Skip to main content

Create, read, modify, and write SunVox files.

Project description

Overview of Radiant Voices

Part of the Metrasynth project.

Radiant Voices provides tools to create, read, modify, and write SunVox files. This includes project files ending in .sunvox, and module/synth files ending in .sunsynth.

SunVox data structures and APIs

Radiant Voices strives toward 100% coverage of all data structures used by SunVox files, exposing APIs for creating and manipulating those structures.

Using these APIs, you can do things not possible with the standard SunVox interface or the SunVox DLL, such as:

Our collective imagination is the limit!

Support for multiple languages

Multiple programming languages are supported.

This is accomplished by combining a structured file format specification with a code generation framework.

Supported languages currently include:

  • Python 3.8+

  • Typescript

Interaction with the SunVox DLL

By combining Radiant Voices with sunvox-dll-python for Python (or the SunVox library wrapper for your language), one can also create alternative editing and performance tools to use alongside, or instead of, the official SunVox app.

The two packages work together to provide convenient high-level APIs for loading project and module objects directly into playback slots managed by the SunVox DLL.

Some possibilities might include:

What can you come up with?

SunVox file format documentation

Radiant Voices intends to serve as a de facto source of documentation about the format, as there is currently no official documentation for the SunVox file format.

The interpretation of SunVox file formats is based on a mix of “clean room” style inspection of what SunVox writes to disk when a file is edited a specific way, the most recent BSD-licensed source code for the SunVox audio engine, and consultation with NightRadio (the author of SunVox).

Documentation is provided in the form of prose in the English language, as well as a specification defined in YAML format.

Requirements

  • A supported programming language.

  • OS and platform supported by sunvox-dll-python, if working with the native SunVox DLL.

About SunVox

From the SunVox home page:

SunVox is a small, fast and powerful modular synthesizer with pattern-based sequencer (tracker). It is a tool for those people who like to compose music wherever they are, whenever they wish. On any device. SunVox is available for Windows, OS X, Linux, Maemo, Meego, Raspberry Pi, Windows Mobile (WindowsCE), PalmOS, iOS and Android.

Changelog

1.0.2

Bug fix release.

Fixes (Python version)

  • Avoids rearranging modules during project loading.

1.0.1

Bug fix release.

Fixes (all versions)

  • Adds correct range information for delay controllers in Delay and Echo modules.

1.0.0

The themes of this release are:

  • New JavaScript version of Radiant Voices.

  • Support for new modules, controllers, options, and curves in SunVox 1.9.6c.

  • Improved general compatibility with the SunVox file format.

  • Improved documentation of the SunVox file format.

  • Radiant Voices API improvements.

Major change: genrv code generator

To support porting Radiant Voices to more languages than just Python, and keeping these ports in sync with new versions of SunVox, this release introduces a new code-generation tool, genrv.

genrv is written in Python, and uses Jinja2 templates to generate source code in various target languages, based on a spec written in YAML.

The generated code is run through code-formatting tools, saved directly into the final package structure, and committed to the git repository as any other code would be.

Base classes for all module types are generated this way. They include controllers, options, and controller value enums. The actual classes for each module inherit from these base classes and add module-specific behavior maintained by hand.

Major change: JavaScript port

Radiant Voices now has a JavaScript port, written in TypeScript. It pairs well with the JavaScript/WebAssembly version of the SunVox library.

It is not a 1-to-1 port from the Python version, although it is designed to allow you to do the same things.

As of this release, each version has strengths and weaknesses compared to the other. Over time, we will work towards parity as we strive toward 100% compatibility with SunVox.

Additions (all versions)

  • Adds Project.receive_sync_midi and Project.receive_sync_other flags.

  • Adds ADSR module.

  • Adds Ctl2Note module.

  • Adds Pitch Detector module.

  • Adds harmonics to enum for AnalogGenerator.waveform controller.

  • Adds new values to enum for SpectraVoice.h_type controller: overtones1+, overtones2+, overtones3+, overtones4+, metal.

