NSI Executive Summary
(This summary document was presented as a poster session at the National Symposium on Educational Applications of
Technology, sponsored by the National Technical Institute for the Deaf in Rochester, New York, July 20-22, 1994.)
The purpose of this study is to improve captioning of "non-speech information" (NSI). NSI is a term to describe aspects of
the sound track, other than spoken words, that convey information about plot, humor, mood, or meaning of a spoken passage.
Currently, captioning companies vary in how they portray this information. Our study showed that consumers do not recognize
features commonly used to represent NSI, when they are presented in isolation. The outcome of the study will be recommendations
to the captioning industry to standardize the way it indicates this information.
Examples of NSI are:
Identification of speaker (off-screen speakers and multiple on-screen speakers)
Music (singing, background music, etc.)
Manner of speaking (whispering, emotion, word emphasis, etc.)
Audience reaction (laughing, groaning, booing, etc.)
Indication of title (book, film, newspaper, play, etc.)
Examples of features that can now be used to identify NSI are:
Double chevrons >>
An advisory committee composed of consumers and caption industry representatives advised the research staff on all aspects
of the study.
Thirty-three caption writers were interviewed for input as to the challenges they face in representing NSI.
Thirty-seven hours of video were analyzed in detail to determine current practice in identifying NSI.
Nineteen video clips from television were selected, giving 19 different examples of NSI. Emphasis was on NSI identified by
the advisory committee and industry as being most important, such as speaker identification.
For each of the 19 examples, two or three different ways of captioning to indicate NSI were selected. New uses of captions,
incorporating new features in the new caption chips, were included as well as conventional features such as italics. In all, 55
uses of caption features were included.
Deaf (n = 106) and hard of hearing (n = 83) consumers viewed the tape and selected their preference from the choices
presented. If any of the choices were unacceptable, consumers were instructed to mark those choices with an X.
Results were analyzed and recommendations to the industry were written. These were circulated for comment to the captioning
industry in the summer and early fall of 1994. The final report will be published in early 1996.
Summary of Results
Consumers wish to see more NSI indicated than is now industry practice. They preferred versions that indicated NSI such as
background music, audience laughter, accents, and sound effects.
Explicit description was most often preferred as the method of indicating NSI. This was particularly true in speaker
identification. The punctuation used--brackets, parenthesis, colon--is relatively unimportant, based on findings of the study,
although there was a slight (but non-significant) preference for parenthesis.
( audience laughing)
( soft, sad background music )
( southern accent )
( female narrator )
Italics were preferred for indicating when a speaker emphasized a word or phrase.
DON'T EVER, EVER DO THAT AGAIN.
Quotation marks were preferred for indicating titles.
The conventional form of indicating singing, with music icons enclosing the words, was preferred over a paint-on style.
Explanation of a pun was preferred over no explanation.
Color was not a preferred method of indicating NSI, although it was tested five times in this study. This was also found in
an earlier study by King and LaSasso (1993).
Flashing captions, paint-on captions, and underlining were not preferred in this study.
The preferences of deaf and hard of hearing consumers followed the same pattern in the vast majority of cases (16 out of 19
clips). The differences were minor and not of consequence for recommendations to the industry.
Back to home page