# Causality

A system is **causal** if the output value at a particular time depends only on the input values at that time or before. For LTI systems, if we examine the convolution sum,

,

then it must be true that *h*(*m*) = 0 for all *m* < 0. Were this not true, then there would be non-zero terms in the summation above where *m* < 0, and those terms would involve a future sample of the input, *x*(*n −
m*). Conversely, it is easy to see that if *h*(*m*) = 0 for all *m* < 0, then the system is causal.

Causality is an important practical property of a system that receives its data in **real time** (physical time). Such systems cannot possibly look ahead in time, at least not until someone invents a time machine. However, there are many situations where causality is irrelevant. A system that processes stored data, such as digital audio off a compact disk or audio files in a computer, has no difficulty looking ahead in "time".