Speech recognition systems began to be developed in the 1950s and 1960s. Computer programs that can recognize the sound of a person’s voice and identify words typically require large memory reserves. Phone speech recognition in the 21st century is included on many mobile phones that have multiple gigabytes of memory. The technology can be used for voice recognition of the phone’s user, to dictate text on a memo or email, or for searching the Web. Some speech recognition programs also enable text messages to be written based on verbal commands.
There is a computer language designed exclusively for using voice recognition to access the Internet on a mobile phone. Users typically do not have to follow commands on the screen, or type on a keypad. Devices that can do this are generally more expensive than most mobile phones. Voice browsers also enable phone speech recognition, and let people go online with a telephone. These can be useful for Web-based call centers, integrated phone systems in automobiles, and even applications in conventional Web browsers.
Speech translation is another type of phone speech recognition. Independent software and Web-based applications are available for translating words in one language to another while a person is speaking. In addition to programs that support text messaging by verbal commands, there are applications for using speech recognition to write and send email.
Phone speech recognition is often used for searching the Web as well. This can be useful for checking movie listings, or with helping a visually impaired person to search the Web on his or her phone. Speech recognition for phones is used in many cases to control device functions. Voice commands can be programmed by the user, by recording his or her voice when prompted during the setup phase. Numbers can be automatically dialed when the appropriate person’s name is spoken into the phone.
The use of phone speech recognition is not only enabled by more memory, but by a connection to servers that store voice data. These servers can be programmed to learn various tendencies in language, and collect data to allow for relatively accurate word recognition. Some concerns with phone speech recognition include interference from background noise. Predictable patterns, however, can be compensated for by some programs used along with recognition software. Voice user interfaces based on the Web are advantageous because they are often accessible from anywhere, just as many smart phone features and services are.