This is my first cloud-robotics application and although there are many possible improvements, it is immediately useful and show the power of cloud-robotics.
I created a behaviour for Nao that used the same speech-to-text services found in Android devices and google chrome to translate spoken words in to text. It works better than the standard speech recognition engine and can be used for many more things. I’m eager to see what other Nao developers do with it.
See a quick demo of the speech recognition:
Nao developers can find the code here.
For those who want to try it on their (Linux) computers, here you have a command that will record 5s of sound, encode it in Flac format and send it to google. Then it will right Google’s response in a txt file.
arecord -f cd -t wav -d 5 -r 16000 | flac – -f –best –sample-rate 16000 -o out.flac; wget –post-file out.flac –header=”Content-Type: audio/x-flac; rate=16000″ -O speech.txt http://www.google.com/speech-api/v1/recognize?lang=en
I have to thank Mike Pultz to be the inspiration for this.