v4l2loopback - a kernel module to create V4L2 loopback devices

This module allows you to create "virtual video devices". Normal (v4l2) applications will read these devices as if they were ordinary video devices, but the video will not be read from e.g. a capture card but instead it is generated by another application. This allows you for instance to apply some nifty video effects on your Zoom video call... It also allows some more serious things (e.g. I've been using it to add streaming capabilities to an application by the means of hooking GStreamer into the loopback devices).

NEWS

To get the main features of each new release, see the NEWS file. You could also have a look at the ChangeLog (which gets automatically generated and might only be of limited use...

ISSUES

For current issues, checkout https://github.com/umlaeute/v4l2loopback/issues Please use the issue-tracker for reporting any problems.

Before you create a new ticket in our issue tracker, please make sure that you have read this document and followed any instructions found within.

Also, please search the issue-tracker before reporting any problems: it's much better to add your information to an existing ticket than to create a new ticket with essentially the same information.

SEEKING HELP

The issue tracker is meant to track specific bugs in the code (and new features). However, it is ill-suited as a user support forum.

If you have general questions or problems, please use the v4l2loopback tag on Unix & Linux instead: https://unix.stackexchange.com/questions/tagged/v4l2loopback

DEPENDENCIES

In order to build (compile,...) anything, you must have a working build-environment (compiler, GNU make,...). The kernel can be somewhat picky if you try to load a module that was compiled with a different compiler as was used to compile the kernel itself. So make sure to have the right compiler in place.

The v4l2loopback module is a kernel module. In order to build it, you must have the kernel headers installed that match the linux kernel with which you want to use the module (in most cases this will be the kernel that you are currently running). Please note, that kernel headers and kernel image must have exactly the same version. For example, 3.18.0-trunk-rpi is a different version than 3.18.7-v7+, even though the first few numbers are the same. (Modules will be incompatible if the versions don't match. If you are lucky, the module will simply refuse to load. If you are unlucky, your computer will spit in your eye or do worse.)

There are distribution-specific differences on how to get the correct kernel headers (or to install a compilation toolchain). Documenting all those possibilities would go far beyond the scope of v4l2loopback. Please understand that we cannot provide support for questions regarding dependencies.

BUILD

To build the kernel module, run:

$ make

This should give you a file named "v4l2loopback.ko", which is the kernel module

Build again

You cannot load a module built for a specific version of the kernel into another version of the kernel. So, if you have successfully built the module previously and have updated your kernel (and the matching headers) In the meantime, you really must clean the build before re-compiling the module. So run this before starting the build again:

$ make clean

Afterwards re-run make to do the actual build.

Build for a different kernel

By default a simple make will (try to) build the module for the currently active kernel (as determined by uname -r). If you want to build for a different kernel, youcan do so by providing the kernel version via the KERNELRELEASE variable:

$ make KERNELRELEASE=6.11.7-amd64

(Of course you must have the kernel-headers for the specified kernel available in the /lib/modules/${KERNELRELEASE}/build/ directory.)

INSTALL

To install the module, run "make install" (you might have to be 'root' to have all necessary permissions to install the module).

If your system has "sudo", do:

$ make && sudo make install
$ sudo depmod -a

If your system lacks "sudo", do:

$ make
$ su
(enter root password)
# make install
# depmod -a
# exit

(The depmod -a call will re-calculate module dependencies, in order to automatically load additional kernel modules required by v4l2loopback. The call may not be necessary on modern systems.)

See below for distribution-specific build instructions or when using frameworks like DKMS.

RUN

Load the v4l2loopback module as root :

# modprobe v4l2loopback

Using sudo use:

$ sudo modprobe v4l2loopback

You can check which loopback devices are created by listing contents of /sys/devices/virtual/video4linux directory. E.g. if there are two v4l2loopback devices /dev/video0 and /dev/video3 you would get:

$ ls -1 /sys/devices/virtual/video4linux
video0
video3

These devices are ready to accept contents to show.

Tested feeders:

• GStreamer-1.0: using the "v4l2sink" element
• Gem(>=0.93) using the "recordV4L2" plugin

In theory most programs capable of writing to a v4l2 device should work.

