The limit on precision wasn't the size of the lookup table, it was the size of the integers. The CPU's integers were only 8-bits. So if you're using native in

The limit on precision wasn't the size of the lookup table, it was the size of the integers. The CPU's integers were only 8-bits. So if you're using native integers (i.e.: [0,256) == [0°,360°) ), you can't get angles more precise than 1.4°.