Directed Speech Recognition

Some of the most popular kinds of speech recognition software are open-ended — they interpret the full range of speech through audio. There are, however, inherent challenges built into an open-ended speech recognition system. For example, the system has to be able to handle all of the different sounds, which generally requires a large algorithmic lexicon and other resources.

With directed speech recognition, the system only has to interpret from a few different choices. One of the best and most common examples of this is in the interactive voice response (IVR) tools that callers encounter in a call center environment. These tools do not anticipate a full range of speech; they look for simple options like "yes" or "no," or phrases like "talk to a representative" or "find balance."

As a result, directed speech recognition often works better and creates more accurate results for more affordable software packages for these kinds of software products.