> TrueType MUL does (ab)/64, not ab This is how fixed-point arithmetic works. For multiplication of Q26.6 numbers you clear the squared scaling factors (2^6) b

> TrueType MUL does (ab)/64, not ab This is how fixed-point arithmetic works. For multiplication of Q26.6 numbers you clear the squared scaling factors (2^6) by dividing/shifting one of them away.