By skinned mesh tortured for a long time, began to feel skinindices wrong, because the PIX display is full of 0
And then followed to find out that the uchar integers were shown in the format of float by the PIX (shown as 0.0000000 instead of the life of the uint shown as 0)
Turns into a very small number, so it's 0.00000.
So Skinindices is right, and weight looks right, because adding up equals 1.
Oh, yes, Phyre's d3d11 to put shader inside the semantic
Skinindices:blendindices;
Skinweights:blendweights;
To write like this, you can't just write a color0 that looks like it was determined by the sementic behind it.
CHR's skin movement looks like a spider, and does not stretch to infinity, nor very chaotic, the explanation is still relatively correct, just where. For example, no unormalize, or the range is enlarged.
Skinned mesh Spider-like