7.4 KiB
webrtcsink
All-batteries included GStreamer WebRTC producer, that tries its best to do The Right Thing™.
Use case
The webrtcbin element in GStreamer is extremely flexible and powerful, but using
it can be a difficult exercise. When all you want to do is serve a fixed set of streams
to any number of consumers, webrtcsink
(which wraps webrtcbin
internally) can be a
useful alternative.
Features
webrtcsink
implements the following features:
-
Built-in signaller: when using the default signalling server, this element will perform signalling without requiring application interaction. This makes it usable directly from
gst-launch
. -
Application-provided signalling:
webrtcsink
can be instantiated by an application with a custom signaller. That signaller must be a GObject, and must implement theSignallable
interface as defined here. The default signaller can be used as an example.An example project is also available to use as a boilerplate for implementing and using a custom signaller.
-
Sandboxed consumers: when a consumer is added, its encoder / payloader / webrtcbin elements run in a separately managed pipeline. This provides a certain level of sandboxing, as opposed to having those elements running inside the element itself.
It is important to note that at this moment, encoding is not shared between consumers. While this is not on the roadmap at the moment, nothing in the design prevents implementing this optimization.
-
Congestion control: the element leverages transport-wide congestion control feedback messages in order to adapt the bitrate of individual consumers' video encoders to the available bandwidth.
-
Configuration: the level of user control over the element is slowly expanding, consult
gst-inspect-1.0
for more information on the available properties and signals. -
Packet loss mitigation: webrtcsink now supports sending protection packets for Forward Error Correction, modulating the amount as a function of the available bandwidth, and can honor retransmission requests. Both features can be disabled via properties.
It is important to note that full control over the individual elements used by
webrtcsink
is not on the roadmap, as it will act as a black box in that respect,
for example webrtcsink
wants to reserve control over the bitrate for congestion
control.
A signal is now available however for the application to provide the initial
configuration for the encoders webrtcsink
instantiates.
If more granular control is required, applications should use webrtcbin
directly,
webrtcsink
will focus on trying to just do the right thing, although it might
expose more interfaces to guide and tune the heuristics it employs.
Building
Make sure to install the development packages for some codec libraries beforehand, such as libx264, libvpx and libopusenc, exact names depend on your distribution.
cargo build
Usage
Open three terminals. In the first, run:
WEBRTCSINK_SIGNALLING_SERVER_LOG=debug cargo run --bin server
In the second, run:
python3 -m http.server -d www/
In the third, run:
export GST_PLUGIN_PATH=$PWD/target/debug:$GST_PLUGIN_PATH
gst-launch-1.0 webrtcsink name=ws videotestsrc ! ws. audiotestsrc ! ws.
When the pipeline above is running successfully, open a browser and point it to the http server:
gio open http://127.0.0.1:8000
You should see an identifier listed in the left-hand panel, click on it. You should see a test video stream, and hear a test tone.
Configuration
The element itself can be configured through its properties, see
gst-inspect-1.0 webrtcsink
for more information about that, in addition the
default signaller also exposes properties for configuring it, in
particular setting the signalling server address, those properties
can be accessed through the gst::ChildProxy
interface, for example
with gst-launch:
gst-launch-1.0 webrtcsink signaller::address="ws://127.0.0.1:8443" ..
Enable 'navigation' a.k.a user interactivity with the content
webrtcsink
implements the GstNavigation
interface which allows interacting
with the content, for example move with your mouse, entering keys with the
keyboard, etc... On top of that a WebRTCDataChannel
based protocol has been
implemented and can be activated with the enable-data-channel-navigation=true
property. The demo implements the protocol and you can easily test this
feature, using the wpesrc
for example.
As an example, the following pipeline allows you to navigate the GStreamer documentation inside the video running within your web browser (in http://127.0.0.1:8000 if you followed previous steps of that readme):
gst-launch-1.0 wpesrc location=https://gstreamer.freedesktop.org/documentation/ ! webrtcsink enable-data-channel-navigation=true
Testing congestion control
For the purpose of testing congestion in a reproducible manner, a simple tool has been used, I only used it on Linux but it is documented as usable on MacOS too. I had to run the client browser on a separate machine on my local network for congestion to actually be applied, I didn't look into why that was necessary.
My testing procedure was:
-
identify the server machine network interface (eg with
ifconfig
on Linux) -
identify the client machine IP address (eg with
ifconfig
on Linux) -
start the various services as explained in the Usage section (use
GST_DEBUG=webrtcsink:7
to get detailed logs about congestion control) -
start playback in the client browser
-
Run a
comcast
command on the server machine, for instance:$HOME/go/bin/comcast --device=$SERVER_INTERFACE --target-bw 3000 --target-addr=$CLIENT_IP --target-port=1:65535 --target-proto=udp
-
Observe the bitrate sharply decreasing, playback should slow down briefly then catch back up
-
Remove the bandwidth limitation, and observe the bitrate eventually increasing back to a maximum:
$HOME/go/bin/comcast --device=$SERVER_INTERFACE --stop
For comparison, the congestion control property can be set to disabled on webrtcsink, then the above procedure applied again, the expected result is for playback to simply crawl down to a halt until the bandwidth limitation is lifted:
gst-launch-1.0 webrtcsink congestion-control=disabled
Monitoring tool
An example server / client application for monitoring per-consumer stats can be found here.
License
All the rust code in this repository is licensed under the Mozilla Public License Version 2.0.
Parts of the JavaScript code in the www/ example are licensed under the Apache License, Version 2.0, the rest is licensed under the Mozilla Public License Version 2.0 unless advertised in the header.