E-Speech uses voice recognition to teach reading.
By: A.J. Moore
Finding better methods to teach reading and speaking in English is an age-old challenge faced by parents and educators.
With a global economy and an influx of immigrants working in the United States, the need for instruction in English as a second language has risen dramatically.
Because of such demand, E-Speech, a company with local connections, is beginning to gain attention for its work in developing interactive software that employs voice recognition technology to teach and improve reading and speaking skills.
E-Speech has already been awarded an initial $100,000 National Science Foundation Small Business Innovation Research grant. Currently, the company’s two founders, Marian Macchi of Princeton and Dan Kahn, based in Morris County, are in the process of applying for a phase-two grant to continue their work toward making it easier for people of all ages and backgrounds to learn and speak English via the Internet.
The two researchers somewhat stumbled into creating this type of educational product with it eventually becoming Ms. Macchi’s pet project. E-Speech is primarily a technology and consulting firm that has developed text-to-speech software, speech synthesizers and voice recognition capabilities that allow a computer controlling voicemail or directory assistance to understand the human voice.
"We got into this because we taught a computer to speak English, so we wondered if we could use that same software to help people speak English," said Ms. Macchi, who works and resides in Princeton.
Developing this type of interactive technology is extremely complicated as the two have created algorithms and codes that enable a computer to understand and voice the spoken word. E-Speech not only has experience creating software that understands English, but other European and Asian languages as well.
Mr. Kahn said the most difficult process is creating rules for the computer to understand and then speak names and places, words not commonly found in the dictionary and which often have numerous pronunciations.
Before creating E-Speech in 1997, Ms. Macchi and Mr. Kahn were researchers at Bell Laboratories in Murray Hill. With their technological expertise and the knowledge of how to establish rules for a computer to understand human language, the two saw an alternative use for E-Speech’s services in the educational sector.
With so many foreign nationals in the United States needing assistance in learning English as a second language, E-Speech’s co-founders think they can provide a more cost-effective solution than what is currently offered.
"Adults or immigrants typically enroll in a course, but that requires a room and teacher," said Mr. Kahn, who received a doctorate from the Massachusetts Institute of Technology. "There are advantages of a computer-based system where someone can just go to our Web site and practice their words to a computer which then helps them and gives them the different rules."
A user of E-Speech’s software can look on a computer screen to see words written in English and then click on them to hear the text voiced. Essentially, it teaches by the use of phonics.
This type of software is also designed to help younger English-speaking school children, along with those with learning and physical disabilities.
Because it is a Web-based service, Mr. Kahn envisions an international clientele.
"Learning English is becoming important for countries all over the world," he said. "There has been an increase of people in China learning English for example. Because of the Internet, they (people in China) can learn English from us."
Both Ms. Macchi and Mr. Kahn believe using the Web as a platform to teach English can be particularly helpful for adults because they can practice in the privacy of their own homes and not feel self-conscious about making mistakes in front of others.
Since this business plan is such a large undertaking for a small company that has just two full-time employees, Ms. Macchi says E-Speech is currently in discussions with larger companies, namely publishing houses and educational software firms, in hopes of forming a partnership.
The two have also drawn on Dr. Dorothy Rubin, professor of language and communication sciences in the reading and language arts program at The College of New Jersey’s School of Education as a consultant.
E-Speech’s present roster of clients for their other voice software services includes IBM, Educational Testing Service, Telcordia Technologies and Eastman Kodak.
Ultimately, Ms. Macchi and Mr. Kahn believe their new product could be marketed to companies looking to help teach English to their workers and to school systems that have an open mind when it comes to using technology.
"Some people feel you absolutely need a human to teach English," said Ms. Macchi, who earned a doctorate in linguistics from New York University. "Some don’t think you need one especially if you can make the software engaging for human beings."
The two see the software as a valuable classroom supplement for elementary and middle school teachers as it can help students practice English pronunciation during their free time.
E-Speech has used a prototype of the software with test and control groups at Montclair State University, and according to Ms. Macchi, the results have been encouraging.
Even if the company’s application for another grant from the NSF is not awarded, Ms. Macchi and Mr. Kahn will continue their work developing interactive software to teach English. The goal for E-Speech is to have its educational language software available to consumers in two years.
For more information about E-Speech visit www.espeech.com.

