I was just going through the crouching normals video and counting the startup, active and recovery frames and then comparing the times against the frame data thread. It looks like the recovery times seem to differ from the the times stated in the frames thread. cr.hp and cr.hk were spot on but nearly everything else I counted had a longer recovery time.
Counting the frames starting from when the red hitbox disappears, until when the wireframe hingepoint bounding boxes resettle back to the idle animation; cr.hp was 25, cr.hk was 21… fine… but cr.lk was 16!!! seems impossible imo. Does anyone have any idea how the original recovery frame data was arrived at?
I want to know, because I wouldn’t mind helping out with some of the as yet unextracted moves.