  • Adds Amplifier.gain controller.

  • Adds AnalogGenerator.true_zero_attack_release option.

  • Adds new values to enum for AnalogGenerator.waveform controller: noise_with_spline_interpolation, white_noise, pink_noise, red_noise, blue_noise, violet_noise, grey_noise.

  • Adds LFO.freq_scale_pct controller.

  • Adds LFO.smooth_transitions controller.

  • Adds Sound2Ctl.send_only_changed_values option.

  • Adds new options to MetaModule: receive_notes_from_keyboard, do_not_receive_notes_from_keyboard.

  • Adds PitchShifter.bypass_if_pitch_eq_0 controller.

  • Adds new value to enum for Compressor.mode controller: peak_zero_latency.

  • Adds new values and value names to enum for Distortion.type controller: clipping is the new name for lim. foldback is the new name for sat. foldback2, foldback3, overflow are new values.

  • Adds new value to enum for Lfo.waveform controller: random_interpolated.

  • Adds DrumSynth.bass_panning controller.

  • Adds DrumSynth.hihat_panning controller.

  • Adds DrumSynth.snare_panning controller.

  • Adds MultiCtl.response controller.

  • Adds MultiCtl.sample_rate_hz controller.

  • Increases the maximum value of Gpio.pin_in and Gpio.pin_out controllers to 256.

  • Renames Kicker.vol_addition controller to Kicker.boost, to reflect naming in SunVox 1.9.4.

Additions (Python version)

  • Ports all test cases from JavaScript version. Note: A limited number of test cases are not yet completely ported.

  • Adds Project.restart_position attribute.

  • Adds Project.detach_module(module) method.

  • Adds patch_decompiler tool.

  • Adds Pattern.project attribute, set once attached to a project.

  • Adds Note.pattern and Note.project to allow notes to be project-aware.

  • Adds Note.module_index property, converts Note.module to 0-based index.

  • Adds Note.mod property, allows setting a note’s module via an actual Module instance (instead of an int).

Changes (All versions)

  • Updates module option reading, writing, and setting to reflect changes in SunVox 1.9.6.

Changes (Python version)

  • Uses dict instead of OrderedDict, as modern Python’s built-in dict maintains key order.

  • Renames incoming_links to in_links.

  • Renames controller_number to ctl_index, and gain_percentage to gain_pct, to more closely reflect the naming in the JavaScript version.

  • Adds type annotations to support static analysis tools and IDEs.

  • Renames “dirty waveform” to “drawn waveform”.

  • Updates default sunvox_version and based_on_version of newly-created Project containers to reflect file format version 1.9.6.1.

  • Updates MetaModule.behaviors to include sends_notes.

  • Now uses black to format all Python modules.

  • Project.attach_module now raises ModuleOwnershipError if the module is already attached to a project.

  • Project.attach_module now re-uses empty module indexes, instead of always appending to the end.

  • Project.connect now raises ModuleOwnershipError if modules do not share a common parent.

  • Python 3.8 is now required.

  • Project.attach_pattern now returns the index of the attached pattern.

Fixes (all versions)

  • Correctly reads and writes SLnK chunks, thus correctly keeping the connection order between modules that have more than one connection going in or out.

  • Renames Sample.loop_end to Sample.loop_len.

  • Fixes how effects embedded into Sampler.effect are serialized.

  • Updates out_controller of Pitch2Ctl, Sound2Ctl, and Velocity2Ctl to have correct range of 0..255.

Fixes (Python version)

  • Fixes writing of controller values to use signed ints instead of unsigned.

  • Fixes reading/writing of VorbisPlayer.finetune controller values.

  • Updates the helloworld example to use correct APIs.

  • Parses chunk IDs in a case-sensitive way, to prevent incorrect parsing of chunks such as SLnK.

0.4.0.dev2 (2018-03-11)

Fixes

  • Corrects a packaging error that included unnecessary cache data from documentation builds.

0.4.0.dev1 (2018-03-11)

Additions

  • Adds documentation about the SunVox file format.

