How speech recognition system work
The voice input to the microphone produces an analog speech signal.
This speech signal is then converted into binary words compatible with
a digital computer by an analog to digital converter.
The binary version of the input voice is then stored in the system and
compared to previously store binary representations of words.
When a match occurs, recognition is achieved.
Then, the spoken work appears on a video screen or is passed along
to an NLP for meaning analysis.
A typical work recognizer
The voice input is applied to a microphone.
The electrical analog signal from the microphone is fed to an amplifier,
where it is increased in level.
The amplifier contains some kind of Automatic gain control to provide
an output signal in a specific voltage range.
The analog signal representing a spoken work is a complex waveform
that contains many individual frequencies.
The way to recognize the spoken work is to break that complex input
signal into its components parts.
This is usually done with a set of filters.
A filter is an electronic circuit that passes or rejects frequencies in a
certain range.
Voice synthesis
Voice synthesis or response is the technology by which computers
speak.
Tools and applications of voice technologies
Speech recognition is used in four major areas.
1. Command and control – including voice control of machine
operations and voice activated dialing
2. Data entry to forms – quality control system, or databases
3. Direction – including creation of documents and reports
4. Data access / Information retrieval – such as directory
assistance and retrieval of online data
Voice technology application sampler – Application
Research on User Interfaces in
Hwang and Wu identified the following indecent and dependent
variables.
Independent Variables
Human user Demographics – Age, education, experience
Psychological – Cognitive style, intelligence, risk attitude
Decision
environment
Decision structure, organizational level
Others – stability, time pressure, uncertainty
Task Decision support – complexity level
Inquiry / information retrieval , data entry, word processing and
Computer aided instruction
Interface
characteristics
Input / Output media, Dialog type, Presentation format and
Language characteristics
Dependent variable : Human / Computer effectiveness
This was measured by usefulness, perceived ease of use and
performance (time, error, task, completion, profit), user attributes
(satisfaction, confidence) and use of system option (high. low).
Development Strategies
Write a customized DSS in general purpose programming language
such as COBOL or PASCAL.
Use a fourth generation language (4GL). Such as data oriented
languages, spreadsheets, and financial oriented languages.
Use a DSS integrated development tool. An integrated package
eliminates the need to use multiple 4GLs. The best known examples
for the PC are Excel, Lotus 1-2-3, and Quattro Pro.
Use a domain specific DSS generator: Domain specific DSS
generators are designed to build a highly structured system, usually in
a functional area.
DSS Development Phase A : Planning Phase E : Construction
Phase B : Research Phase F : Implementation
Phase C : System analysis and
conceptual design
Phase G : Maintenance and
documentation
Phase D : Design Phase H : Adaptation
Phase A : Planning
Planning deals mainly with need assessment and problem diagnosis.
A crucial step in the planning effort is determining the key decisions to
be supported by the DSS.
Example _ in a portfolio selection system, a key decision might be
selecting the correct stocks for a particular customer’s needs.
Phase B : Research
This phase involves the identification of a relevant approach for
addressing user needs and available resources. (Hardware, software,
vendors, systems, studies or related experience in other organizations,
and review of relevant research).
Phase C : System analysis and conceptual design
This phase includes the determination of the best construction
approach and specific resources required implementing it, including
technical, staff, financial, and organizational resources.
Phase D : Design
The detailed specifications of the system components, structure and
features are determined.
Phase E : Construction
A DSS can be constructed in different ways depending on the design
philosophy and the tools being used.
The construction is the technical implementation of the design.
Phase F : Implementation
The implementation phase consists of the following tasks: testing,
evaluation, demonstration, orientation, training and deployment.
Several of these tasks are performed simultaneously.
Testing Data on the system’s outputs are collected & compared against
the design specifications.
Evaluation The implemented system is evaluated to see how well it meets
user’s needs.
Technical & organizational loose ends are also identified.
Testing & evaluation usually result in changes in the design &
construction.
Demonstration Demonstration of the fully operational system capabilities to the
user community is an important phase.
Orientation This involves instruction of users in the basic capabilities &
operation of the system
Training Operational users are trained in system structure and functions.
Deployment The full system is operationally deployed for all members of the
user community.
Phase G : Maintenance and documentation
Maintenance involves planning for ongoing support of the system and
its user community.
Phase H : Adaptation
Adaptation requires recycling through the earlier steps on a regular
basis to respond to changing user needs and thus as several variables.
