Demonstrate simple use of built-in video capture engine with DV consumer cameras.
VideoDVCamCaptureDemo([fullscreen=0][, fullsize=0][, roi][, depth][,deviceId=0][, moviename])
NOTE: As of October 2014, DV video camera capture has not been tested at
all on MS-Windows, as [GStreamer](GStreamer)-1 currently doesn’t ship with video
capture plugins for MS-Windows. On Apple OSX 10.9.5 Mavericks with our
new [GStreamer](GStreamer)-1 video capture engine, video capture from DV did not work,
neither with Psychtoolbox, nor with [GStreamer](GStreamer) command line tools or other
independent open source video capture applications. Only Apples Facetime
app managed to get a marginally useable video stream from the DV camera.
Testing on Linux with this specific demo showed mixed results. However on
Linux there is a slightly hacky way that makes it work somewhat reliably
with the new backend by exposing the camera as a regular video source, so
all regular video capture/recording/processing demos can be used for DV
capture, without any need for special treatment like in this demo. Read
the section below for further instructions for a reliable DV setup on
‘fullscreen’ If set to non-zero value, the image is displayed in a
fullscreen window, as usual, otherwise a normal GUI window is used.
‘fullsize’ If set to 1, the cameras image is scaled up to full screen
resolution, ie. so it fills the maximum amount of display area, but
preserving the original aspect ratio.
‘roi’ Set to [0 0 720 576] for a PAL-DV camera and [0 0 720 480] for a
NTSC-DV camera. Default is PAL if omitted.
‘deviceId’ Device index of video capture device. Defaults to system
default. You can also specify a gst-launch style string to define a
videosource here. Or you can set the special string deviceId = ‘X’ so
builtin spec strings suitable for each operating system will be used.
‘moviename’ Name string for selection of filename of a target movie file
to which video should be recorded. Defaults to none,ie., no video
VideoDVCamCaptureDemo also allows you to test out video capture from
special video sources other than Consumer-DV cameras, ie. sources which
require use of a custom built [GStreamer](GStreamer) video source bin. You test such
setups by specifying the bin spec-string and other parameters as parameters
for the demo. As an example, the following call would try to capture a video
stream that is encoded as H264 video and that is transmitted over the network
via TCP-IP protocol, ie., a H264 video stream encapsulated in TCP:
VideoDVCamCaptureDemo(, , [0 0 320 240], 6, ‘tcpclientsrc port=8554 host=localhost ! h264parse ! avdec_h264 name=ptbdvsource’);
This would receive the TCP stream from port 8554 on the machine with
the name ‘localhost’ (the local machine). The stream would decode from
H264 to color format 6 - YUV-I420 - with video frames of 320 x 240 pixels
size. For testing purpose you could enter the following [GStreamer](GStreamer) command
line into a terminal window to generate a test video stream and send it from
the local machine ‘localhost’ on port 8554, as H264 encoded TCP-IP stream,
then receive it via VideoDVCamCaptureDemo as shown above:
gst-launch-1.0 videotestsrc horizontal-speed=5 ! x264enc tune=”zerolatency” threads=1 ! video/x-h264,stream-format=byte-stream ! tcpserversink port=8554
One application of such a custom setup can be seen in the discussion thread
containing message #18807 on the Psychtoolbox forum. There the video source
is an IP camera attached to a robot, streaming H264 video over the network for
consumption by a machine running Psychtoolbox.
Loopback setup on Linux for use with new [GStreamer](GStreamer)-1 video backend:
This specific configuration was shown to work at least on Ubuntu 14.04
LTS with a Sony PAL-DV firewire camera. After following the setup steps,
demos like our standard VideoCaptureDemo, VideoRecordingDemo, … worked
without any special configuration or treatment of DV cameras.
Here you need to install a Video4Linux2 loopback kernel module. It will
allow to create virtual video sources, from which Psychtoolbox can
read/capture/process record live video. Then some external application
can feed video into those virtual sources. You then attach an external
command line DV capture session as video source.
Install the package “v4l2loopback-dkms” to get the kernel module installed and
loaded. A “sudo apt-get install v4l2loopback-dkms” on Ubuntu 14.04-LTS
and later distributions should do the trick. The package is probably
also available on Debian, other Debian/Ubuntu derived distros etc. Or
you get the most recent version to compile and install from source
code from the homepage of the project:
You may or may not need to “sudo modprobe v4l2loopback” on first use.
Then you use a [GStreamer](GStreamer) video capture pipeline launched from a terminal
window to connect to your DV camera, capture live video and feed it
into the virtual video loopback device. An example launch line can
look like this:
gst-launch dv1394src ! dvdemux ! dvdec ! v4l2sink device=/dev/video0
This would make live video from the first connected DV camera
available on /dev/video0. See
https://github.com/umlaeute/v4l2loopback/wiki for more detailed
If this doesn’t work for you with [GStreamer](GStreamer)-1 you may need to install
good old [GStreamer](GStreamer)-0.10 in addition to the already installed
[GStreamer](GStreamer)-1 and instead use the gst-launch-0.10 command instead of the
gst-launch command to select for the old implementation.
Psychtoolbox video capture functions should now report and be able to
use a new virtual video capture device with a name like “Dummy video
device 0000” or some name defined by you. Psychtoolbox should be able
to video capture or record video from that device aka your DV video
The Wiki of v4l2loopback describes more elaborate setups, e.g., for
capturing from multiple video DV cameras.