Comparing ‘real time’ Captioning Services for the hard of hearing/oral deaf/late deaf population
Facts, Reviews and Opinions on these services on a thread:
"C-PRINT is a system that’s very similar to TypeWell.
"C-PRINT is a system that’s very similar to TypeWell.
They both use computer keyboards and omit vowels when they type so that the software can guess the words they're trying to write. They're both really notetaking or outlining programs and not true verbatim transcription systems. C-PRINT and Typewell have a top speed of about 120 words per minute with a very fast and accomplished operator. CART providers have to be able to write at least 225 words per minute, and the best of them can get up to 240 or even 260. The average rate of speech is 180, but some people speak much more quickly than that, and the complexity of the material can really affect the accuracy of the transcription….”
"I am a CART Provider who uses my voice instead of a steno machine.”
"Meaning-for-Meaning vs. Verbatim
One thing to know is that TypeWell & C-print are meaning-for-meaning transcription systems. They are not designed to be verbatim transcripts.
What is "meaning-for-meaning" transcribing? It is taking the spoken English (which is not grammatically correct) and putting it down in grammatically correct written English. Meaning-for-meaning transcripts put all of the meaning of the words into the transcript, without the needless utterances, stammers and stuttering that sometimes happens…
The normal delay from the time the transcriber hears a sentence to the time the words start to appear on the screen should be about 3 - 5 seconds. In some cases where the transcriber is very slow, it may be that they are new, they are unfamiliar with the material, or the material is too technical for TypeWell and would be better served by CART.
In most cases, TypeWell can cover nearly any subject.Classes that are very technical such as medicine or law are best served by CART because they are able to capture all of the vocabulary required for the course quickly.
Bottom Line There is an appropriate place for all of these service types. If you don't like one of them, try another if it is available in your area.
In places that don’t have on-site providers available, remote CART and TypeWell providers are available.”
"Typewell and C-Print, as I understand, both work with regular computer keyboards and proprietary software that's relatively inexpensive. Their training sessions involve teaching people to type words without vowels and to learn how to abbreviate common words with a few keystrokes. Most people graduate from the training in a month or two. The software then expands the abbreviations so that the words appear in regular English. Using this technique, people can write around 100-130 words per minute. The average rate of speech, however, is 180-200 words per minute, with some people getting up into the 240-260 range for short bursts. Obviously these services that use the regular computer keyboard can never get up into verbatim territory, so they advertise themselves as providing "meaning for meaning" rather than "word for word".
In my experience as a CART provider, I've sometimes noticed a radical increase in speed when I redefine a word that I formerly wrote in two strokes into a single stroke. Even so, when people suddenly rocket off into high speeds, I find myself occasionally paraphrasing and leaving out inessential words and phrases to keep up and get the important part of what's being said. Considering that everything written in Typewell or C-Print is at the very least two strokes, but more commonly three or four -- more for very complex technical words -- I'm quite skeptical that the entirety of "meanings" expressed in a given lecture can be conveyed by someone whose speed is nearly half that of a certified CART provider."
With voice to text services such as CaptionMic, the issue is usually not of verbatim versus non-verbatim, but of accuracy rates. It is possible to get to a very high degree of accuracy using voice recognition software, as long as the rate of speech stays below about 180 to 200 words per minute and the voice writer has spent many, many hours training their software and themselves to speak accurately and precisely, disambiguating all soundalike words and phrases."
"Typewell and C-Print, as I understand, both work with regular computer keyboards and proprietary software that's relatively inexpensive. Their training sessions involve teaching people to type words without vowels and to learn how to abbreviate common words with a few keystrokes. Most people graduate from the training in a month or two. The software then expands the abbreviations so that the words appear in regular English. Using this technique, people can write around 100-130 words per minute. The average rate of speech, however, is 180-200 words per minute, with some people getting up into the 240-260 range for short bursts. Obviously these services that use the regular computer keyboard can never get up into verbatim territory, so they advertise themselves as providing "meaning for meaning" rather than "word for word".
In my experience as a CART provider, I've sometimes noticed a radical increase in speed when I redefine a word that I formerly wrote in two strokes into a single stroke. Even so, when people suddenly rocket off into high speeds, I find myself occasionally paraphrasing and leaving out inessential words and phrases to keep up and get the important part of what's being said. Considering that everything written in Typewell or C-Print is at the very least two strokes, but more commonly three or four -- more for very complex technical words -- I'm quite skeptical that the entirety of "meanings" expressed in a given lecture can be conveyed by someone whose speed is nearly half that of a certified CART provider."
With voice to text services such as CaptionMic, the issue is usually not of verbatim versus non-verbatim, but of accuracy rates. It is possible to get to a very high degree of accuracy using voice recognition software, as long as the rate of speech stays below about 180 to 200 words per minute and the voice writer has spent many, many hours training their software and themselves to speak accurately and precisely, disambiguating all soundalike words and phrases."
No comments:
Post a Comment