The purpose of this document is to outline and explain the specific requirements that the LISTEN system will fulfill, as well as the constraints under which the system is operated. Readers of this document should have a general knowledge of computers and computer architecture.

1.2 Scope

The Americans with Disabilities Act was a landmark piece of legislation requiring government and private business to make appropriate facilities and services available to all Americans with disabilities. Its scope is vast, requiring that obstacles in the path of participation for all Americans be removed. While compliance from a physical standpoint for disabilities of a mobility nature is easy to gauge, compliance is much less simple for those with hearing disabilities. Sign language interpreter is not a common choice in profession, and given the small number of deaf Americans when compared to the populace in general, it’s no surprise that there’s often no interpreter available.

Arguments have been put forth that literate deaf citizens can make use of printed and written materials, without suffering any ill affects, or disadvantage when compared to their hearing able counterparts. While this is indeed true in the case of forms and other written materials, it fails to take into account situations where an interpreter would be most useful, such as public meetings, lectures, and civil gatherings. Lack of an interpreter in these cases can lead to deaf citizens being excluded from public participation in events that affect their lives.

Also often ignored is the psychological effect that the lack of an interpreter, can have on an individual. Forced to rely on written communication or an interpreter, a deaf individual is forced into a situation where personal autonomy must be sacrificed as a cost of entrance into society. This is both inconvenient, and more importantly unjust.

LISTEN will not be a panacea for deaf-hearing communications. The limited scope of the project necessarily means that the communication is one way. But as a piece of software utilized to allow a deaf individual the opportunity to provide himself with the ability to provide his own interpretation, LISTEN will provide both personal utility, and personal empowerment to deaf individuals.

1.3 Definitions

GUI: Graphical User Interface.

LISTEN: Language Interpreter and Sign Translator for Educational Needs.

ASL: American Sign Language. This version of Sign Language does not have the same grammatical structure as the conversational English used by the non-deaf community.

Signed English: Signed English is a version of Sign Language that shares the same grammatical structure as the conversational English used by the non-deaf community.

Fingerspelling: The act of translating a word using the sign language alphabet.

Interpreter: A member of the non-deaf community whose occupation is to translate spoken word into either Signed English or ASL for a member of the deaf community.

1.4 Overview

This document outlines the LISTEN project with respect to functionality and context. It details the system entities, data flow, and human interfaces of the project. Interface demonstrations have been created, though are subject to change through the full development process.

1.5 Roles and Responsibilities

Jennifer Crowell - Project Lead, System Designer. Currently employed as a developer for Lockheed Martin, Jennifer brings to this project both her extensive experience as developer and her personal connection to the deaf community.

Abraham Evangelista - Video Producer, Documentation Lead. Having lived and worked with language disparate groups, Abraham is all too familiar with the difficulties that language barriers present. His work in television industry will provide video production expertise to this project.

Sugnesh Patel - Integration Specialist, Tester. As a speaker of several languages, Sugnesh is aware of the difficulties involved in multilingual communication. Along with his experience with the .net development platform, Sugnesh also brings to this project his extensive experience as a tester and system integrator.

Susan Philip - Lead Tester, Quality Assurance. Susan was responsible for developing the prototype and testing the projects during her co-op with Blue Cross. She also holds a bachelor‘s degree in Computer Application from Marian College in India and was part of the senior project team. She brings to the project her interest in application development, and a keen eye for vigorous testing.

Kulvir Singh - Interface specialist. Kulvir worked in the Software Integration and Testing group at Comcast his last co-op as a GUI Tester and MFC application bug fixer. His responsibilities include the implementation of the GUI interface for the LISTEN system and creating the help files.

2 Overall Description

2.1 Product Perspective

LISTEN software is not self-contained and uses third party software for the speech to text conversion.

2.2 Product Functions

LISTEN aids the physically challenged by converting the vocal speech to a more familiar language sign. There are 4 modes for the application.

