- Change sample rate type to plain f32 - Remove namespacing of buffer, audio_param and periodic_wave - Reduce AudioRenderQuantum public API surface - Remove ConcreteBaseAudioContext from public API docs
0.19.0
- Added baseLatency and outputLatency attributes - Audio processor callback now has access to AudioWorkletGlobalScope-like env - Performance optimization for sorting large audio graphs - Use default sample rate for output devices instead of highest to prevent insane values - Fixed incorrect channel mixing for AudioNode input buffers
0.18.0
- Implement BaseAudioContext state - AudioContext can now change the number of output channels while running - Microphone input stream is now configurable - Microphone can properly pause, resume and close - Consistently use `usize` for channels, inputs, lengths
0.17.0
- Simplify AudioNode's required channel methods - Apply all channel count/mode/interpretation constraints - AudioContext can now update channel count while playing - Improve AudioContext constructor - Validate more input values for AudioParam events
0.16.0
- AudioBufferSourceNode can now resample - Add MediaStreamAudioDestinationNode - Performance improvement when no PannerNodes are present - More consistent method/argument names - Added benchmark program - Removed MediaElement and MediaElementSourceNode for now, will reimplement
0.15.0
- Allow method chaining on AudioParams - Some fallible methods will now panic instead of returning a Result - Sub sample scheduling for OscillatorNode - Implement AudioContext.close to free resources - Rename BaseAudioContext trait and concrete type - Fix spec deviations for node methods and constructor options and defaults - Rename some functions to `_sync` to denote they do not return a Promise