Same as, In One Ear Out the Other, except switch sensory organs and direction ;)
KMC mentioned he was having difficulties getting the IPCamera service working. A little while ago there was an effort to merge OpenCV's IPCamera FrameGrabber and IPCameras method of getting data. In order to make this work one of the challenges was getting a valid test. This was made more difficult since my laptop's camera no longer works do to a windows update :P
Step 1. I figured I need a video source so I started OpenCV with a mkv file. It worked quite well, no sound, but that's not OpenCV's purpose ;)
Step 2. I started a VideoStreamer which can take a video source (OpenCV) and publish it as MJPEGS on a IP Port. This was accomplished with 2 lines of Python.
Step 3. I wanted to verify that it was streaming video so I started a modern browser (Chrome / Firefox - I think IE still can't do this !) and went to http://localhost:9090/output - Yay! Worky !
It worked, but I was surprised a little regarding the choppyness of Chrome.
Step 4. After a couple of bug fixes, I started another instance of MRL with OpenCV2 - and instead of an OpenCV frame grabber, I used the IPCamera frame grabber. AND IT WORKED - YAY ! It was very smooth, and less choppy than Chrome ! At this point, it should be equivalent to attach to ANY MJPEG server ! But wait I had more to test.
Step 5. I started the IPCamera service and set the URL to http://localhost:9090/output - YAY ! Normalized code is a good thing ! Of course the direction buttons don't work, because there is no "real" IP camera, but overall it was a WORKY day !
BRAVO!!!
Nice to know I helped inspire a WORKY kind of day. Now I'll have to try it out.
BTW I like your choice of video files to test on. Brian Cox "Wonders of the (Universe|Solar System)"
Worky!!
I got it to work with both of my foscam clones (Vonnic and Wansview). Screenshot shows the camera that watches the print progress on my Prusa Air 2.