I Cant Understand Visual Systems for Face-to-Face Communication
V isual systems have been developed
to ensure an effective delivery of Have you found yourself in a situation where you have to capture spoken words, although more recently they have also been used to complete information to people whose understand information being provide realtime captioning of the hearing loss could significantly impact presented in a lecture-style news, sports and celebrity events. on the reception of the information. format? Even if you were to hear CART writers use a stenotype most of the information while machine to capture verbatim what has Note-Taking using a personal listening or been said. In order to pass the United One of the most commonly used States Reporter test, a trained court re- sound field speaker system, the strategies has been the use of note- porter or captioner must type at a speed effort over a sustained period of takers. However, the ability of note- approximating 200-225 words per time could lead to fatigue and a takers to effectively capture the infor- minute. The stenotype machine consists mation can vary greatly. In the business decreased ability to process and of phonetic-based keys which means world, this can vary from effective note- absorb the information. that the keys represent the sounds taking by secretaries who have been heard, not the alphanumeric characters trained in shorthand (almost obsolete advantage of this method is that it can representing those sounds. Multiple today) to colleagues doing their best be used for one individual (using a keys can be pressed simultaneously to take notes. small desk-type monitor or tablet) or (initial and final consonant keys as In school settings, note-takers can for many. Another advantage is that it well as vowels) allowing for sound be teacher aides or designated fellow is not as costly as realtime captioning, combinations, and in turn, allowing students. The ability of note-takers to although it only provides just enough the stenographer to spell out syllables, capture all the presented information information to stay on topic. whole words, or phrases. is limited. Instructors generally speak Modern stenotype keyboards 170-220 words per minute, while note- Voice-to-Text Options have more in common with computers takers generally write 20-30 words per in that they contain microprocessors minute or up to 50-60 words per Communication Access and are connected to computers with minute if they type the notes. This Realtime Translation (CART) software that can convert the stenotype might allow for the capture of key con- CART encompasses realtime caption- code to English, and stored onto user- tent but it does not allow for enough ing and involves the use of intensively specific dictionaries. That is, CART information to be captured for real- trained stenographers. time interaction. Stenographers have Although note-taking is one typically been used in of the least costly ways to transcribe courtroom settings orally presented information to written where their role is to text, it is an ineffective way to follow information in classes, meetings, or one-to-one exchanges. Communication Access Realtime Computer-Assisted Note-Taking Translation (CART) This consists of a typist synthesizing is used at all HLAA Conventions. Right: the key information and typing into Richard Einhorn spoke a computer connected to an LCD at Convention 2014 in projector or other monitor. One Austin, TX.
28 Hearing Loss Magazine Visit us at hearingloss.org and follow @HLAA on Twitter
captioners must build up their own vocabularies. This is the reason why it is beneficial to provide text of a presentation prior to an event since this enables the CART provider to enter special words ahead of time into the dictionary stored on their computer. If there is no matching word in the computer dictionary, the software will try and make the best phonetic match possible which often leads to some of Dragon Dictationvoice recognition comes to the iPhone, allowing for a voice-to-text message the comical errors that we might see in realtime captioning of the news, sports Education Act (IDEA). In other interpreter usually have their own or celebrity events. situations, the costs might be covered laptops when the service is provided The captioned text is conveyed under the Americans with Disabilities on-site. onto a computer projection panel, Act. In addition, because there is not an There are a number of available projecting onto a screen or combined abundance of these trained individuals options such as C-Print and TypeWell. into one unit (merged with other they might not always be available for C-Print was developed by the National media, such as a television). Realtime conferences or presentations unless Technical Institute for the Deaf captioning can be done either with reserved ahead of time. (NTID). In this case, the C-Print the CART provider located on-site Recently, voice-to-text technology captioner (usually located in the same or off-site. If the provider is doing has become refined enough to use with class as the student, but can be done remote captioning, there is an online realtime captioning (e.g., captioned remotely) uses a software application connection to the Internet that captures telephone services). In order to achieve such as the C-Print Pro, which uses the transcribed text and interfaces with high accuracy, this presently requires an abbreviation system based on the projector panel in the setting where that the speaker train the software phonetics. The C-Print captioner the presentation is taking place. to recognize his or her voice. Obvious condenses and summarizes the key However, the CART captioner advantages to this technique are that meaning and content in realtime and must be able to hear the presenters the costs are significantly less than types into his computer, which is then voice. This is done by having the pre- when using stenographers; one is also relayed to the students computer. senter talk into a microphone, his voice not constrained by their availability. When a student has a question, he can is transmitted to the on-site computer Although this technology has improved type the question which is then relayed and, in turn, transmitted via the significantly, it still is not as accurate to the C-Print captioner who then Internet to the captioners computer as