CL-RemiAudio

CL-RemiAudio is a collection of high-performance audio format I/O routines, codecs, processors/effects, and other audio-related code written entirely in Common Lisp. It is split up into multiple ASDF systems for ease of use.

There is also a version for the Crystal programming language if you'd like available here.

Wanna support this project and the developer? Buy Remilia a coffee on Ko-Fi, or support her through Liberapay.

Features

A list of supported features. Unless specified, the feature is implemented in pure Common Lisp.

• WAV and Au format support
  • Reading and writing of files both all-at-once and as a stream of samples
  • Sample rates from 1hz to whatever your CPU/RAM can handle
  • Channels from 1 to whatever your CPU/RAM can handle
  • Integer and IEEE Floating Point sample support
  • Min/Max sample calculations and RMS calculations for input files
• Encoders/Decoders
  • MPEG-1 layers I, II, and III (e.g. mp3, via libmpg123)
  • FLAC (decoding only)
  • QOA codec
  • Extended QOA Format codec
  • WavPack (via libwavpack)
  • Vorbis (decoding only)
  • YM2610 ADPCM-A codec
  • High quality μ-law and A-law codec
• Processing
  • Bit-depth conversion with optional TPDF dithering
  • Integer -> Floating Point sample conversions
  • High quality resampler (a native port based on libsamplerate)
• Effects
  • Parametric EQ with low/high shelves and unlimited bands.
  • Soft clipping
  • Stereo enhancement
  • Zita-Rev1 reverb effect (native port)
  • MVerb reverb effect (native port)
  • Chorus effect inspired by an 80s synth with model number ending in "-60"
  • DC filter
  • Generic lowpass filter
  • CEM-style filter
  • Wasp-style filter
  • MS-20-style filter
  • SSM-style filter
• Other
  • Windowing functions (Blackman, Blackman-Harris, Hamming, Nuttall, Blackman-Nuttal, Hann)
  • CUE sheet reading and writing
  • XSPF and JSPF reading and writing
  • Raw Vorbis Comments reading and writing
  • Raw APEv1 and APEv2 tag reading and writing
  • ReplayGain calculation support.
• Output
  • Abstract "AUDIO-DRIVER" protocol to smooth-over the differences in various output backends
  • ALSA backend
  • libao backend
  • out123 backend
  • TCP backend - send audio over a TCP connection

Building and Development

A recent version of SBCL is highly suggested (CL-RemiAudio is known to work with SBCL 2.3.x and above).

If you want to enable SIMD code, then SBCL v2.3.x or higher is required. You will also need to push :CL-REMIAUDIO-SIMD into the *FEATURES* variable before compiling/loading CL-RemiAudio.

Packages

The base package is cl-remiaudio is kind of like a "meta package' that loads all of the sub packages. The sub packages can also be loaded individually, and all have nicknames for shorter usage:

• cl-remiaudio-core (cl-ra): Base package containing various common routines.
• cl-remiaudio-codecs (cl-ra/codecs): Base package for all codecs.
• cl-remiaudio-codecs-mpeg1 (cl-ra/mpeg1): MPEG-1 decoder.
• cl-remiaudio-codecs-flac (cl-ra/flac): FLAC decoder.
• cl-remiaudio-codecs-qoa (cl-ra/qoa): QOA and Extended QOA Format encoders and decoders.
• cl-remiaudio-codecs-wavpack (cl-ra/wv): WavPack encoder and decoder.
• cl-remiaudio-dithering (cl-ra/dith): Dithering and sample format conversions.
• cl-remiaudio-dsp (cl-ra/dsp): Various effects/processors.
• cl-remiaudio-replaygain (cl-ra/rg): ReplayGain calculation support.
• cl-remiaudio-replaygain-parallel (cl-ra/rg-parallel): Additional parallel-processing support for ReplayGain calculations.
• cl-remiaudio-formats (cl-ra/formats): WAV and Au file I/O.
• cl-remiaudio-ogg (cl-ra/ogg): Ogg demuxer and Vorbis Comments parser.
• cl-remiaudio-cue (cl-ra/cue): CUE file reading and writing functions.
• cl-remiaudio-xspf (cl-ra/xspf): XSPF and JSPF file reading and writing functions.
• cl-remiaudio-ape (cl-ra/ape): Raw APEv1 and APEv2 reading and writing functions.
• cl-remiaudio-drivers (cl-ra/drivers): Base package for backend audio output drivers.
• cl-remiaudio-drivers-ao (cl-ra/driver-ao): Audio output via libao.
• cl-remiaudio-resampler (cl-ra/resamp): Fast, high quality resampler.

Style info

I use a slightly different style for my code.

• Keep lines 118 characters or shorter. Obviously sometimes you can't, but please try. Use 80 characters for Markdown files, though.
• New type declarations should have their names prefixed with T/, for example, t/uint32-vector. This is only for DEFTYPEs and DEFINE-PSEUDO-ENUMs.

How do I contribute?

1. Go to https://chiselapp.com/user/MistressRemilia/repository/CL-RemiAudio/ and clone the Fossil repository.
2. Create a new branch for your feature.
3. Push locally to the new branch.
4. Create a bundle with Fossil that contains your changes.
5. Get in contact with Remilia.

Contributors

• Remilia Scarlet - creator and maintainer
  • Homepage: https://remilia.sdf.org/
  • Fediverse: @remilia@nanako-さめ.mooo.com
  • Email: zremiliaz@postzeoz.jpz My real address does not contain Z's

Links and Licenses

• MVerb by Martin Eastwood (GPLv3 License)
• Zita-Rev1 by Fons Adriaensen (GPLv3 License)
• YK Chorus by Spotlightkid (GPLv2 License)
• libsamplerate by Erik de Castro Lopo (2-clause BSD License)
• The ReplayGain calculation code is based on code from WavPack (3-clause BSD License)
• The Vorbis decoder is a pure Common Lisp port of the Public Domain stb_vorbis code.

CL-RemiAudio itself is mostly under the GNU Affero General Public License version 3, except for the items listed above. Copies of the licenses are in the licenses/ directory.