[Feature]: Stream host audio to connected clients

[imxade]

Problem Statement

Rein mirrors the host screen to connected clients, but it does not send host/system audio. This makes videos, presentations, games, and other media sessions incomplete on the remote device.

This is an implementation-focused follow-up to #163 now that screen mirroring is available.

Proposed Solution

Capture host/system audio and add it as an audio track to the existing single WebRTC peer connection for each viewer.

The implementation should:

• Support the platform capture backends used by Rein, with PipeWire/PulseAudio on Linux, WASAPI on Windows, and the appropriate macOS capture API.
• Keep audio synchronized with the mirrored video while prioritizing low latency.
• Add a setting to enable or disable audio sharing and a client-side mute/volume control.
• Reuse the existing WebSocket signaling and single RTCPeerConnection architecture.
• Stop capture and release platform resources when hosting stops or the final viewer disconnects.
• Report capture or permission failures through the existing debug/logging surfaces.

Alternatives Considered (Optional)

• Sending encoded audio over a WebSocket would avoid another WebRTC track, but would require custom buffering, synchronization, and playback logic.
• Microphone capture is not equivalent because the goal is host/system audio.

Additional Context (Optional)

Suggested acceptance criteria:

• Host audio is audible on a LAN client and remains synchronized with video.
• Audio sharing can be disabled without disabling screen mirroring.
• Multiple viewers can receive audio through their existing peer connections.
• Missing permissions or unavailable capture APIs fail gracefully and are visible on /debug.
• Relevant unit checks and platform documentation are added.

Checklist

• I have searched for existing feature requests to avoid duplicates