SDK Release Notes

iOS SDK


Release 2.2.3 (01/06/2017)

  • Improved documentation with sample app to indicate how to free up network resources when application goes in the background
  • Fixed Wait for the start of speech to trigger a recognition transaction
  • Fixed Error types and error codes are exposed in onError callbacks

Release 2.2.2 (20/10/2016)

  • Fixed Swift bindings

Release 2.2.1 (30/09/2016)

  • Ability to set subscriber_id in SKSession options
  • Fixed issue where 64 consecutive ASR transactions would freeze your app

Release 1.4.21 (10/12/2015)

  • Fixed bug preventing Sample App from opening on iOS 9
  • iOS 9 bitcode enabled in SpeechKit

Release 1.4.19 (06/24/2015)

  • Improved documentation for SSL hostname and keychain verification
  • Fixed bug on iOS 8 where speech recognition would prevent another audio channel which was paused from restarting

Release 1.4.14 (03/19/2015)

  • Added support for SSL hostname and keychain verification

Release 1.4.12 (09/11/2014)

  • Update to fix an issue compiling for ARM64 and XCode 6
  • Update to fix an issue related to arm64 compatibility for simulators and a bug where TTS audio would not play when the ringer was muted

Release 1.4.10 (04/29/2014)

  • Update to fix an issue related to arm64 compatibility for simulators

Release 1.4.9 (11/20/2013)

  • SpeechKit is now compatible with ARM64 architectures in Xcode
  • Fixed an issue where background audio, such as from the music app, would be stopped when initializing SpeechKit. It now continues playing after a 1s interruption
  • Fixed an issue where sending an app to the background after doing a recognition with the SKRecognizer class would interrupt the background music when bringing back the app to the foreground
  • Lacking the microphone permission on iOS 7 will now give the message "Microphone input permission refused - will record only silence" rather than a generic "Speech not recognized"
  • Updated the Xcode deployment target from 3.0 to 4.3
  • Releases between 1.4.5 and 1.4.9 internal only

Release 1.4.5 (10/4/2012)

  • Fixed issue that background music is not stopped after SpeechKit started in iOS6 environment

Release 1.4.4 (internal release)

  • Remove SpeechKit setter override compiling warning

Release 1.4.3 (internal release)

  • Build for ARM v7s in Xcode 4.5 this adds support for the ARM v7s instruction set and resolves error messages generated in Xcode 4.5 in previous releases

Release 1.4.2

  • Save UUID into keychain instead of app storage. UUID will be same even app is reinstalled

Release 1.4.0

  • Replaced use of iPhone Unique Device ID with an anonymous per-application UUID generated the first time each application is run
  • Improved handling of Bluetooth headsets
  • Fixed issue with VoiceOver

Release 1.3.2

  • Updated audio system for iOS Updates to documentation and sample code

Release 1.3.1

  • Fixes a bug for a specific use case where repeated losing of the network connection while doing recognition will eventually crash the SpeechKit

Release 1.3.0

  • Fixed a few timing issues related to audio recording
  • Correctly exposed the SpeechError.Codes values
  • Added SSL support

Release 1.2.1

  • Implemented new audio management system resulting in faster initialization/audio record start/stop
  • Increased support for backgrounding (i.e. if the home key is pressed during a recording, there are no longer any ill effects)
  • Increased support for background audio (i.e. recording and playback now coexist more effectively with iPod music playback)
  • Implemented new [Speechkit destroy] function and an associated [SpeechKitDelegate destroyed] callback enabling users to tear down the framework and re-call [SpeechKit setupwithID] if necessary
  • Additional memory management improvements

Release 1.1.2

  • Updated release

Release 1.1.1

  • Patches minor memory leaks in the "setupWithID()" and "initWithType()" methods that occur when the connection to the network servers is initiated

Release 1.1

  • Added SKEarcon class and setEarcon:forType: method to provide automatic earcon playback on recording, finish and cancel stages of the recognition process
  • Added custom data property to SKRecognition for custom server-end models and processing
  • Alternatives list only lists results other than the top item in Recognizer demo app

