go-aac

A pure Go implementation of an AAC (Advanced Audio Codec) decoder supporting ADTS format.

Features

• Pure Go implementation with no external dependencies
• Supports AAC-LC (Low Complexity) profile
• ADTS (Audio Data Transport Stream) format parsing
• Decodes AAC frames to PCM audio
• Full implementation of AAC decoding pipeline including:
  • Huffman decoding
  • Inverse quantization
  • Temporal Noise Shaping (TNS)
  • Mid/Side stereo
  • Intensity stereo
  • Channel coupling
  • Modified Discrete Cosine Transform (MDCT)
  • Filterbank processing

Installation

go get github.com/skrashevich/go-aac

Command-Line Tool

The repository includes a command-line tool for converting AAC files to WAV or raw PCM:

go install github.com/skrashevich/go-aac/cmd/aac2pcm@latest

Usage

# Convert AAC to WAV
aac2pcm input.aac output.wav

# Output raw PCM to stdout
aac2pcm --raw input.aac > output.pcm

# Verbose mode with details
aac2pcm -v input.aac output.wav

See cmd/aac2pcm/main.go for a complete example of how to use the library.

Library Usage

The library provides a simple API for decoding AAC audio. See the aac2pcm implementation for a practical example demonstrating:

1. Probing ADTS data with adts.Probe()
2. Creating a decoder with decoder.New()
3. Reading ADTS headers with adts.ReadHeaderFromBytes()
4. Decoding frames with decoder.DecodeFrame()
5. Converting float32 samples to int16 PCM
6. Writing WAV files

Architecture

The decoder is organized into the following packages:

• decoder: High-level AAC decoder interface
• adts: ADTS header parsing and validation
• ics: Individual Channel Stream processing
• cpe: Channel Pair Element (stereo channels)
• cce: Coupling Channel Element
• huffman: Huffman decoding with spectral tables
• filterbank: IMDCT filterbank processing
• mdct: Modified Discrete Cosine Transform
• fft: Fast Fourier Transform
• tns: Temporal Noise Shaping
• tables: Lookup tables for AAC decoding

Supported Profiles

• AAC Main (partially)
• AAC-LC (Low Complexity) - recommended
• AAC-LTP (Long Term Prediction) (partially)

Limitations

• SBR (Spectral Band Replication) / HE-AAC is not supported
• AAC Main and LTP prediction features are not fully implemented
• PCE (Program Config Element) is not implemented
• Gain control is not implemented

License

This project is licensed under the LGPL v3 license, matching the original AAC.js implementation.

Credits

• Author: Sergei Krashevich
• Based on AAC.js by Devon Govett and the Audiocogs team

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.