no need to parse any glsl, opengl has an introspection api for this exact purpose. see glGetProgramiv and glGetActiveUniform
I moved into my new house, fun and all getting some mi lights and hooking up my raspberry pi up with domoticz, opentherm and maybe some sonoff stuff. Now I wanted to do voice recognition and the google cloud speech to text api is amazing :
https://cloud.google.com/speech-to-text/
It works so good for my language (dutch), only streaming to it 24/7 isn't going to work due to costs so I only want it to listen to me when I want it to. Obviously a button connected to the wifi with a ESP Module or maybe a simple RF transceiver will do the trick. But I want it to go full voice control, I thought what if I program a specific application that only focuses on recognizing my voice (or similar doens't really matter) and a single word like computer, or Patrick or James or whatever. After it hears it it triggers the google speech to text api. If it only has to recognize that specific instance with a specific voice it should be doable right?
Anyone has some idea's / resources on how to do this? Maybe some audio samples with FFT or a simple neural net? Oke wel neural nets aren't simple but anyways idea's ?
Just got my first paycheck and now I don't care about how much I hate the technology here
Have you seen Snowboy?
Alright who's going to compile the highlights?
Sorry, you need to Log In to post a reply to this thread.
What are you working on? v67 - March 2017
