QUESTION about API working with the Android client

[xvoland]

Thanks for your interesting project.

I was trying to figure out how your Windows client works (I’m not a C# developer), specifically how it controls the Android app over Wi-Fi — changing the resolution, switching the camera, and so on.

As far as I understand, it switches the app into some kind of simplified mode and then sends commands.

I noticed the following structures:

// PacketType.kt
enum class PacketType(val code: Byte) {
    HELLO(0x01),
    ACK(0x02),
    VIDEO(0x03),
    CONTROL(0x04),
    HEARTBEAT(0x05),
    DISCONNECT(0x06)
}

// ControlPacket.kt
enum class ControlPacket(val code: Byte) {
    SWITCH_CAMERA(0x01),
    TOGGLE_FLASH(0x02),
    SET_ZOOM(0x03)
}

How exactly are the TCP packets sent — what’s the algorithm and the sequence?

Could you help me by describing it step by step, for example:

Windows client → command: (1 byte)(1 byte) → Android

Command: c0

Thanks in advance for your reply.

[darusc]

The windows app doesn't use any C# or dotnet libraries. Networking is done using ASIO for C++.

The app is controlled the same way on both WiFi and over USB (ADB). If connection is over USB adb reverse tcp:6969 tcp:6969 is used to route the request to 127.0.0.1 made by the android phone to the windows server.

The windows app opens a TCP server and waits for a connection from the android app. Upon connecting the android app sends an initial packet with config identifiers (DeviceDescriptor.kt) and starts listening for messages/commands received from the server. (these are all the commands the server can send PacketType.kt)

The android app can talk back to the windows app, sending errors if a command could not be executed.

You can take a look at https://github.com/darusc/VCamdroid/blob/main/windows/src/rtsp/manager.cpp to see how each command is sent by the server.

[xvoland]

Thank you, I'll try to figure it out further.

[xvoland]

I don’t know why, but I still can’t get an RTSP stream from the Android app like it says on your github.

Also, I can’t fully control the app on my old SM-G950F.
Maybe it’s just too outdated and has some limitations or compatibility issues. (???)

Right now, for Wi-Fi, my setup looks like this:
Android → TCP (8554) → JPEG decoding → frameBuffer → HTTP (8080) → browser

The video is choppy at higher resolutions (anything above 640x480).

Some commands work, for example: I can switch to the front camera, but I can’t switch back.
Video filters work fine.

Here’s how it currently works on my side:

• My app starts a TCP server on port 8554
• Then it starts an HTTP server on port 8080
• After scanning the VCamdroid QR code (<local_IP>:8554)
• Android VCamdroid opens a TCP connection to port 8554
• I read the device descriptor and send back confirmations: bytes 0x01 and 0x02

Then this happens:

• Android VCamdroid sends a continuous stream of JPEG data (not RTSP?)
• I look for JPEG markers (0xFFD8 / 0xFFD9) in the stream
• I decode each frame and store it in a frameBuffer to serve as an MJPEG stream (multipart/x-mixed-replace)

Any ideas what I might be doing wrong or where I should look?

[darusc]

I don't really understand what you're trying to do. Build an app that receives the incoming stream from the android app and stream it over HTTP?

In VCamdroid, on the android side, streaming is done using RootEncoder and the RTSP server plugin.

The video stream is H264 encoded by default, and streamed over RTSP protocol on port 8554. Maybe opening another connection on the same port blocks the RTSP stream.

From what you wrote, I think you got the "confirmation" part, after connecting, wrong. When you want the android app to start streaming you need to send the "ACTIVATION" command (0x02) together with the stream options - line 36-39 in manager.cpp.

Regarding your phone, I don't know what you say, maybe some android APIs, through which control is done, are not supported on the phone's android version. I didn't test it on versions lower than 11.

[xvoland]

Thanks a lot for your detailed reply and for explaining how it works.

What I’m trying to do is receive the RTSP stream in my own application so I can process it there. I’m not 100% sure yet how exactly I’m going to process it (it’s more of an experiment at this point) but I’d really like to have that flexibility and control over the stream.

Yeah, my phone is pretty old, and it sometimes goes to sleep while the camera is on. So it’s very likely just an Android version compatibility issue. That makes sense.

I’ll keep digging and experimenting and I still hope I can get this working the way I originally planned.

Thanks again for your time and help, I really appreciate it.