What does "CART" stand
- What does "CART" stand for?
- How long has CART been
- Who are you?
- How quickly do stenographers
- How does a steno machine
- What's the difference between a CART
provider and a court reporter?
- What's the difference between
CART and captioning?
- What are the pros and cons of CART
versus ASL interpretation?
- Why hire a CART provider directly,
rather than through an agency?
- Will speech recognition make CART
- Can you provide CART for people with
- Why do you stop writing when your
client leaves the room?
- What do the various NCRA Certifications mean?
- Where can I learn more about
CART is a technological service that provides Deaf or hard
hearing people with verbatim realtime transcription of live speech.
Depending on who you ask, it stands for "Communication Access Realtime
Transcription" or "Computer-Assisted Real-Time". I prefer the former,
because it distinguishes CART from the computer-assisted realtime
provided by legal stenographers for hearing clients.
How long has CART been
The precursor of today's steno machine was invented around
but stenotype machines that can output raw steno code to computers
have only existed since the 1980s. The first live captioned television
broadcast appeared in 1981, and CART was introduced in the mid-1980s
as an onsite complement to broadcast captioning.
Who are you?
I'm Mirabai Knight, a Certified CART Provider in New York
and sole proprietor of StenoKnight CART Services. I work on a
semester-long basis in the NYC higher educational system for
Deaf and hard of hearing university students, and on a per diem basis
for individuals and groups needing CART in a variety of settings.
I also provide CART remotely via the internet throughout the United
I specialize in medical, technical, and graduate level subjects.
How quickly do stenographers
A good qwerty touch typist probably writes around 80 to 90
per minute. The best typists have an upper limit of about 110 to 120
words per minute. The average rate of English speech generally ranges
around 180 words per minute. In order to graduate from steno
school, a court reporter must pass tests at 225 words per minute. The
world record for a one-minute take at 95% non-realtime accuracy is 360
words per minute, currently held by Mark Kislingbury, RMR, CRR.
Certification speed for a CART provider is 180 words per minute in
uncorrected instant translation (also known as realtime) with an
accuracy of 96%. The CART provider is not allowed to look over their
notes to correct any errors or guess at any missing words; their
output is judged exactly as it comes through the software's
translation engine, including punctuation and spelling.
How does a steno machine
While ordinary keyboards (often called "qwerty" keyboards
first six letters on the top row) write words letter by letter, steno
keyboards write them syllable by syllable -- more or less. Nearly
every word sound (called a phoneme) in the English language can be
constructed from the initial consonants on the left-hand side of the
keyboard, the vowels located in the middle and operated by the thumbs,
and the ending consonants on the right-hand side. There is no space
bar; before the computer age, stenographers would read back their
notes to themselves and resolve the jammed-together syllables into
words based on context. When the computer came along, offering with it
a much more efficient form of transcription, that method couldn't be
relied upon anymore.
What's the difference between a CART
provider and a court reporter?
Computers don't know English and can't comprehend meanings of words,
so stenographic theory had to be significantly restructured. Now every
word a stenographer wants to be able to write on the machine must be
entered in a steno "dictionary", a lexicon of English words paired
with their corresponding steno strokes. For instance, the word "glue"
is written in steno as "TKPWHRAOU". This looks like a lot of letters,
but it's actually only one stroke on the steno keyboard, as opposed to
the four strokes it takes to write the word letter by letter. To
distinguish the sentence "put some glue on the cardboard" from the
sentence "a gluon is an elementary particle", a stenographer must have
separate strokes for the word "on" and the "-on" suffix.
Another technique stenographers employ is the use of abbreviations, or
"briefs". If I were CARTing a quantum physics course, the word "gluon"
might come up so often that I won't want to have to stroke two
syllables whenever I write it. In that case I might decide to define
"gluon" in my steno dictionary as "TKPWHRAOUPB" (pronounced "gloon"),
which takes only one stroke to write. That new definition would be
considered a brief, since it doesn't strictly correspond to the sound
of the word it represents. The number of briefs in a stenographer's
dictionary is limited only by their ability first to devise them and
then to recall them to their fingers at realtime speeds.
CART providers and court reporters both use the same
software, but CART providers always work in realtime for a primarily
Deaf and hard of hearing clientele, whereas court reporters work for
attorneys or judges in courts of law or deposition rooms. They
sometimes employ realtime, but their primary responsibility is to take
down a full and complete record to be edited, printed, and sold to
clients. They are paid a fee per appearance and a somewhat higher fee
if they offer realtime feeds, but most of their income comes from
What's the difference between
CART and captioning?
