[FFmpeg-user] [Bulk] Re: Status of two prores decoders/encoders

[Tim Nicholson]

On 17/05/13 13:41, Paul B Mahol wrote:
> On 5/17/13, Tim Nicholson <nichot20 at yahoo.com> wrote:
>> On 17/05/13 11:05, Carl Eugen Hoyos wrote:
>>> Robert Krueger <krueger <at> lesspain.de> writes:
>>>
>>>> is there a place where I can find out why there are
>>>> two prores decoders (both LGPL) and two encoders
>>>> (also both LGPL)
>>>
>>> http://blog.pkh.me/p/13-the-ffmpeg-libav-situation.html
>>> http://en.wikipedia.org/wiki/Not_invented_here
>>>
>>>> and what the pros and cons of each are?
>>>
>>> The "Anatoliy" encoder is faster.
>>>
>>
>> But I think only the Kostya supports interlace...
> 
> Proof?

If you look at the code you will see that Kostya inspects the field
order settings, and Anatoly does not.

To confirm the consequences of this behaviour I made transcodes to mov
using both versions, and with interlaced and non interlaced material.

I then inspected the mov's with both Baton, Atom Inspector and ffmbc's
ffprobe.

in all cases the Kostya files flags adhered to the field order settings
I had made, but all the Anatoly files were flagged as "progressive".

Whilst I appreciate that the stream material may well have been
interlaced in the Anatoly samples, they were not correctly flagged as
such, and this led to problems of interpeting the files correctly in
other processes, and rejection by broadcast industry standard testing
equipment.

Of course I may have missed a patch that fixed this behaviour, but it
was sufficent proof for me at the time I conducted the tests.


> [..]
> 


-- 
Tim