Week 5: March 1st 2023

Working with: p5 Speech, tfjs Speech Command Model, ml5js Sound Classifier

idea:

• verbal sudoku - num, directions
• array for each block
  • L/R inc/dec array index by 1
  • up/down inc/dec array index by number of col in a row
  • highlight current block
• entering number sets array value to that number
  • loop conditional. for every i, if recognized word is num, set array[counter/pos] to that num
• confirm number with yes to enter, erase with no. go to go back to start
• fill green if right
  • array position 1s shouldn’t be the same (array1[0] =! array2[0] =! array3[0])

applications:

• competitive co/op - color coded for multiple voices?
• project onto screen, group sudoku

Start with 3 by 3 for code, 1-3.

Setting up sound classifier:

let soundClassifier;

function preload(){
  let options = { probabilityThreshold: 0.95 };
  soundClassifier= ml5.soundClassifier('SpeechCommands18w', options);
}

function setup() {
  createCanvas(400, 400);
  soundClassifier.classify(gotResults);
 }

function gotResults(error, results) {
  if (error) {
    console.log(error);
    return;
  }
  if(results){
	  console.log(results[0].label, results[0].confidence);
	  // console.log(results)
  }
}

function draw(){
}