In this code we are using the ml5 library to listen 

    to the user commands and try to understand them.

    The model is trained to recognize pre-trained commands, such as,

    the numbers from zero to nine, up, down, left, right, go, stop and

    yes.

    Lets go over each line 🤓   

​

​

   'options' is a variable that will set how sure (confident) we want the 

   machine learning model to be on what is hearing, before responding with

   what it thinks it heard */

let options = {

  probabilityThreshold: 0.9

};

​

/* 'classifier' is a variable that stores the result of calling the

   ml5 sound Classifier with the options we provided.  In this case, the

   'options' we provided is the probabilityThreshols on line 15 and when 

   it all that is processed, it runs a function called 'modelReady'

   this function is on line 24 */

let classifier = ml5.soundClassifier('SpeechCommands18w', options, modelReady);

​

/* this function takes the variable 'classifier' and runs an ml5 function

   called 'classify' passing the 'gotResult' function to it */

function modelReady() {

  // classify sound

  classifier.classify(gotResult);

}

​

/* this function has two parameters, 'error' and 'result'.  The error

   will return an error if one happens, and the result will have the

   result of classifying (identifying a sound) the input from the user */

function gotResult(error, result) {

  /* this if statement is checking if there is an error or not.  If 

     there is an error it will show it in the 'console' and stop. */

  if (error) {

    console.log(error);

    return;

  }

​

  /* if there is not an error the following code will run

     set the colour of the background, colour of the text and

     the size of the text */

  background(42);

  textSize(50);

  text('🎙', 175, 100);

  fill('white');

  textSize(18);

​

  /* show on the canvas the text between quotes '', add to it

     the content of the array 'result' at position 0, and from that 

     information extract the content of a 'variable' called 'label' */

  text('🦻 The computer heard: ' + result[0].label, 20, 190);

​

  /* show on the canvas the text between quotes '', add to it

     the content of the array 'result' at position 0, and from that 

     information extract the content of a 'variable' called 'confidence' */

  text('🤔 The confidence is: ' + result[0].confidence, 20, 240);

}

​

function setup() {

  createCanvas(400, 400);

}

​

/*  Play with the probabilityThreshold on line 16 and see how it

    affects the results on the canvas */