Skip to content
This repository has been archived by the owner on May 6, 2022. It is now read-only.

Releases: spokestack/spokestack-ios

Spokestack ASR, Improved Error Messages, Partial ASR Results

17 Sep 21:08
928679c
Compare
Choose a tag to compare
  • New Spokestack ASR speech pipeline component
  • Improved error messages that include error description instead of enum index
  • ASR components deliver partial transcript results via new optional recognizedPartial API
  • stopSpeaking API for TTS

Internal VAD improvements

04 Sep 18:34
9378009
Compare
Choose a tag to compare
13.1.4

release 13.1.4

Fix for Apple ASR sending two recognize events

01 Sep 15:50
90bd3b9
Compare
Choose a tag to compare

Also improves the performance of tflite wakeword encoder.

Fix com.apple.coreaudio.avfaudio error -10868 in AppleSpeechRecognizer

28 Aug 17:16
2df1639
Compare
Choose a tag to compare

fix duplicate double recognize/deactivate events; restore timeout event in AppleSpeechRecognizer

20 Aug 19:59
5634a1f
Compare
Choose a tag to compare

Fix tracing in pipeline builder and correct tracing bug in TFLiteWakewordRecognizer

19 Aug 13:40
31c6cb4
Compare
Choose a tag to compare
  • fix bug in pipeline builder that prevented tracing level from being set
  • fix incorrect equality check for some TFLiteWakeword tracing

SpeechContext is required in the SpeechPipeline stages constructor

17 Aug 18:35
8d6904e
Compare
Choose a tag to compare

New Speech Pipeline featuring stage components and configuration profiles

13 Aug 15:03
4760a12
Compare
Choose a tag to compare

This is a major release that features:

  1. Testability for previously under-tested components
  2. The pipeline utilizes stages that are completely configurable
  3. The pipeline may be instantiated from a set of profiles that encapsulate configuration complexity

See https://www.spokestack.io/docs/iOS/speech-pipeline for how to utilize the new stages and profiles!

Clients are no longer required to declare use_frameworks! in their Podfile.

09 Jun 21:28
6d7b22f
Compare
Choose a tag to compare

fix: update filter_audio dep to 0.5.0 to fix the cocoapods warning: The Swift pod depends upon , which does not define modules.

feat: Clients are no longer required to declare use_frameworks! in their Podfile.

feat: llvm can optimize Spokestack framework loading as a static library.

Faster NLU classification with additional slot information

03 Jun 22:00
210eaf2
Compare
Choose a tag to compare

feat: nlu classification result includes raw input in slot
feat: nlu classification output included all slots defined by the classified intent

refactor: remove wakePhrases from SpeechConfiguration
refactor: set default for nluMaxTokenLength to 50 for objc interop

fix: NLU returns an empty slot instead of crashing