Version 0.82, December 26th, 2018
- Enhancement Updates to most recent iOS (v1.63) and Android (v0.82) SDKs.
- Bug Fix Android: RecognizerState was always returing RecognizerStateNeedsDecodingGraph in Android plugin for Unity, reglardless of the real state of the recognizer.
- Bug Fix iOS: background/foreground state change of the app was triggering either crash or problems with audio playing. The root cause of this problem was Unity bug, which has been resolved in Unity 2018.1. iOS plugin has been updated to take advantage those fixes in Unity.
Version 0.81, June 19, 2018
- Enhancement Updates to iOS and Android SDKs to support new ASR Bundle
- Enhancement Changed default ASR Bundle that comes with the plugin to keenB2mQT-nnet3chain-en-us, which is robust in noisy environments.
Version 0.8, May 30, 2018
- Enhancement Plugin now uses the latest version of KeenASR SDK for Android (0.8) which has many optimizations and some bug fixes.
- Enhancement Changed default ASR Bundle that comes with the plugin to keenB1-nnet3chain-en-us, which has lower CPU and memory requirements.
Version 0.7, April 19, 2018
- Feature Plugin now supports Android platform.
- Enhancement Plugin now uses the latest version of KeenASR SDK for iOS (1.61).
- Enhancement ASR Bundle is stored in Assets/StreamingAssets directory so that it’s automatically included in the app on all platforms.
- Enhancement A more compact version of the librispeech ASR Bundle is now included with the plugin, reducing on-disk size to about 38MB (26MB when compressed).
- Feature Plugin now provides a callback which is triggered when initialization is completed. NOTE: while this change currently does not affect developers who are using plugin only for iOS platform (since iOS plugin will still init synchroniouslly), we recommend you modify the logic so that all post-init work is done in the callback.
Version 0.6, November 24, 2017
- Bug Fix We introduced regression bug in version 0.4 where InputLevel() was always reporting 0 value.
Version 0.5, November 2, 2017
- Feature Support for automatic upload of speech recognition results and audio recordings to Dashboard, our cloud-based development tool. See CreateDataUploadThread() and SetRemoveDataAfterUpload(), PauseUpload(), and ResumeUpload(), SetRemoveDataAfterUpload() static methods for details. Typically, you will create data upload thread after you initialized the recognizer. You will also need to make sure that recognizer is setup to create audio files via SetCreateAudioRecordings(true). Dashboard cloud tool is currently available only to our customers.
Version 0.4, Septembar 26, 2017
Enhancement We reduced the memory footprint of the recognizer by about 20%. In addition, version 1.3 of the SDK supports quantized ASR Bundles which take 3-3.5x less disk space. Currently, this 3-3.5x reduction in ASR Bundle size only applies for on-disk storage (i.e. affecting your app download size). Future releases will follow similar reduction in memory footprint.
Enhancement We introduced recognizerState readonly property for KeenASR instances, which you can access via GetRecognizerState method. This is replacing IsListening method which was not providing sufficient information.
Feature you can now direct SDK to perform echo cancellation, if available on the device, to remove audio played by the app speaker from the signal captured by the microphone. For details see PerformEchoCancellation: and IsEchoCancellationAvailable methods of the KeenASR class. This feature is experimental.
Bug Fix Cleanup of the audio interruption logic that was causing crashes on older iOSes and prevented compilation with older versions of XCode. The current implementation takes complete control of AVAudioSession management and provides two callback methods that allow developers to: 1) cleanup audio playing resources (stop playing audio, remember app state as necessary) before the app goes to the background. This allows us to properly deactivate AVAudioSession before app goes to background, 2) setup UI, etc. after audio interrupt is over and the app comes to foreground. Currently, when the app is about to go to the background the SDK will deactivate audio session, and then reactivate it when the app comes back to the foreground. See RecognizerReadyToListenAfterInterruptHandler callback method, which you can use to setup UI after the app comes to the foreground.
Bug Fix SDK crashes if PrepareForListeningWithCustomDecodingGraph is called while the app is listening. This is now fixed.
Version 0.3, August 14, 2017 **
Enhancement Added PostBuildProcess.cs script which automates XCode settings updates
Enhancement Added stub methods for the Editor build, so that you can compile editor version of your app (KeenASR SDK will not function, but you will be able to create builds)
Enhancement Reorganized structure of the plugin, so that everything related to KeenASR plugin is stored in
Enhancement Added default ASR Bundle in the plugin so that it does not need to be downloaded separately
Enhancement Changed SetLogLevel method to be a static method of KeenASR class. This allows you to set the log level before the framework is initialized
Bug Fix Fixed issue with KeenASR class (and sharedInstance) persisting across multiple scenes