ml5/demos/06-sound-classifier-vanilla-js.html

<!DOCTYPE html>
<html lang="en" class="loading">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  <title>Sound Classifier</title>
  <link rel="stylesheet" href="css/normalize.css">
  <link rel="stylesheet" href="css/style.css">
  <style>
    .hidden {
      display: none;
    }
    .loading .display-loading {
      display: block;
    }
    .running .display-running {
      display: block;
    }
    #results {
      background-color: white;
      width: 20rem;
      min-height: 6rem;
      display: none;
      justify-content: center;
      align-items: center;
      font-weight: bold;
      font-size: 2rem;
    }
    .running #results {
      display: flex;
    }
  </style>
</head>
<body>
  <h1 class="h1">Sound Classifier</h1>
  <h2 id="state"></h2>
  <div id="results" class="fancy-shadow hidden display-running"></div>
  <script src="https://unpkg.com/ml5@1/dist/ml5.js"></script>
  <script>

    const $results = document.querySelector("#results");
    const $state = document.querySelector('#state');
    let classifier;
    let classification = [];
    const words = [
      "zero",
      "one",
      "two",
      "three",
      "four",
      "five",
      "six",
      "seven",
      "eight",
      "nine",
      "up",
      "down",
      "left",
      "right",
      "go",
      "stop",
      "yes",
      "no",
    ];

    const STATE_LOADING = "loading";
    const STATE_RUNNING = "running";
    const ALL_STATES = [STATE_LOADING, STATE_RUNNING];
    let state = STATE_LOADING;

    const setState = (value) => {
      console.log('setState', value);
      state = value;
      $state.textContent = state;
      document.documentElement.classList.remove(...ALL_STATES);
      document.documentElement.classList.add(state);
    };

    const preload = async () => {
      setState(STATE_LOADING);
      requestAnimationFrame(draw);
      const options = { probabilityThreshold: 0.7 };
      classifier = ml5.soundClassifier("SpeechCommands18w", options);
      await classifier.ready;
      console.log('model ready');
      setup();
    }

    const setup = async () => {
      console.log('setup');
      // no need for video stream
      // classifier contains its own
      // logic to set up audio
      // start classification
      classifier.classifyStart((results) => {
        // store in global
        classification = results;
      });
      // start the app
      setState(STATE_RUNNING);
    }

    const draw = () => {
      if (state === STATE_RUNNING) {
        let predictedWord = "no word detected";
        if (classification[0]?.confidence > 0.7) {
          predictedWord = classification[0].label;
        }
        $results.textContent = predictedWord;
      }
      requestAnimationFrame(draw);
    }

    preload();
  </script>
</body>
</html>