[Feature] Audio streaming for large files

[vmarcella]

Overview

Add streaming playback for large audio files (music, ambient tracks) that
decodes audio on-the-fly instead of loading entire files into memory.

Current State

Sound loading loads entire files into memory. Large music files (5+ minutes)
would consume excessive RAM.

Scope

Goals:

• Stream audio from file during playback
• Buffer decoded audio in small chunks
• Support seeking within streamed audio
• Support OGG and WAV streaming

Non-Goals:

• Network audio streaming
• Gapless playback between tracks

Proposed API

pub struct StreamingSound {
  // internal decoder + buffer
}

impl StreamingSound {
  pub fn from_file(path: &Path) -> Result<Self, AudioError>;
  pub fn play(&mut self);
  pub fn pause(&mut self);
  pub fn stop(&mut self);
  pub fn seek(&mut self, position_seconds: f32);
  pub fn position(&self) -> f32;
  pub fn duration(&self) -> f32;
  pub fn set_volume(&mut self, volume: f32);
  pub fn set_looping(&mut self, looping: bool);
}

Acceptance Criteria

• Stream a 5+ minute audio file with <10MB memory usage
• Seeking works without audible glitches
• Looping works without gaps
• CPU usage remains reasonable during playback
• Example streaming background music

Affected Crates

lambda-rs, lambda-rs-platform

Notes

• Buffer size tradeoff: larger = more latency, smaller = more CPU
• Consider 100-200ms buffer as default
• Decoder runs on separate thread or async task