1 “ Speech ” EMPOWERED COMPUTING Greenfield Business Centre, 20 th September, 2006.

1

“Speech” EMPOWERED COMPUTING

Greenfield Business

Centre, 20th September, 2006

2“Speech” EMPOWERED COMPUTING

Objective

Describe what computer speech recognition is, what it can do and how it

could help you be more productive


Agenda1. Our heavy reliance on computers nowadays

2. How do you manage your e-mail, report and letter writing commitments?

3. Demonstration – Microsoft Word

4. Speech recognition on the computer What is it? How does it work? What software is available? History of speech recognition What can it do? Demonstration – Microsoft Excel, Microsoft Outlook What do you need to get started? How easy is it to set up, what results can I expect? How much does it cost? Demonstration – Internet browsing (recorded)

5. Which users can speech recognition help?

6. Transcription from digital recorder (from a training CD)


Our heavy reliance on computers

Are you like this?


Our heavy reliance on computers

Or like this?


How do you manage your e-mail, report and letter writing commitments?

• Reply to e-mails immediately• Write letters promptly• Write reports before they are due

• “I’ll reply in a few days time”• “I’ll write that letter when I get a chance!”• “I’ll write that report when I get a minute”

• “I’ll reply one of these days!”• “I never seem to have time to write letters”• “I can’t remember the last time I typed a report”


Demonstration

Microsoft Word


Speech recognition- what is it? Speech recognition (in many contexts also known as

automatic speech recognition, computer speech recognition or erroneously as voice recognition) is the process of converting a speech signal to a set of words, by means of an algorithm implemented as a computer program. Speech recognition applications that have emerged over the last years include voice dialling (e.g., Call home), call routing (e.g., I would like to make a collect call), simple data entry (e.g., entering a credit card number), and preparation of structured documents (e.g., a radiology report).

Voice or speaker recognition is a related process that attempts to identify the person speaking, as opposed to what is being said.

Source: http://en.wikipedia.org/wiki/Speech_recognition


Speech recognition- how does it work?


What software is available?

Microsoft Windows 2000 and XP

iListen (Apple Mac)

Dragon NaturallySpeaking

IBM ViaVoice


History of speech recognition


Speech recognition- what can it do?

Dictate, punctuate, format, correct recognition errors, edit text.

Open programs, navigate menus, and click buttons

Open, close and switch between applications

E-mails Access files and folders Commands – built-in and custom (macros)

to automate your work Surf the web Roaming Digital recorder


Demonstration

Demonstration – Microsoft Excel, Microsoft Outlook



VERY ACCURATE - Up to 99% accuracy, never makes a spelling mistake, gets smarter the more you use it

FASTER THAN TYPING! - Most people speak over 120 words per minute, but type less than 40 words a minute. Create letters and e-mails about three times faster than typing by hand!

EASY TO USE - dictate letters, e-mails and surfing the web by voice very quickly.

USE WITH ANY WINDOWS PROGRAM - Use voice to dictate, edit and control applications like Microsoft® Word, Microsoft® Excel, Microsoft Internet Explorer, and Corel® WordPerfect®.

WIRELESS/BLUETOOTH SUPPORT - Use with certain Wireless and Bluetooth headsets.

MOBILE - Dictate into a handheld device for automatic transcription when you synchronise with the PC.



It can never be 100% accurate Need to take time to:-

Learn the software Speak clearly Correctness recognised words Add custom words and phrases

Cannot transcribe conference recordings Can’t recognise a casual new user Caution over the claimed 120 words per

minute wpm – see next chart

“Watch outs”



Chart 1: Speed (seconds) when working by voice, compared to keyboard and mouse - transcription task

0

100

200

300

400

500

1st dictation, all by voice 2nd dictation, all by voice 3rd dictation, correction byhand

All by hand

Tim

e (s

econ

ds)

Proofreading and correcting stage

Dictation/typing stage

“Watch outs”


What you need to get started?

Computer – minimum of:- P4 processor with speed of 1.8 GHz 512 Mb RAM memory

Windows 2000 SP4, Windows XP (SP1 or SP2), XP Home

High accuracy, noise-cancelling microphone

A good quality sound card Low background noise Recommend 0.5 - 1 day’s training


How easy is it to set up?

1. Create a new user < 5 mins

2. Run audio set up < 5 mins


How easy is it to set up?

3. Choose and read an enrolment training text (20 mins)

Total time to be up and running – approx 30 mins!


What results can I expect?

Dragon NaturallySpeaking accuracy over time

95

99

90

91

92

93

94

95

96

97

98

99

100

After initial enrolment Later e.g. 3 - 4weeks

Acc

ura

cy (

%)

After use of vocabulary editor, correction of mistakes, use of document analyser etc

• Transcription speed up to 120 wpm

• Actual speed less as need to proof/correct

• Composition speed depends on user


How much does it cost?

Dragon software Preferred version - £119

Professional version - £495

Hardware Noise cancelling microphone – typical price

£40 - £110

USB sound pod - £50 - £60

Training - varies according to provider - expect to pay £250 for a half day


Demonstration

Internet Explorer demonstration

(recorded)


Which users can speech recognition help?

• Corporate teams

• Office Workers

• Secretaries

• Administrators

• Lecturers

• Teachers

• Students

• Lawyers

• Solicitors

• Quantity surveyors

• Financial Advisers

Users with disabilities

Mobile users


Demonstration

Transcription from digital recorder (from training CD)

1 “ Speech ” EMPOWERED COMPUTING Greenfield Business Centre, 20 th September, 2006.

Documents

computer speech recognition

speech recognition help

automatic speech recognition

speech signal

voice recognition

microsoft outlook slide

productive slide

speaker recognition