web-speech-cognitive-services - npm Package Compare versions

Comparing version 5.0.2-master.d388ae9 to 5.0.2-master.e462e9e

CHANGELOG.md

@@ -9,2 +9,10 @@ # Changelog
 
+### Fixed
+
+- Speech recognition: Removed extraneous finalized `result` event in continuous mode, by [@compulim](https://github.com/compulim), in PR [#79](https://github.com/compulim/web-speech-cognitive-services/pull/79)
+
+### Added
+
+- Speech recognition: New `loosenEvents` option, default is `false`. When enabled, we will no longer follow observed browser event order. We will send finalized `result` event as early as possible. This will not break conformance to W3C specifications. By [@compulim](https://github.com/compulim), in PR [#79](https://github.com/compulim/web-speech-cognitive-services/pull/79)
+
 ## [5.0.1] - 2019-10-25
@@ -11,0 +19,0 @@

lib/BingSpeech/SpeechToText/createSpeechRecognitionPonyfill.js

@@ -46,3 +46,3 @@ "use strict";
 
-var VERSION = "5.0.2-master.d388ae9";
+var VERSION = "5.0.2-master.e462e9e";
 
@@ -49,0 +49,0 @@ function buildSpeechResult(transcript, confidence, isFinal) {

lib/SpeechServices.js

@@ -66,4 +66,4 @@ "use strict";
 meta.setAttribute('name', 'web-speech-cognitive-services');
-meta.setAttribute('content', "version=".concat("5.0.2-master.d388ae9"));
+meta.setAttribute('content', "version=".concat("5.0.2-master.e462e9e"));
 document.head.appendChild(meta);
 //# sourceMappingURL=SpeechServices.js.map

lib/SpeechServices/SpeechToText/createSpeechRecognitionPonyfill.js

@@ -140,2 +140,4 @@ "use strict";
       enableTelemetry = _ref3$enableTelemetry === void 0 ? true : _ref3$enableTelemetry,
+      looseEvent = _ref3.looseEvent,
+      looseEvents = _ref3.looseEvents,
       referenceGrammars = _ref3.referenceGrammars,
@@ -157,2 +159,7 @@ _ref3$region = _ref3.region,
 
+  if (typeof looseEvent !== 'undefined') {
+    console.warn('web-speech-cognitive-services: The option "looseEvent" should be named as "looseEvents".');
+    looseEvents = looseEvent;
+  }
+
   var _onAudibleChunk;
@@ -618,11 +625,24 @@
                                   }));
+                                } // If it is continuous, we just sent the finalized results. So we don't need to send it again after "audioend" event.
+
+
+                                if (_this3.continuous && recognizable) {
+                                  finalEvent = null;
+                                } else {
+                                  finalEvent = {
+                                    results: finalizedResults,
+                                    type: 'result'
+                                  };
                                 }
 
-                                finalEvent = {
-                                  results: finalizedResults,
-                                  type: 'result'
-                                };
-
                                 if (!_this3.continuous) {
                                   recognizer.stopContinuousRecognitionAsync();
+                                } // If event order can be loosened, we can send the recognized event as soon as we receive it.
+                                // 1. If it is not recognizable (no-speech), we should send an "error" event just before "end" event. We will not loosen "error" events.
+
+
+                                if (looseEvents && finalEvent && recognizable) {
+                                  _this3.dispatchEvent(new SpeechRecognitionEvent(finalEvent.type, finalEvent));
+
+                                  finalEvent = null;
                                 }
@@ -629,0 +649,0 @@ } else if (recognizing) {

package.json

 {
   "name": "web-speech-cognitive-services",
-  "version": "5.0.2-master.d388ae9",
+  "version": "5.0.2-master.e462e9e",
   "description": "Polyfill Web Speech API with Cognitive Services Speech-to-Text service",
@@ -68,2 +68,3 @@ "keywords": [
     "microsoft-speech-browser-sdk": "^0.0.12",
+    "prettier": "^1.19.1",
     "rimraf": "^2.6.3"
@@ -70,0 +71,0 @@ },

