When making ouput to image sequence in other models its easy to stop and resume. We just change starting frame and going forward.
But with SlowMo its different.
We process some frames but get more image files if frame rate is increased.
If we change starting frame after stop to continue, file names numbering will not continue, but will start from frame number selected.
This makes a problem. So to join results we need to start second processing a little before first was ended (looking by image\scene content). Then we need to find same images to find place where we need to make join. And at last we need to rename all images, processed second time.
We need a more easy approach. Better built in.
I suggest that you add a button “Continue” that will appear when output to image sequence is selected and output folder specified. And that it will read the last saved image number from folder (not from save.json), calculate the frame number from which to continue and start a little before it if necessary.
When increasing the frame rate by 2.5 times, there is a problem with the number of output frames. (23.976> 59.94 24> 60)
The output is about 1.086 times more than the total number of frames originally expected.
Also, the change is not constant, and it seems as if acceleration and deceleration are repeated.
I tried all patterns of short clip, long clip, mp4 input, mkv input, Prores output, png output but all didn’t work. (The number of output frames was the same for all formats.)
Thanks for reaching out!
There’s currently an issue with some videos where duplicated frames are present indeed.
We don’t currently drop duplicated frames so the effect becomes more pronounced,
and we hope this is something we can improve in the future.
2.5x frame conversion (23.974> 59.97 24> 60) should be desired by so many people as many video sources are not 29.97 30 and we look forward to improvement.
