Lines Matching full:bits
15 // quantize the full eight bits of grayscale to three bits. The simplest way to do
16 // this is to take the top three bits of the grayscale value. However, this does not
19 // of grayscale values, and so when we go from three bits to eight bits, we replicate
20 // the three bits into the lower bits of the eight bit value. Below are two different
39 // top three bits, the function computes the best three-bit value such that its
60 // Moves the top three bits of each byte in the 32-bit argument to the least
61 // significant bits of their respective byte.
