Dave Long writes: > Just for completeness, from another Kragen thread: > > > From John Canny's "The Future of Human-Computer Interaction" in the > > July/August ACM Queue, some optimistic notes about current speech > > recognition. The URL is > > http://www.acmqueue.com/modules.php?=20 > > name=3DContent&pa=3Dprinter_friendly&pid=3D402&page=3D1 > > even though VUIs don't seem relevant for a cheap device.
Assuming for the sake of argument that a voice user interface has the same quality as a graphical or visible-text UI, VUIs are extremely relevant for cheap devices: everywhere I've seen so far, an LCD big enough to display more than a couple of words on costs $20, and you can get eleven 70MHz 32-bit LPC2101 processors for that price. There's good voice-recognition software that supposedly needs about 15MIPS (ViaVoice) to run in real-time, and you may not even need real-time. I don't know how much bandwidth to how much memory speech-recognition needs, though. Speech synthesis is even CPU-cheaper than speech recognition. Audio I/O hardware (earphones, speakers, microphones, ADCs) is almost free compared to graphical LCD panels.
