What are the advantages and disadvantages of using continuous-speech recognition systems?
- speech is a very natural way of interacting and provides that appeal to a user (instead of typing or using a remote control to execute an action),
- the mostly self-explanatory method of handling a device utilizing a speech system (by simply speaking into a system while it records one’s voice), and
- the usefulness in its flexibility of catering to many different languages and dialects for its users and even for those who can phonetically speak but cannot spell (hence why a keyboard is not really required for speech systems)
- the low signal-to-noise ratio (a program must hear a person speak clearly with little to no noise in the background or else, it won’t accurately decipher a user’s speech),
- overlapping speech (some programs may have trouble differentiating between one user’s voice and another’s if more than one person is speaking at once), and
- trouble discerning homophones within vocabulary being said (speech-recognition systems have no way of telling the difference unless context is given between words such as there / they’re / their and raise / raze / rays because they all sound the same and can make for a troubling task for both the user to say what they mean and the system to process what’s it thinks it hears being said).