Siri, the Apple (News 
- Alert) iPhone 4S personal assistant, looks to be one of those "They did it again" moments where Apple has taken a couple of basic concepts that have been kicking around for years -- a front voice recognition process feeding into a natural language user interface -- and refined them into a simple, user-friendly tool. But if you aren't getting good voice into the application, you're out of luck. With voice services moving to a hosted/cloud model, HD voice and its associated benefits have become more relevant now than ever before.
It all goes to a simple principle dating back to the very beginnings of computing: garbage in, garbage out. In the case of voice, poor quality voice in means a slower and poorer customer experience when interacting with any voice recognition process, be it a "simple" natural language IVR process, speech-to-text or Siri.
Rolling out HD voice will provide two advantages over PSTN narrowband. The obvious one is providing more voice -- more data -- to a voice recognition process. Narrowband delivers voice in a range of 300 Hz to 3.4 kHz. Baseline HD voice codecs such as G.722 or AMR-WB deliver 50 Hz to 7 kHz, so you get roughly double the sound information, allowing humans and machines to easily distinguish frictive consonants in a conversation. Comprehension is enhanced, so the human doesn't have to repeat or spell out particular words or phrases to a listener... Read More
