You pick up your remote to turn on the TV and do a little channel-surfing. For many people with hearing loss, this is also the time to activate the TV's closed caption (CC) function. Closed-captioning makes otherwise inaudible speech readable: words spoken on the nightly news, favorite sitcoms, a documentary or sporting event. Text displayed on screen brings these programs to life for people with hearing loss. As for movies, we've come to expect the option to see subtitles in English as well as other languages. Clearly, people with hearing loss aren't the only ones to benefit. Subtitles and captions aid comprehension for everyone in noisy environments, like sports bars or airports. The case for text on screen has been irrevocably established.
Now imagine how helpful it would be to a student with hearing loss to be able to read everything that is spoken in a typical classroom lecture or discussion. Believe it or not, the technology is already available to achieve just that. Known as speech-to-text services, this technology captures auditory information (speech) and translates it directly into a readable format (text), in real time.
A major obstacle for students who are deaf or hard of hearing is not having complete access to auditory information presented in the classroom. Noisy classrooms, group discussions and even the limited ability of younger students to sit still can make it difficult to hear the teacher and other students. Another problem with noise is that teachers tend to talk louder to compensate for the loud noise. But talking louder is not only physically exhausting, it can also exaggerate or distort the sounds of speech. Over the course of the day, as teachers get tired of speaking, they will normally begin to talk more softly, which makes it even more difficult for a student with hearing loss to understand them.
In many cases, a personal assistive listening device (ALD) or classroom speaker system is provided as an accommodation for these students so that they can gain better access to auditory information. However, if the student depends heavily on lip-reading or speech-reading, auditory information may only be of limited help. Depending on the age group or the type of class, the teacher may need to move around the classroom and the student with hearing loss will lose visual contact with the teacher's face if he or she turns around or walks away. Students who do not use hearing aids or other "auditory" devices may use a sign language interpreter but this option is only beneficial for those who communicate fluently in sign language. Students with hearing loss often have the added challenge of trying to take their own notes, and the moment they take their eyes off of the teacher or the interpreter, they may miss important information.
Speech-to-text services solve the glitches in current classroom accommodations and provide benefits for all students. Spoken auditory information is transcribed and displayed on a projector screen or laptop display almost simultaneously as it is being spoken (see Figure 1). An electronic or paper copy of the entire lecture as it was produced is usually available afterwards potentially benefiting even hearing students.
Speech-to-text services are generally categorized into three broad groups: steno-based Communication Access Realtime Translation (CART), text interpreting (C-Print® or Typewell™) and automatic speech recognition. Each of these services requires a specially trained professional who knows how to properly set up and operate the software and equipment. CART and text interpreting services require that the provider listen and type while automatic speech recognition requires that the provider listen and speak (see Table 1 for a comparison of services). Providers undergo many hours of training and practice to develop and improve their efficiency and accuracy in listening and typing, or listening and speaking, at the same time for fairly long durations. For extended classes or sessions, there may be two providers who work as a team, enabling each one to take breaks at regular intervals. Additionally, most if not all of these services can be provided remotely without a provider being physically present. This is ideal for rural or smalltown schools where resources may be lacking.
CART providers use a stenograph or stenotype machine. Unlike the traditional computer keyboard, stenographs have two rows of keys without markings. Rather than typing each word exactly as it is spelled, the provider uses a shorthand method in which various keys are pressed in particular clusters to construct each word. CART has the advantage of being practically verbatim that is, every word that is spoken, regardless of relevance to the lecture, is captured and typed. A potential disadvantage, however, is that a lecture may produce as much as 20 pages of transcript!
With text interpreting (C-Print and Typewell) the provider uses a laptop computer and employs a partial shorthand method. Although all letter keys are available on a laptop computer keyboard, abbreviations of common words are typically used to form whole words. Long words can often be written without the vowels and the software will select the closest word match possible and fi ll in the vowels. The output of text interpreting typically uses a meaning-to-meaning approach, rather than a verbatim approach. Unlike CART, where every single word is transcribed, only the main points of the information are included and any repetitions and corrections that the teacher makes are not included. Text interpreting is a lot like American Sign Language interpreting because the full meaning of the message is relayed without having to translate every word. Because of the meaning-to-meaning approach, a hard copy of the lecture will take up fewer pages than a CART transcript. Certain versions of C-Print and Typewell allow students a more interactive experience. For example, if the output of C-Print or Typewell is projected to another laptop, then the student may have the option of adding their own notes, or deleting anything that they feel is irrelevant.
Finally, automatic speech recognition involves the provider speaking into a special mask with a built-in microphone which enables everything the provider says to be transcribed automatically by computer software. Because of the speech recognition capabilities of the software, little to no typing is involved. For automatic speech recognition to work well, the provider sets up a profile and "trains" the software, correcting common errors for that user. Whenever the computer misunderstands a word, the provider can correct it immediately, or later when reviewing the transcript. Like CART, automatic speech recognition is practically verbatim.
Whereas none of these services is perfect, speech-to-text providers are trained to provide the highest quality service and to ensure that their equipment is working properly. Nonetheless, the providers are human beings who may not always feel their bestevery day, may be ill or might need to miss work. And, though rare, all technology is prone to glitches. Therefore, having a backup plan is wise.
At present, speech-to-text services are most commonly used in educational settings at larger universities, though they are also provided in secondary schools when included in a student's individualized education plan (IEP). In these cases, the university or school usually bears the cost of the service, which is comparable to that of certified sign language interpretation. In a school setting, the speech-to-text provider follows the student to each class where the service is needed. Thus, the provider is a team member in the education of and advocacy for the student. However, the provider is not personally responsible for the welfare of the student, is typically not involved in educational decisions, and should not be used as a personal messenger to communicate with the student. The provider uses time between classes to set up and pack the equipment for transport.
Speech-to-text is useful in other settings as well, such as for transcribing legal proceedings in the courtroom. And it is not unusual to see these services at large conferences. The output of the speech-to-text service is prominently displayed on a large screen so that anyone in the audience is able to read the text.
Here in Arkansas, each summer we host an annual mock college camp for students with hearing loss. Students are exposed to various kinds of assistive technology, including speech-to-text service. Many of these high school campers have neither seen nor used speech-to-text services and are often amazed when they realize how much they might have been missing. Some of our campers request this accommodation when they go on to college.
Speech-to-text services are a great resource for people with hearing loss, as well as the public in general. Each of the services described above is an attractive option for visual and readable access to spoken auditory information. Not only is the information displayed in real time, but the same information is likely to be just as easily accessed later for comparing notes and studying. They powerfully contribute to making information accessible to all.
Samuel R. Atcherson, Ph.D., is an assistant professor with a joint faculty appointment at the University of Arkansas at Little Rock and the University of Arkansas for Medical Sciences (UAMS) and is the director of the Auditory Electrophysiology and (Re)habilitation Laboratory located in the Department of Audiology and Speech Pathology. E-mail Dr. Atcherson at SRAtcherson@ualr.edu.
Rachel Smith, M.S., CF-SLP, a recent UAMS graduate, focused her research project on speech-to-text awareness in high school students with hearing loss. She is a speech-language pathologist with Children's Therapy TEAM in Fayetteville, Ark.