  • Adds equality checking to Range.

  • Adds Module.midi_in_always attribute, defaulting to False. When True, the module will respond to MIDI events regardless of whether it’s selected in the SunVox UI.

  • Adds Module.midi_in_channel attribute, defaulting to 0 (all channels). Set to 1-16 to make the module respond to only a specific MIDI channel.

  • Adds Project.time_grid2 attribute.

  • Adds MultiSynth.curve2_influence controller.

  • Adds MetaModule.event_output option (default: True).

  • Adds MultiSynth.trigger option (default: False).

  • Adds ModuleFlags and VisibleModuleFlags enums for reading/writing Module.flags.

  • Adds PatternFlags and PatternAppearanceFlags enums for reading/writing Pattern.flags and Pattern.appearance_flags.

  • Adds Visualization, LevelMode, Orientation, and OscilloscopeMode for reading/writing the Module.visualization structure.

  • Adds Project.selected_generator attribute.

  • Adds Lfo.Waveform.triangle constant.

  • Adds Lfo.generator controller.

  • Adds Reverb.random_seed controller.

  • Adds Sampler.pitch_envelope and Sampler.effect_control_envelopes[] containing new envelopes from SunVox 1.9.3.

  • Adds Sampler.effect to contain an optional SunSynth instance which in turn contains the effect being modulated by the effect control envelopes.

  • Adds Sampler.Sample.loop_sustain flag.

  • Adds Sampler.ignore_velocity_for_volume option.

  • Adds Container.clone() method.

  • Adds Project.pattern_lines() method, which iterates over a range of project lines and yields information about the active pattern lines for each project line.

  • Adds Pattern.source_method() and PatternClone.source_method(), to determine the source pattern for any given pattern or pattern clone.

Changes

  • Python 3.6 is now required.

  • While writing files, CHFF and CHFR chunks will not be written if they always have a value of 0.

  • AnalogGenerator.unsmooth_frequency_change option is now inverted to .smooth_frequency_change.

  • Updates Sampler and Sampler.Envelope to support SunVox 1.9.3 format. When a pre-1.9.3 formatted Sampler is loaded, it will be upgraded to 1.9.3 format.

  • More detailed exception message when attempting to set an out-of-range value to a controller.

  • Ignores chunk types no longer used by modern SunVox versions: PSYN, PCTL, and PAMD.

  • Ignores value of CHNK when reading module-specific chunks.

  • Does not write the optional -1 to the end of SLNK chunks.

  • Uses the value mapping curve when converting a MultiCtl.value to downstream controllers.

  • Updates MetaModule.play_patterns, which is now of type MetaModule.PlayPatterns instead of bool. This introduces support for the new MetaModule.PlayPatterns.on_no_repeat value.

  • Uses the Fruchterman-Reingold layout algorithm from NetworkX for auto-layout of modules, not PyGraphviz neato algorithm. This affects the arguments accepted by Project.layout().

  • Improves variable names generated from MetaModule user defined controller labels.

Fixes

  • SMIN (module MIDI output device name) is now correctly read and written.

  • SMII and SMIC chunks are now encoded as unsigned int32 (was previously signed).

  • Strings now use UTF-8 encoding.

  • All module types now have a correct default .flags attribute.

  • After MetaModule.update_user_defined_controllers() is called, user defined controllers will have correct value_type set. (This avoids errors such as 44100 being out of the 0..32768 range.)

  • In MetaModule, when a user defined controller mapping points to a non-existent module, the mapping will be ignored. (Was throwing an AttributeError)

  • When Module options are loaded from older projects, assume 0 as a default value if there are not enough bytes. (Was throwing an IndexError)

  • Range validation for Lfo.freq and Vibrato.freq now depends on the value of the frequency_unit controller. Some modules created in earlier versions of SunVox have out-of-range values. These are only warned about using logging, instead of the standard behavior of raising an exception.

  • Writes correct value of CHNK when writing module-specific chunks.

  • Default to signed 8-bit int when a CHFF value was 0.

