Voice relay for the Deaf with Asterisk

By

Docas Dudu Zulu

A thesis submitted in partial fulfilment of the requirements for the degree of Bachelor of Science Honours in the Department of Computer Science, University of the Western Cape.

Date:02 April 2009 University of the Western Cape

ABSTRACT

Voice relay for the Deaf with Asterisk

BY DOCAS ZULU

Chairperson of the Supervisory Committee:William D. Tucker Department of Computer Science

Deaf people have restricted services when it comes to mobile phones. My project will explore the conversion of Text-to-Speech using Festival and Speech-to-text using Sphinx where Asterisk will serve as a media gateway. The focus of this project is to provide a communication bridge between a Deaf user and a hearing user, a hearing user will be able to call a Deaf user via a mobile phone and the Deaf user will receive a text message instead of voice and vice versa. This will be achieved through the use of SIP (Session Initiation Protocol) on mobile phones in Voice Over IP solution.

ii

Table of Contents Contents Abstract...... ii TABLE OF CONTENTS...... iii GLOSSARY...... iv CHAPTER 1...... 1

Introduction...... 1 Introduction ...... 1 CHAPTER 2...... 2 Users Requirement Document...... 2 Introduction ...... 2 User’s view of the problem...... 2 Software expectations...... 3 Not expected from the software...... 3 CHAPTER3...... 4 Requirements Analysis Document...... 4 Introduction...... 4 Interpretation of user’s requirements...... 4 PROJECT PLAN...... 6 REFERENCES...... 7

iii

Glossary

Session Initiation Protocol (SIP) Voice Over IP (VOIP) Private Branch Exchange (PBX) Deaf Community of Cape Town (DCCT) iv Chapter1 Introduction

Communication is a way of interchanging thoughts, opinions or information by means of speech, writing or signs. Mobile phones are used as means of communication where users can text, video conferencing, play games, make calls, receive calls etc. These services that are provided by mobile network operators are however not designed for people with disabilities like the Deaf community. In the Deaf community there are Deaf people with severe auditory impairment and Deaf people that are prelingually deaf, those are people that are hearing impaired or have difficulties in hearing. This implies that it is really not easy for Deaf people to communicate with their loved ones who are not Deaf but could sign (Tucker, Glaser & Penton, 2002).

There are organizations around the world that provides the needs of Deaf people like Deaf Community of Cape Town (DCCT). Some Deaf people can write and read English, so they can use SMS or Instant Messaging as form of communication. This means that Deaf and hearing people can primarily use SMS or Instant Messaging to communicate. The idea of this project is to design a communication bridge between a Deaf and hearing user using Asterisk. This means that a Deaf user will send text to a hearing user and a hearing user will receive voice instead of text.

1

Chapter 2 Users Requirements Documentation Introduction

The User Requirements Document discusses the user’s point of view, what the users require. It broadly describes the problem domain, the reason why this project has to be designed. The introduction project on this documentation I described the problem at hand, why Deaf and hearing users need this communication bridge. I also discuss what the users are expecting from this software, i.e. the functionality of the software system and what the system should not do, which are things that are not covered on this project. The user’s view of the problem The current communication between a Deaf user and a hearing user is through asynchronous communication. If a hearing user would like to communicate with a Deaf user he/she has no choice but to SMS or possibly use an email. Hearing users would like to make and receive voice calls to/from Deaf users. Deaf users would like to send and receive text either using Instant Messaging or SMS. This is possible with a relay operator where Deaf users are able to request a relay operator to call a hearing user by typing a text message using a Deaf telephony , the relay operator will then relay the Deaf user’s message to the hearing user(See Figure 1)(Tucker, 2008). The honours project focuses on mobile network technology where Deaf users will be able to communicate with hearing users using their VOIP phones but without the presence of a human relay operator.

