It was all going so well. The four python scripts
for automating Blender
rendering were working.
I rendered a full screen animation in just under three days. But the result was too noisy. I updated my new render_finally.py to include clamp
values and had started another 3 day render.
Turning my attention to the fact that I wanted the animations to be staggered e.g. first 2 seconds from the first video then the next 2 seconds from the next. Staggered so that to the viewer the model just kept turning.
made github repository for the script video-sequence-splicer
This went really well right up to the testing of the finished videos, you see there was a glitch. The animation would revert a few degrees and then start turning again.
I found the last video was slow than what I remembered from viewing the first few. Sure enough the animation was "delayed" and further investigation should I had inadvertently set the frame
to 11 in all the Blender
files from 52 onward. What was worse was, that all of those had the window layout set to hide the time line so I might have never seen it.
The fix is easy but painful. I updated the track.py script to include a "set frame to 0" before doing anything.
bpy.context.scene.frame_current = 0
Now I have to go back and re-render 80 videos. At least my script will make it painless from a interaction point of view.