Module decoder 

AMBE-to-PCM-to-MP3 audio decode pipeline.

D-STAR voice transmissions encode speech using the AMBE 3600x2450 codec at 3600 bits/second (2450 bps voice + 1150 bps FEC). Each voice frame is 9 bytes, transmitted at 50 frames/second (one every 20 ms). The decoder pipeline converts these compressed frames into standard MP3 audio:

[u8; 9] x N           mbelib-rs           mp3lame-encoder
AMBE frames  ------>  PCM i16 @ 8 kHz  ------>  MP3 bytes
             decode_frame()             encode + flush

Pipeline stages

1. AMBE decode (mbelib_rs::AmbeDecoder): Each 9-byte AMBE frame is decoded into 160 signed 16-bit PCM samples at 8000 Hz (20 ms of audio). The decoder carries inter-frame state for delta prediction and phase-continuous synthesis, so frames must be fed sequentially.

2. PCM accumulation: All 160-sample chunks are concatenated into a single Vec<i16> buffer. For a typical 3-second D-STAR transmission (150 frames), this is 24,000 samples.

3. MP3 encoding (mp3lame_encoder): The accumulated PCM buffer is encoded in one pass using LAME configured for mono, 8000 Hz input sample rate, CBR at the requested bitrate. A final flush writes any remaining LAME internal buffer to complete the MP3 stream.

Enums

DecodeError 🔒

Errors that can occur during the AMBE-to-MP3 decode pipeline.

Functions

bitrate_from_kbps 🔒

Maps a bitrate in kbps (u32) to the mp3lame_encoder::Bitrate enum.

decode_to_mp3 🔒

Decodes a sequence of 9-byte AMBE voice frames into an MP3 byte buffer.