What's New
Audio Processing Overhaul
- Full broadcast-quality audio processing chain for conversion modes: HPF (120 Hz) → 4:1 compressor → FFT denoiser → EQ cut at 250 Hz (−3 dB) → presence boost at 3 kHz (+5 dB) → LPF at 3.2 kHz → loudnorm → brick-wall limiter at −1 dBFS
- Mode 2 (normalization): −14 LUFS, LRA 11 (natural dynamics)
- Mode 3 (loud normalization): −14 LUFS, LRA 3 (tight broadcast leveling)
- Output bitrate updated to 48 k AAC
Transcription Audio Enhancement
- New admin toggle to pre-process audio before transcription using noise reduction and dynamic compression
- Outputs optimal 16 kHz mono WAV to transcription providers (Whisper, Google, Azure, AssemblyAI)
- Independent of the storage/streaming conversion pipeline
UI Updates
- ThinLine Radio logo moved to the top-right header alongside Admin and Sign Out buttons
- Volume slider relocated into the toolbar right zone; redesigned as a flat inline control (no border box) — Volume label → speaker icon → slider → percentage
See CHANGELOG.md for full details.