Being somewhere between a buzzword and a standard computer feature, 'multimedia' has escaped all attempts of definition. The label is attached to any software or hardware trying to achieve symbiotic relationships between two or more of the following monomedia: text, graphics, images, audio and video. In the context of computers, this symbiosis takes place on a computer screen, preferably with the user being able to interact at will. In many current systems, the symbiosis never occurs, and 'interactivity' turns into an endless chain of mouse clicks.

Only recently have computers reached a point where they can claim to handle multimedia. The presentation of audio and video is demanding on the part of the computer for several reasons. First, the amount of memory required to represent an utterance as a digital audio recording is thousands of times larger than the textual representation of the utterance. Adding video further increases demand for computer resources. Secondly, both audio and video are 'continuous' media types, i.e. their playback requires a constant data stream - something a computer often finds hard because of other duties.

One challenging area of multimedia is the transcoding from one medium to another. E.g., a speech synthesizer attempts to transcode from ASCII to human speech, and speech recognition systems go the other way. Transcoding is hard to get right, and some of the theoretically possible conversions are yet far from being realised - e.g. transcoding from a still image to text, or from text to video.