[FFmpeg-devel] [PATCH 5/6] avcodec/qtrle: return last frame even if unchanged

[Michael Niedermayer]

On Sun, Aug 25, 2019 at 08:04:03PM -0300, James Almer wrote:
> On 8/25/2019 6:46 PM, Michael Niedermayer wrote:
> > On Sun, Aug 25, 2019 at 01:22:22PM -0300, James Almer wrote:
> >> On 8/24/2019 3:18 PM, Michael Niedermayer wrote:
[...]

> > if every filter following needs to process frames twice 2x CPU load
> > and after the filters references would also not be the same anymore
> > the following encoder would encoder 2x as many frames 2x CPU load,
> > bigger file lower quality per bitrate. Also playback of the resulting
> > file would require more cpu time as it has more frames.
> 
> What if the filter the user injected is meant to affect each and every
> frame in a different way? Think for example a constant fade out. If you
> remove the duplicate ones from what's meant to be a several seconds long
> completely static scene, would that effect be altered? Between dts x and
> dts y there are meant to be 100 frames, but now there's 1.
> I'm not sure how libavfilter works here, but if it tries to fill out the
> missing frames on its own, then the work of inserting the duplicate
> frames for the fade out effect would simply move from lavc to lavfi.

If you run a filter that requires a minimum number of frames per second
then you will need to ensure that to be there.
Theres a wide range of reasons why your input might not have frames
there.
So unless you only work with filter chains designed for a single input
container and codec simply assuming that there would be 25 fps at least
might once in a while bite you with no frame drop patches.
you might even with a single container and codec be hit by a 1fps or
lower slide show once in a while

thx

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Old school: Use the lowest level language in which you can solve the problem
            conveniently.
New school: Use the highest level language in which the latest supercomputer
            can solve the problem without the user falling asleep waiting.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20190826/20faaab3/attachment.sig>