What Are Speech Recognition Systems?

Natural language understanding systems usually accept typed input but for a number of applications this is not acceptable. Spoken language is a more natural from of communication in many human computer interfaces. Speech recognition systems have been available for some time, but their limitation has prevented widespread use. Below are major designs issues in speech systems.  These issues also provide dimensions along which systems can be compared with one another.

Speaker dependence versus speaker independence, a speaker independent system can listen to any speaker and translate the sounds into written text. Speaker independence is hard to achieve because of the wide variations in pitch and accent.

It is easier to build a speaker dependent system, which can be trained on the voice patterns of a single speaker. The system will only work for that one speaker. It can be retrained on another voice but then it will no longer work for the original speaker.
Continuous versus isolated word speech interpreting isolated word speech, in which the speaker pauses between each word, is easier than interpreting continuous speech. This is because boundary effect cause words to be pronounced differently in different contexts.

Real time versus offline processing highly interactive applications requires that a sentence be translated into text as it is being spoken.