May show Apple interest in iOS speech recognition

Apple has filed for a patent on a text-to-speech and speech-to-text conversion system, designed to work within noisy surroundings, documents show. Handling the conversion would be an iPhone's logic board; the technique stands in contrast to a common one, used by companies like Nuance, in which conversion is processed through a remote service. The Apple system would be able to detect ambient noise using a microphone, and prompt a person answering a call to use text-to-speech instead of their own voice.

Alternately conversion could be used in a two-way texting scenario. As one person's text is converted to speech, they might also see the other person's voice translated to text, resulting in a text-based conversation on one end and a vocal exchange on the other. Although not ideal, the scheme would help to deal with calls in loud venues, or potentially serve as an aid for the disabled.

The patent application becomes more significant in the light of rumors of an Apple deal with Nuance. An arrangement could be announced in early June at WWDC, and the result may rely on Apple's new North Carolina data center. In contrast with the patent, though, there may be at least some remote processing of speech, and the purpose of the partnership might include general voice commands in iOS 5, alongside (or in place of) any communications enhancements.

by MacNN Staff