indicating that resultsBySpeaker does not work for recognizeMicrophone due to broadband/narrowband mismatch

nfriedly · nfriedly · commit 80542c517220 · 2016-12-15T10:53:28.000-05:00
[ci skip]
diff --git a/speech-to-text/recognize-microphone.js b/speech-to-text/recognize-microphone.js
@@ -49,7 +49,7 @@ var bitBucket = new Writable({
  * @param {Boolean} [options.keepMicrophone=false] - keeps an internal reference to the microphone stream to reuse in subsequent calls (prevents multiple permissions dialogs in firefox)
  * @param {String|DOMElement} [options.outputElement] pipe the text to a [WriteableElementStream](WritableElementStream.html) targeting the specified element. Also defaults objectMode to true to enable interim results.
  * @param {Boolean} [options.extractResults=false] pipe results through a ResultExtractor stream to simplify the objects. (Default behavior before v0.22) Requires objectMode.
- * @param {Boolean} [options.resultsBySpeaker=false] pipe results through a SpeakerStream. Causes each data event to include multiple results, each with a speaker field. Automatically enables objectMode and speaker_labels. Adds some delay to processing.
+ * @param {Boolean} [options.resultsBySpeaker=false] Not currently functional because microphone input provides broadband audio, but during the current beta release, speaker recognition only works on narrowband models.
  *
  * @returns {RecognizeStream|SpeakerStream|FormatStream|ResultStream}
  */

Original file line number	Diff line number	Diff line change
`@@ -49,7 +49,7 @@ var bitBucket = new Writable({`
`49`	`49`	`* @param {Boolean} [options.keepMicrophone=false] - keeps an internal reference to the microphone stream to reuse in subsequent calls (prevents multiple permissions dialogs in firefox)`
`50`	`50`	`* @param {String\|DOMElement} [options.outputElement] pipe the text to a [WriteableElementStream](WritableElementStream.html) targeting the specified element. Also defaults objectMode to true to enable interim results.`
`51`	`51`	`* @param {Boolean} [options.extractResults=false] pipe results through a ResultExtractor stream to simplify the objects. (Default behavior before v0.22) Requires objectMode.`
`52`		`- * @param {Boolean} [options.resultsBySpeaker=false] pipe results through a SpeakerStream. Causes each data event to include multiple results, each with a speaker field. Automatically enables objectMode and speaker_labels. Adds some delay to processing.`
	`52`	`+ * @param {Boolean} [options.resultsBySpeaker=false] Not currently functional because microphone input provides broadband audio, but during the current beta release, speaker recognition only works on narrowband models.`
`53`	`53`	`*`
`54`	`54`	`* @returns {RecognizeStream\|SpeakerStream\|FormatStream\|ResultStream}`
`55`	`55`	`*/`