperability. With
> previous version I had no problem, so I think the b-frame coding has
> been modified since the last time I checked out cvs.

That depends on _when_ you checked out the last time. Please check which
bitstream version of XVID that was. (it's in xvid.h)

As far as I know only padding has been changed lately (a few weeks ago),
and that was because is was broken _before_ (MPEG-4 padding is rather
strange business...), and now isn't anymore. 

> Here you can find a small test movie. I used quant 8 in order to gain
> bytes (I've a normal slow analogic modem, no adsl) but also using
> quant 2 the situation is the same.
>  
> The avi movie (no visible problem if played with Xvid or Divx filter) 
> http://www.neowebsite.it/xvid/test.avi <http://www.neowebsite.it/xvid/test.avi>  
> 
> The iso mp4 movie obtained from the above avi (with the problem, played with Envivio TV, www.envivio.com <http://www.envivio.com>  ) 
> http://www.neowebsite.it/xvid/test.mp4 <http://www.neowebsite.it/xvid/test.mp4>  

Hm, there is no problem with ffmpeg decoder, so maybe it's not
encoder related but rather the decoder. Can you describe _what_ you
see with Envivio? Can you maybe check with real MS FDAM and provide output
of that? 
Apart from the decoder, if it's really a bug, we need your encoding
settings and hardware/software plattform. 

 
> I think that should be better for XviD to test also mpeg4 iso
> interoperability on corrections/new implementations... obviously not
> on aplha implementation, but isn't this the case.

Envivio is not perfect mpeg4 iso, it's Envivio. I'm pretty sure it 
has as many bugs as other decoders. And even MPEG-4 ISO Specs are changed
(called "clarified") from time to time. 

gruel 


> The mine would be just an advice, I use XviD only for mp4 iso streams,
> I would continue to use it if possible ....
> 
> Sorry for my english that is far to be perfect, 
> Thanks in advance, 
> 
> FunMan
> 
> >>
> 
> Hoping in your answer,
> 
> Greeting
> 
> Daniele.
> 
> 



More information about the XviD-devel mailing list