Jump to content

Speech-to-text reporter

From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by FouryintheHouse (talk | contribs) at 19:34, 3 February 2021 (Made a word lowercase). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

This article is about a human occupation. For computer systems, see speech recognition and predictive text.

A speech-to-text reporter (STTR), also known as a captioner, is a person who listens to what is being said and inputs it, word for word (verbatim), using an electronic shorthand keyboard or speech recognition software and a CAT software system. Their keyboard or speech recognition software is linked to a computer, which converts this information to properly spelled words.[1] The reproduced text could then be read by deaf or hard-of-hearing people.

The latest in speech-to-text reporters are called voice writers. Voicewriting is a method used for court reporting, and it is also used by some medical transcriptionists. Using the voicewriting method, a court reporter speaks directly into a stenomask or voice silencer—a hand-held mask containing one or two microphones and voice-dampening materials. A voice writing system consists of a stenomask, an external sound digitizer, a laptop, speech recognition software, and CAT software. A foot pedal can plug into a computer's USB port.

A real-time voice writer's words go through the mask's cable to an external USB digital signal processor, From there the words go into the computer's speech recognition engine, for conversion into streaming text. The reporter can send the streamed text to a) the Internet; b) a computer file; c) a television station for subtitling; d) to an end-user who is reading the captions via their laptop, tablet, smart phone, or e) software which formats the results in a way most familiar to judges, attorneys, or subtitling consumers.

Voice writers enjoy very high accuracy rates, based upon pure physiology. The route taken by a person's words goes from the mouth to the reporter's ear, brain, and "inner" voice. This form of repetition is naturally effortless; it is what we all do in our daily conversation, as we listen to a person speak, or when we read a book. So the most natural extension of this process is to psychologically switch the repetition mechanism from "inner voice" to the physiological "spoken voice." Therefore, we minimize the introduction of cognitive overhead in our task of routing the spoken word to its permanent destination as printed words. This streamlined process allows voice writers to achieve excellent performance for many continuous hours and greater than 98 percent accuracy at speeds as high as 350 words per minute.

Voice writers produce the same products as their stenotype colleagues, including transcripts in all electronic and printed formats. Realtime verbatim reporters connect their laptops to captioning equipment, real-time viewer programs, and provide attorneys or other clients with computer files at the end of the sessions. Only the physical way of capturing speakers' words differentiates voice writing from other methods of court reporting. Every other aspect of this profession is the same, with the exception of the time required to learn the skill, which is much shorter with voice writing.

Other ways of producing speech-to-text: There are two types of keyboard used in the UK: the Palantype system and the Stenograph system. Hence STTRs are also sometimes referred to as palantypists and stenographers. Unlike a QWERTY keyboard, not every letter in a word is pressed, but several keys will be pressed at once which represent whole words, phrases or shortforms. Specially designed computer software will then convert these phonetic chords back into English which can then be displayed for someone to read.

The computer software can use a pre-programmed vocabulary specific to the context, information that matches syllable clusters to written forms, and may suggest alternative captions from which the STTR chooses. Errors occur from the STTR mishearing the words and from the need for the STTR to make a decision before an ambiguous statement is made clear by what is said next.

The professional association for STTRs is the Association of Verbatim Speech-to-Text Reporters. The Council for Advanced Communication with Deaf People and the Royal National Institute for the Deaf also give more information about STTRs.

What will a service user see on the screen?

Every word that is spoken will appear on the screen in an accessible format, although one can request a change in the color and font size. As well as every word spoken, the words "NEW SPEAKER:" will typically appear to denote when the speaker changes. If one sends the STTR (voice writer/palantypist/stenographer) the names of people attending the conference or meeting before the event, they, too, can be programmed into the computer, making it easier for one to recognize who is speaking. Other phrases, in curly brackets, may also appear, such as {laughter} or {applause}, to denote relevant events.

Occasional mondegreen errors may be seen in closed-captions when the computer software fails to distinguish where a word break occurs in the syllable stream. For example, a news report of a "grand parade" might be captioned as a "grandpa raid". Mondegreens in this context arise from the need for captions to keep up with the fast pace of live television broadcasts.

History

Many STTRs began their careers as court reporters. In the courts, the system was used to record proceedings and provide transcripts when requested. The skills developed in this area have also made them invaluable in the field of communication with deaf people, as they are used to producing work with an extremely high degree of accuracy and acting with complete discretion at all times. An STTR expects to reach consistent levels of accuracy of 98.5% and above.

Training

In order to become an STTR one needs extensive training, typically two years, if a Palantype/Stenograph and the associated software, plus at least a further two years of practice, building up speed, accuracy, dictionary/vocabulary and gaining experience.

If becoming a voice writing STTR, the training is somewhat shorter, though nevertheless, just as strenuous. Two years to learn the art of voice writing includes the associated software and required practice, building up speed, accuracy, dictionary/vocabulary and gaining experience.

Only then is one ready to undertake the USA and/or UK Examinations for Certification. In the USA, NVRA and NCRA offer certifications that demonstrate the necessary training.

There are many levels of certification. NVRA certifies voice writers as court reporters, realtime reporters, CART providers, and broadcast captioners. Beyond any licensing considerations, NVRA certification clearly demonstrates that the voice writer has attained a level of professionalism and skill well above that of others in the field.

The NVRA certification testing program is available to all members. Tests are held at regular intervals throughout the year in various locations across the country.[2]

In the UK, Unitised CACDP Examinations and membership with the CACDP Register, confirms that one has reached the required minimum standard. The majority of Registered STTRs are also Members of the Association of Verbatim Speech-to-Text Reporters.

See also

References

  1. ^ "Closed Captioning Web". Captions.org. 2006-02-13. Retrieved 2009-06-11.
  2. ^ "NVRA certifications". nvra.org. Retrieved 2017-01-05.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Speech-to-text_reporter&oldid=1004666094"