A resilient RTSP client based on the .net framework

A .net library for receiving raw audio/video streams using RTSP.

Main features

• Support Unicast TCP (interleaved mode) / Unicast UDP / Multicast Streaming transports
• Support multiple authentications: Basic, MD5, SHA256, SHA512
• Support auto reconnection in case of network failures
• Support message reordering when multiple requests are sended and responses arrive in a different order
• Easy to used
• Provide Event Handlers for connection loss, receiving packet, etc...
• Provide classes to access to the SDP informations
• Thread safe, except the sdp classes
• Handle large streams with a high bitrate like 50 MBits/seconds
• Force the creation of ports used for receiving packets in case if the ports are temporaly used by some others applications

➡️ The current implementation used the .net Framework 4.8 and not .net core

➡️ The classes used for RTP layer was already implemented in a another project. Here a new implementation is actually in progress.

➡️ The classes used for decoding was already implemented in a another project. Here a new implementation is actually in progress.

How to receive raw rtp packets using the rtsp client ?

Like this :

using ( var client = new RTSPClient() )
{
    // Fired when a successfull connection or when the communication has been recovered after a lost
    client.Connected += (sender, e) =>
    {
        Console.WriteLine("Client connected - " + client.Configuration.Uri);
    };

    // Fired when the communication has been lost
    client.Disconnected += (sender, e) =>
    {
        Console.WriteLine("Client disconnected - " + DateTime.Now);
    };

    // Fired when a raw media data has been received 
    client.PacketReceived += (sender, e) =>
    {
        var interleavedPacket = e.Packet as RTSPInterleavedPacket;

	if ( interleavedPacket != null && interleavedPacket.Channel > 0 )
	    return;
	
	if ( RTPPacket.TryParse( e.Packet.Data , out RTPPacket packet ) )
            Console.WriteLine( "DataReceived {0}" , packet.Payload.Length );
    };

    client.Configuration.Uri = Constants.LocalServer;
    client.Configuration.UserName = Constants.UserName;
    client.Configuration.Password = Constants.Password;
    client.Configuration.KeepAliveType = RTSPKeepAliveType.Options; 
    client.Configuration.ReceiveTimeout = TimeSpan.FromSeconds(3);
    client.Configuration.SendTimeout = TimeSpan.FromSeconds(5);

    client.Configuration.DeliveryMode = RTSPDeliveryMode.Tcp;
    client.Configuration.MediaFormat = RTSPMediaFormat.Video;

    client.StartCommunication(); 

    Console.WriteLine("Press any keys to close the application");
    Console.ReadKey();
}

About the connection class

The RTSPConnection class will give you more features and a direct access to the protocol messaging layer. The following code demonstrate how to list the supported methods available on a security camera:

using ( var connection = new RTSPConnection() )
{
    // Events subscriptions

    connection.Opened               += (sender, e) => Console.WriteLine("Connected");
    connection.MessageReceived      += (sender, e) => Console.WriteLine("Message received");
    connection.MessageSended        += (sender, e) => Console.WriteLine("Message sended");
    connection.DataReceived         += (sender, e) => Console.WriteLine("Data received");
    connection.AuthenticationFailed += (sender, e) => Console.WriteLine("Authentication failed");
    connection.Closed               += (sender, e) => Console.WriteLine("Connection closed");
    connection.Error                += (sender, e) => Console.WriteLine("Error occurs");

    // Connect to RTSP server (happytime-rtsp-server.exe)
    
    connection.Open("rtsp://192.168.1.11/city1.mp4");
     
    // Request the available methods from a server

    connection
            .GetOptions()
            .Invoke()
            .Response
            .GetHeaderPublicOptions()
            .ToList()
            .ForEach( Console.WriteLine )
            ;
}

Below, on the pseudo code, the remote method invocation works like this:

var bodyResult =

 connection

        .Play()

        .AddHeader( new RTSPHeaderContentType( RTSPMimeType.ApplicationText ) )
        .AddHeader( "X-Header1" , "my value 1")
        .AddHeader( "X-Header2" , "my value 2")
        .AddHeader( "X-Header3" , "my value 3")
        .AddHeader( "X-Header4" , "my value 4")
        .AddHeader( "X-Header5" , "my value 5")
        .AddHeader( "X-Header6" , "some_values={0}-{1};other_param={2};" , 1234 , 4321 , 5000 )

