Quote:
Originally Posted by knc1
[...] If you don't mind talking to your Kindle - this might be another choice. (I don't think VR can be done on a DX - not enough horsepower - GM might have an opinion on that - getting fast response out of a Kindle is one of his "things")
|
Of course Voice Recognition can be done on a DX. It has an ARM. That is WAY more powerful than the stuff we had when I was a kid (i.e. relays and vacuum tubes). And even THEN, folks did "voice feature extraction" (early form of voice recognition) with vibrating glass fibers cut to various lengths and a simple analog "perceptron" circuit to "recognize" phonemes and such from the fluctuating light focused on photocells (after passing through a "light mask" gate). People got more "physical" back in the day, before ASICs and "really fast" software (with multicore pipelining and multilevel cache and such) were all the rage. It is mostly about imagination (and necessity) -- folks are trained to give up much too early, before their favorite TV program starts...
I have a long history of doing things even the "experts" claimed was "impossible" (even after they saw my code actually working) on the constrained hardware it was running on. "
Clark's Third Law" in action. In fact, to me, "impossible" was just a challenge to think "outside the box" where there always seemed to be plenty of "obvious" solutions just laying around out in the open where others did not bother looking...
If not obvious, any old ARM processor (like the one in a DX) beats an Arduino all to hell and back, and yet even that can do this:
AN ARDUINO WITH BETTER SPEECH RECOGNITION THAN SIRI
And an Arduino has tiny firmware storage capacity and a miniscule amount of RAM (compared to a DX). So, Voice Recognition on a DX, hell yeah!
The problem for me is TIME. I need to start doing things that pay "real money", to keep pesky "bill collectors" from bothering me...