Hi,
I tried to combine two Metahuman animations, via baking them to control rig. Both of the animations consist of baked mocap data and manual animation layers. After the bake everything looks fine, and even after fusing the two layers (by setting in and out weight-keyfremes), in viewport everything looks fine, too. But when rendered things get crazy, the Metahumans eyes loose camera contact, the face emotion gets washed out, or is maybe even a few frames off… I rendered it a lot of times and everytime I get a different result, with different errors.
It seems like the current frame on the timebar has an impact on the result, but neither starting at Frame 0 or 1 or in something in the middle, gives something correct.
Do you have any thoughts or solutions?