Yeah. I figured that one out :) Do you know of any particular reason that googles speech api would refuse to work with the microphone in headless mode when other microphone functions work? For instance, if I go here:
https://www.google.com/intl/en/chrome/demos/speech.html in headless mode and allow the microphone using --use-fake-ui-for-media-stream recognition never starts...it behaves as though the mic is not enabled. But if I run the same setup and head to
https://watson-speech.mybluemix.net/microphone-streaming-text-to-console.html mic recognition begins immediately. So I think I've narrowed it down to just googles speech api not liking headless. I can't find any specific info on this though.