Recording a media element

<p>
  Click the "Start Recording" button to begin video recording for a few seconds.
  You can stop recording by clicking the "Stop Recording" button. The "Download"
  button will download the received data (although it's in a raw, unwrapped form
  that isn't very useful).
</p>
<br />

Let's start by looking at the key bits of the HTML. There's a little more than this, but it's just informational rather than being part of the core operation of the app.

<div class="left">
  <div id="startButton" class="button">Start Recording</div>
  <h2>Preview</h2>
  <video id="preview" width="160" height="120" autoplay muted></video>
</div>

We present our main interface in two columns. On the left is a start button and a <video> element which displays the video preview; this is the video the user's camera sees. Note that the autoplay attribute is used so that as soon as the stream starts to arrive from the camera, it immediately gets displayed, and the muted attribute is specified to ensure that the sound from the user's microphone isn't output to their speakers, causing an ugly feedback loop.

<div class="right">
  <div id="stopButton" class="button">Stop Recording</div>
  <h2>Recording</h2>
  <video id="recording" width="160" height="120" controls></video>
  <a id="downloadButton" class="button">Download</a>
</div>

On the right we see a stop button and the <video> element which will be used to play back the recorded video. Notice that the playback panel doesn't have autoplay set (so the playback doesn't start as soon as media arrives), and it has controls set, which tells it to show the user controls to play, pause, and so forth.

Below the playback element is a button for downloading the recorded video.

<div class="bottom">
  <pre id="log"></pre>
</div>

body {
  font:
    14px "Open Sans",
    "Arial",
    sans-serif;
}

video {
  margin-top: 2px;
  border: 1px solid black;
}

.button {
  cursor: pointer;
  display: block;
  width: 160px;
  border: 1px solid black;
  font-size: 16px;
  text-align: center;
  padding-top: 2px;
  padding-bottom: 4px;
  color: white;
  background-color: darkgreen;
  text-decoration: none;
}

h2 {
  margin-bottom: 4px;
}

.left {
  margin-right: 10px;
  float: left;
  width: 160px;
  padding: 0px;
}

.right {
  margin-left: 10px;
  float: left;
  width: 160px;
  padding: 0px;
}

.bottom {
  clear: both;
  padding-top: 10px;
}

Now let's have a look at the JavaScript code; this is where the majority of the action happens, after all!

We start by establishing some global variables we'll need.

let preview = document.getElementById("preview");
let recording = document.getElementById("recording");
let startButton = document.getElementById("startButton");
let stopButton = document.getElementById("stopButton");
let downloadButton = document.getElementById("downloadButton");
let logElement = document.getElementById("log");

let recordingTimeMS = 5000;

Most of these are references to elements we need to work with. The last, recordingTimeMS, is set to 5000 milliseconds (5 seconds); this specifies the length of the videos we'll record.

Next, we create some utility functions that will get used later.

function log(msg) {
  logElement.innerText += `${msg}\n`;
}

The log() function is used to output text strings to a <div> so we can share information with the user. Not very pretty but it gets the job done for our purposes.

function wait(delayInMS) {
  return new Promise((resolve) => setTimeout(resolve, delayInMS));
}

The wait() function returns a new Promise which resolves once the specified number of milliseconds have elapsed. It works by using an arrow function which calls setTimeout(), specifying the promise's resolution handler as the timeout handler function. That lets us use promise syntax when using timeouts, which can be very handy when chaining promises, as we'll see later.

The startRecording() function handles starting the recording process:

function startRecording(stream, lengthInMS) {
  let recorder = new MediaRecorder(stream);
  let data = [];

  recorder.ondataavailable = (event) => data.push(event.data);
  recorder.start();
  log(`${recorder.state} for ${lengthInMS / 1000} seconds…`);

  let stopped = new Promise((resolve, reject) => {
    recorder.onstop = resolve;
    recorder.onerror = (event) => reject(event.name);
  });

  let recorded = wait(lengthInMS).then(() => {
    if (recorder.state === "recording") {
      recorder.stop();
    }
  });

  return Promise.all([stopped, recorded]).then(() => data);
}

startRecording() takes two input parameters: a MediaStream to record from and the length in milliseconds of the recording to make. We always record no more than the specified number of milliseconds of media, although if the media stops before that time is reached, MediaRecorder automatically stops recording as well.

