Sphinx voice recognition + Max-msp + Arduino + Processing

Added Open Sound Control capabilities to Sphinx voice recognition to make it useful as there is no sockets/network facilitation to route recognition data, which makes it very un-useful. Not sure why OSC or any UDP sockets are not used, as this weakens the application for integration into 3rd party apps. I used the JAVAOSC library in Eclipse, video shows how to set it up and proof that voice can control Arduino, Max/msp, Processing or any OSC app. Results are 90% ish, could actually control things like DMX lights, or use in an installation, also a good controller for people with limited ability.
Also this could be implemented with MSDOS/TERMINAL to run apps from command prompt, e.g. open browser, etc and even used with JAVA ROBOT class to highjack keyboard shortcuts, so your voice could trigger F5 to do something etc,

Category:

Howto & Style

Tags:

• Sphinx4
• Arduino voice recognition
• java voice recognition
• max/msp voice recognition

[interactiveartsNI]

You create your own grammar files , look at 2:01 in the video, pin is the word with option on/off, these are sphinx grammar files, so you can make up anything you like to trigger audio, visuals, control itunes, lights etc.

Sphinx is not a continuous voice recognition system like dragon or something commercial, as a lot of control systems just need a few well recognised words.

[carlandj]

I have a team of engineering undergraduates that i want to make a voice control system for lighting a house. Do you have a tutorial for getting the various software systems to communicate, whether video or a write-up. We are all very new to this. I would like to delegate this task to them but they are a little lost. Is max-msp required? Is there an open-source equivalent for starting out.

[interactiveartsNI]

Sorry, i don't have a tutorial, but it would be worth reading around OSC and create sockets for each program, not surprised your students are lost cause they need to be taught by you !, reminds me of my time at University :)

[joaorexinho]

what is the library to processing?

[interactiveartsNI]

sphinx4 library

[TheCheshireCat86]

Song's name?

[interactiveartsNI]

sigur ros, von

[mudithead]

Hey can you please help me out here//

Well, i have a code written in java, whose output is like keyboard and sounds when you press any key, normal pre defined sounds,, which is done in eclipse only... now what i wanna do is send these sound to max msp and add effects or manipulate those sounds through the max msp patch!! How can i do this?

HELP!

[tehuster]

I'm 3rd year student, and i;m very interested in learning how to control max/msp, arduino with my voice. Is there any way you could learn me or give me the Sphinx 4 demo voice recognition via open sound control. Or can you give me any tips where to start. I can find my way around max/msp and my arduino. I'm no profesional but i;m very eiger to learn. Thanks in advance

[dvwood66]

Have you tested in terms of accuracy, with long sentences (read gibberish) that have nothing to do with the "pin off/pin on", and does your software try to recognize what's being said, but still reacts and outputs the "pin off/pin on" at times?

[TheCheshireCat86]

@interactiveartsNI

Thanks a bunch, it's a very nice song.