A data encoding process that allows the representation of very large numbers with fewer bits. Used in computers that have floating point registers, and where complex graphics, scientific or audio processes need to be executed rapidly. As far as audio is concerned, a debate is presently centered around whether floating point or fixed-point computation is better for the end product, in terms of sonics. So far there does not seem to be a consensus on this issue, since there are advantages and disadvantages to both systems.
(See Bit Depth)