[FFmpeg-devel] [RFC] scale filter can create output which scale filter cannot accept on its input

[Michael Niedermayer]

On Sun, Sep 04, 2011 at 10:51:31PM +0100, Mark Himsley wrote:
> I know this is a bit of a bizarre filter chain, It's constructed to  
> demonstrate the point I'm hoping to make, without using any of my code.
>
> Clock10.mov is just a PAL DV, but I think any _interlaced_ file will do.
>
> ./ffmpeg -loglevel debug -i ~/Movies/Clock10.mov -vf  
> "slicify=32,scale=0:0:interl=1,format=yuv420p,scale=0:0:interl=1,format=yuv422p,scale=0:0:interl=1,format=yuv420p" 
> -pix_fmt yuv420p -vframes 1 -an -f rawvideo -y /dev/null
>
> Essentially, the 'y' parameter in the slices of the output of the second  
> scale is not aligned by 4 (the first slice is, in fact, 26 lines high,  
> so the second slice starts at 26) so it is not valid for the input of  
> the third scale.
>
> I don't think the output of this filter chain should throw an assert.
>
> I stumbled upon this when trying to make the fieldorder filter do slices.
>
> Sorry to have not included my suggestion for a patch! My brain is fried.

Please open a bugreport on trac so this issue is not forgotten in
case noone has time to look into this now

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

No human being will ever know the Truth, for even if they happen to say it
by chance, they would not even known they had done so. -- Xenophanes
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20110909/d64b2de8/attachment.asc>