The data sent to the v4l2loopback device can then be read by any v4l2-capable application.

You can find a number of scenarios on the wiki at http://github.com/umlaeute/v4l2loopback/wiki

Troubleshooting

If you have a secure-boot enabled kernel, you might not be able to simply build a kernel module and insert it. (You will get SSL errors when building the module.) This is actually a security feature (as it prevents malicious code to be inserted into kernel-space).

If you are not allowed to insert the kernel module (running modprobe, or insmod), you have a few options (consult your distribution's documentation on how to perform any of these steps):

• disable secure-boot and reboot
• sign the module binary with a whitelisted key (this probably only applies if you are creating a distribution)

You could also just try building the module via DKMS, and hope that it does all the magic for you.

OPTIONS

If you need several independent loopback devices, you can pass the "devices" option, when loading the module; e.g.

# modprobe v4l2loopback devices=4

Will give you 4 loopback devices (e.g. /dev/video1 ... /dev/video5)

You can also specify the device IDs manually; e.g.

# modprobe v4l2loopback video_nr=3,4,7

Will create 3 devices (/dev/video3, /dev/video4 & /dev/video7)

# modprobe v4l2loopback video_nr=3,4,7 card_label="device number 3","the number four","the last one"

Will create 3 devices with the card names passed as the second parameter:

• /dev/video3 -> device number 3
• /dev/video4 -> the number four
• /dev/video7 -> the last one

If you encounter problems detecting your device with Chrome/WebRTC you can try 'exclusive_caps' mode:

# modprobe v4l2loopback exclusive_caps=1

This will enable 'exclusive_caps' mode that only reports CAPTURE/OUTPUT capabilities exclusively. The newly created device will announce OUTPUT capabilities only (so ordinary webcam applications (including Chrome) won't see it). As soon as you have attached a producer to the device, it will start announcing CAPTURE capabilities only (so applications that refuse to open devices that have other capabilities apart from capturing can open it too.)

CHANGING OPTIONS

Options that you provided when loading the module (e.g. via modprobe) cannot be easily changed on the fly. In order to change these options, you must first unload the module with rmmod (which will only work if no application is any longer accessing one of the loopback devices) and then load it again (with the new options).

See also the section about DYNAMIC DEVICE MANAGEMENT.

ATTRIBUTES

you can set and/or query some per-device attributes via sysfs, in a human readable format. See /sys/devices/virtual/video4linux/video*/

also there are some V4L2 controls that you can list with

$ v4l2-ctl -d /dev/video0 -l

• keep_format(0/1): while set to 1, once negotiated format will be fixed forever, until the setting is set back to 0
• sustain_framerate(0/1): if set to 1, nominal device fps will be ensured by means of frame duplication when needed
• timeout(integer): if >0, will cause a timeout picture (a null frame, by default) to be displayed after (value) msecs of missing input
• timeout_image_io(0/1): if set to 1, the next opener will write to timeout frame buffer

CHANGING THE RUNTIME BEHAVIOUR

FORCING FPS

$ v4l2loopback-ctl set-fps /dev/video0 25

or

$ echo '@100' | sudo tee /sys/devices/virtual/video4linux/video0/format

FORCING FORMAT

$ v4l2loopback-ctl set-caps /dev/video0 "UYVY:640x480"

Please note that GStreamer-style caps (e.g. video/x-raw,format=UYVY,width=640,height=480) are no longer supported!

SETTING STREAM TIMEOUT

You can define a timeout (in milliseconds), after which the loopback device will start outputting NULL frames, if the producer suddenly stopped.

$ v4l2-ctl -d /dev/video0 -c timeout=3000

Requires GStreamer 1.0, version >= 1.16: You can provide a timeout image, which will be displayed (instead of the NULL frames), if the producer doesn't send any new frames for a given period (in the following example; 3000ms):

$ v4l2loopback-ctl set-timeout-image -t 3000 /dev/video0 service-unavailable.png

DYNAMIC DEVICE MANAGEMENT

You can create (and delete) loopback devices on the fly, using the add (resp. delete) commands of the v4l2loopback-ctl utility.

When creating a new device, module options might be ignored. So you must specify them explicitly.

To create a new device /dev/video7 that has