After some very focused work, I am happy to show off a working prototype of the VI Terminal I promised in my post "On Interactive Virtual Intelligence".
This prototype of the Virtual Intelligence Terminal implements five separate functional modules:
1. A working Speech Recognition system based on the CMU Sphinx library
2. A simple working Text-to-Speech (TTS) module
3. An animation code animating the termina's Mouth shapes for the text
4. A Serial communication module transmitting the frames to the Arduino and LED Matrix
5. A sync module which coordinates the synchronization between the audible speech and the text displayed on the LED Matrix
Here is a video demonstration:
A couple of notes:
* The speech recognition is currently not ideal, and still needs to be refined (some commands are not recognized, or confused for other words)
* Background noise tends to confuse the speech recognition, so I must manually Mute and Un-Mute the microphone before and after each command
* The synchronization between the spoken and written words displayed on the LED Matrix is not perfect, and needs more work done on it
These kinks will be fixed for the next prototype version update.
In the future you can expect a documentation of the process up to the next point.
All the VI Terminal posts will be titled: Building a Virtual Intelligence Terminal
Kommentare