Interpolation issues and wishes

[Morgan Turner]

The ability to select a only subset of frames to interpolate could be useful. Some gaps in tracking (like when markers are out of frame) don't need to be interpolated. Thanks!

[Benjamin Knorlein]

Should interpolated points be overridden when tracked?

[Morgan Turner]

For our uses, spline interpolation is being used to either fill in gaps when markers are present and have difficulty tracking, or when we have no markers and therefore can't track markers. Once issues or gaps in tracking have been identified, we spline. Given that workflow, we wouldn’t be using tracking after splining, and we wouldn’t want the ability for tracking to overwrite interpolating.

I think ultimately this comes back around to assigning a frame range to a certain type of tracking.

Chatting with Steve, I think what would solve many of our problems is simply being able to assign a range of frames (or multiple ranges) as ‘no data’. Tracking can’t continue through it; splining wouldn’t span across it. Like our alligator on a treadmill example, markers cyclically go off screen and aren't part of a rigid body, and we need a way to keep those frames empty.

[Benjamin Knorlein]

fixes #60 : Reworked interpolation : Interpolation can be assigned per frame, interpolation method is drawn in the status, interpolations are executed after each other, e.g. first repeat then linear, then cubic. Changed cubic spline interpolation to a method which takes irregular spacing of datapoints into account.

→ <<cset 833689b2485b>>