The invention provides a method of generating a value of a pixel, and apparatus for accomplishing the method. The method includes the steps of (a) providing a floating-point representation of a pixel value; and (b) applying the floating-point representation as an index into a lookup table, the lookup table including a memory device having a plurality of entries. Each of the entries is uniquely indexed by one value of a floating-point representation of a pixel value, and each of the entries stores a corresponding integer representation of the pixel value. A third step (c) outputs from the lookup table the integer representation of the pixel value that corresponds to the applied floating-point representation. It is shown that only a sub-set of the entire floating-point representation need be employed as a table index value. In the presently preferred embodiment of the invention, only 16 bits of a 32-bit single-precision floating-point value are shown to be relevant to the result; specifically the sign bit, the exponent, and a portion of the mantissa.