One thing I have never understood, why do we have a field for signed_value? For the most part that seems to relate to if there is a negative integer in either R, G or B. I thought that basically your min and max RGB was 0,0,0 and 255,255,255?

The table is also used for the adjustments made to the face like ears, nose, cheek, etc. Those values seem to be from -1.0 to 1.0, so the signed value just lets us know how we need to load the next 3 values either as an unsigned int8 or a float.

Ah, I see. That makes way more sense now that I am looking at it. I will just mask out my query with a where statement then. You are a gentlemen and a scholar.