A client of ours has a mobile web cam placed in a forest that is streaming video on a public IP address. Since the web cam has a limited bandwidth (and it is streaming with a format that often requires clients to install a codec), the stream needs to be re-broadcast by a server on a landline, preferably as streaming FLV.
What components can be used to write a client/server that can do this? It would be written using C#.
(Software solutions would be fine too, but we're on a limited budget so it can't be something very expensive...)
What's the format that the camera is sending you?
Rebroadcasting is easy using off-the-shelf servers - which means no programming as such, no C#.
camera -> ffserver -> flash players
ffserver is part of ffmpeg.
Related
I'm trying to make a remote desktop app where user controls his pc from a webapp (as in logmein).
I achieved that with C# for the desktop part, and NodeJS for the webapp, the communication was made using Socket.IO.
My first attempt was capturing the screenshot (only 5 fps), then comparing it to the previous screenshot and sending only the difference in 8-bit image color which resulted - in a 800 * 600 resolution virtual desktop - in a 100kb first image, then from 5kb to 60Kb depending on the changes on the screen.
With my local machine controlling a virtualbox, everything was perfect, but when I hosted the webapp online, the result was catastrophic, an improbable lag was taking place.
After a few researches It turned out this kind of app was impossible to achieve with my way, and that I have to use a real-time protocol and make a live streaming out of the client screen.
My questions are :
Is there any free / open-source RTP libraries that is ready-to-use ?
How would-I transfer a live streaming from the desk app to the webapp since it's coming from the client side which has no open port ? I was thinking of another desktop app that will run on the server (hosting the webapp) and then it will stream the same content again, and then the webapp can simply display the content by acceding to the local ip with the RTP port, but this doesn't solve the mystery of transferring a live streaming from the client to the server ?
Is there any free / open-source RTP libraries that is ready-to-use ?
live555 - I've used and is excellent, but C++ so you would have to interop.
gstreamer - also native requiring interop.
Managed Media Aggregation I've not used but it is completely managed.
How would-I transfer a live streaming from the desk app to the webapp
since it's coming from the client side which has no open port ? I was
thinking of another desktop app that will run on the server (hosting
the webapp) and then it will stream the same content again, and then
the webapp can simply display the content by acceding to the local ip
with the RTP port, but this doesn't solve the mystery of transferring
a live streaming from the client to the server ?
This would be tricky. All the libraries above follow strict RTSP/RTP specification which requires opening a listening port on your host side, which is undoubtedly going to be behind a nat'd address. I would stick with each end being a client and reaching 'up' to your webservice. You also need to guarantee delivery of your frames (because your delivering incremental deltas) so RTP (which is traditionally over UDP) would be challenging.
Some thoughts
At the end of the day RTP is just a standardized 12 byte header and packetization rules for compressed media. It's not going to help with latency. The real benefit would allow you to connect to the endpoint with in a standards compliant way, like with a VLC client.
You could tune your sockets and that will help a bit but what I would focus on to be honest is compression and screen capture efficiency. What image compression are you using? VNC has traditionally used zlib and some others lossy like jpeg. The smaller you get those frames the better.
Also another thought which may help - Microsoft has an API for getting 'dirty screen areas'. It is called Desktop Duplication API and it performs incredibly fast. It is Win8 and up however.
All the best on your endeavor!
I'm working on living Stream to server in UWP using MediaCapture, but I can't find any useful solution about it.
Microsoft's library, but it only supports Azure.
https://github.com/MicrosoftDX/AzureRTMPIngestLib
I can play RTMP live video streaming from server but can't send video streaming to server, I want to know if there has any solution or library can send RTMP live streaming in UWP?
Following example below uses STSP. I've tested this example on a local network with using ipv4 addresses of two different computers. These computers are transmitting and receiving data at the same time to each other. Client and server sides of your app have to support the same protocol. And it gives too much properties about video recording and streaming processes.