Release 1.0

  • Ensure all appropriate delegate messages are sent for all vocalizer requests Support for simulator
  • Original string pointer passed to speakString is provided in delegate callbacks for direct pointer comparison Delegate error methods are called in response to cancel
  • Delegate error methods report no connection
  • Updated documentation to reflect new APIs
  • Updated documentation to reflect recent changes to sample applications

Release 0.9.6

  • Recognition sample app allows selection of Search or Dictation recognizer types

Release 0.9.5

  • Add sessionID method to SpeechKit class to return session ID
  • Add audioLevel property to SKRecognizer class to provide on demand access to audio power level
  • Add VUMeter to Recognizer demo app, providing a sample usage of the audioLevel property

Release 0.9.4

  • Release SDK

Android SDK


Release 2.2.4 (01/06/2017)

  • Improved documentation with sample app to indicate how to free up network resources when application goes in the background
  • Fixed OutOfMemoryException when performing several requests in batch
  • Fixed Wait for the start of speech to trigger a recognition transaction
  • Fixed Stop/pause are functional

Release 2.2.3 (10/04/2017)

  • Fixed Enable obfuscation

Release 2.2.2 (26/03/2017)

  • Fixed Error types and error codes are exposed in onError callbacks
  • Fixed Retry result is returned when progressive results were enabled and poor audio was sent

Release 2.2.1 (30/09/2016)

  • Ability to set subscriber_id in Transaction.Options

Release 1.4.12 (06/24/2015)

  • Provided guidance in Dev Guide on how to use SSL hostname and keychain verification
  • Added support for x86 and armeabi-v7a

Release 1.4.11

  • Added support for SSL hostname and keychain verification

Release 1.4.7 (11/20/2013)

  • Fixed an issue that prevented SpeechKit from working correctly on the HTC One V device
  • Fixed a latency issue when starting recognition on Android 4.2 phones when Bluetooth is enabled
  • Added a SpeechKitReleaseListener and a setReleaseListener method to SpeechKit.java to allow checking if the SpeechKit instance has been released
  • Releases between 1.4.2 and 1.4.7 internal only

Release 1.4.2

  • Added "BLUETOOTH" permission to the SpeechKit sample application

Release 1.4.1

  • Fixed the "java.lang.RuntimeException Can't create handler inside thread that has not called Looper.prepare()" problem on calling "BluetoothAdapter.getDefaultAdapter()" function inside our non-looper message system thread (only for 3.1 devices, apply the same fix for other 3.x devices as well)

Release 1.4.0

  • Replaced use of Android getDeviceID() with an anonymous per-application UUID generated the first time each application is run
  • Improved handling of Bluetooth headsets

Release 1.3.1

  • Updated documentation and sample code

Release 1.3.0

  • Fixed a few timing issues related to audio recording
  • Correctly exposed the SpeechError.Codes values
  • Added SSL support

Release 1.2.1

  • Implemented enhancement to Vocalizer.Listener.onSpeakingBegin() enabling it to be called when audio is actually being spoken, rather than when the player starts playing (potentially silence)
  • Implemented corrections for circumstances in which Recognizer or Vocalizer is terminated immediately after starting Improved recorder performance on certain devices
  • Additional minor corrections and optimizations

Release 1.1

  • Initial public release of Android Speech Kit
  • Recognition features:
    1. Three endpointer modes: disabled, short silence, or long silence
    2. Earcon and vibration prompts configurable for "listening", "done listening", "result", and "error"
    3. Two recognition types: dictation and search
  • Vocalization features:
    1. Plain text or SSML formatted text
    2. Easy queuing of multiple requests
    3. Configurable languages and voices
  • Sample Voice Application features:
    1. Dictation activity demonstrates basic recognition (dictation or search)
    2. TTS activity demonstrates basic vocalization (English only, Tom or Samantha voices)
  • Supports the same major features as the iPhone Speech Kit 1.1.0

Windows Phone 7 SDK


Release 1.4.0

  • Replaced use of WinPhone 7 "DeviceUniqueId" value with an anonymous per-application UUID generated the first time each application is run
  • Brought version number in-line with Android and iOS versions to indicate comparable feature set

Release 1.0.2

  • Updated documentation and sample code

Release 1.0.1

  • Initial release

Windows Phone 8 SDK


Release 1.4.0

  • Initial release
  • Bluetooth not supported (WP8 does not support streaming BT audio into applications)