tflite_audio 0.3.0

Audio classification Tflite package for flutter (iOS & Android). Can support Google Teachable Machine models.

0.3.0

• BREAK CHANGE: Recording bufferSize now takes in 2x the number of samples. To keep the same recording length, simply divide your previous bufferSize by 2.
• Experimental: Support MFCC, melSpectrogram and spectrogram inputs
• Feature: Can automatically or manually set audio length
• Feature: Can automatically or manually transpose input shape
• Improvement: Stability of asyncronous operations with RxJava and RxSwift
• Improvement: (iOS) Removed meta info when extracting data from audio file.
• Improvement: (Android) Splicing algorithm passes all test case. Audio recogntion should now be more accurate.
• Fixed: (iOS) Duplicate symbol error. Set version of TensorFlowLite to 2.6.0. Problem found here.
• Fixed: (Android & iOS) Incorrect padding when splicing audio file. All test cases have passed.

0.2.2+4

• Handled NaN exception for raw output on swift

0.2.2+3

• Hot fixed iOS issue where it will record indefintetly.

0.2.2+2

• Hot fixed missing AudioProcessing class.

0.2.2+1

• Hot fixed issue with unresponsive forced stop recognition.

0.2.2

• Feature: Added ability to recognise stored audio files
• Breaking Change: RecordingLength will no longer be required as a parameter.
• Fixed: NaN output for bufferRates that are non divisible to audioLength
• Fixed: android permission error when granted outside app.

0.2.1+2

• Fixed NaN raw score output for Android.

0.2.1+1

• Fixed inaccurate numOfInference count for iOS and android.

0.2.1

• Improved recognition accuracy for Google Teachable Machine models
• Fixed memory crash on android
• Improved memory performance on iOS
• Added feature to output raw scores
• moved inputType to loadModel() instead of startAudioRecognition()

0.2.0

• Fixed crashed on Android when force stopping recognition
• Improve recognition latency on android by reducing number of event calls.

0.1.9

• Added support for android V2 embedding
• Breaking change - no longer supports deprecated versions of Android (pre 1.12)

0.1.8+2

• Fixed null safety incompatability with example

0.1.8+1

• Fixed the problem with bridge NSNumber to Float
• Merged rawAudioRecognize() and decodedWavRecognize() on native platforms
• Set detection parameters to 0 for better performance.

0.1.8

• Added null safety compatability

0.1.7+1

• Hotfixed iOS crash when casting double to float for detectionThreshold

0.1.7

• Fixed iOS bug where stream wont close when permission has been denied.
• Added feature where you can adjust the detection sensitivity of the model

0.1.6+2

• Fixed podsec error
• Fixed iOS incompatability with fluter 2.0.0

0.1.6+1

• Hotfixed missing value for recording.

0.1.6

• bufferSize no longer needs to be divisible to recording length.

0.1.5+3

• Fixed major android crash, where forcibly stopping the stream causes recorder.stop() to be prematurely called.
• Fixed minor iOS crash, where forcibly stopping the stream during recognition returns a nil exception.
• Cleaned up example for easy switch between decodedWav and Google's Teachable Machine model

0.1.5+2

• Disabled Google's Teachable Machine by default to reduce app footprint. (This can be enabled manually)
• Adjusted example's values to improve inference accuracy

0.1.5+1

• Added documentation
• Added example model from Google's Teachable Machine.
• Fixed iOS crash when loading text file with empty elements.

0.1.5

• Added support for Google Teachable Machine models.
• Fixed inaccurate reading with recording
• Added feature to switch between decodedwav and Google's Teachable machine model.

0.1.4

• Added a new feature where you can run multiple inferences per recording.
• Replaced future with stream when getting results from inferences
• Added a button to cancel the stream / inference
• Removed unnecessary code for easier reading.

0.1.3+1

• Used reusable widgets for easier to read code.
• Added some documentation

0.1.3

• Hotfix for crash when permission has been denied.
• Added the key 'hasPermission' for the future startAudioRecognitions().
• Added feature in example where it'll show inference times

0.1.2

• Instead of returning a single string value, the future startAudioRecognition() now returns a map with the following keys:
  • recognitionResult
  • inferenceTime
• Fixed issue in example where pressing the record button multiple times will crash the app.
• Added feature in example where pressing the recording button changes color.

0.1.1

• Made some fixes with making options explicit
• Added alert dialog when permission is denied.

0.1.0

• Added iOS support

0.0.4

• Added the following arguments into the future: startAudioRecognition()
  • sampleRate
  • audioLength
  • bufferSize

0.0.3

• Merged permission and audio recognition futures into one future.

0.0.2

• Fixed image url

0.0.1

• Initial release.