Sample times do not need to be uniformly spaced, but they must be finite and
strictly increasing. Values are flattened in sample order. Translation and
scale channels use three values per sample; rotation channels use four
quaternion values [x, y, z, w] per sample. Morph-weight channels use
valueSize values per sample.
Sampled scalar/vector/quaternion animation data.
Sample times do not need to be uniformly spaced, but they must be finite and strictly increasing. Values are flattened in sample order. Translation and scale channels use three values per sample; rotation channels use four quaternion values
[x, y, z, w]per sample. Morph-weight channels usevalueSizevalues per sample.The default interpolation mode is
LINEAR.