Inside Macintosh: Speech Recognition Manager /

Glossary

active language model The language model that is currently associated with a recognizer. You can call the SRSetLanguageModel function to set the active language model.

Apple event A high-level event that adheres to the Apple Event Interprocess Messaging Protocol. An Apple event consists of attributes (including the event class and event ID, which identify the event and its task) and, usually, parameters (which contain data used by the target application for the event).

Apple event handler An application-defined function that extracts pertinent data from an Apple event, performs the action requested by the Apple event, and returns a result.

Apple Event Manager The collection of routines that allows client applications to send Apple events to server applications for the purpose of requesting services or information.

Backus-Naur Form (BNF) A standard form for representing formal grammars or other language models.

BNF See Backus-Naur Form.

callback routine See speech recognition callback routine.

callback routine parameter structure A data structure that contains information about a speech recognition callback routine. A callback routine parameter structure is defined by the SRCallBackParam data type.

cleanup mode A property of a recognition system that determines whether the recognition system and all other objects it has created are disposed of when your code exits.

container See container object.

container object Any speech object that contains distinguishable subitems. For example, a phrase may contain distinguishable words.

embedded language model A language model that occurs in the path or set of paths that define some other language model. Compare top-level language model.

embedded speech command In a buffer of input text, a sequence of characters enclosed by command delimiters that provides instructions to a speech synthesizer.

feedback and listening modes A property of a recognition system that determines whether a feedback window is displayed for any recognizer associated with that system and whether the recognizer uses the listening modes selected by the user in the Speech control panel.

feedback character The picture (often of a head) displayed in a feedback window.

feedback services A feature of the Speech Recognition Manager that provides a standard set of cues and responses during speech recognition and the ability for your application to alter those cues and responses.

feedback window A floating window that provides audiovisual cues and responses to a PlainTalk user. A feedback window is divided into a status pane and a transcript pane. See also feedback character, output text, recognized text.

key word The word that must precede utterances when the recognizer is in toggle-listen mode.

key word mode See toggle-listen mode.

language model A language object that represents a list of zero or more words, phrases, or paths. A language model is defined by the SRLanguageModel data type. See also embedded language model, top-level language model.

language object Any speech object that belongs to a subclass of the SRLanguageObject class. See also language model, path, phrase, word.

listening mode indicator A word or phrase that indicates which keyword must be uttered or which key must be held down in order for a recognizer to start listening. The current listening mode indicator appears at the bottom of the status pane in a feedback window.

listen key The key (or key combination) that either must be held down in order for a recognizer to start listening or that, when pressed, toggles listening on and off.

listen key mode A recognizer property that determines whether the listen key operates in push-to-talk or toggle-listening mode.

notification See recognition notification.

object See speech object.

object reference See reference.

output text Any text drawn into the feedback window that represents a response to a recognized utterance. Compare recognized text.

path A language object that represents a sequence of zero or more words, phrases, or embedded language models. A path is defined by the SRPath data type.

phrase A language object that represents a sequence of zero or more words. A phrase is defined by the SRPhrase data type.

PlainTalk A collection of operating system managers, control panels, and other software that enables Macintosh computers to speak written text and to respond to spoken commands. See also Speech Recognition Manager, Speech Synthesis Manager.

property An item of data associated with an object. Properties control some of the object's behavior.

property selector A 4-byte value passed to SRGetProperty or SRSetProperty that specifies which property of an object to get or set.

property type The type of a property. A property type is specified by a property selector.

property value The value of a property.

push-to-talk mode A listening mode that requires a key to be held down in order for a recognizer to start listening. Compare toggle-listening mode.

recognition notification A message sent to your application by a recognizer when certain events occur in the recognizer.

recognition result A speech object that describes a recognized utterance. A recognition result is defined by the SRRecognitionResult data type.

recognition system A speech object that defines certain global characteristics of the speech recognition process. A recognition system is defined by the SRRecognitionSystem data type.

recognition system ID A value passed to the SROpenRecognitionSystem function that specifies a recognition system.

recognized speech The product of converting speech into digitally-stored words or phonemes or into computer actions. See also Speech Recognition Manager, synthesized speech.

recognized text Any text drawn into the feedback window that represents a recognized utterance of the user. Compare output text.

recognizer A speech object that recognizes utterances and sends recognition results to your application. A recognizer is defined by the SRRecognizer data type.

reference A 4-byte value associated with a speech object.

reference constant A 4-byte value defined by your application that can be associated with a language object as the value of its kSRRefCon property.

reference count The number of references an application has to a speech object.

reject To deem an utterance unrecognizable.

rejectable Said of an object if a recognition system can return the value kSRRejectedWord instead of that object.

rejected word A word that is passed to your application in a recognition result whenever a recognizer hears some sound but cannot recognize it.

separation character A character used to separate the words in a phrase. By default, a recognition system's separation character is the character " ".

speech The process or product of speaking.

speech class A structure for the data that characterize speech objects, together with a set of properties for those objects. Compare speech object.

speech class hierarchy The hierarchical arrangement of speech object classes.

speech command See embedded speech command.

Speech Manager See Speech Synthesis Manager.

speech object Any instance of a speech class.

speech recognition The process of listening to and interpreting spoken utterances. The Speech Recognition Manager provides speech recognition services for Macintosh computers.

speech recognition callback routine An application-defined routine called by the Speech Recognition Manager when certain recognizer events occur (for instance, when the recognizer has recognized an utterance).

speech recognition callback structure A data structure that contains information about a recognition result. A speech recognition callback structure is defined by the SRCallBackStruct data type.

Speech Recognition Manager The part of the Macintosh system software that provides a standardized method for Macintosh applications to recognize speech. Compare Speech Synthesis Manager.

Speech Synthesis Manager The part of the Macintosh system software that provides a standardized method for Macintosh applications to generate synthesized speech. Previously called the Speech Manager. Compare Speech Recognition Manager.

status pane The portion of a feedback window that provides information about the status of a recognizer. The status pane includes the feedback character and the listening mode indicator. Compare transcript pane.

subitem An item in a container.

synthesized speech The product of converting nonaural tokens (such as written or digitally-stored words or phonemes) into speech. See also recognized speech, Speech Synthesis Manager.

text-to-speech See synthesized speech.

toggle-listening mode A listening mode that interprets presses on the listen key as a toggle to turn listening on or off. Compare push-to-talk mode.

top-level language model A language model that does not occur in the path or set of paths that define any other language model. Compare embedded language model.

transcript pane The portion of a feedback window that contains a readable transcript of the few most recent recognized utterances and feedback. Compare status pane. See also output text, recognized text.

value See property value.

voice recognition See speech recognition.

word A language object that represents a single spoken word. A word is defined by the SRWord data type. The properties of a word include its spelling, its pronunciation, and its reference constant.