Name: Distributed Speech Transcription System

Text: Distributed Speech
Transcription System

Description
! Nestor is a Network based
distributed transcription and
documenting system, designed
to make transcription faster,
more efficient giving better
quality.
! It provides automated workflow
to multiple transcriptionists for
fast documenting of large
volume of speech.

Applications
Nestor is used for real-time
transcription of
!
!
!
!
!
!

Lectures,
Conferences,
Meetings,
Negotiations,
Speeches,
Interviews.

Applications
Typical customers are:
!
!
!
!
!

Parliaments,
Congress Halls,
Documentation
Departments,
Journalists and News
Agencies,
Academic Institutions.

Structure
Negotiations

Digital media
& Internet

Lectures

Printouts
Meetings

Operational Procedures
All incoming recordings are
automatically cut into fragments of
a given duration, then evenly
distributed among the operators,
who transcribe the speech
fragments for text production.
The editor’s workstation then
combines the separate text
segments from various operators
into the whole document, which
can be saved or printed out

SPEECH

Legend

Nestor
workflow

Sound signal
Text

The speech signal is
recorded by Sound
Recording Stations.
Digitized sound is
automatically sent to
Sound Server.

Conference audio
equipment

Channel №1

Sound
recording
stations

Sound
Server

Sound + Text
Control

SPEECH

Nestor
workflow

Conference audio
equipment

Sound Server
Channel №1
automatically
distributes
sound fragments between
Operators. The
Sound
distribution process
is
recording
controlled by Editors.
stations

Sound
Server

Operators in groups

Legend
Sound signal
Text

Editors

Sound + Text
Control

SPEECH

Nestor
workflow

Conference audio
equipment

Operators listen to the
Channel №1
speech fragments
and type
the corresponding text.
Ready text fragments are
SoundThe
returned to the Editor.
Sound
Editor workstationrecording Server
stations
automatically connects
these fragments in a whole
document.
.

Operators in groups

Legend
Sound signal
Text

Editors

Sound + Text
Control

SPEECH

Nestor
workflow

Conference audio
equipment

Channel №1

Sound
recording
stations

Editors check transcription
quality using hyperlink
feature between text and
sound, make correction,
spell checking.

Sound
Server

Operators in groups

Legend
Sound signal
Text

Editors

Sound + Text
Control

SPEECH

Conference audio
equipment

Nestor
workflow
Within several minutes after
completion of speech the text
of the speech is ready to be
printed or published in the
Internet or on digital media.
.

Channel №1

TEXT

Printout or
web-publishing

Sound
recording
stations

Sound
Server

Operators in groups

Legend
Sound signal
Text

Editors

Sound + Text
Control

SPEECH

Conference audio
equipment

Nestor
workflow

Channel №2

TEXT

Printout or
web-publishing

Sound signal
Text

Editors

Nestor’s workflow engine can process
speech coming from more than 30
channels simultaneously in several
different languages by different
groups of transcriptionists.

Channel №1

Channel №3

Legend

Sound
recording
stations

Sound
Server

Operators in groups

Sound + Text
Control

SPEECH

Legend

Nestor
workflow

Sound signal
Text

Editors

Administrators

Conference audio
equipment

Channel №1
Channel №2
Channel №3

TEXT

Sound
recording
stations

Sound
Server

Administrator workplace is needed
in order to manage groups in case
of big number of operators.
Printout or
web-publishing

Backup
Server

Operators in groups

Sound + Text
Control

SPEECH

Legend

Nestor
workflow

Sound signal
Text

Editors

Administrators

Conference audio
equipment

Channel №1
Channel №2
Channel №3

TEXT

Sound
recording
stations

Sound
Server

After completing the work, the sound
files and the corresponding texts are
stored in the archive if necessary.
Printout or
web-publishing

Backup
Server

Operators in groups

Sound + Text
Control

The main goal of Nestor is to receive the full text of
speech in several minutes after it is finished and
with absolutely perfect quality.

Documentation Speed
A group of five skilled
Operators can handle
transcription of one speech
event (meeting, lecture, etc.)
To manage three events
simultaneously, three groups
of Operators, i.e. 15 people
are necessary.

Comfort features
Operator’s software set of
features to transcribe effectively:
"Variable playback speed
"Use of Hotkeys
"Tone correction
"Loop playback

Data protection
Only the Administrator and the
Editors have access to the full
sound recordings plus texts and
are allowed to edit and modify
them.
Operators work with short,
separate speech fragments that
can not be linked together. Thus
the Operators will not know the
contents of the whole event or
tamper with the source audio or
text.

Data transmission

The system is distributed over
TCP/IP LAN

System components
• Administrator’s workstation (optional)
• Editor’s workstation(s)(s)
• Operator’s workstation(s)
• Sound recording station(s)
• Backup server (optional)
• Sound server

System Requirements
# Administrator, editor, operator workplaces–
MS Windows NT/2000/XP OS, sound card,
LAN
# Sound recording station, Sound Server,
Backup Server – PIII or higher, MS
Windows NT/2000/XP OS, LAN
# Multi-channel professional I/O sound
boards STC-H200

Contact us
Speech Technology Center
St.Petersburg, Russia
http://www.speechpro.com
email: [email protected]

$ Tel. +7 812 3310665

% Fax +7 812 3279297

German office:
Speech Tech GmbH
Saarbrücken, Germany
http://www.speech-tek.com
email: [email protected]

$ Tel. +49 681 9655709

% Fax +49 69 25577077

Document Path: ["1135-speech-technology-center-presentation.pdf"]

e-Highlighter

Click to send permalink to address bar, or right-click to copy permalink.

Un-highlight all Un-highlight selectionu Highlight selectionh