[Feature] Sound file loading (WAV, OGG)

[vmarcella]

Overview

Add the ability to load audio files from disk into memory buffers suitable for
playback. Support common formats starting with WAV and OGG.

Current State

No audio file loading exists. The engine cannot read any audio formats.

Scope

Goals:

• Load WAV files into audio buffers
• Load OGG Vorbis files into audio buffers
• Provide SoundBuffer type holding decoded audio data
• Support loading from file path and from memory bytes

Non-Goals:

• MP3 support
• Streaming large files
• Audio playback

Proposed API

// crates/lambda-rs-platform/src/audio/buffer.rs
pub struct SoundBuffer {
  samples: Vec<f32>,
  sample_rate: u32,
  channels: u16,
}

impl SoundBuffer {
  pub fn from_wav_file(path: &Path) -> Result<Self, AudioError>;
  pub fn from_wav_bytes(bytes: &[u8]) -> Result<Self, AudioError>;
  pub fn from_ogg_file(path: &Path) -> Result<Self, AudioError>;
  pub fn from_ogg_bytes(bytes: &[u8]) -> Result<Self, AudioError>;
  
  pub fn sample_rate(&self) -> u32;
  pub fn channels(&self) -> u16;
  pub fn duration_seconds(&self) -> f32;
}

Acceptance Criteria

• Load mono and stereo WAV files (16-bit, 24-bit, 32-bit float)
• Load OGG Vorbis files
• Return meaningful errors for unsupported formats
• Unit tests with small test audio files
• Example loading a sound file

Affected Crates

lambda-rs-platform

Notes

• Consider hound crate for WAV decoding
• Consider lewton crate for OGG Vorbis decoding
• Store samples as f32 internally for mixing flexibility