README.md

@@ -5,4 +5,2 @@ # web-speech-cognitive-services
 
-> This scaffold is provided by [`react-component-template`](https://github.com/compulim/react-component-template/).
-
 [![npm version](https://badge.fury.io/js/web-speech-cognitive-services.svg)](https://badge.fury.io/js/web-speech-cognitive-services) [![Build Status](https://travis-ci.org/compulim/web-speech-cognitive-services.svg?branch=master)](https://travis-ci.org/compulim/web-speech-cognitive-services)
@@ -137,5 +135,16 @@
       <td><code>undefined</code></td>
-      <td>Pass-through option to enable or disable telemetry for Speech SDK recognizer as <a href="https://github.com/Microsoft/cognitive-services-speech-sdk-js#data--telemetry">outlined in Speech SDK</a>. This adapter does not collect any telemetry.<br /><br />By default, Speech SDK will collect telemetry unless this is set to <code>false</code>.</td>
+      <td>
+        Pass-through option to enable or disable telemetry for Speech SDK recognizer as <a href="https://github.com/Microsoft/cognitive-services-speech-sdk-js#data--telemetry">outlined in Speech SDK</a>. This adapter does not collect any telemetry.<br /><br />By default, Speech SDK will collect telemetry unless this is set to <code>false</code>.
+      </td>
     </tr>
     <tr>
+      <td><code>looseEvents: boolean</code></td>
+      <td><code>"false"</code></td>
+      <td>
+        Specifies if the event order should strictly follow observed browser behavior (<code>"false"</code>), or loosened behavior (<code>"true"</code>). Regardless of the option, the package will continue to <a href="https://wicg.github.io/speech-api/#eventdef-speechrecognition-result">conform with W3C specifications</a>.
+        <br /><br />
+        You can read more about this option in <a href="#event-order">event order section</a>.
+      </td>
+    </tr>
+    <tr>
       <td><code>ponyfill.AudioContext:&nbsp;<a href="https://developer.mozilla.org/en-US/docs/Web/API/AudioContext">AudioContext</a></code></td>
@@ -151,3 +160,5 @@ <td><code>window.AudioContext&nbsp;||</code><br /><code>window.webkitAudioContext</code></td>
       <td><code>undefined</code></td>
-      <td>Reference grammar IDs to send for speech recognition.</td>
+      <td>
+        Reference grammar IDs to send for speech recognition.
+      </td>
     </tr>
@@ -177,3 +188,3 @@ <tr>
       <td><code>speechSynthesisOutputFormat:&nbsp;string</code></td>
-      <td><code>audio-24khz-160kbitrate-mono-mp3</code></td>
+      <td><code>"audio-24khz-160kbitrate-mono-mp3"</code></td>
       <td>Audio format for speech synthesis. Please refer to <a href="https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/rest-text-to-speech#audio-outputs">this article</a> for list of supported formats.</td>
@@ -434,2 +445,38 @@ </tr>
 
+## Event order
+
+According to [W3C specifications](https://wicg.github.io/speech-api/#eventdef-speechrecognition-result), the `result` event can be fire at any time after `audiostart` event.
+
+In continuous mode, finalized `result` event will be sent as early as possible. But in non-continuous mode, we observed browsers send finalized `result` event just before `audioend`, instead of as early as possible.
+
+By default, we follow event order observed from browsers (a.k.a. strict event order). For a speech recognition in non-continuous mode and with interims, the observed event order will be:
+
+1. `start`
+1. `audiostart`
+1. `soundstart`
+1. `speechstart`
+1. `result` (these are interim results, with `isFinal` property set to `false`)
+1. `speechend`
+1. `soundend`
+1. `audioend`
+1. `result` (with `isFinal` property set to `true`)
+1. `end`
+
+You can loosen event order by setting `looseEvents` to `false`. For the same scenario, the event order will become:
+
+1. `start`
+1. `audiostart`
+1. `soundstart`
+1. `speechstart`
+1. `result` (these are interim results, with `isFinal` property set to `false`)
+1. `result` (with `isFinal` property set to `true`)
+1. `speechend`
+1. `soundend`
+1. `audioend`
+1. `end`
+
+For `error` events (abort, `"no-speech"` or other errors), we always sent it just before the last `end` event.
+
+In some cases, loosening event order may improve recognition performance. This will not break conformance to W3C standard.
+
 # Test matrix
@@ -436,0 +483,0 @@

No alert changes

Improved metrics

Total package byte prevSize

4283397

increased by0.21%

Lines of code

9028

increased by0.21%

Number of lines in readme file

516

increased by10.02%

Worsened metrics

Dev dependency count

14

increased by7.69%

No dependency changes