Page 1 of 1

Simple question on naming (what's it streaming TO?)

Posted: 24 Mar 2009 22:10
by deadletter
Okay, I'm a very similar issue.


1) I have a variety of cameras around the building, some running off of their own terminal, some are web-ready with their own internal IP feed.

For example, one of the feeds (which I can open in VLC no problem) is http://10.1.10.104/img/video.mjpeg

2) I'm running a local webserver (currently w2k server, soon an apache server on an ubuntu build). The local address of the webserver is 10.1.10.197

3)What I want to have happen is this:

VLC opens the web stream from the internal ip addresses
VLC feeds each video stream to a local IP reference - WHAT IS THAT REFERENCE? It would be fine if the local cameras were something like,

http://10.1.10.197/cam1
http://10.1.10.197/cam2
http://10.1.10.197/cam3

or something similar - I have gone through the streaming menu and it seems to be working (I can set local playback and see what it is sending), I simply can't figure out what name I should be referencing in my browser to see this stream.

Questions: should I be using http or UTP or some other setting?
What encapsulation should I be using, knowing that it is coming in as a mjpeg?
how do I name it properly?

Thanks!

Re: Simple question on naming (what's it streaming TO?)

Posted: 31 Mar 2009 20:22
by RĂ©mi Denis-Courmont
If you did not specify it, it would be the root of the web server.

UTP is a cable type, not a streaming protocol.

There is no efficient way to stream out multipart JPEG. Because multipart JPEG is a bullsh*t invention from cheap camera vendors.

Re: Simple question on naming (what's it streaming TO?)

Posted: 01 Apr 2009 07:01
by cocoapepper
Motion JPEG isn't a very good for streaming. If you do the math you will find the bitrate is pretty high for even low quality, low framerate videos.

Many camera companies have used M-JPEG because, well, their cameras already support JPEG. But it was once a professional video editing format. I was doing full resolution, full framerate, video editing in 1992 on a 25/50 MHz 68040 (similar to a 25/50 MHz Intel 486) Macintosh Quadra with only 12 MB of system ram. How was this possible? The Mac had a video capture card that used hardware JPEG compression for each frame. The system simply had to save those sequences of JPEG images to the hard drive and/or output them to the video-our port. The quality was very nice , and since it used JPEG, it was possible to adjust the quality setting quite easily. With a 2 drive RAID it was possible to have 25 MB/sec video, which works out to about a 850 KB JPEG for each 640x480 pixel frame! The only problem was rendering transitions and titles. Even with fancy video cards, those old processors were SLOOOOW at rendering anything more complex than a basic wipe or fade!