uSpeech

by arjo129

arjo129 /uSpeech

Speech recognition toolkit for the arduino

441 Stars 104 Forks Last release: almost 7 years ago (v4.1.1) MIT License 207 Commits 6 Releases

Available items

No Items, yet!

The developer of this repository has not created any items for sale yet. Need a bug fixed? Help with integration? A different license? Create a request here:

uSpeech library

WARNING!!! DO NOT USE THIS BRANCH. IT IS FOR EXPERIMENTATION. DOWNLOAD IT FROM THE RELEASES PAGE!!!!

Also I do not have the time to maintain this library and given the price of Raspberry Pis you probably want to use them for speech recognition instead of an arduino.

The uSpeech library provides an interface for voice recognition using the Arduino. It currently produces phonemes, often the library will produce junk phonemes. Please bare with it for the time being. A noise removal function is underway.

Minimum Requirements

The library is quite intensive on the processor. Each sample collection takes about 3.2 milliseconds so pay close attention to the time. The library has been tested on the Arduino Uno (ATMega32). Each signal object uses up 160bytes. No real time scheduler should be used with this.

Features

  • Letter based recognition
  • Small memory footprint
  • Arduino Compatible
  • Up to 80% accuracy with words
  • Novel algorithm based on simple calculus
  • Plugs directly into an
    analogRead()
    port

Documentation

Head over to the wiki and you will find most of the documentation required.

Algorithm

The library utilizes a special algorithm to enable speech detection. First the complexity of the signal is determined by taking the absolute derivative of the signal multiplying it by a fixed point saclar and then dividing it by the absolute integral of the signal. Consonants (other than R,L,N and M) have a value above 40 and vowels have a value below 40. Consonants, they can be divided into frictaves and plosives. Plosives are like p or b whereas frictaves are like s or z. Generally each band of the complexity coeficient (abs derivative over abs integral) can be matched to a small set of frictaves and plosives. The signal determines if it is a plosive or a frictave by watching the length of the utterance (plosives occur over short periods while frictaves over long). Finally the most appropriate character is chosen.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.