[theora] ffmpeg2theora bug: generation of a video from a series of images

sdaau sd at imi.aau.dk
Mon Aug 9 06:52:18 PDT 2010


Hi all,

Good I found this list has a (read-only) newsgroup interface in gmane - 
so hopefully I'm replying with the right thread headers :)

> On 03/05/2010 09:13 AM, Basil Mohamed Gohar wrote:
>> On 03/05/2010 01:26 AM, Michael Opdenacker wrote:
>>
>>> Hi J.,
>>>
>>> We are using ffmpeg2theora to produce an introduction video from a
>>> series of images:
>>>
>>> ffmpeg2theora-0.26.linux32.bin images/img-%03d.png -o intro.ogv
>>>
>>> The video that we get is huge and horrible, with lots of blinking colors:
>>> http://free-electrons.com/issues/ffmpeg2theora-mar5-2010/intro.ogv
>>>
>>> Here's what we had in the past:
>>> http://free-electrons.com/issues/ffmpeg2theora-mar5-2010/intro-correct.ogv
>>>
>>> This issue was introduced in ffmpeg2theora-0.25, and a work-around is to
>>> revert to ffmpeg2theora-0.24
>>> [snip]
>>>
>>>
>
> Okay, I just ran all the PNG images through optipng after noticing
> that they had an alpha channel.  Also, ffmpeg reported they were in
> bgra format (the "a" is for alpha, but I don't know why they were
> bgr order).  I ran it through png2theora again, after this lossless
> (for the purposes of this test) processing, and I got a smaller video
> file without the corruption.  So, my first advice to you is to run
> the images through optipng and try again.


I would just like to say, that I tried this - and optipng doesn't seem 
to be the solution.

So I produced "a script that produces non-optimized (and then optimized) 
PNG's on the fly, and then uses ffmpeg2theora to encode them to videos; 
which demonstrates there may a bug with ffmpeg2theora 0.27, when using 
it to make videos of a PNG image sequence", which can be found in the 
second post in this thread:

10.04: Video overlay/composite editing with Theora ogv
http://ubuntuforums.org/showthread.php?p=9697655#post9697655

I have included the results I get when running the script, as comments 
at the end of the script:"

# For both converted and non-converted pngs, 0.24 shows OK videos; 0.27 
shows extra colors for both;
# for all videos, vlc says:
# vlc (command line): "swScaler: pal8 is not supported as output pixel 
format""

And all this is with the latest source of ffmpeg and theora (maybe some 
week old)

Could anyone please look into this problem a bit more closely? Or should 
I try to file a bug somewhere (if so, where?) I'd love to be able to use 
the latest ffmpeg2theora - also with PNG image sequence :)

Thanks,
Cheers!



More information about the theora mailing list