The voice input to the microphone produces an analog speech signal.
This speech signal is then converted into binary words compatible with
a digital computer by an analog to digital converter.
The binary version of the input voice is then stored in the system and
compared to previously store binary representations of words.
When a match occurs, recognition is achieved.
Then, the spoken work appears on a video screen or is passed along
to an NLP for meaning analysis.
A typical work recognizer
The voice input is applied to a microphone.
The electrical analog signal from the microphone is fed to an amplifier,
where it is increased in level.
The amplifier contains some kind of Automatic gain control to provide
an output signal in a specific voltage range.
The analog signal representing a spoken work is a complex waveform
that contains many individual frequencies.
The way to recognize the spoken work is to break that complex input
signal into its components parts.
This is usually done with a set of filters.
A filter is an electronic circuit that passes or rejects frequencies in a
certain range.
Voice synthesis
Voice synthesis or response is the technology by which computers
speak.
Tools and applications of voice technologies
Speech recognition is used in four major areas.
1. Command and control – including voice control of machine
operations and voice activated dialing
2. Data entry to forms – quality control system, or databases
3. Direction – including creation of documents and reports
4. Data access / Information retrieval – such as directory
assistance and retrieval of online data
Voice technology application sampler – Application
Research on User Interfaces in
Hwang and Wu identified the following indecent and dependent
variables.
Independent Variables
Human user Demographics – Age, education, experience
Psychological – Cognitive style, intelligence, risk attitude
Decision
environment
Decision structure, organizational level
Others – stability, time pressure, uncertainty
Task Decision support – complexity level
Inquiry / information retrieval , data entry, word processing and
Computer aided instruction
Interface
characteristics
Input / Output media, Dialog type, Presentation format and
Language characteristics
Dependent variable : Human / Computer effectiveness
This was measured by usefulness, perceived ease of use and
performance (time, error, task, completion, profit), user attributes
(satisfaction, confidence) and use of system option (high. low).
Development Strategies
Write a customized DSS in general purpose programming language
such as COBOL or PASCAL.
Use a fourth generation language (4GL). Such as data oriented
languages, spreadsheets, and financial oriented languages.
Use a DSS integrated development tool. An integrated package
eliminates the need to use multiple 4GLs. The best known examples
for the PC are Excel, Lotus 1-2-3, and Quattro Pro.
Use a domain specific DSS generator: Domain specific DSS
generators are designed to build a highly structured system, usually in
a functional area.
DSS Development Phase A : Planning Phase E : Construction
Phase B : Research Phase F : Implementation
Phase C : System analysis and
conceptual design
Phase G : Maintenance and
documentation
Phase D : Design Phase H : Adaptation
Phase A : Planning
Planning deals mainly with need assessment and problem diagnosis.
A crucial step in the planning effort is determining the key decisions to
be supported by the DSS.
Example _ in a portfolio selection system, a key decision might be
selecting the correct stocks for a particular customer’s needs.
Phase B : Research
This phase involves the identification of a relevant approach for
addressing user needs and available resources. (Hardware, software,
vendors, systems, studies or related experience in other organizations,
and review of relevant research).
Phase C : System analysis and conceptual design
This phase includes the determination of the best construction
approach and specific resources required implementing it, including
technical, staff, financial, and organizational resources.
Phase D : Design
The detailed specifications of the system components, structure and
features are determined.
Phase E : Construction
A DSS can be constructed in different ways depending on the design
philosophy and the tools being used.
The construction is the technical implementation of the design.
Phase F : Implementation
The implementation phase consists of the following tasks: testing,
evaluation, demonstration, orientation, training and deployment.
Several of these tasks are performed simultaneously.
Testing Data on the system’s outputs are collected & compared against
the design specifications.
Evaluation The implemented system is evaluated to see how well it meets
user’s needs.
Technical & organizational loose ends are also identified.
Testing & evaluation usually result in changes in the design &
construction.
Demonstration Demonstration of the fully operational system capabilities to the
user community is an important phase.
Orientation This involves instruction of users in the basic capabilities &
operation of the system
Training Operational users are trained in system structure and functions.
Deployment The full system is operationally deployed for all members of the
user community.
Phase G : Maintenance and documentation
Maintenance involves planning for ongoing support of the system and
its user community.
Phase H : Adaptation
Adaptation requires recycling through the earlier steps on a regular
basis to respond to changing user needs and thus as several variables.
No comments:
Post a Comment