- Posts: 34
- Joined: Sun Feb 04, 2024 12:01 am
- Real Name: Steve Wellens
Technically frame numbers in comp almost always have to be handled as float or double "for reasons", because there are multiple situations where time samples have to be non-integer. Transform motion blur for example, where transform time samples are taken around current frame time and as such all parameter curves have to be evaluated at those fractional times too. Even with float precision comp can run into issues if frame numbers are high (mapping from timecode to frame number can produce these easily), motion blur samples won't have enough granularity anymore and produce stair-stepped look.
I think this is more of an UI issue because if user can't set frames to be fractional values, they shouldn't be presented as such either.