A friend pointed me to this patent filing from Apple which shows that the company has big plans for spoken I/O for iPods, iPhones and Apple TV (although I think the iPad must figure into the formula as well). A key concept in the patent is “multitiered approach to speech recognition that takes into account the “context” of a spoken word in order to arrive at the “focus” or the utterance.

Conversely, looking at spoken output, the patent filing describes “speakable strings” which are employed to provide audio feedback associated with the media that is being displayed or played.

The patent filing is clearly a product of inventors in Apple’s labs, so the concepts are presented at a fairly high level. Still it shows how thoughtfully speech will be baked into Apple’s audio user interface.

Categories: Articles

Tags: speech processing, Apple, iPhone, Recombinant Communications