2.2.1 Classroom Mode

This mode is used by students in class. The instructor’s voice is converted to sign and then displayed in the students monitor.

2.2.2 Presentation Mode

This mode is used by speakers during their presentation, in conjunction with external presentation software such as Microsoft’s Office Power Point, or Open Office Impress. The presenter’s speech is converted to sign and be displayed.

2.2.3 Tutor Mode

This mode is used by students that wish to learn sign language. The user clicks on a letter or word and the sign’s graphic is displayed.

2.2.4 Playback

This mode plays back a previously saved session.

2.3 User Characteristics

The users for the software are the hearing-impaired, prospective students of sign language, and presenters.

2.4 Constraints

There are no constratins that need to be taken into consideration for this project.

2.5 Assumptions and Dependancies

LISTEN software depends upon third party software package to convert sound into text.

2.6 Development Model

The Waterfall development model will be used for the development of the LISTEN software.

3 Functional Requirements

3.1 Speech to Text Converter

LISTEN requires third party software to convert from speech to text. This software should at least encompass the English Sign Language vocabulary. This software must run on all Microsoft Windows 2000 and up.

3.2 Graphical User Interface

This section describes the look and feel of the LISTEN GUI.

3.2.1 Four Selection Mode Buttons

At the start of the application, the following buttons are displayed:

3.2.1.1 Classroom Mode

Selecting this mode displays the Graphic Display Window, and begins interpretation from the audio source.

3.2.1.2 Tutor Mode

Selecting this mode displays the Tutor Window.

3.2.1.3 Presentation Mode

This mode is used to integrate with presentation software such as Microsoft Power Point, or Open Office Impress. Selecting this mode displays the Graphic Window Display with only the upper pane, overlaying the presentation software, and begins interpretation from the audio source.

3.2.1.4 Playback Mode

This mode displays the full Graphic Display Window, opens a file selection window, and prompts the user to select a previously saved session.

3.2.1.5 Exit

Selecting this button will terminate the program, and the third party speech-to-text application.

3.3 Graphic Display Window

The Graphic Display Window has two panes. The upper window displays the graphic output. The lower window displays the English text that is being translated to Signed English. There is a control bar containing the Play, Pause, Rewind, Forward, Replay buttons and Speed Slider.

3.3.1 Tutor Window

The tutor window contains Graphic Display Window, in addition to a list of all the English words in the current dictionary.

3.3.2 Presentation Display

The graphic output is displayed on the presentation screen.

3.4 Graphic Display Window Requirements

The window has the following parts

3.4.1 Upper Window Pane

This pane displays the video or graphic clip. The window is large enough to see the complete output clearly.

3.4.2 Lower Window Pane

This window pane displays based on the mode any two of the following

· English text that is being translated to Signed English.

· Text Description of the video or graphic clip when in Tutor mode.

· Hide Button which hides the Lower Window Pane

3.5 Control Bar

The control bar is displayed on the bottom of the Graphic Display Window. The bar contains the following components:

3.5.1.1 Play and Pause Button

The play button plays the Video or Graphic clip. While the video or graphic clip is playing, the Play text changes to Pause. The Pause button pauses the graphic display output.

3.5.1.2 Rewind Button

The rewind button stops playback, and then resume displaying from a previous sign. The user determines how many signs previous that play resumes from, by how many times he clicks.

3.5.1.3 Forward Button

The forward button stops playback and then resume playing from a later sign. This button is only active while the sign being displayed is not the last sign.

3.5.1.4 Stop Button

The stop button ends graphic output. When the user clicks the play button, play resumes from the beginning of the session.

3.5.1.5 Speed Slider

The user uses the speed slider to slow the speed of graphic output. The default speed is the fastest speed.

3.6 Window Dimensions

The window dimensions are as follows:

· The Graphic Display Window is divided horizontally.

· The Upper Window Pane fills 80 percent of the window area.

