I\'m working on an audio library in which the sine of a number needs to be calculated within a very tight loop. Various levels of inaccuracy in the results might be tolerabl