Skip to content
This repository has been archived by the owner on May 29, 2023. It is now read-only.

Latest commit

 

History

History
111 lines (84 loc) · 4.12 KB

README.md

File metadata and controls

111 lines (84 loc) · 4.12 KB
Raw

Attention! This repository is archived and the library has been moved to tinkoff/ng-web-apis monorepository

ng-web-apis logo Web Speech API for Angular

Part of Web APIs for Angular

npm version npm bundle size Travis (.com) Coveralls github angular-open-source-starter

This is a library to use Web Speech API with Angular.

Install

If you do not have @ng-web-apis/common:

npm i @ng-web-apis/common

Now install the package:

npm i @ng-web-apis/speech

How to use

Web Speech API consists of speech synthesis and speech recognition.

  1. Use SpeechSynthesisModule to gain access to TextToSpeechDirective and UtterancePipe. Use them like the example below for speech synthesis functionality:
<textarea
    class="textarea"
    [(ngModel)]="text"
    [waTextToSpeech]="text | waUtterance: options"
    [waTextToSpeechPaused]="paused"
    (waTextToSpeechEnd)="onEnd()"
/>
  1. For speech recognition use SpeechRecognitionService in supporting browsers (only Chrome at this point)

Operators

SpeechRecognitionService provides access to speech recognition in familiar RxJS Observable model. To work with the stream there are certain operators included in this library:

  1. confidenceAbove to filter recognitions to desired level of confidence
  2. continuous to enable continuous mode of recognition
  3. final to only include final recognition results
  4. firstAlternative to quickly retrieve first alternative (which typically is the only one anyway)
  5. skipUntilSaid to ignore stream until certain phrase is said
  6. takeUntilSaid to stop listening to stream upon certain phrase
  7. isSaid utility function to check if a phrase is in SpeechRecognitionResult[]

You may want to use repeat() and retry() in your stream to restart speech recognition. It is stopped after some time and error is thrown if nothing was said for a while.

Here are a few examples:

// Record speech after "Okay Angular" is said
this.stream$ = this.speechRecognition$.pipe(
    retry(),
    repeat(),
    skipUntilSaid('Okay Angular'),
    takeUntilSaid('Thank you Angular'),
    repeat(),
    final(),
    continuous(),
);
// Fire photon torpedoes with a voice command
this.torpedoes$ = this.speechRecognition$.pipe(
    retry(),
    repeat(),
    filter(isSaid('Fire photon torpedoes')),
);

Tokens

There are also a couple of tokens included in this library:

  1. SPEECH_RECOGNITION_MAX_ALTERNATIVES to configure number of alternatives presented in SpeechRecognitionResult
  2. SPEECH_RECOGNITION_SUPPORT to check if browser supports speech recognition
  3. SPEECH_SYNTHESIS_SUPPORT to check if browser supports speech synthesis
  4. SPEECH_SYNTHESIS_VOICES to get the list of available voices for speech synthesis

See also

Other Web APIs for Angular by @ng-web-apis

Open-source

Do you also want to open-source something, but hate the collateral work? Check out this Angular Open-source Library Starter we’ve created for our projects. It got you covered on continuous integration, pre-commit checks, linting, versioning + changelog, code coverage and all that jazz.