Skip to main content

earshot_classifier_darwin 1.0.0 copy "earshot_classifier_darwin: ^1.0.0" to clipboard
earshot_classifier_darwin: ^1.0.0 copied to clipboard

PlatformiOSmacOS

The iOS and macOS implementation of earshot_classifier, backed by Apple's on-device SoundAnalysis.

Changelog #

1.0.0 #

  • First stable release. The public API follows semantic versioning from now on: breaking changes only in a major version (ADR 0021).
  • Breaking: SoundAnalysis failures are delivered as a ClassifierException instead of a CaptureException with CaptureErrorCode.unknown: unsupported when SoundAnalysis cannot run on the OS or at the sample rate, and modelUnavailable when it fails while classifying (ADR 0024).

0.2.0 #

  • SoundAnalysis now classifies PCM from earshot_capture, so classification and capture share one microphone tap (ADR 0018). The plugin no longer runs an AVAudioEngine, reads the microphone permission or touches the iOS audio session.
  • Add classify(pcm, config), which accepts every sample rate. detections(config) classifies 16 kHz PCM from earshot_capture, so it fails with sampleRateConflict while capture runs at another rate.
  • Detection timestamps follow the PCM clock and jump over interruptions. Capture errors reach the stream unchanged.
  • Remove the classifier's copy of EarshotAudioSession.swift.

0.1.0 #

  • Follow the new EarshotClassifierPlatform API (isAvailable, detections(ClassifierConfig)).
  • Implement the Apple SoundAnalysis backend on iOS 15+ and macOS 12+: SNClassifySoundRequest (version 1) on an AVAudioEngine input tap, through Pigeon. auto now picks it on iOS and macOS.
  • Each listen() gets its own analyzer session; concurrent sessions share one input tap. Cancelling the last one releases the microphone.
  • Stream errors are CaptureExceptions (permissionDenied, audioUnavailable, unknown). Classification never prompts for the microphone permission.
  • Handle interruptions as earshot_capture does: the engine restarts after calls, Siri, media services resets and device changes.
  • On iOS, share the app's audio session with the other earshot plugin: stopping capture or classification no longer stops the other (ADR 0017).

0.0.1 #

  • Initial API preview.
0
likes
160
points
380
downloads

Documentation

API reference

Publisher

unverified uploader

Weekly Downloads

The iOS and macOS implementation of earshot_classifier, backed by Apple's on-device SoundAnalysis.

Repository (GitHub)
View/report issues
Contributing

Topics

#audio #sound-classification

License

MIT (license)

Dependencies

earshot_capture_platform_interface, earshot_classifier_platform_interface, flutter, meta

More

Packages that depend on earshot_classifier_darwin

Packages that implement earshot_classifier_darwin