[RAW-dee-oh] (rhymes with "audio")
An additive synthesizer for expressive music.
This is a synth designed specifically for Monic Theory and is intended to be applied in non-real time to render monic playbooks as audio files.
The application currently accepts Monic playbooks as input. (See test-druidic-score.json).
As Monic Theory is a superset of MIDI, raudio is also capable rendering any MIDI composition.
raudio supports inline syn-midi (a lightweight MIDI alternative). It does not accept .mid files, but it could accept .mid as input if that is important to you.
This project has a few goals. It also has some explicit no-gos of conventional audio tasks. It also features demos!
Some code writes audio files to the audio/demo directory. This is intended to represent applied music. Music which many people consider "good".
Current demos:
To render the demos, use
cargo test demo
You will then have files output to directory audio/demo! Play the files with your favorite audio player, such as with sox:
play audio/demo/beat.wav repeat 3 reverb 20
Some code writes audio files to the audio/music directory. This is for original (new) music!
An example application for this synthesizer, here is a video with four hours of music rendered in a "Baroque" style.
https://www.youtube.com/watch?v=mFipUHqXrw0
The Primary Objective is to have a synthesizer that creates dynamic harmonic content for any musical context.
All of the music rendered by raudio should feel natural or organic. It should "sound good" to a lot of people.
The supplementary objective is to provide a high level API for describing sound as a small (manageable) set of input parameters.
As a list,
- Provide a high level API for creating and modulating sound in the frequency domain
- Provide a secure and predictable environment for rendering audio files
- Aliasing is easy to manage. Just don't include the frequencies you don't want to render!
- Support for per-sample paramter render
- Highest resolution output with respect to synth control parameters. In other words, all computations are a-rate (no k-rate).
- Provide support for time domain signal analysis (but not modulation)
- Melt your heart and glowup your soul with lush sounds
- No support for creating compositions (bring your own score).
- No support for time domain signal modulation
- "Conventional" lowpass and highpass filters not welcome here!
- Reverb? more like stretchy harmonics
To run tests:
cargo test
If it didn't already exist, this will crate a new directory dev-audio for writing test results.
Many tests will create files. It is intentional that the files are not deleted so the tester can manually inspect the generated asset at their leisure.
Please make sure you have 1-3GB of disk space available before running the tests.
You can remove the assets by using rm -rf dev-audio.
A lot of tests will render audio samples featuring a speciifc feature of synthesis (ex. amplitude modulation, sequencing notes, stacking chords).
Some tests may be non-deterministic in that they produce a variant in the set of all available frequency spaces of the input test.
A new and possibly different version of the "same" sample will overwrite the previous version.
Some tests take a long time to run because they enumerate all the available parameters.
This repository is licensed under a custom license that encourages contribution and education.
Contributors who have made significant contributions by completing an approved GitHub issue or an approved independent pull request (PR) are granted additional ability to use the Software for commercial purposes. For more details, see the LICENSE file.
Contributions are welcome and encouraged.
Non-contributors are permitted to use and view the code for personal or educational purposes only. Commercial use is strictly prohibited for non-contributors.