[MEncoder-users] Telecine and B frames for DVD

[Walter R. G. Hachmann]

Rich Felker escreveu:
> On Thu, Feb 14, 2008 at 07:07:56PM -0200, Walter R. G. Hachmann wrote:
>> On Tue, 12 Feb 2008 21:19:21 -0500, Rich Felker wrote:
>>
>>
>>
>>>> Do we have a mpeg muxer bug here???
>>>   
>>> Probably.. Can you try with max_b_frames=1 and any other variants you
>>> can think of?
>>>
>>> Rich
>>
>> Hi, Rich
>>
>> I repeated the tests with vmax_b_frames=1, 2, 3, 4, 5 and 6. The 
>> interesting thing is that the problem arises with 2, 4 and 5 B frames, 
>> but not with 1, 3 and 6. With 6, I start to have additional problems 
>> when muxing (mplex error messages). I also tested it again with 
>> vmax_b_frames=2 and mbd=2, but the same problem happened again.
> 
> OK, I suspect the mpeg muxer was only tested with 1 B frame...
> Hopefully nico or someone will pick up on this report and fix the
> issue but for now I think your best bet is to stick to 1. IIRC this is
> what most commercial discs do anyway and it might even be part of the
> (not-so-strictly-followed) DVD spec. Most of the old PSNR tests done
> with mencoder also showed the best results with only 1 B-frame, I
> think.
> 
> Rich
> _______________________________________________
> MEncoder-users mailing list
> MEncoder-users at mplayerhq.hu
> https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> 
Ok, let's wait the DV guys. I wanted to use 2 B frames just because most 
of the commercial DVD's I've seen used the sequence IBBPBBPBBP...
But, after several tests, with bitrates ranging from 1500 to 6000 kbps, 
I couldn't see any improvement from using B frames. As a matter of fact, 
by including "psnr" in the lavcopts chain the numbers got even worst 
then with no B frames at all, although the differences are usually 
smaller than 0.3 dB.

Thanks,

Walter