Same number of frames should be in transcoded output about lpms HOT 2 CLOSED

Looked into that today, and I believe the 6th frame is produced by the FPS filter. The test asks for "P720p30fps16x9" profile, but the input video is 25FPS (which I checked ffprobe -show_frames -unit: frame duration is 0.04s or 40ms). Now converting from 25FPS to 30FPS requires adding 5 frames per second, or one additional frame every five original frames. This is what I believe happens here. To be more precise, these are original video timestamps:
pkt_pts_time=1.480000 s
pkt_pts_time=1.520000 s
pkt_pts_time=1.560000 s
pkt_pts_time=1.600000 s
pkt_pts_time=1.640000 s
So original video has 5 frames over 0.16s
And transcoded video timestamps:
pkt_pts_time=1.466016 s
pkt_pts_time=1.499349 s
pkt_pts_time=1.532682 s
pkt_pts_time=1.566016 s
pkt_pts_time=1.599349 s
pkt_pts_time=1.632682 s
So transcoded video has 5 frames over 0.166666s. Not sure why the timestamps are shifted (I mean why the first frame of transcoded video doesn't repeat the timestamp of the first frame of original video), but I guess this is how filter works.

I also verified my theory by invoking the test with "P720p25fps16x9" profile, thus getting rid of FPS filter. The results were as expected, 5 output frames. My conclusion is: this is not really a bug, transcoder is behaving properly.

What do you think @AlexKordic ?

from lpms.

Good point! This is then false report from my side

from lpms.