· The Lower Window Pane fills the remaining 20 percent of the window area.

· The Pixel Size of the Graphic Display Window is determined by the current display resolution of the system running LISTEN.

· The Graphic Display Window has a fixed 4:3 height-to-width ratio.

· Users cannot alter the dimensions of the graphic display window, except to select full screen mode.

· The Graphic Display Window can be full-screened by double clicking the graphic display. Doing this causes the Graphic Display Window to increase in size to the current display resolution and to overlay all other existing windows.

· LISTEN can be minimized.

· The Graphic Display Window occupies 33% of the screen.

3.7 Window Behavior

The window behaves as follows:

· The Graphic Display Window is not resizeable.

· The graphic display window will ‘snap’ to corners of the display.

· Upon closing, if the program is operating in Classroom or Presentation mode, the Graphic Display Window will prompt the user to save the session to the file. Saving the session will write the tokens to a comma delimited text file.

3.8 Tutor Window Requirements

The tutor window contains the following elements:

3.8.1 Graphic Display Window

Tutor Window has all the Graphic Display Window Requirements. The graphic clip is displayed for the selected word form the Expandable Word List.

3.8.2 Vertical Window Pane

This is on left side of the Graphic Display Window. This contains the following elements:

3.8.3 Options Combo Box

This combo box contains selections for Category, Sign English Dictionary, and Alphabets and Numbers.

3.8.4 Expandable Words List

This contains the list of words. The list displays words depending on the selected option from the combo box. These words are drawn from the word/symbol database. The Options contained in the combo box selection are:

3.8.4.1 Category

Selecting this option in the Options Combo Box displays all the Signed English categories.

3.8.4.2 Signed English Dictionary

Selecting this option in the Options Combo Box displays all the English words that exist in Signed English.

3.8.4.3 Alphabetic and Numeric Characters

Selecting this option in the Options Combo Box displays the English alphabet and numbers

3.9 Presentation Display Requirements

3.9.1 Presentation mode

Presentation mode displays only the Graphic Display Window.
The Graphic Display Window overlays the external presentation software window.

3.10Failure Mode and Error Handling

3.10.1 Third Party Software

LISTEN relies on third party software to convert from spoken language to text formatted data. Failure of this software will not affect LISTEN.
LISTEN handles the launching and closing of the speech-to-text application.
LISTEN monitors the status of the speech-to-text application.
If the speech-to-text application crashes or hangs, LISTEN stops the application’s process and restarts the speech-to-text application.

3.10.2 Error Handling

If the speech-to-text application produces a symbol that is not in the LISTEN dictionary, and cannot be fingerspelled, LISTEN can handle this token.
LISTEN displays a symbol indicating an unknown token and continues to parse the stream.
Listen writes the token to the database for later analysis.

3.11 Non-Functional Requirements

3.11.1 Hardware Requirements

Microsoft Windows 2000 or later.
Pentium-4 1.5Ghz or equivalent.
Minimum memory of 256 MB.
“Supported” Microphone.

3.11.2 Hardware Recommendations

Monitor capable of displaying at a resolution of 1280x800 or better.
2.0Ghz or Multiprocessor system.
1.0GB RAM.

4 System Evolution

System evolution refers to the dynamic behavior of software systems as they are maintained and enhanced over their lifetimes and is particularly important as systems in organizations become longer-lived. Following are the areas we have identified so far where system evolution has a scope.

4.1 Other Languages

Currently the application supports Signed English. There is scope to extend the software to support other sign languages.

4.2 Platform Independence

Due to system calls within the application, the current application is not fully platform independent. But future updates using platform independent techniques are possible.

4.3 Real time Learning

The application can be further developed as a real time learning tool, which does not require the student to be physically present in the class and thus facilitates distance learning.

4.4 Finger Spelling for Words

English Sign Translation Mode can be further extended to include the finger spelling graphics for the included words, rather than only for words that are not in the dictionary.