The trick, whether it is instructing a television set, or dictating the "Great American Novel" is not so much in recognizing a limited command set as it is in being able to correctly interpret the various regional, racial, ethnic, native language accents, pronunciation, intonations, voice patterns, etc//i]. Even gender can make a difference. AFIK Dragon Naturally Speaking is about the only consumer grade VR software that runs on a Mac or PC and it costs anywhere from $150 to $999 depending on the complexity of the vocabulary.
I understand the problem with Siri, which you're liable to ask virtually anything, but I was thinking that the command set for a TV is limited enough that minimal, built-in, multi-user software with
([i]) "
n" commands, and "
this" is how each user speaks each one, would make remote access, even interpretation, unnecessary.
