Speech Solutions

SpeechMagic Glossary

A
accession number	Unique number assigned to an individual test or procedure by either the Hospital Information System (HIS) or Radiology Information System (RIS).

acoustic adaptation	Process which continuously improves an author’s acoustic references. This is done by analysing dictations and automatically updating the ARF to understand an author’s voice better.
acoustic feedback	Function which enables the user to hear parts of the recorded speech while forwarding or rewinding a dictation. This feature can also be referred to as ‘monkey chatter’.
acoustic references	Statistical data which describes the voice characteristics of an individual user. This can include the user’s accent, pronunciation, input device, etc. Speech Recognition uses this data to interpret an author's speech input; it is stored in an ARF.
adaptation	Process which updates either an author's ARF (Acoustic Adaptation) or their ConText (ConText Adaptation).
ARF (acoustic reference file)	File which stores the acoustic references of an individual user. It is updated by the Acoustic Adaptation process.
ASR (Automatic Speech Recognition)	Automatic conversion of speech to text via a software program.
asynchronous playback	Function whereby the recognized text is not highlighted during playback of the sound file. See also synchronous playback.
audio system	All components (both hardware and software) used for recording and playing back sound: sound card, microphone, speakers, etc., as well as audio settings for the recording volume and playback volume. Also known as the sound system.
audio wizard	Tool for adjusting the settings of your audio system to ensure the best possible sound quality for speech recognition. These include setting recording volume, playback volume, and the VA level.
author	User who records the sound file for a dictation.
auto-backspace	Setting which determines the amount of time that a dictation is rewound when playback is restarted.
auto-text	Pre-defined text which can be inserted into a document via command. Also known as a Macro.
back to top
B
Background Lexicon	Dictionary containing between 300,000 and 800,000 words (depending on the language), whose usage is not considered frequent enough for inclusion in a specific ConText Lexicon. Used for reference when unknown words are added to the ConText during ConText Adaptation.
background recognition	Speech recognition which takes place during the dictation but where the recognized text is not displayed to the author on-screen. Can also be referred to as ‘hidden online’ recognition.
backend	Any process which takes place on a server rather than on a client PC (e.g. recognition, adaptation, etc.). Can also be referred to as ‘central’ or ‘offline’. Opposite of frontend.
batch recognition	Automatic backend recognition of multiple dictations after the author has recorded them. Can also be referred to as ‘offline recognition’. The opposite of online recognition.
back to top
C
Cache Lexicon	Dictionary containing all the words added to a ConText (during ConText Adaptation). A Cache Lexicon exists for each installed Language Package, and is shared by a number of users. It is used to add words automatically to other installed ConTexts of the same language.
CELP	Audio compression format (CELP stands for Code Exited Linear Prediction) of 19 kbits/second. This format requires considerably less disk space than the PCM 16 kHz audio compression format.
command grammar	Collection of words or phrases which are not recognized as text but upon recognition a partner integration can execute a specific action. For example, text formatting commands and document navigation commands will be grouped into their own respective command grammars.
connected mode	Work mode where a user works on a client PC which is connected to a central server.
ConText	Nuance Communications Austria product which contains the vocabulary (ConText Lexicon) and the default language model for speech recognition. It contains words and word combinations, as well as information on pronunciation. ConTexts are specific to one language and one field of expertise.
ConText Adaptation	Process which updates an author's language model and ConText Lexicon in order to improve the recognition rate.
ConText Adaptation Task	SpeechMagic System Task which analyzes corrected dictations for words and grammar items not contained in the ConText, adds unknown words to the ConText and updates the user's language model in order to improve the recognition rate. It can use dictations which have been recognized and corrected, or documents not created using speech recognition (e.g. imported via the ConText Tuner).
ConText administrator	User who performs ConText Adaptation for authors.
ConText grammar	Set of words or phrases which the recognizer knows to display in a certain way. For example, predefined section headings may be formatted bold, or numerals displayed according to standard conventions (dosages, dates, etc.).
ConText Lexicon	Dictionary containing words and phrases specific to a particular field and language, e.g. Radiology (US); used for speech recognition.
ConText Tuner	ConText Adaptation function which enables the import and analysis of documents or files for words or phrases not contained in a ConText.
correction	Editing a document created by speech recognition in order to replace incorrectly recognized words or phrases. Correction may be performed by the author or a designated correctionist, and can be aided by functions such as synchronous playback and correction alternatives.
correction alternative	Alternative recognition results which are offered by the ConText Lexicon and can be used to replace incorrectly recognized words or phrases.
correction editor	SpeechMagic application which enables the user to make corrections to the recognition results or to transcribe sound files (digital dictations ) manually. The Correction Editor uses a third-party text editor and provides additional functions to aid the correctionist.
Correction Manager	SpeechMagic application which lists the dictations scheduled for correction. The Correction Editor can be started from the Correction Manager for a selected dictation.
