Transcript Document

Text-to-Speech for the less fortunate: from talking cellular phones
to augmentative alternative communication devices
Luisa Cordano
Speechtek West 2007
Speechtek West 2007
February 21, 2007
Luisa Cordano
1
Overview







What is AAC?
AAC users and requirements
AAC solutions and devices
The importance of speech technologies
The role of the speech technology designer
Loquendo’s experience
Case studies and real deployments
Speechtek West 2007
Luisa Cordano
2
What is AAC?
•
AAC is augmentative and alternative communication
Defined as: 1) The supplementation or replacement of natural speech and/or writing
using aided and/or unaided symbols...
2) The field or area of clinical/educational practice to improve the
communication skills of individuals with little or no functional speech."
(Lloyd, L.L., Fuller, D.R., & Arvidson, H.H. (1997) Augmentative and Alternative Communication
•
•
•
AAC is a field of endeavor addressing the expressive communication
needs of people with significant speech disability.
AAC interventions range from no technology (gestures, signs) to low
technology (communication board, wallet) to high technology (voice
output communication aids).
Individuals may be recommended several AAC systems over a
lifetime. Therefore, how systems handle this transition and maintain
consistency when change occurs is important to achieving long-term
effective communication.
Speechtek West 2007
Luisa Cordano
3
Who needs AAC? The Users
• Users
Individuals with severe communication disorders and for whom gestural,
speech and/or written communication is temporarily or permanently
inadequate to meet all of their communication needs, for example:
– Blind, visually impaired and elderly
– Speech-impaired (e.g. Stroke (i.e. aphasic, dysartric), accident, brain
tumor, anoxia, Amytrophic Lateral Sclerosis (ASL ), Lou Gehrig’s
Disease,…)
– Developmental causes (e.g. prenatal, perinatal, congenital
or genetic problems)
– Parkinson disease
– Spinal cord injury
Speechtek West 2007
Luisa Cordano
4
Who needs AAC? The Use Requirements
• User requirements depends on impairement
 No use of hands
mouth sticks, headsticks, switches, eye-pointing
 No read and spell
picture and graphic symbols e.g.PCS + speech
generation
 On wheelchair
large communication devices
 Walking
small and light aids
Speechtek West 2007
Luisa Cordano
5
Scope and Goals of AAC
• AAC in everyday life
Speech
gesture and body language
+ facial expressions
writing
manual signs
• AAC supporting users with impairments or disabilities
- To express themselves, their needs and requests, their
ideas and feelings
- Compensate severe difficulties in natural expression
- Assist in maintaining existing competencies
- Empower life skills and independence
- Improve social networking and inclusion
Speechtek West 2007
Luisa Cordano
6
Scope and Goals of AAC
The ultimate goal of AAC is:
The highest performance communication
possible
The ATIA reports that people who rely on AAC
indicate that the two most important things for
them are:
 Saying EXACTLY what they want to say, and
 Saying it as fast as they can
Speechtek West 2007
Luisa Cordano
7
Types of AAC solutions
• Topic Vocabulary
– Organized set of symbols and/or text that allows a quick generation of
speech
• Symbolic Aid
– Communication via symbols such as PCS, PICTO and photos
• Text-based aid
– Communication via text (e.g. word prediction, completion etc.)
• Dynamic
– Communication via screen changes on selection basis
Speechtek West 2007
Luisa Cordano
8
Types of AAC devices: communication aids
• Growing variety and complexity of communication and
computing devices
• Target: full age range, no economic, ethnic or gender
boundaries
• Communications Devices:
– SGD: Speech Generating device, used by an individual to
communicate more effectively with others, through printed words,
speech output, picture or combinations of these.
– From basic components and fundamental language functions to
computer-like functions and control household appliances.
– Primary or supplementary mode of communication, customizable
Speechtek West 2007
Luisa Cordano
9
AAC devices
• Features and categorization
– Speech Output - Digitized (recorded human speech) or Synthesized
(electronic conversion of text into speech).
– Message Type - Prerecorded (messages that are stored)
or Message Formulation (can spontaneously create novel messages).
– Recording Time - less than or equal to 8 minutes, greater than 8 minutes, or
based on the memory size of the device in general (e.g., 64 MB card).
– Access Method - Direct Physical Contact with Device or Multiple Access
Methods (e.g., switch, mouse, joystick, etc.).
– Message Formulation Technique - Spelling only or Spelling and Other
Methods.
Speechtek West 2007
Luisa Cordano
10
The role of the Speech Technology Supplier
• Improve quality and naturalness of speech generation:
– Improve intelligibility
– Improve intonation and expressivity
– Improve flexibility of the overall application
• Study user requirements with speech therapists and with
selected sample of users
• Identify aspects that could be neglected in usual applications,
but could be critical for particular populations of user (e.g.
