Merges the video of multiple MediaStreams. Also merges the audio via the WebAudio API.
- Send multiple videos over a single WebRTC MediaConnection
- Hotswap streams without worrying about renegotation or delays
- Crop, scale, and rotate live video
- Add crazy effects through the canvas API
Check out WBS, which uses this package.
npm install video-stream-merger
or
<script src="dist/video-stream-merger.js"></script>Let's first get two media streams. One from the webcam, and another a screen capture.
var getusermedia = require('getusermedia')
var screenRecord = require('screen-record')
getusermedia({video: true, audio:true}, function (err, webcamStream) {
screenRecord(window, function (err, sourceId, constraints) {
getusermedia(constraints, function (err, screenStream) {
// We now have 2 streams: webcamStream, screenStream
})
})
})We want to overlay the webcam stream in the corner of the screen stream.
var VideoStreamMerger = require('video-stream-merger')
var merger = new VideoStreamMerger()
// Add the screen capture. Position it to fill the whole stream (the default)
merger.addStream(screenStream, {
x: 0, // position of the topleft corner
y: 0,
width: merger.width,
height: merger.height,
mute: true // we don't want sound from the screen (if there is any)
})
// Add the webcam stream. Position it on the bottom left and resize it to 100x100.
merger.addStream(webcamStream, {
x: 0,
y: merger.height - 100,
width: 100,
height: 100,
mute: false
})
// Start the merging. Calling this makes the result available to us
merger.start()
// We now have a merged MediaStream!
merger.resultCreate a new video merger.
Optional opts defaults to the below:
{
width: 400, // Width of the output video
height: 300, // Height of the output video
fps: 25, // Video capture frames per second
clearRect: true, // Clear the canvas every frame
audioContext: null, // Supply an external AudioContext (for audio effects)
}
Add a MediaStream to be merged. Use an id string if you only want to provide an effect.
The order that streams are added matters. Streams placed earlier will be behind later streams (use the index option to change this behaviour.)
Optional opts defaults to the below:
{
x: 0, // position of the top-left corner
y: 0,
width: <width of output>, // size to draw the stream
height: <height of output>,
index: 0, // Layer on which to draw the stream (0 is bottom, 1 is above that, and so on)
mute: false, // if true, any audio tracks will not be merged
draw: null, // A custom drawing function (see below)
audioEffect: null // A custom WebAudio effect (see below)
}
Remove a MediaStream from the merging. You may also use the ID of the stream.
If you have added the same MediaStream multiple times, all instances will be removed.
A convenience function to merge a HTML5 MediaElement instead of a MediaStream.
id is a string used to remove or update the index of the stream later.
mediaElement is a playing HTML5 Audio or Video element.
opts are identical to the opts for addStream.
Streams from MediaElements can be removed via merger.removeStream(id).
Update the z-index (draw order) of an already added stream or data object. Identical to the index option.
If you have added the same MediaStream multiple times, all instances will be updated.
Start the merging and create merger.result.
You can call this any time, but you only need to call it once.
You will still be able to add/remove streams and the result stream will automatically update.
The resulting merged MediaStream. Only available after calling merger.start()
Never has more than one Audio and one Video track.
Clean up everything and destroy the result stream.
Get the WebAudio AudioContext being used by the merger.
Get the MediaStreamDestination node that is used by the merger.
This library makes it easy to change streams in a WebRTC connection without needing to renegotiate.
The result MediaStream will appear to be constant and stable, no matter what streams you add/remove!
...
getusermedia({video: true, audio:true}, function (err, webcamStream) {
let merger = new VideoStreamMerger()
merger.start()
players[0].srcObject = merger.result
players[0].play()
var peer1 = new SimplePeer({initiator: true, stream:merger.result})
var peer2 = new SimplePeer()
peer1.on('signal', function (data) {
peer2.signal(data)
})
peer2.on('signal', function (data) {
peer1.signal(data)
})
peer2.on('stream', function (stream) {
players[1].srcObject = stream
})
var clones = []
shareWebCamStream.addEventListener('click', function () {
clones.push(webcamStream.clone())
merger.addStream(clones[clones.length-1])
})
removeWebCamStream.addEventListener('click', function () {
merger.removeStream(clones.pop())
})
})
...If sizing and positioning aren't enough, you can directly draw the video frames by passing a function to the draw option.
merger.addStream(mediaStream, {
draw: function (ctx, frame, done) {
// You can do whatever you want with this canvas context
ctx.drawImage(frame, 0, 0, merger.width, merger.height)
done()
})
})See the bottom example of the Live Demo to see this in action.
You can also take direct control over how audio streams are merged, and apply effects.
merger.addStream(mediaStream, {
audioEffect: function (sourceNode, destinationNode) {
// sourceNode is the input streams audio (microphone, etc)
// destinationNode is the output streams audio
sourceNode.connect(destinationNode) // The default effect, simply merges audio
})
})Both the draw and audioEffect options can be used without any MediaStream at all. Just pass a string instead.