Add per-channel hybrid audio spread spectrum and env feature toggles

Spread spectrum v2: independent per-channel embedding with round-robin
bit distribution, preserving spatial stereo/surround mix. Adaptive chip
tiers (256/512/1024) trade capacity for lossy codec robustness. LFE
channel skipped for 5.1+ layouts. v2 header (20B) with backward-
compatible v0 decode fallback.

Environment toggles (STEGASOO_AUDIO, STEGASOO_VIDEO) gate audio/video
features for minimal builds (e.g. Raspberry Pi image-only). Values:
auto (default, detect deps), 1/true (force on), 0/false (force off).

Web UI fixes: accordion defaults to step 1 on load, chevron arrow
styling, required attribute toggling for audio carrier type switch,
"Images & Mode" renamed to "Reference, Carrier, Mode".

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

CHANGELOG.md

@@ -5,6 +5,25 @@ All notable changes to Stegasoo will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org).
 
+## [4.3.0] - 2026-02-27
+
+### Added
+- **Audio Steganography** — Hide messages in audio files (WAV, FLAC, MP3, OGG, AAC, M4A)
+  - LSB mode: Direct least-significant-bit embedding in audio samples
+  - Spread Spectrum mode: Noise-resistant encoding using pseudo-random spreading
+  - Automatic format transcoding to WAV for embedding
+  - Full CLI support: `stegasoo audio-encode`, `audio-decode`, `audio-info`
+  - REST API endpoints: `/audio/encode`, `/audio/decode`, `/audio/info`
+  - Web UI: Unified encode/decode pages with carrier type selector (Image/Audio)
+- New `AudioCapacityInfo`, `AudioEmbedStats`, `AudioInfo` model classes
+- Audio-specific exceptions: `AudioError`, `AudioValidationError`, `AudioCapacityError`, `AudioExtractionError`, `AudioTranscodeError`, `UnsupportedAudioFormatError`
+- Subprocess isolation for audio operations (crash protection)
+- `debug.py` module for structured logging across all steganography operations
+
+### Changed
+- Encode/Decode web pages now have a "Carrier Type" step to switch between Image and Audio
+- Version bumped to 4.3.0
+
 ## [4.1.5] - 2026-01-07
 
 ### Added
@@ -201,6 +220,7 @@ and this project adheres to [Semantic Versioning](https://semver.org).
 - CLI interface
 - Basic PIN authentication
 
+[4.3.0]: https://github.com/adlee-was-taken/stegasoo/compare/v4.2.1...v4.3.0
 [4.1.5]: https://github.com/adlee-was-taken/stegasoo/compare/v4.1.3...v4.1.5
 [4.1.3]: https://github.com/adlee-was-taken/stegasoo/compare/v4.1.0...v4.1.3
 [4.1.0]: https://github.com/adlee-was-taken/stegasoo/compare/v4.0.2...v4.1.0