MMIR Framework

Class: WebspeechAudioInput

WebspeechAudioInput

new WebspeechAudioInput()

env/media/webspeechAudioInput.js, line 4

Members

_pluginName

_prevResultString

field for storing the previous (main) recontion result (this is used for calculating "unstable" parts, see helper_extract_results)

aborted

active

currentFailureCallbackfunction

currentSuccessCallbackfunction

DEFAULT_ALTERNATIVE_RESULTS

default_error_functionfunction

DEFAULT_LANGUAGE

constantEVENT_RESULT_FIELD

constantEVENT_SCORE_FIELD

final_recognition_result

helper_error_handlerfunction

intermediate_results

loggermmir.Logger

max_error_retry

Maximal number of errors-in-a-row for trying to restart recognition in repeat-mode.
Default Value:
  • 5
See:
  • error_counter

micLevelsImplFile

pluginExportswebspeechAudioInput

recognitionSpeechRecognition

recording

constantRESULT_TYPESEnum

Result types (returned by the native/Cordova plugin)

SpeechRecognitionImplSpeechRecognition

constantUNSTABLE_LIMIT

Methods

cancelRecognition()

env/media/webspeechAudioInput.js, line 943
See:

default_error_function()

env/media/webspeechAudioInput.js, line 408

helper_error_handler(){Boolean}

env/media/webspeechAudioInput.js, line 334
default helper for error-events: determines, if RESTART is allowed/possible (in case of RECORDing mode), AND otherwise triggers the current failure-callbacks. SIDE-EFFECTS: sets private field aborted:=true if RESTART is NOT possible.
Returns:
Type Description
Boolean true, if the function could process the error (i.e. return false for unknown errors; these should be handled by the invoking code of this helper function)

helper_extract_results()

env/media/webspeechAudioInput.js, line 192
create callback-arguments for ASR-result callback:
Returns:
Array with [ String result, Number score, String type ["INTERIM" | "FINAL" ], Array alternatives, //OPTIONAL String unstable //OPTIONAL ]

initialize()

env/media/webspeechAudioInput.js, line 8

recognize()

env/media/webspeechAudioInput.js, line 834
Start speech recognition with end-of-speech detection: the recognizer automatically tries to detect when speech has finished and triggers the status-callback accordingly with results.

NOTE: no end event, if recognize() is stopped via stopRecord()

See:

asyncstartRecord(options, statusCallback, failureCallback)

env/media/webspeechAudioInput.js, line 585
Start speech recognition (without end-of-speech detection): after starting, the recognition continues until stopRecord is called.
Name Type Description
options PlainObject optional OPTIONAL options for Automatic Speech Recognition: 
{
			  success: OPTIONAL Function, the status-callback (see arg statusCallback)
			, error: OPTIONAL Function, the error callback (see arg failureCallback)
			, language: OPTIONAL String, the language for recognition (if omitted, the current language setting is used)
			, intermediate: OTPIONAL Boolean, set true for receiving intermediate results (NOTE not all ASR engines may support intermediate results)
			, results: OTPIONAL Number, set how many recognition alternatives should be returned at most (NOTE not all ASR engines may support this option)
			, mode: OTPIONAL "search" | "dictation", set how many recognition alternatives should be returned at most (NOTE not all ASR engines may support this option)
			, eosPause: OTPIONAL "short" | "long", length of pause after speech for end-of-speech detection (NOTE not all ASR engines may support this option)
			, disableImprovedFeedback: OTPIONAL Boolean, disable improved feedback when using intermediate results (NOTE not all ASR engines may support this option)
		}
statusCallback function optional OPTIONAL callback function that is triggered when, recognition starts, text results become available, and recognition ends. The callback signature is: 
				callback(
					text: String | "",
					confidence: Number | Void,
					status: "FINAL"|"INTERIM"|"INTERMEDIATE"|"RECORDING_BEGIN"|"RECORDING_DONE",
					alternatives: Array<{result: String, score: Number}> | Void,
					unstable: String | Void
				)
Usually, for status "FINAL" | "INTERIM" | "INTERMEDIATE" text results are returned, where 
			  "INTERIM": an interim result, that might still change
			  "INTERMEDIATE": a stable, intermediate result
			  "FINAL": a (stable) final result, before the recognition stops
If present, the unstable argument provides a preview for the currently processed / recognized text.
NOTE that when using intermediate mode, status-calls with "INTERMEDIATE" may contain "final intermediate" results, too.
NOTE: if used in combination with options.success, this argument will supersede the options
failureCallback function optional OPTIONAL callback function that is triggered when an error occurred. The callback signature is: callback(error)
NOTE: if used in combination with options.error, this argument will supersede the options
See:

stopRecord()

env/media/webspeechAudioInput.js, line 720
See: