I'm trying to remove a track from a MediaStream. MediaStream.removeTrack() removes the track from the stream, but the camera light is left on indicating that the camera is still active.
https://developer.mozilla.org/en-US/docs/Web/API/MediaStreamTrack?redirectlocale=en-US&redirectslug=DOM%2FMediaStreamTrack
This references a stop() method which I suppose would stop the camera completely, In chrome however I get "Object MediaStreamTrack has no method 'stop'"
Is there a way around this or do I have to stop the whole stream and then recreate it with the tracks I don't want gone? As an example, I want to remove the video track while the audiotrack is still there.
MediaStreamTrack.stop() is now added to the Chrome.
MediaStream.stop() is deprecated in Chrome 45.
You should use MediaStream.getVideoTracks() to get video tracks and stop the track using MediaStreamTrack.stop()
for stopping specific media stream, Maybe this help: (Link)
function stopStreamedVideo(videoElem) {
const stream = videoElem.srcObject;
const tracks = stream.getTracks();
tracks.forEach(function(track) {
track.stop();
});
videoElem.srcObject = null;
}
You need to call stop() on the MediaStream, not a MediaStreamTrack.
Take a look at simpl.info/gum. From the console, call stream.stop(): recording stops and the video camera light goes off.
It looks like the proper way to deal with this issue is to stop your MediaStream, recreate (and reattach) it as an audio-only one and then renegotiate the PeerConnection session. Unfortunately, Firefox currently doesn't support renegotiation mid-session.
The only viable hack is thus to also recreate the PeerConnection with the new MediaStream as suggested here (see "Adding video mid-call").
Related
Unfortunately, from time to time when making a one-on-one video call using react-native-webrtc one of the two video streams freezes or becomes black. Is there a way to detect when that happens programmatically? Thx in advance!
It looks like each video track has a listener that fires as soon as the stream freezes.
In react native it's the onmute listener:
stream.getVideoTracks().forEach(videoTrack => {
videoTrack.onmute = () => {
console.log("Frozen video stream detected!");
};
});
Note that in React Native detecting frozen streams with this method only seems to work for remote tracks!
To detect if a stream is currently frozen I use the muted property on the video track:
console.log(videoTrack.muted); // true when frozen
Another way I've found but haven't explored further is the getStats() method on the RTCPeerConnection. It returns a promise with a huge amount of data that can be used to detect frozen video streams and a lot more I suppose.
I am creating a demo application in which user can record a video using ReactJs. I am able get the list of devices, and record the video.
I am stuck and couldn't find the solution to let the user switch the camera while recording is on.
as of now this is how i am getting the video and audio Stream
function getStream() {
const stream = navigator.mediaDevices.getUserMedia({audio: {deviceId: 'audioDeviceId'}, video: {deviceId : 'videoDeviceId'})
}
by using above I am able to get the stream for the selected devices when Recording is not started.
The problem is when user switched audio or video device while recording.. i get a new Stream every time, so I am losing the previous stream and only the newly generated stream is getting recorded .
I will really appreciate if any one can help me out in handling the scenario where user can switch the devices while the recording is on. or if there is any other approach for switching the devices .
Thanks for the help anyway
Did you try and add/remove tracks? I stumbled upon a similar situation. In my scenario, If the user switches to a different video device, I get the new stream and from it, extract the video track. I then remove the older video tracks from the older stream and add this newly gotten track. If this makes sense, I can post my code snippet here.
Since Apple disabled the ability to autoplay audio via HTMLMediaElement.play()
in javascript without user interaction, I am not sure how I should play a sound when a user gets a chat message before interacting with the DOM after the page loads.
socket.on("receive message", data => {
const receiveSound = new Audio("1.mp3");
messages.push(data);
receiveSound.play();
});
I tried playing the audio element on a mousemove event. I also tried to fake a click() through an element on a React ref to initially activate it. Neither solutions worked.
Is there a way to autoplay an audio element if there is a message coming in? It must be possible since YouTube can autoplay videos without interaction.
Every time I try to play the audio, I get this error:
Unhandled Rejection (NotAllowedError): The request is not allowed by the user agent or the platform in the current context, possibly because the user denied permission.
There is a way that you can make Safari play a sound without necessarily having to allow it there in the site settings.
According to the webkit documentation, it makes explicit that the user needs to interact to play the audio/video, in this case by the click, but he does not say that after you play some audio he lets you play any other audio then without any problem. With this in mind, you can for example run some script on your index.html where it will play your notification audio without any volume and then you can run it again without any problem, as in the example below:
function unlockAudio() {
const sound = new Audio('path/to/your/sound/notification.mp3');
sound.play();
sound.pause();
sound.currentTime = 0;
document.body.removeEventListener('click', unlockAudio)
document.body.removeEventListener('touchstart', unlockAudio)
}
document.body.addEventListener('click', unlockAudio);
document.body.addEventListener('touchstart', unlockAudio);
To run your code after this workaround, just do it this way:
function soundNotification() {
const sound = new Audio('path/to/your/sound/notification.mp3');
const promise = sound.play();
if (promise !== undefined) {
promise.then(() => {}).catch(error => console.error);
}
}
Remembering that the above example is just to show you an alternative, there are several ways you can solve this problem, just keep in mind that you will need to play some sound before...
What solved it for me was using HowlerJS
from Docs:
howler.js is an audio library for the modern web. It defaults to Web
Audio API and falls back to HTML5 Audio. This makes working with audio
in JavaScript easy and reliable across all platforms.
I have this in the window load listener function.
var audio = new Audio();
audio.src = "assets/silence.mp3";
audio.load();
document.getElementById("body").addEventListener(
'touchstart',
function(evt){ audio.play(); audio = 0; },
{capture:false,once:true,passive:true}
);
In chrome on android (with remote debugging open), I touch the screen and it (correctly) triggers trying to play the audio. However, it fails and logs this to the console:
(index):65 Uncaught (in promise) DOMException: play() failed because the user didn't interact with the document first.
How else am I supposed to trigger audio in a way that ensures the user initiates it, if not through a user interaction event listener?
aha. turns out, touchstart doesn't count as a "user interaction" from the perspective of the mechanisms preventing auto-play.
solution: swap it with click.
one thing I'm still frustrated with:
I'm using this specific snippet as a means of very quickly allowing audio to be played on the page so that the underlying html5 game can play audio without worrying about it being blocked by the browser.
however- because it needs to wait for a full "click" event (as opposed to a "touchstart" event), any audio that should be triggered on "touchstart" is going to be passed over (for the first time it would otherwise be triggered).
it seems like 1. there's no way to get around this to use it how I'd like, and 2. if you're able to just listen to the whole document for a "click" anyways, it's not actually doing anything from stopping unwanted audio.
If anyone can come up with a better solution, I'll swap the "correct answer" to that.
I had the same problem when I was making a web piano. The browser charged the piano on load, so you could press(actually touch the screen) the piano keys before you click something. What I did was to add a starting screen with a button to start the game and this solved it because you had to click something before you touch the piano keys.
I am currently working with webRTC and therefore use the getUserMedia() function. When doing this a red icon appears inside the tab-header. Is there a way to remove this symbol? I set my microphone stream to undefined after a certain task, but the symbol is still shown there.
You have to stop the MediaStream (MediaStreamTracks) you've acquired with getUserMedia
stream.getTracks().forEach(track => track.stop());
stream.stop(); // deprecated in FFox