• We first create the MediaRecorder that will handle recording the input stream.
• data is an array, initially empty, that holds the Blobs of media data provided to our ondataavailable event handler.
• The ondataavailable assignment sets up the handler for the dataavailable event. The received event's data property is a Blob that contains the media data. The event handler pushes the Blob onto the data array.
• We start the recording process by calling recorder.start(), and output a message to the log with the updated state of the recorder and the number of seconds it will be recording.
• We create a new Promise, named stopped, which is resolved when the MediaRecorder's onstop event handler is called, and is rejected if its onerror event handler is called. The rejection handler receives as input the name of the error that occurred.
• We creates another new Promise, named recorded, which is resolved when the specified number of milliseconds have elapsed. Upon resolution, it stops the MediaRecorder if it's recording.
• Finally, we use Promise.all to create a new Promise which is fulfilled when both of the two Promises (stopped and recorded) have resolved. Once that resolves, the array data is returned by startRecording() to its caller.

The stop() function stops the input media:

function stop(stream) {
  stream.getTracks().forEach((track) => track.stop());
}

This works by calling MediaStream.getTracks(), using forEach() to call MediaStreamTrack.stop() on each track in the stream.

Now let's look at the most intricate piece of code in this example: our event handler for clicks on the start button:

startButton.addEventListener(
  "click",
  () => {
    navigator.mediaDevices
      .getUserMedia({
        video: true,
        audio: true,
      })
      .then((stream) => {
        preview.srcObject = stream;
        downloadButton.href = stream;
        preview.captureStream =
          preview.captureStream || preview.mozCaptureStream;
        return new Promise((resolve) => (preview.onplaying = resolve));
      })
      .then(() => startRecording(preview.captureStream(), recordingTimeMS))
      .then((recordedChunks) => {
        let recordedBlob = new Blob(recordedChunks, { type: "video/webm" });
        recording.src = URL.createObjectURL(recordedBlob);
        downloadButton.href = recording.src;
        downloadButton.download = "RecordedVideo.webm";

        log(
          `Successfully recorded ${recordedBlob.size} bytes of ${recordedBlob.type} media.`,
        );
      })
      .catch((error) => {
        if (error.name === "NotFoundError") {
          log("Camera or microphone not found. Can't record.");
        } else {
          log(error);
        }
      });
  },
  false,
);

When a click event occurs, here's what happens:

• MediaDevices.getUserMedia is called to request a new MediaStream that has both video and audio tracks. This is the stream we'll record.

• When the Promise returned by getUserMedia() is resolved, the preview <video> element's srcObject property is set to be the input stream, which causes the video being captured by the user's camera to be displayed in the preview box. Since the <video> element is muted, the audio won't play. The "Download" button's link is then set to refer to the stream as well. Then, we arrange for preview.captureStream() to call preview.mozCaptureStream() so that our code will work on Firefox, on which the HTMLMediaElement.captureStream() method is prefixed. Then a new Promise which resolves when the preview video starts to play is created and returned.

• When the preview video begins to play, we know there's media to record, so we respond by calling the startRecording() function we created earlier, passing in the preview video stream (as the source media to be recorded) and recordingTimeMS as the number of milliseconds of media to record. As mentioned before, startRecording() returns a Promise whose resolution handler is called (receiving as input an array of Blob objects containing the chunks of recorded media data) once recording has completed.

• The recording process's resolution handler receives as input an array of media data Blobs locally known as recordedChunks. The first thing we do is merge the chunks into a single Blob whose MIME type is "video/webm" by taking advantage of the fact that the Blob() constructor concatenates arrays of objects into one object. Then URL.createObjectURL() is used to create a URL that references the blob; this is then made the value of the recorded video playback element's src attribute (so that you can play the video from the blob) as well as the target of the download button's link.

  Then the download button's download attribute is set. While the download attribute can be a Boolean, you can also set it to a string to use as the name for the downloaded file. So by setting the download link's download attribute to "RecordedVideo.webm", we tell the browser that clicking the button should download a file named "RecordedVideo.webm" whose contents are the recorded video.

• The size and type of the recorded media are output to the log area below the two videos and the download button.

• The catch() for all the Promises outputs the error to the logging area by calling our log() function.

The last bit of code adds a handler for the click event on the stop button using addEventListener():

stopButton.addEventListener(
  "click",
  () => {
    stop(preview.srcObject);
  },
  false,
);

This calls the stop() function we covered earlier.

When put all together with the rest of the HTML and the CSS not shown above, it looks and works like this:

You can view the full demo here, and use your browsers developer tools to inspect the page and look at all the code, including the parts hidden above because they aren't critical to the explanation of how the APIs are being used.

• Media Capture and Streams API landing page
• MediaDevices.getUserMedia()