0.3.0 (2017-04-18)

Additions

  • Adds propagate argument to MultiCtl.reflect(). Defaults to True which causes the new MultiCtl.value to immediately propagate to all mapped controllers, including the one that was just reflected.

    Set to False if you only want to set MultiCtl.value without propagating to mapped controllers.

  • Pass a value for initial when calling MultiCtl.macro() to set and propagate an initial value. Default behavior is to not set a value.

Changes

  • The repr of a CompactRange instance now shows that class name, instead of Range.

Fixes

  • Fixes algorithm for propagating MultiCtl.value changes to mapped controllers.

  • Fixes algorithm for reflecting mapped controllers back to MultiCtl.value.

0.2.0 (2017-04-02)

Additions

  • Adds Controller.pattern_value() instance method, to map a controller’s value to a pattern value in the range of 0x0000-0x8000.

  • Adds ALL_NOTES constant to see if a NOTECMD is a note or a command. (Example: if some_note in ALL_NOTES: ...)

  • Adds tabular_repr() instance methods to Note and Pattern, returning a tabular representation suitable for inclusion in text documents.

  • Adds behaviors attribute to all module classes, describing the types of information each module can send and receive.

  • Adds package-specific exception base classes to rv.errors.

  • Adds support for reading, writing, and modifying controller MIDI mappings.

  • Adds a MultiCtl.macro() static method, for quickly creating a MultiCtl that controls several similar controllers on connected modules.

  • Adds a MultiCtl.reflect() instance method, for setting a MultiCtl’s value based on the destination controller mapped at a given index.

  • Adds # TODO: ... notes to indicate unimplemented features.

  • Allows property-style access to user-defined controllers on MetaModule``s using a ``u_ prefix. For example, if there’s a user-defined controller named “Attack”, it will be accessible via the .u_attack property.

  • Adds ArrayChunk.set_via_fn() method, for setting various curves using the output of a function.

  • Adds DRUMNOTE, BDNOTE, HHNOTE, and SDNOTE enumerations to DrumSynth class, providing note aliases for easier programming of drum sequences.

  • Adds Pattern.set_via_fn() and .set_via_gen() instance methods, for altering a pattern based on the output of a function or generator.

Changes

  • Renames Output module’s module group to "Output".

  • When using Project.layout(), default to using dot layout engine.

  • Uses a direct port of SunVox’s algorithm for mapping MultiCtl values to destination controllers.

  • Uses 1.9.2.0 as SunVox version number when writing projects to files.

  • Allows using separate x/y offsets and factors during Project.layout()

Fixes

  • Uses same sharp note notation as used by SunVox (lowercase indicates sharp).

  • Honor prog keyword arg when passed into Project.layout() method.

  • Does not require pattern x or y to be divisible by 4.

  • Assigns correct controller number to user-defined controllers on MetaModule.

  • Corrects the max value allowed in a MultiSynth velocity/velocity curve.

  • Moves pygraphviz from requirements/base.txt to .../tools.txt to be more Windows-friendly.

0.1.1 (2016-11-09)

  • Fixes upload to PyPI.

0.1.0 (2016-11-09)

  • Initial release.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

radiant-voices-1.0.2.tar.gz (145.8 kB view details)

Uploaded Source

File details

Details for the file radiant-voices-1.0.2.tar.gz.

File metadata

  • Download URL: radiant-voices-1.0.2.tar.gz
  • Upload date:
  • Size: 145.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.3.0 pkginfo/1.5.0.1 requests/2.23.0 setuptools/49.6.0 requests-toolbelt/0.8.0 tqdm/4.54.0 CPython/3.8.5

File hashes

Hashes for radiant-voices-1.0.2.tar.gz
Algorithm Hash digest
SHA256 7a2780ad7ee7f9403002583503945c2362785b472404daff6378d8ca26b41610
MD5 6f7862107b10e255753b2a5348c50109
BLAKE2b-256 832136283385f86d0eac7478853cf60c86e622137d85429ac0b626db7523d701

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page