Real-time communication sample
A simple end-to-end video call client that demonstrates the low latency mode of the Windows Runtime capture engine. This is enabled using the msRealTime the video tag or RealTimePlayback on the MediaElement. The sample uses a custom network source and a custom sink extension to send and receive captured audio and video data between two computers.
A demonstration of the end-to-end latency of video captured using the Media Capture API and displayed using a video and MediaElement with low latency mode enabled. Two output windows are displayed. The first shows a camera preview window of the raw output from your camera. The second is a local host client window that shows the video from the camera when compressed, streamed, and received over machine's loopback network interface. This window demonstrates the end-to-end latency of video captured, streamed to, and displayed by a remote client minus network latency.
Now it's your turn. Please inform us about results.
I have a situation where videos reside on an FTP server, and I need to stream them through my Website Project.
Using a very crude method of including the FTP username and password in the URL, I can just drop the formed URL in as a link in the HTML video player.
http(s)://username:password#server
I am a bit stuck on how to proceed with consuming the video from this remote secured FTP site. There is no web server running to "serve" the videos over http. It is a dedicated FTP server.
Initially, I have played around with making a physical FTP connection from code, but the streaming seemed to be a problem using this method. I just temporarily used the URL authentication method, and it is time to revisit.
Unfortunately, I do not have the original code where I attempted to make the FTP connection through code.
I need to re-visit this, and would like some input before I proceed.
Video streaming is quite a specialist area, especially if you want to use techniques like Adaptive Bit Rate streaming (ABR - see note below) to give your users the best possible user experience.
Given this I think the best approach might be for your to FTP the videos to a streaming server where they can be prepared and made available for streaming properly.
There are open source and commercial streaming servers available, some cloud SaaS based, which means you don't have to invent and develop all this yourself. Some examples are:
https://gstreamer.freedesktop.org (opensource)
https://www.wowza.com (commercial with free trial)
Note: ABR - this essential means you have multiple versions of your video available on the server. Each is a different bit rate, and all are broken into (for example) 10 second chunks. The client requests the next chunk of the video from whatever is the most appropriate bit rate for the current network conditions. Many clients will also request a low bit rate to start the video to ensure a quick start and then 'step up' through the bit rates to the most appropriate one. You can see this when you start a new video on sites like Netflix etc.
I'm trying to generate a live video stream and surface it via a UPnP framework.
I'm using the UPnP framework that was originally developed by intel available here. It seems to have rolled it's own lightweight webserver. I'm using FFMPEG to generate my video stream from images, I can set it up to feed it frames on a timer. But how I manage the data that's generated? How do I use send an HTTP response that could be a stream of unlimited length?
Is there a well worn technology to do this that I'm not aware of?
Any input would be great.
https://trac.ffmpeg.org/wiki/StreamingGuide possibly ffmpeg could listen on a tcp port, though that would require you to restart ffmpeg each time a client exits...and would only serve one client at a time. If you want more than one client at a time you would have to use some type of real server...
My task is to pick up a video stream from an AXIS M1011 IP-camera and relay it to multiple clients with a REST web service.
The reason I'm using such a thing is because the camera can hold only up to 15-20 connections, and will have a bit more. Also there is some security and authentication issues, some additional features and so on. Basically the ideal thing is to have a "singlton" video streaming from the camera relayed to multiple clients.
The camera can stream MJPEG and H.264 stream and static images. I'm having troubles capturing the H.264 stream or MJPEG with the service. Since the H.264 stream is served via RTSP protocol, and the MJPEG has poor compression and I'm not clear how to resend it to the client.
So one of my ideas is to pull the current static image, for example, 25 times a second and somehow create an H.264 stream which would be streamed to the user. Is there any way to create one H.264 stream object which would be shown to all clients or not? Because what would a new client see if he connects later is the stream from the start of capturing. Does that mean I should start separate encoding processes for every connected client? That sounds like it is resource-hungry solution (reading and encoding images).
What would be the best path to take for this? Every advice would be a great help, code snippets even better.
Tnx in advance.