nmul

// normalized byte multiplication:
// range [0, 255] maps to [0.0, 1.0]
// examples:
// 0xff * 0xff = 0xff   (255 * 255 = 255)
// 0xff * 0x7f = 0x7f   (255 * 127 = 127)
// 0x7f * 0x7f = 0x3f   (127 * 127 =  63)
// 0x01 * 0xff = 0x01   (  1 * 255 =   1)
// 0x01 * 0x7f = 0x00   (  1 * 127 =   0)

inline byte nmul(byte a, byte b)
{
    return (((uint16)a + 1) * b) >> 8;
}