Common Voice Industry Terms

In order to understand the many components of voice recognition, industry terms are defined below:

Barge-in - The ability of a caller to interrupt a recorded prompt or message to give a voice command

Dialog - The interaction between the speech system and the caller. Includes the call flow (organization of conversation), the prompts (everything the system can say) and the grammars (everything the user can say).

DTMF - Dual Tone Multi-Frequency, also known as touch-tone or keypad entries.

Grammar - The set of caller utterances allowed at any given time on the dialog.

IVR Unit - Interactive Voice Response Unit, a computer system that evokes and responds to a user’s touch-tone or voice commands.

NLP - Natural Language Processing, the ability of a speech recognition system to translate many different utterances with the same meaning into a single result.

Persona - The consistent character of the system as reflected by voice, language use and the audio environment. A persona is typically defined and or created for each application and is based on the target audience for that application.

Prompt - A speech system’s audio output. Prompts are used for asking questions, presenting options, giving direction or conveying information.

Speech Recognition System - A system that interacts with the caller to elicit speech, understands the speaker’s meaning and performs some pre-defined action based on their understanding.

SpeechObjects - Open, standards-based reusable speech application components that can be used independently or with VoiceXML to create robust speech applications.

Usability - The extent to which the voice application or service can be used by specified callers with effectiveness, efficiency and satisfaction; also, the proactive of studying a system’s usability.

Utterance - The word or collections of words that a user speaks in response to the voice application’s prompts.

Voiceprint - A unique audio pattern that is created by the sound of the person’s voice. Like a thumbprint, this can be used to identify individuals. It can replace the need for a user ID and PIN when trying to access secure areas of the voice application.

Voice Authentication - Also known as speaker verification. The process of matching a caller’s voice to a voiceprint on file, providing secure access to systems and services through voice print matching.

VIA - Voice Interface Applications, AnyTransactions’ customized voice applications accessed by any telephone and driven by the spoken word creating a new interface to existing Internet business systems.

VIA ToolSet - A real-time software development, content integration, application management platform.

VoiceXML - An extensible markup language that is an emerging standard for the VoiceWeb.

VUI - Voice User Interface, a broad term for the experience a caller is presented with when using the system. It is analogous to a GUI (graphic user interface) for a website. It includes the dialog components as well as the persona and the audio files used to build the application.