Figure 1 shows communication between a Deaf user and hearing user wtih a relay operator. 2

Deaf user Hearing user  Send/receive text  Make and receive calls  Instant Messaging  IP phone or Cellphone Figure 2 the table shows the user requirements

Software expectations

A hearing user should be able to place voice calls to a Deaf user and a Deaf user should be able to receive a text message instead of voice. The system should be understandable and easy to use so that users would not find it difficult to register their new numbers. A hearing user should be able to choose that which voice would he/she like to listen the new message with and this should not need telecommunication expect. The system should give the option of repeating a message in the hearing user’s side and a Deaf user should be able to read a message more than once. Not expected from the software

The system is only used as a calling and texting for hearing and Deaf users. It is not expected to work when a user send a video to a Deaf user. This means that the system is not for video calls but for voice calls only. The system is not expected to work when a Deaf user is texting another Deaf user or a hearing user is calling another hearing user.

3

Chapter 3 Requirements Analysis Document Introduction

The RAD is the interpretation of the user’s requirements from the designer’s point of view. This document focuses on what is needed so that the project may be successfully developed, how we go about implementing the user’s requirements. The RAD breaks URD into modules and defines their purpose and relationships.

Interpretation of the user’s requirements

As soon as development starts software tools will be required, therefore open source tools Sphinx and Festival will be used. Festival is a tool that we definitely need in this project because it is a tool that will convert the voice message to text (The University of Edinburg, 2009), where as Sphinx is a tool that will help us convert text to voice(Aksyonoff, 2009).This tools will be integrated in Asterisk. The system should provide a communication bridge between a Deaf user and a hearing user. The users should have SIP phones as we are exploring the VOIP solution, this means that the project will have to use VOIP functionality to send SIP based communication to the Asterisk server. The system should allow a Deaf user to send text to the hearing user via Asterisk server, the hearing user will then “hear” the Deaf user’s typed text; the text string is converted to speech using Festival and when the hearing user replies the speech is converted to text using Sphinx.

In addition, the system should have administrative control so as to register the user’s SIP numbers to Asterisk and also the administrator should be able to setup the system such that a Deaf user is identified and also a hearing user is identified. Since there will be an integration of Sphinx and Festival, the system the Asterisk server specifically will have dial plans that will give the extensions of each and every user and also how is Asterisk server going to manage the Deaf and the hearing users when they are communicating. So in case of Festival there will be dial plans that will call Festival, this means that in all users that will require Festival their extensions will have a way of invoking Festival same applies to Sphinx, because the dial plans will be in a way that those that are Deaf their dial plans will call Sphinx when their receiving.

4 The Asterisk has been identified as a potential media gateway for the system; it is a software “Private Branch Exchange” that has support in Automatic Speech Recognition and text-to-speech. Therefore the Asterisk PBX has to detect if the message is text or voice, if it is text Asterisk has to invoke Festival to convert to speech and if it is voice Asterisk has to invoke Sphinx to convert it to text (see figure 2). In addition, a hearing user should be able to choose which voice he/she will like to use for each and every contact person in the dial plan.

Figure 2 The diagram shows the communication between a hearing user and a Deaf user. The Deaf user send text to a hearing user this message will go via Asterisk server where sphinx will convert text-to-speech and also when the hearing user replies the message goes via Asterisk where Festival convert hearing user’s speech to text.

5

Project plan 6 References

Tucker,W. D. (2008). Softbridge:a social aware framework for communication bridges over digital divides. Published Doctor of Philosophy’s thesis, University of Cape Town, Cape Town, South Africa.

Tucker,W. ,Merly, G. & Jason,P. (2002). A Bridge for the Problem of Deaf Telephony. Retrieved june , 2002 , from http: www.scienceinafrica.co.za The University of Edinburg. The festival speech synthesis system. Retrieved April 01,2009, from http: // www.cstr.ed.ac.uk/projects/festival/ Aksyonoff, A.(2009). Free open-source SQL full-text search engine. Retrieved April 01, 2009,from http: //www.sphinxsearch.com/docs/current.html#author 7