2

Well, if no connection with internet is a requirement, I would say that the best option could be using Julius with a self trained corpus model. In this case you can use your own language and intonation. As an advantage, the success percentage of the ASR unit will be higher than using a public trained corpus. As a disadvantage, this corpus will be useful just ...


2

You might want to examine different microphone pickup patterns and determine which is best for your application: Most headset microphones are designed using a cardioid or hypercardioid pattern for their noise-cancelling properties. They probably won't be as good for recording sound at a distance of 1 meter, but you can still get cardioid microphones that ...


2

I would consider using a Raspberry Pi. It certainly has the required memory, already has audio output and enough performance to perform image processing with. You can easily interface a simple 2x16 character LCD with it and the camera would not be a problem either. To control the motors you will need PWM, which is available on the PI, for example using this ...


2

Yes, it's possible to tweak the ponetic proncunciation, see the documentation here: tts.setLanguage("English") tts.say("\\toi=lhp\\‘zi.R+o&U \\toi=orth\\") # Same as tts.say("zero")


1

From a practical perspective, usage of Beaglebone Black is amazing (although you may end up with image capture issues). It has two PRUs and quite a lot of GPIOs if hardwares are to be interfaced. If image recognition can be done with costly cameras which support h264 compression format (like Logitech c920), Beaglebone black is the clear 'winner'. Else, ...


1

Speech recognition is a pretty vast problem, if you are simply wanting to command a robot using your voice, a lot of ROS users use Pocket Sphinx. You'll have to teach the program all the words, sentences or phrases you want it to recognize, but there are built-in tools that make this easy. It's also straight forward to teach it multiple phrases that mean the ...


1

This does exist (or will, both eventually and inevitably). The most relevant project to what you are talking about is Eugene who recently sort of passed the Turing test. Update: hopefully non-rotting link with up-to-date information https://en.wikipedia.org/wiki/Eugene_Goostman


1

a couple of years ago I tried Julius and it worked quite well. It is documented here, hope it helps: http://achuwilson.wordpress.com/2012/01/11/speech-recogition-using-julius-in-linux/ and http://achuwilson.wordpress.com/2012/01/11/chippu-speech-recognition/


1

Also for completeness, I did find a Julius model for Brazilian Portuguese, but was rather underwhelmed with its performance – Julius would take several seconds to process any utterance, and usually get what I said hilariously wrong. A second model that promised improved performance for dictation would require me to recompile Julius in order to work, but at ...


1

robotergarten gave some really interesting tips, but in the end I decided to roll my own Google Speech client, using sox to record the user's voice. The ability to work without an Internet connection would be nice but is not really required for my demo, and response time is short enough.


Only top voted, non community-wiki answers of a minimum length are eligible