        .WriteBodyLine( "timestamp:{0}" , DateTime.Now )
        .WriteBodyLine( "computer:{0}" , System.Environment.Machine )
        .WriteBodyLine( "uuid:{0}" , Guid.NewGuid() )

        .Invoke()
	
        .Response

        .GetBody()
        ;

You will be able to decorate each request by adding customs headers, because some cameras can not reply to a request that just contains only standard headers or if there the message contains incomplete headers. If you want to invoke a method on a particular server, you MUST read the server documentation especially the SETUP method. For instance, the SETUP are used to ask to the camera to create a streaming session that relies on a specific transport layers.

⚠️: The only thing to known about the connection class is that the current implementation does not support SSL/TLS. But not yet.

About Session Description Protocol layer

That an exemple of how to use this .net implementation of SDP, please not that is implementation provide custom validation classes which are not explain in this example.

var sessionDescriptor = new SessionDescriptor();

sessionDescriptor.Origin.UserName = "John";
sessionDescriptor.Origin.Address = "192.168.1.23";
sessionDescriptor.Origin.AddressType = AddressType.IPV4;
sessionDescriptor.Origin.NetworkType = NetworkType.Internet;
sessionDescriptor.Origin.Version = "V1";
sessionDescriptor.Origin.SessionId = "123456789";
sessionDescriptor.Version.Value = 1;
sessionDescriptor.SessionName.Value = "My session name";
sessionDescriptor.Repeats.Add(new RepeatField(new ValueTime(1, 2), new ValueTime(3, 4)));
sessionDescriptor.Repeats.Add(new RepeatField(new ValueTime(10, 20), new ValueTime(30, 40)));
sessionDescriptor.Attributes.Add(new AttributeField("myAttribute1", "myValue1"));
sessionDescriptor.Attributes.Add(new AttributeField("myAttribute2", "myValue2"));
sessionDescriptor.Attributes.Add(new AttributeField("myAttribute2", "myValue3"));

sessionDescriptor.Phones.Add(new PhoneField("+33 1 12 34 56 78"));
sessionDescriptor.Phones.Add(new PhoneField("+33 1 12 34 56 79"));
sessionDescriptor.Uri.Value = "rtsp://192.168.1.11:554/video/channel/1";

for ( int i = 1; i <= 10; ++ i )
{
    var mediaDescription = new MediaDescriptionField();

    mediaDescription.Payload = 1 + i;
    mediaDescription.Port = 10 + i;
    mediaDescription.Profile = ProfileType.AVP;
    mediaDescription.Protocol = ProtocolType.RTP;
    mediaDescription.Type = MediaType.Video;

    mediaDescription.Encryption.Key = "myKey"+i.ToString();
    mediaDescription.Encryption.Method = "myMethod"+i.ToString();
    mediaDescription.Connection.Address = "192.168.1."+i.ToString();
    mediaDescription.Connection.AddressType = AddressType.IPV4;
    mediaDescription.Connection.NetworkType = NetworkType.Internet;
    mediaDescription.Bandwiths.Add(new BandwithField("modifier", i));
    mediaDescription.Bandwiths.Add(new BandwithField("modifier"+i.ToString(), i+i));
    mediaDescription.Attributes.Add(new AttributeField("myAttribute1", "myValue1"));
    mediaDescription.Attributes.Add(new AttributeField("myAttribute2", "myValue2"));
    mediaDescription.Attributes.Add(new AttributeField("myAttribute3", "myValue3"));

    sessionDescriptor.MediaDescriptions.Add(mediaDescription);
}

Console.WriteLine( sessionDescriptor.ToString() );

if ( SessionDescriptor.TryParse( sessionDescriptor.ToString() , out SessionDescriptor descriptor ) )
{
    Console.WriteLine("Session descriptor parsed");
}

⚠️: Please notes, that this implementation has been tested ONLY with many security cameras models, but NOT with VoIP devices. It may possible with VOIP device add extra fields.

Getting more details ?

If you want to get more details, send me an email to "a.sahnine@netcourrier.com"