StreamChannelAssembler

Create and publish streams from a folder of images, videos and livefeeds. Just place your arbitrary images and videos in the ./data/contents/ folder and run the application. By utilizing gStreamer and OpenCV the images will be livestreamed to a channel or locally. JSON files with links to webcam or other live feeds can also be used. The stream can then be sent to various media servers. The current implementation is only a MVP (Minimum Viable Product) and more features will be added further on. Check out the ./data/config.json file to change settings.

license size commit

demo // TODO:

Motivation

To be fair, I am quite rusty in C++, which motivates the decision to create a new project in C++. The focus with the project is to create clean and good practice code towards high performing computing. Video streaming is a computer resource heavy task which fit perfectly. The application is strongly coupled to gStreamer and OpenCV which are great libraries to learn when renewing one's code language skills. This was also a great opportunity for me to try out the MediaMTX project.

Running

The general usercase is as follows:

Place all wanted content in ./data/contents/ folder. (Optionally remove the examples)
Update ./data/config.json to change general settings such as output location, playlist order, overlays and animations.
Execute the code locally or docker. (See more below.)
View the stream. (If chosen to play locally, the stream will be shown as soon as the application starts running.)
Enjoy!

NOTE: For fast run use ./run.sh it will trigger docker-compose and then open VLC or gStreamer to visualize the stream if installed.

Executing locally

Firstly install dependencies as in the Dockerfile. Then compile, build and run the application by running the following command:

cmake -S . -B build && cmake --build build && ./build/StreamChannelAssembler

NOTE: The main advantage with running locally is that the stream can be viewed locally without an X-server. Using docker though, the project build will be volumed, which means the project can then be run locally but built in docker container.

Executing Docker

This repository has been dockerized. Install docker and run command below. It is recommended to use mediaMTX in docker, which means the video will be streamed to an url, which can be collected using vlc or gStreamer. This can be accomplished by running the commands below or using the ./run.sh script. The build will also be volumed, which means you can run the application locally too.

# Build and run the container and stream
docker-compose up
# View Stream in vlc (if installed!) by running
vlc rtsp://localhost:554/livestream

The built project is volumed to the local folder ./build. The data folder is also volumed so as to share the content with the docker container easily.

Testing

Unit Tests are written using Catch2, they are linked to github actions. Tests will be built with the rest of the project. //TODO: update this

Examples

These are some examples of streams that can be added to the content folder except for videos and images. //TODO:

{
    "name": "Test Video Example",
    "type": "video",
    "target": "videotestsrc ! videoconvert ! appsink",
    "length": 20, // how many seconds the live stream should play
    "start": 0 // where in the stream we should start
}

NOTE: Comments using "//" are not allowed in json syntax, so remove the comments if you are trying this example out.

{
    "name": "Local Webcam Example",
    "type": "video",
    "target": 0,
    "length": 20,
}

These JSON objects can be seen as defining the input to the application.

Supported Data Types

This list show supported datatypes in a table.

FileType	ContentType	Explanation
.png	Image	Std. Type
.jpg	Image	Std. Type
.jpeg	Image	Std. Type
.json	GStreamer	Define a gStreamer command to appsink. See gstreamer_example.json
.json	FFMPEG	Define a FFMPEG command to app. See ffmpeg_example.json
.json	Video	Define a file or command as video. See webcam_example.json
.json	Audio	Define a file or command as audio. See radio_example.json
.json	Image	Define a file or command as image.
.mp4	Video	Std. Type
.avi	Video	Std. Type
.mov	Video	Std. Type
.gif	Video	Gif:s are seen as video.
.mkv	Video	Std. Type
.mp3	Audio	Std. Type

How the formats are handled by the media processors are illustrated in the figure below. system

Relevant links during execution

These links are relevant too know when running the docker-compose file.

http://localhost:8888/livestream/ - View FLV HLS
http://localhost:80 - Example standalone WebView hosted by Nginx

Known issues

This is a list of known issues that hopefully will be solved in near future:

readme need more examples and helpful tips of how to run and so on.
demo added to readme
add scheduler
add a couple of playlist orders, such as random
fix local view in a manageable windowform
no animations

Some major sources

License

This project is under the MIT license, read more here.

grebtsew / StreamChannelAssembler

readme