2003 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2003 Table of Contents 


The C-Print Speech-to-Text System for Communication Access and Learning

Presenter
Pamela Francis and Michael Stinson
National Technical Institute for the Deaf
Rochester Institute of Technology
52 Lomb Memorial Dr
Rochester, NY 14623-5604
Email: pggncp@rit.edu

C-Print is a speech-to-text transcription system developed primarily to help support students who are deaf or hard of hearing in general education classes with hearing students. The C-Print system includes specialized software and training. The specialized software produces real-time text through automatic speech recognition (ASR) and/or a computerized keyboard-based abbreviation system. An intermediary operator, or "captionist" in the classroom with the deaf/hard of hearing students dictates into a microphone linked to a laptop computer containing the ASR software (and/or types with the abbreviation system). A text display of the message appears on the monitor of a student laptop or other display approximately 3 seconds after the words have been spoken by the teacher. The specialized software, called "C-Print Professional (Pro)," also has computer networking capabilities and displays for the captionist (host) and student (client) computers. This software enables students to participate in class discussion, to highlight the text produced by the captionist, and to easily take their own notes. Given current ASR and keyboard technology, it is desirable to have both the ASR and abbreviation approaches because this provides greater flexibility than a single approach.

An extensive program of research has provided evidence that the C-Print system works effectively in public school and postsecondary educational settings (Elliot, Foster, & Stinson, 2002; Elliot, Stinson, McKee, Everhart, & Francis, 2001). More than 500 individuals through out the United States and internationally have completed the 5-week training program to become a C-Print captionists (Elliot, McKee, & Stinson, 2002). The presentation will demonstrate the C-Print system and discuss its use in educational settings.

Equipment. The C-Print system uses laptop computers compatible with Windows 2000 or greater as server computers that captionists may dictate or type into. The client or student computers are compatible with Windows 2000, or greater, or with Mac OS9 or greater. A variety of local area network (LAN) cards that fit into computer slots are used for radio-frequency based wireless communication among computers. The system uses IBM's ViaVoice (Version 9 or greater) as the ASR software.

ASR. In producing text with ASR, as the instructor (or student) speaks, the captionist dictates continuously into the dictation mask, saying each word distinctly, at a pace that keeps up with the instructor. The captionist adds punctuation and formatting. An example of formatting is moving to a new line and indicating whether the instructor or a student is speaking, as needed. Since it is not possible for the captionist to capture lectures verbatim, s /he must summarize or "condense" information. The goal of condensing is to capture the important points while using fewer words than the original speaker. The captionist occasionally switches from dictation to typing, such as when a word is not in the dictionary or when a mistranslated word needs to be corrected. In addition, the captionist performs preparation and editing activities. Before the first session of a class the captionist creates a voice file specifically for that class, and enters specialized vocabulary as well as available information that will help ViaVoice construct a model of how language may be used in that class. After class, the captionist corrects errors in the text for distribution of the text to students. In addition, words that ViaVoice mistranslated during class are trained so that it will translate them correctly in later classes. The procedures for using ASR to produce text as a support service have just been developed and are being further refined.

Word-abbreviations. With the keyboard-based computerized word-abbreviation system, as the instructor talks, the captionist types a series of abbreviations. The interface software includes a dictionary and word-expansion component. The software searches the dictionary for the equivalent full word and displays it on the screen. (e.g. Captionist types the abbreviation "slvg" and this expands to full word "solving" on the computer screen.) The captionist does not memorize all the abbreviations in the system, but rather uses a set of phonetic rules which are then applied to any English word that has been added to the software's dictionary. The general dictionary contains approximately 12,000 abbreviation-word pairs. The captionist also condenses information in a manner similar to that for ASR. In addition, the captionist performs preparation and editing activities. These include addition of new abbreviations for words that are not in the dictionary that will be used in class and correction of errors in the transcript before distributing it.

User interface software. This software, C-Print Pro, includes separate text displays for captionist and student, allows keyboard and ASR input, provides networking capabilities, and has educational tools. The separate text displays permit the captionist and student to adapt their computer screens to individual needs such as different font size or type. For example, deaf-blind students may enlarge the font size on their computer. The software provides general formatting capabilities similar to word processing software; e.g. "open," "close," "cut," and "copy." With ASR input, the interface software accepts and displays text that is produced with the IBM ViaVoice engine. With keyboard input, the software provides automatic expansion of word abbreviations, and it supports standard typing as well. The networking component supports the communication between the captionist (server) and student (client) computers. In addition to sending text from the captionist to students, it includes a chat feature that enables student and captionist to communicate with each other separate from the stream of text. The software provides for identification of students registered on the network and ability of the captionist to disconnect students on the network. The educational tools include an optional split screen for the student to take their own notes. A "tag" feature connects these notes to specific lines in the text section of the screen. Students can also high-light important points in the text section. The file can be saved with this highlighting so that the student can note information that is important to remember later for assignments, tests, etc.

References

Elliot, L., Foster, S., & Stinson, M. (2002). Student study habits using notes from a speech-to-text support service. Exceptional Children, 69, 25-40.

Elliot, L., Stinson, M., McKee, B.G., Everhart, V. & Francis, P. (2001). College students' perceptions of the C-Print speech-to-text system. Journal of Deaf Studies and Deaf Education, 6, 285-298.

Elliot, L., McKee, B.G., & Stinson, M. (2002). C-Print captionist survey. National Technical Institute for the Deaf, Rochester, NY.


Go to previous article 
Go to next article 
Return to 2003 Table of Contents 
Return to Table of Proceedings


Reprinted with author(s) permission. Author(s) retain copyright.