back to top
D
Database Recovery	SpeechMagic application which provides access to information stored in the SpeechMagic Database, e.g. dictations and running tasks (applications and/or system task). It can be used for reviewing this information.
dictation	The act of speaking into a recording device, or the sound file resulting from this process.
Dictation Editor	SpeechMagic application in which authors create dictation documents by means of online recognition. The Dictation Editor uses a third-party text editor and provides added functionality for the author (e.g. correction alternatives).
digital dictation	Sound file created by an author without speech recognition being enabled, usually sent for transcription.
digital input device	Input device which converts speech into digital audio data signals, e.g. DPM, microphone, etc.
disconnected mode	Work mode where a user works on a client PC which has been disconnected from a central server (and will be reconnected for data synchronization later). Also known as island mode.
distributed installation	Type of SpeechMagic installation where dictation, recognition, correction and other SpeechMagic System Tasks are performed on several PCs in a network. These PCs are set up as SpeechMagic Workstations during installation. In addition to the Workstations, there is a PC which is set up as the SpeechMagic File Server, and a Database Server.
DPM (Digital Pocket Memo)	Mobile input device developed by Philips for recording dictations. These dictations are stored digitally and can be downloaded to a PC for further processing. DPMs can also contain barcode scanners to assign job information to dictations.
DSS file	Sound file based on the Digital Speech Standard. This standard aims at establishing compatibility and exchangeability between recording and playback products from multiple vendors.
back to top
E
EOL (End-of-Letter)	Button on a device (SpeechMike or DPM), or any function which closes a sound file and marks a dictation as finished.
back to top
F
Fast Job Propagation	ConText Adaptation mode which updates an author’s language model immediately but collects unknown words and grammar items for handling at a later stage.
Field	Placeholder in a SpeechMagic InterActive document template. These can also be filled in automatically with data from the database (Smart Fields) or navigated to by name (Named Fields). Also known as jump-points.
Foot Control	Philips foot pedal device which provides functions for navigation in a sound file during correction or transcription. The three Foot Control pedals have the default functions Playback, Fast Forward and Fast Rewind, but can in some applications be configured otherwise.
frontend	Process which takes place on the client rather than the server, (e.g. frontend recognition, adaptation, etc). Opposite of backend.
back to top
G
General Correspondence ConText	Philips Speech Recognition Systems ConText which contains vocabulary for general dictation purposes, for example, letters, reports, memos. Compare: professional ConText.
grammar item	Word or phrase which the recognizer knows to treat differently, with reference to a ConText grammar rather than the language model. For example, numbers which represent a date or a dosage are typical grammar items.
back to top
H
HIS (Hospital Information System)	Computer system storing patient, order and examination data for a hospital or clinic.
back to top
I
Initial Training	Process during which an author-specific ARF is created from training texts read aloud. This can take from as little as two minutes to over an hour, depending on the user’s standard and pronunciation of English, and is important to establish a high initial recognition rate.
input channel	Acoustic property of a dictation which takes into consideration the different input devices which may be used for recording sound files, e.g. microphone, mobile input device, telephone. Different ARFs are used for recognition depending on the input channel used.
input device	Peripheral device (e.g. a keyboard) which enables users to give input to a computer. In a speech recognition environment, 'input device' usually refers specifically to a device for recording speech, e.g. a microphone.
Insert mode	Recording mode which inserts new speech at the current position in the sound file but does not overwrite existing parts of the sound file.
back to top
J
back to top
K
back to top
L
language model	Statistical model which represents word usage and sequences of words. The language model is specific to an author and a ConText.
language package	All necessary data which allows to recognition for a specific language. A Language Package contains a Background Lexicon, initial ARFs and other language-dependent statistics. The ConText used for speech recognition is always based on a specific Language Package.
back to top
M
modality	Imaging machine in a radiology department. Different modalities are used for different procedures, e.g. CT, MR, X-Ray, etc.
multimodal input	Accepting different forms of user input, e.g. voice commands and mouse simultaneously.
back to top
N
back to top
O
online recognition	Recognition of speech to text which takes place during the dictation, and where the recognized text is displayed simultaneously to the author. Opposite of batch recognition. Can also be referred to as ‘frontend’ or ‘simultaneous’ recognition.
Overwrite mode	Recording mode which inserts new speech at the current position in the sound file and overwrites existing speech for the duration of the recording.
back to top
P
phonetic description	Literal spelling of a word which represents how that word is actually pronounced (e.g. ‘etc’ becomes ‘etcetera’).
phonetic transcription	Representation of a word in phonetic symbols which describe how a particular word is pronounced. This is used by the recognizer to transform sound into text.
playback volume	Volume at which dictated text is played back through the speakers or headphones.
professional ConText	ConText for specific professional fields, e.g. medical, legal, insurance. Compare: General Correspondence ConText.
