[FFmpeg-user] Questions about w3fdif deinterlacing

Seth Parker csparker247 at gmail.com
Fri Jan 17 17:33:11 CET 2014

Carl Eugen Hoyos <cehoyos <at> ag.or.at> writes:
>The reason is that I was wrong;-)
>The filter "works" exactly as you expected because the
>"magic" is there, and for the sample you provided, it
>works perfectly: All frames are correctly identified.
>Note that there are many streams for which the detection
>will not work so nice.

So idet DOES report frame info to w3fdif? Of course, as you mentioned and
as I've seen elsewhere, idet isn't the most reliable, so it's not a great
solution. Still, good to know that it will work in certain situations.

Andy Furniss <adf.lists <at> gmail.com> writes:
> AIUI DV* is specified as bff, so I guess this sample was not direct from a
camera but has been coded from elsewhere?

It would have originally been DVCPro HD footage recorded to P2 cards (back
before P2 supported AVC). The only conversion of any sort would have been
the remux to MOV that FCP7 did for P2 media. Now, this assumes that my
predecessor followed the correct workflow for ingestion, so I honestly
can't say. ;)

Thanks for all the help! This has cleared up quite a lot!

*Seth Parker*
Video Editor | The University of Kentucky Vis Center
www.csethparker.com | www.vis.uky.edu

More information about the ffmpeg-user mailing list