when one uses CART providers. verbalizes the questions to the class. whereby he is able to hear what is being Thus, for professional settings, CART NTID indicates C-Print is said. The captioner then types what he is preferred. for students whose preferred mode has heard via the stenotype machine is English; who have a hearing loss which is connected to the computer, Text Interpreting significant enough to make it difficult with the transcribed captions being Unlike CART, text interpreting to follow spoken English in the relayed back to the on-site computer involves capturing the key content classroom; whose reading level allows and projected for everyone to see. of what is being presented. This for reading the text of the lesson at least The obvious advantage of service is specifically designed for the at a fourth grade reading level; and, CART is capturing verbatim what has educational setting. Text interpreting who know little or no sign language. been said. The disadvantage is that is an electronic note-taking system Because training for text interpreting by involving highly-trained CART that is designed to provide meaning- can be done online and requires only captioners, CART is costly (ranging for-meaning transcription of spoken a few months of training, C-Print from $60 to as high as $350 an hour). words into text, versus the verbatim or TypeWell are less expensive than The party responsible for payment transcription provided by CART realtime captioning. The disadvantage depends on the situation. In some stenographers. The text interpreter of text interpreting when compared cases, if used in a school setting the listens to the presenter, condenses to CART is that only the meaning is costs might be covered by Section the text to derive the key meaning, conveyed, thus, relying on the typists 504 of the Rehabilitation Act of 1973 then types in the text via a shorthand derivation and interpretation of the key or the Individuals with Disabilities method. The student and the text content. continued on page 30
Like HearingLossAssociation on Facebook March/April 2016 29
Face-to-Face Communication FaceTime, Google Hangout, and Skype continued from page 29 use a webcam and a desktop, laptop, table Automatic Speech Recognition or smartphone to Development of voice-to-text software see the other party. has allowed for significant innova- This allows for voice, speechreading and tions for use by people with hearing sign language. loss. Examples are Dragon Speech recognition software, Speechlogger, Crescendo Speech Processing. Go to capterra.com/speech-recognition- software/ for a review of some of the best speech recognition software products on the market. These soft- ware products can be downloaded RogerVoice an individual with In addition to training the software, onto computers and mobile devices. hearing loss calls or receives calls the CA must also be trained to shadow Voice-to-text products can be on a smartphone that are instant (repeat) in realtime what he or she used in various situations depending transcriptions of what the other party hears. Once the CA has trained the on the specific goal of the software. is saying, regardless of the spoken voice-to-text software and is able to One possible application is for those language (this software is meant to shadow other talkers at an acceptable who wish to avoid typing. In addition, work with any voice, thus errors can level, he or she is ready to serve as a CA. there are a number of applications that occur in the transcription) have been developed for people whose Video Face-to-Face hearing might not allow them to easily Embedded Voice Recognition Telephone Services follow spoken language. Depending on programs included with Windows 10, FaceTime, Google Hangout, and Skype the specific product, the voice-to-text Google Chrome, Apple Siri, which can use a webcam and a desktop, laptop, software could allow the individual to be used to dictate information into tablet or smartphone to see the other engage in one-on-one conversation or Word software or to control various party, thereby allowing for either voice follow a presentation. A remote mic computer programs plus speechreading or for those who worn by the presenter can be used use sign language. to transmit the speakers voice to an An important application of intermediary device connected to a voice-to-text software in recent years Summary computer with voice-to-text software has been in use with Voice Carry Over It is amazing how technological ad- installed which, in turn, allows the (VCO) relay services as provided by vances in visual systems have allowed user to easily follow the presenter. The CapTel or CaptionCall. In this case, people with all levels of hearing loss accuracy of the transcribed text will people with hearing loss conduct to overcome obstacles that previously be greatest if the presenter has trained telephone calls using a relay service. made communication in meetings, the software to recognize their voice. The person with hearing loss speaks on classes, or on the phone to be so However, in reviewing online a number the phone to a hearing person at the difficult that they were discouraged of voice-to-text products, a number of other end. What the hearing person from participating. I hope I have providers have expressed that they can says is not only transmitted directly provided information of use to the provide high quality text without much but is also relayed to a third party, a reader. Heres to seeing a brighter training. Communication Assistant (CA) via the future. HLM Internet. Rather than typing what the Here are a few examples: hearing person has said and relaying Larry Medwetsky, Speech Assistant a speech it to the individual with hearing loss, Ph.D., associate recognition software for enterprises, voice recognition software has been professor, can allowing for communication between adapted for use to relay the text. Prior be contacted customers and employees to being able to serve as a CA, the at Gallaudet individual must train the voice-to-text University at larry. Dragon Naturally Speaking software to recognize his or her voice medwetsky@ used by professionals, students for to the point that minimal transcription gallaudet.edu. dictation and transcription errors occur, which takes many hours.
30 Hearing Loss Magazine Visit us at hearingloss.org and follow @HLAA on Twitter