Speech to Text in Special Education

What is it?

Speech to text, or voice recognition technology, is a computer application that can be downloaded to a computer, smart phone or compatible tablets. The application allows the user to transfer speech into text simply by speaking. Dependent upon that program, some software even allows the user to control all computer functions simply from speech.

How does it work?

How does it work? For a user to successfully use the software, he/she must learn how to use "discrete speech," or dictate word-by-word. This allows the computer to recognize the individual words. The program must then be "taught" to recognize the user’s voice. This occurs through a series of tests where the speaker is asked to pronounce a set of words. After the test, the program will create a specific file for that user that contains the information on that particular voice. The software will then use this information when attempting to translate words during dictation.

For further information: http://www.rehabtool.com/forum/discussions/97.html

History

Speech to Text technology seems like a fairly new development; however, most would be surprised to find that the first synthetic speech was developed as early as the 18th century. A machine created by Wolfgang von Kempelen, was built by wood and leather but was till able to generate audible speech. Eventually in the 20th century, electricity made it possible to create the first electric speech synthesis called, Voder. Voder was created by Homer Dudley and was revealed in 1939 at the world fair in New York. Other creations soon followed such as Gunnar Fant’s, OVE and Walter Lawrence’s PAT, or Parametric Artifical Talker. Although the idea came way before it’s time, it can be said that the most improvements have come in the last ten years. Software now allows for easy speech recognition, computer commands and reads back with fluency. These programs and moving away from machine, and more into humans.

For further information: http://www.rehabtool.com/forum/discussions/97.html

Speech to Text: Examples

This is a video showing a recreation of Kemplen’s first speaking machine.

A great example of Voder’s speaking machine and how improvements were made from Kemplen’s to this one.

What’s Available for Classrooms

Communication Access Real-time Translation (CART): This captioning method employs a trained stenogra- pher using a special keyboard to produce a practically verbatim script of what is said.

PROS: *Employs highly trained professionals who must have court reporter skills and beyond *Produces a “live” highly accurate transcription *Transcribes all auditory information in the classroom.

CONS: *Requires court reporter skills plus additional training *Produces 15-20 pages of transcript for an hour’s lecture *Transcribes every word regardless of its relevance.

Text Interpreting: C-Print or Typewell are examples of this technology. In these systems a hearing transcriber uses a notebook computer with the C-Print Pro or TypeWell abbrevia- tion software to transcribe a condensed text of what is said in lectures and discussions. The deaf person reads the transcript in real-time from a second computer.

PROS: *Eliminates extraneous information *Is a translation of the speaker’s intended meaning into concise sentences *Allows for notes to be added by the student *Is less costly than CART

CONS: *Requires the transcriber to understand the topic being translated *Produces a script that is not a verbatim transcription *Requires 35 to 60 hours of training to become certified

For further information: http://www.typewell.com or http://www.ntid.rit.edu/cprint/what_cprint.php

Automatic Speech Recognition: Ex. Dragon Naturally Speaking; CART stenomask provider

This technology involves a provider speaking into a special mask with a built-in microphone. Everything the provider says is transcribed automatically by computer software. All punctuation, new paragraphs, etc. must be voiced for the computer. The goal of this system is to produce a verbatim script; however the computer often translates words incorrectly, and the script must be edited by the provider. At this time, automatic speech recognition does not appear to be appropriate for public school use.

For further information: http://www.nuance.com/dragon/index.htm

Speech to Text in Special Education and General Education Classrooms

There are many different ways to use Speech to Text software in general education classrooms as well as special education classrooms. In a general education classroom it can be extremely difficult for deaf or hard of hearing students to really understand what is going on. While they can be provided other services like interpreters or FM systems these can do more harm than good. In a class where the students need to be working while a lecture is going on it’s hard for a student who is deaf to be able to pay attention to their interpreter and be working. Text interpreting or speech recognition software can really be beneficial for these students.

Using speech recognition software will require some cooperation from the general education teacher but if done properly it can be beneficial to not only those students with disabilities but also the rest of the students in the class. When using software like dragon dictation it’s important that the teacher can practice with it beforehand so it can ‘learn’ there voice. After the teacher has gotten acquainted with the software they can use it when giving lectures in class. The software will make a transcript of what the teacher says allowing students who are hard of hearing or deaf to pay more attention to their interpreters instead of taking notes or they can read the transcript as it’s being typed so they are capable of doing work and knowing what’s happening at the same time. The teacher can decided to make these transcripts available to everyone in the class which might benefit those students who aren’t disabled but still need help.

You can also use the system TypeWell when deemed appropriate. When dealing with harder subjects interpreters can’t always be counted on to transcribe the information correctly. You can use the TypeWell system in hard classes where the information might need to be translated into an easier description then what a verbatim account from an interpreter can provide.

Speech dictation can also be extremely important both in the actual classroom and at home when doing homework. There are those who have hard times using their hands or using a computer. When you have students who have these problems it may seem daunting and you may feel that the only thing you can do is change the things you want to teach in class or have the students do. BUT you don’t have to change your lessons to something completely different. As a teacher you can try speech dictation software and show your students how to use it. Instead of changing your lesson so that those students who have issues using their hands don’t feel left out you can include them completely and allow them to be independent, which is extremely important. Teaching disabled students how to use their voice to govern the computer will allow them to do almost everything that a non-disabled student could do.

For Further Information

http://www.readingonline.org/articles/art_index.asp?HREF=balajthy2/index.html

http://www.rehabtool.com/forum/discussions/97.html