Reputation: 53
I'm trying to take a snapshot from my webcam with the navigator.mediaDevices.getUserMedia()
and canvas.getContext('2d').drawImage()
functions.
When I do it like this, it works perfectly:
function init(){
myVideo = document.getElementById("myVideo")
myCanvas = document.getElementById("myCanvas");
videoWidth = myCanvas.width;
videoHeight = myCanvas.height;
startVideoStream();
}
function startVideoStream(){
navigator.mediaDevices.getUserMedia({audio: false, video: { width: videoWidth, height: videoHeight }}).then(function(stream) {
myVideo.src = URL.createObjectURL(stream);
}).catch(function(err) {
console.log("Unable to get video stream: " + err);
});
}
function snapshot(){
myCanvas.getContext('2d').drawImage(myVideo, 0, 0, videoWidth, videoHeight);
}
<!DOCTYPE html>
<html>
<head>
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<script src="debug.js"></script>
</head>
<body onload="init()">
<div id="mainContainer">
<video id="myVideo" width="640" height="480" autoplay style="display: inline;"></video>
<canvas id="myCanvas" width="640" height="480" style="display: inline;"></canvas>
<input type="button" id="snapshotButton" value="Snapshot" onclick="snapshot()"/>
</div>
</body>
</html>
The thing is, I don't want to use a button click for taking the snapshot, but take the snapshot as soon as the the camera stream is loaded. I tried calling the snapshot() function directly after setting the video source:
function init(){
myVideo = document.getElementById("myVideo")
myCanvas = document.getElementById("myCanvas");
videoWidth = myCanvas.width;
videoHeight = myCanvas.height;
startVideoStream();
}
function startVideoStream(){
navigator.mediaDevices.getUserMedia({audio: false, video: { width: videoWidth, height: videoHeight }}).then(function(stream) {
myVideo.src = URL.createObjectURL(stream);
snapshot();
}).catch(function(err) {
console.log("Unable to get video stream: " + err);
});
}
function snapshot(){
myCanvas.getContext('2d').drawImage(myVideo, 0, 0, videoWidth, videoHeight);
}
But it doesn't work. My canvas stays white. I guess it's because the the camera stream is not fully loaded at this point.
So is there any other event getting fired, which I could use for drawing the snapshot as soon as the camera feed is loaded? Or am I totally on the wrong track?
Thanks in advance!
Upvotes: 5
Views: 3149
Reputation: 42480
Wait for the loadedmetadata
event:
navigator.mediaDevices.getUserMedia({video: true})
.then(stream => {
video.srcObject = stream;
return new Promise(resolve => video.onloadedmetadata = resolve);
})
.then(() => canvas.getContext('2d').drawImage(video, 0, 0, 160, 120))
.catch(e => console.log(e));
<video id="video" width="160" height="120" autoplay></video>
<canvas id="canvas" width="160" height="120"></canvas>
The above should work in all browsers (that do WebRTC).
In Chrome you can also do this - but play()
doesn't return a promise in any other browser yet.
Also note that URL.createObjectURL(stream)
is deprecated. Use srcObject
.
Update: Thanks to @KyleMcDonald in comments for pointing out the importance of registering the loadedmetadata
listener synchronously with setting the srcObject
!—Code updated.
Upvotes: 5