A data reduction technique which takes advantage of the fact that a loud sound affects the perception of quieter signals both before and after it. If, for example, a relatively quiet signal occurs 10-20ms before a louder one, it may still be masked by the louder signal. This is called backwards masking. The hearing mechanism also takes time to recover from a relatively louder sound, and this creates a masking effect which extends up to 100-200ms after the masking signal signal has ceased, called forward masking. The length of the masking is related to the relative amplitude of the masking signal. To reduce data by using this technique, the input signal is divided up into blocks of samples usually around 10ms in length, and each block is analyzed for transients which act as temporal maskers. Most systems vary the length of the block to take advantage of both backwards and forwards masking. Also called time-masking. See perceptual coding.