flexibility in speech rate, articulation, etc…)
• Be able to evaluate and compare the speech technology not only
with respect to usual applications (such as navigation,
advertising, etc…), but also with respect to more critical
requirements
Speechtek West 2007
Luisa Cordano
11
The importance of speech technologies
• Scope
– Allow the most effective communication possible for the individual
enabling them to say EXACTLY what they want
• Important Features
–
–
–
–
–
Different languages
Expressivity
Natural sounding
High Intelligibility
Control
proposals
Speechtek West 2007
Luisa Cordano
12
Loquendo’s experience
Thanks to
• Extensive and detailed
research on the sector
• User trials
• Close partnership with
our customers
• Continuous feedback
loops
Has become a leading player on this market
Speechtek West 2007
Luisa Cordano
13
Loquendo Language offer
Language
English US
English UK
Spanish (Castillian)
French
German
Italian
Greek
Portuguese
Swedish
Dutch
Polish
Brazilian Portuguese
American Spanish
Mexican
Chilean
Argentinean
Chinese (Mandarin)
Valencian (bilingual)
Catalan (bilingual)
Female
P
PP
P
PPP
PP
PPPP
PP
P
P
P
P
P
Male
PP
P
PP
P
P
PPPPP
P
P
P
P
P
P
P
P
P
P
P
The 2007 roadmap sees the development of 4 new languages
Speechtek West 2007
Luisa Cordano
14
Expressive and Lifelike Voices
Loquendo TTS allows for the highly emotive pronunciation, creating
extremely natural sounding speech. An easy-to-use prompt editing
provides the user a repertoire of "expressive cues".
Common phrases and interactions are pronounced with a natural and
colorful intonation, to express intention (to confirm, doubt, exclaim,
thank, etc.).
The same sentence can be pronounced using different styles and
intonations, from neutral to emphatic, from sad to amazed.
Figures of speech, such as greetings and exclamations ("hello!", "oh
no!", 'I'm sorry!"), interactions ("Oh!", "Well!", "Hum"..) and
paralinguistic events (e.g. breathing, coughing, laughter, etc.)
US
Speechtek West 2007
LA
Luisa Cordano
15
Loquendo Mixed language capability™
Each of our voices is able to speak any other available language! This
means that for example, our Italian voice can speak English with an
Italian accent, or a French voice speak German with a French accent.
This unique and patented feature is extremely useful for reading when
you are out of your home country or simply for reading foreign names
in a book or in any text.
Speechtek West 2007
Luisa Cordano
16
Underlying Feature: Phonetic Mapping
Applies the foreign language grapheme-to-phoneme transcriptionrules to the foreign text, and then maps the transcribed phonemes
onto those of the voice's native language in order to access its
acoustic units
 Approximate Pronunciation (speaker maintains his native-tongue
phonological system when pronouncing foreign words)
German Voice
Spanish Voice
French Voice
Speechtek West 2007
Luisa Cordano
17
Loquendo’s Experience: real multilingual
deployments
• Blind Associations
– ONCE (Spain)
– UIC (Italy)
• Assistive device providers
–
–
–
–
–
–
–
Intaal (Touchspeak)
Oralys
Onewrite (Cyrano communicator)
Code Factory (Mobile Speak)
Voice Systems
BAUM (Poet)
…..
Speechtek West 2007
Luisa Cordano
18
Loquendo’s Experience: blind and visually
impaired
Official suppliers to the Italian and
Spanish National Blind Associations
Italian National
Research Council
Speechtek West 2007
Luisa Cordano
19
Loquendo’s Experience: AAC devices
Pen or fingers
Speechtek West 2007
Text to Speech
Luisa Cordano
20
Oralys: Pocket PC Communicator
Pen or
fingers
Text to Speech
The Oralys Communicator works on the basis of mental imagery
known as, active ideography, combined to multi-sensory elements
using sound and symbols. It enables to communicate quickly and
effectively, with a simple touch. The Communicator holds over
3.500 ideograms with the possibility of 12.000.000 connections/
sentences.
Speechtek West 2007
Luisa Cordano
21
OneWrite: Cyrano communicator
Cyrano, an augmentative communication device, is a hand-held,
portable software application designed to aid individuals with speechimpairments to communicate through a series of customizable images,
text, and synthesized voices.
Speechtek West 2007
Luisa Cordano
22
Intaal: TouchSpeak™ & TypeSpeak™
TouchSpeak is the result of nearly 10 years of on-going
development and continuous refinement of the original 1996
PCAD Consortium. TouchSpeak has been specifically designed
and developed as an AAC communication aid. Although
originally primarily dedicated to people with aphasia, it can
comfortably be adapted for either higher more functioning
clients, or those with lower cognition and functions
Speechtek West 2007
Luisa Cordano
23
E-Values e-book reader
E-library Voice AppLication for eUropean blind, Elderly and
Sight-impaired
The purpose of the eVALUES service is to improve the possibilities for
education and jobs for blind and sight impaired, increasing the
assimilation in the society.
eVALUES does this by providing a Trans-European e-Library service
based on Internet for Visually Impaired, to improve the access to written
material. The service uses Text-To-Speech for the content provisioning
on a portable hand-held PDA, which enables mobility.
Speechtek West 2007
Luisa Cordano
24
Code Factory’s mobile screen readers
Speechtek West 2007
Luisa Cordano
25
Code Factory’s mobile screen readers
• Demo Video
Speechtek West 2007
Luisa Cordano
26
Thank you for your kind attention.
• Questions?
Luisa Cordano
Sales Manager Embedded Technologies
[email protected]
ph: +39 011 2913442
Speechtek West 2007
Luisa Cordano
27