[FFmpeg-user] ffprobe -show_frames and coded_picture_number

[Andy Furniss]

thljcl wrote:

> From the sources, I created both 24-FPS version and 25-FPS version,
> which I uploaded to http://sdrv.ms/14FXYb9. On my display, I could
> not tell the difference with my naked eyes at most parts, other than
> at the end where the credits are rolling. The lack of “motion
> blurring” causes the obvious “jarring unrealistic motion” when the
> credits are rolling for the 25-FPS version. I only used ffmpeg to
> duplicate frames. If I edit the frames by adding some “motion
> blurring” or “transitional effects”, the “jarring unrealistic motion”
> could be somewhat mitigated.

I did have a look and the result was as I expected = very noticeable
"stop" every second viewing on both 50Hz and 60Hz.

Of course not every scene was affected but a lot more that just the
credits. The opening fly in looked terrible.

I accept you may not notice this as much, peoples perceptions are
different when it comes to many different things related to video. I've
read that some don't even notice tearing without vsync, I certainly do.
Some people are more sensitive to flicker that others. Film judder is
another case - some people just don't seem to see that  deinterlacing
50i to 25p looks really juddery on some shots, I am also very sensitive
to that one.

I don't think that motion blur could help the 24 -> 25 dup, I guess it
could in the case of 50i -> 25p.

> A single experiment is far from conclusive to prove anything.
> However, from my experiences as well as this experiment, while
> duplicating/dropping frames is not strictly perfect, it’s certainly
> an acceptable solution for many if not most cases. The reasoning is
> mainly the properties of human visual system. “Motion picture” is
> never “the picture that moves” but our visual system makes us think
> that “it moves”.

I would never dup 24 -> 25, if you need to make something for pal, speed
it up - that's the way it's always been done.