Here's some news: According to a patent filing made available today online, Apple Inc. lays out the way it intends to accomplish speech-to-text for mobile phones, MP3 players (i.e. its iPods) and other PDA devices.
The patent application appears to lay out the method in which Apple is going to implement speech-to-text in its iPhone. The newest 3G S version of the iPhone already has some voice control features, for controlling music and dialing phone contacts. One of the diagrams included in the patent filing gives an example of using the method to create an email.
From the filing:
The speech recognition module can analyze the speech data to derive text data, the text data comprising sequence information associated with each of a plurality of words associated with the speech data. The text composition module can receive the text data and combine the text data with the non-speech data based upon the sequence information. The text composition module can thereby produce combined text data derived from the text data and the non-speech data. The interface can transmit the combined text data to the mobile device for presentation to a user of the mobile device.
It's worth noting that "non-speech data" will include "typeface data, symbols or punctuation" -- meaning you'll be able to control such inputs with your voice. Interesting stuff.
You'll also be able to record speech for later "subsequent processing" into text. And, you'll be able to tag and time-stamp your speech entries. Doubly interesting, no?
So, what it seems Apple is doing is building a fully capable audio-editting interface for speech-to-text. We knew it was coming, no? Thoughts?