Purge Task	SpeechMagic System Task which automatically calculates performance values and deletes dictations and related data no longer required by the SpeechMagic system. The Purge Task runs on an assigned SpeechMagic Workstation.
back to top
Q
Queuing	Process of storing and forwarding dictations between different processes of the workflow. For example, dictations are recorded and then queued for batch recognition.
back to top
R
recognition rate	Percentage of words correctly recognized when dictating. Sometimes referred to as recognition accuracy.
recognition result	Text generated by the recognition of dictated speech, including ancillary data such as sound file position and correction alternatives.
recognition task	SpeechMagic System Task that performs speech recognition of dictations in batch mode.
recognition unit	Separate units within the Recognition Task which perform recognition. On an SMP machine there are multiple recognition units.
recording volume	Loudness of the speech input when recording a dictation; also referred to as microphone input level. The recording volume is set by the Audio Wizard together with the VA level.
redundant phrase	Word or re-occurring words that are recorded but should not appear the final report, e.g. thank you, etc.
RIS (Radiology Information System)	Computer system storing patient, order and examination data for the radiology department of a hospital or clinic.
back to top
S
silence detection level	See: VA level.
sound card	Adapter card in the computer hardware which enables the input and output of audio signals, for example via a microphone and speakers.
sound file	File which contains audio data in a special format, e.g. .wav or .dss.
Speech Recognition	See: Automatic Speech Recognition.
SpeechMagic	Professional speech recognition software system for a large number of users in a network (distributed installation) or as a stand-alone installation.
SpeechMagic Database	Database which is used by SpeechMagic to coordinate the processing of dictations in the SpeechMagic system, to keep track of the status of dictations and to synchronize the various SpeechMagic tasks and applications.
SpeechMagic File Server	Location where all data needed to perform speech recognition is stored (statistic models, job information, etc.).
SpeechMagic job	A dictation and other related data which goes through several processing steps within the workflow of a SpeechMagic system.
SpeechMagic Recorder	Software program with enhanced recording, playback and positioning functions. It enables the user to create, edit and save dictations as sound files for later processing. The SpeechMagic Recorder also includes the Audio Wizard for adjusting the audio settings.
SpeechMagic Server Setup	Setup application for installing the SpeechMagic File Server where all SpeechMagic data is centrally stored, including the files required for the setup of SpeechMagic Workstations.
SpeechMagic System Tasks	Service which performs a task in the SpeechMagic system. System Tasks include: Recognition Task, ConText Adaptation Task, Purge Task.
SpeechMagic Workstation	A PC in the SpeechMagic system which runs SpeechMagic applications and SpeechMagic System Tasks. There may be any number of Workstations in a SpeechMagic system.
SpeechMike	Industry standard audio device designed and built by Philips specifically for dictation and speech recognition. This device includes a microphone and speaker and can also incorporate a trackball or barcode scanner.
split cursor	SpeechMagic feature whereby text can be typed in to a document while synchronous playback is enabled.
SRD file (Speech Recognition Document)	File format for storing speech recognition dictations. An SRD file usually contains sound, text and additional information, including the author's name, the name of the ConText used for speech recognition, the recognized and corrected text, etc.
stand-alone installation	Installation in which all data files and executables are located on the local hard disk of a PC. This PC acts as both server and client and performs all tasks required in the workflow.
Supervised Acoustic Adaptation	Acoustic Adaptation which takes place after dictations have been recognized and corrected.
synchronous playback	Function which highlights every word while playing back sound files. This makes it easier to find errors and to correct them. See also: asynchronous playback.
back to top
T
transcription	Act of manually typing out a dictation recorded by an author.
back to top
U
Unsupervised Acoustic Adaptation	Acoustic Adaptation which takes place after dictations have been recognized, but before correction.
back to top
V
VA level (Voice-activated recording level)	Threshold level of the recording volume which enables the recording program to distinguish between speech (which should be recorded), and silence or background noise (which should not be recorded). Recording stops automatically when the speech input volume is below the VA level, and restarts automatically when the input volume is above this level. It is calculated by the Audio Wizard. Also referred to as silence detection level, voice sensitivity threshold, VOR level or VAR level.
vocabulary	Words contained in the ConText with information on how they are pronounced. New words can be added to the vocabulary via ConText Adaptation or by using the ConText Tuner.
voice recognition	See: speech recognition.
VOR level	See: VA level.
back to top
W
WAV file	Sound file format with a .wav extension. This is a standard sound file format which can be created by most computer recording applications. SpeechMagic supports several sound file formats such as CELP 8 kHz, CELP 16 kHz and PCM 8 kHz.
workflow	Automation of a business process during which documents, information or tasks are passed from one workflow participant to another for action, according to a set of procedural rules. For example, it may be defined how and by whom a document is corrected, reviewed and approved.
back to top
X
back to top
Y
back to top
Z
back to top

Client Login

SpeechMagic Glossary

Client Login