v4l2rtspserver

v4l2rtspserver is a streaming server that captures feeds from Video4Linux (V4L2) devices and ALSA audio devices and serves them via RTSP or HTTP.

Supported Video Capture Formats (V4L2):

• H264, HEVC, JPEG, VP8, or VP9

Supported Audio Capture Formats (ALSA):

• PCM S16_BE, S16_LE, S32_BE, or S32_LE

Supported Streaming Protocols:

• RTSP: RTP/UDP unicast, RTP/UDP multicast, RTP/TCP, and RTP/RTSP/HTTP.
• HTTP: HLS and MPEG-DASH (enabled via the -S option for compatible capture formats).

If your camera provides an uncompressed format, you can use v4l2tools to compress it to an intermediate virtual device using v4l2loopback before feeding it to v4l2rtspserver.

Workflow:

1. Load v4l2loopback.
2. Compress /dev/video0 to /dev/video10.
3. Stream from /dev/video10.

modprobe v4l2loopback video_nr=10
v4l2compress -fH264 /dev/video0 /dev/video10 &
v4l2rtspserver /dev/video10 &

Authentication can be enabled by defining users with the -U flag. You can use plain text passwords or MD5-hashed passwords.

Plain Text: -U user:password

MD5 Password: To use MD5, you must define a realm with -R. The password must be the result of md5(<username>:<realm>:<password>).

Example (MD5):

-R myrealm -U foo:$(echo -n foo:myrealm:bar | md5sum | cut -d- -f1)

You can run the server using a Docker image. To access hardware devices like cameras, you must use the --device flag.

Run with default settings:

docker run -p 8554:8554 -it mpromonet/v4l2rtspserver

Run with a specific V4L2 device:

docker run --device=/dev/video0 -p 8554:8554 -it mpromonet/v4l2rtspserver

Run with custom parameters:

docker run --device=/dev/video0 -p 8554:8554 -it mpromonet/v4l2rtspserver -u "" -H640 -W480

To build the project from source, use cmake and make. The build process will automatically attempt to download and compile live555 if it is not present on your system.

Note on Dependencies:

• libasound2-dev: If missing, audio support will be disabled.
• libssl-dev: If missing, rtsps and srtp will not be available.
• liblog4cpp5-dev: If missing, logging defaults to std::cout.

cmake . && make
# Optional: install to system
sudo make install

To use a Raspberry Pi camera with this server, you must use the open-source V4L2 driver bcm2835-v4l2.

sudo modprobe -v bcm2835-v4l2

When started with the -S argument, the server enables HLS and MPEG-DASH via HTTP.

MPEG-DASH: Use MP4Client <URL>/unicast.mpd.

HLS:

• VLC: vlc <URL>/unicast.m3u8
• GStreamer: gst-launch-1.0 playbin uri=<URL>/unicast.m3u8
• Browser: Use Firefox with [Native HLS addons] or Chrome with [Native HLS playback].

The v4l2rtspserver executable provides a command-line interface to stream video and audio from V4L2 (Video4Linux2) and ALSA devices via RTSP, HLS, or MPEG-DASH.

Basic Usage

To start a basic stream, provide the V4L2 device path. You can also specify an ALSA device for audio using a comma-separated string.

# Stream from /dev/video0
v4l2rtspserver /dev/video0

# Stream from /dev/video0 (video) and hw:1,0 (audio)
v4l2rtspserver /dev/video0,/dev/hw:1,0

Device Configuration

• Video/Audio Split: Use a comma to separate the video device and the audio device: [V4L2 device],[ALSA device].
• Resolution and Framerate: Use -G <w>x<h>[x<f>] to set width, height, and FPS in one command, or use -W, -H, and -F individually.
• Capture Mode: Use -s to use the live555 mainloop instead of the default reader thread.

Use the following flags to tune the capture behavior:

V4L2 Options:

• -r: Use the read interface (default is memory mapped buffers).
• -w: Use the write interface.
• -B: Use blocking mode.
• -s: Use live555 mainloop.
• -f: Use current capture format (ignores -W, -H).
• -fformat: Use a specific format (respects -W, -H).
• -W <width>: Set capture width (default 640).
• -H <height>: Set capture height (default 480).
• -F <fps>: Set framerate (default 25, 0 to disable).
• -G <w>x<h>[x<f>]: Set capture format.

ALSA Options:

• -A <freq>: Capture frequency and channel (default 44100).
• -C <channels>: Capture channels (default 2).
• -a <fmt>: Capture audio format (default S16_LE).

The server is invoked via a command-line interface that allows configuring RTSP, V4L2, and ALSA parameters.

Basic Syntax: ./v4l2rtspserver [options] [device1] [device2]

Session Composition Examples:

• Single Video Session: ./v4l2rtspserver /dev/video0
• Single Audio Session: ./v4l2rtspserver ,default
• Combined Audio and Video (Single Session): ./v4l2rtspserver /dev/video0,default
• Combined Audio and Video (Two Sessions): ./v4l2rtspserver /dev/video0 ,default
• Two Video Sessions: ./v4l2rtspserver /dev/video0 /dev/video1
• Audio associated with specific V4L2 device: ./v4l2rtspserver /dev/video0,/dev/video0

Use these flags for logging, output files, and queue management.

Use these flags to configure how the server interacts with V4L2 video devices.