[Feature] Audio device initialization and enumeration

[vmarcella]

Overview

Add platform-layer audio device abstraction that enumerates available audio
output devices and initializes a default playback device. This is the
foundational building block for all audio functionality.

Current State

No audio support exists in lambda-rs or lambda-rs-platform. The engine cannot
produce any sound output.

Scope

Enumerate available audio output devices on the system

• Initialize a default audio output device
• Provide AudioDeviceBuilder following existing builder patterns
• Abstract platform audio via lambda-rs-platform

Non-Goals:

• Audio input/recording
• Sound playback
• Audio effects processing

Proposed API

// crates/lambda-rs-platform/src/audio/device.rs
pub struct AudioDevice { /* platform handle */ }

pub struct AudioDeviceBuilder {
  sample_rate: Option<u32>,
  channels: Option<u16>,
  label: Option<String>,
}

impl AudioDeviceBuilder {
  pub fn new() -> Self;
  pub fn with_sample_rate(self, rate: u32) -> Self;
  pub fn with_channels(self, channels: u16) -> Self;
  pub fn with_label(self, label: &str) -> Self;
  pub fn build(self) -> Result<AudioDevice, AudioError>;
}

pub fn enumerate_devices() -> Vec<AudioDeviceInfo>;

Acceptance Criteria

• AudioDeviceBuilder creates a working audio output device
• enumerate_devices() returns available system audio devices
• Works on Windows, macOS, and Linux
• Error handling for device initialization failures
• Unit tests for builder configuration

Affected Crates

lambda-rs-platform

Notes

• Consider wrapping cpal for cross-platform audio device access
• Sample rate defaults: 44100 or 48000 Hz
• Channel defaults: stereo (2 channels)