CART providers, on the other hand, work primarily to provide a
realtime feed that is as complete and readable as possible, so they
charge a higher hourly fee than court reporters, but far less for
transcripts, if they choose to provide them. For them, transcripts are
a by-product of their work, and their transcripts, unlike those
produced by court reporters, are not certified verbatim. Often CART
providers will offer complimentary or low-cost transcripts to the
students whose classes they transcribe, with the understanding that
the student will use the transcripts only to study from, and will not
distribute them to anyone else. For reasons of liability, CART
providers for large events will generally not offer transcripts; if a
transcript is required, it's usually advisable to hire a court
reporter in addition to a CART provider.
In an ideal situation, a CART provider's transcript will be
100% verbatim, just like a court reporter's transcript, and both
realtime feeds will be identical. When push comes to shove, though,
the court reporter will let their realtime feed fill up with sloppy or
untranslated steno code to clean up later, rather than risk missing a
word in the final transcript. The CART provider will make sure that
the realtime feed stays legible, even if that means paraphrasing
slightly or omitting redundant words. They're both working toward the
same goal, but the court reporter always gives deference to the final
edited transcript, and the CART provider to the immediate realtime
feed. The National Court Reporters Association offers membership and
certification to CART providers and captioners as well as court
reporters, and many NCRA members provide both captioning/CART and
court reporting services.
There's a significant overlap between CART and captioning,
often consumers don't make any distinction between them, calling
everything "captioning". Properly speaking, captioning refers to words
displayed on a video screen, usually underneath or next to whoever's
speaking. Closed captioning means that the words can be turned on or
off by the viewer. Scripted television programs are closed captioned
by qwerty typists. Live television programs are usually closed
captioned by stenocaptioners. Television stations combine the caption
feed with the video and broadcast it to television sets around the
country. Scripted open captioning is offered by organizations such as
the Caption Coalition, whose captioners use qwerty keyboards to
display the text of plays and musicals for Deaf and hard of hearing
theater patrons. Live open captioning is what's usually referred to
as CART. Clients can read the CART provider's realtime output from a
laptop, tablet, or netbook screen, or the CART display can be
projected onto the wall for the benefit of multiple clients.
What are the pros and cons of CART
versus ASL interpretation?
There's also internet streaming captioning, sometimes called "remote
CART". In that case, the client provides an audio feed of a class or
event to the captioner by connecting a microphone to a phone line or
Skype connection. The captioner listens to the feed and transcribes
the audio on their steno machine, which is then sent back via the
internet to the client's computer. The main differences between remote
CART and broadcast captioning are: First, the words appear on a
computer screen rather than a television screen, so they don't have to
be encoded by special closed captioning equipment; second, that the
audience is usually at the class or event in person rather than
receiving the video and captions together in the same feed; and third,
the audience is usually fairly small, compared to the millions who
watch any given closed captioned television program. As more video
broadcasting moves onto the internet, though, the line between
captioning and remote CART will become increasingly blurry.
CART can be a useful alternative to American Sign Language
interpretation for certain students, such as those whose first
language is English (an estimated 90% of the 25 to 35 million
Americans with hearing loss) or who are enrolled in courses containing
highly technical material. As with translation between any two
languages, nuance can be lost and ambiguities can be introduced in the
translation from English to ASL. ASL interpretation also requires that
a student maintain continuous eye contact, usually necessitating
additional notetakers. Students using CART can take their own notes;
the last dozen or so lines of the lecture are displayed on the CART
screen for easy reference.
Why hire a CART provider directly, rather than through an agency?
Individual ASL interpreters tend to charge less for their services
than CART providers, but since educational interpreters work in teams,
two interpreters are usually required for any class that lasts longer
than one hour. CART providers generally work alone, making the cost of
CART roughly equal to interpretation for most classes. Not all students
with hearing loss are fluent in ASL, and some classes include a great
deal of specialized terminology that can be very difficult to translate
into Sign Language. Some students will also request ASL for interactive
classes such as seminars and labs, while preferring CART for lectures.
In short, ASL interpretation is the most appropriate accommodation for some
students in some classes, and CART is the most appropriate accommodation for others.
The student's preference should be paramount in deciding which service to select.
The biggest advantage of dealing directly with a CART provider is price.
CART providers usually charge the same hourly rate whether they work through an agency
or directly with their clients, so organizations that go through agencies tend to
pay more -- sometimes as much as 30% more -- to cover the agency's markup fee.
On the other hand, agencies tend to have several providers available, so they can
sometimes cover more work with less notice than individual CART providers, whose
schedules fill up fairly quickly. I've worked with a number of firms over the
years and still take per diem jobs from them now and then, but I also pride myself
on offering flexible, customized service to my direct clients, which
they often seem to enjoy as much as they value the comparatively lower cost.
Will speech recognition make CART
Not until true artificial intelligence is developed, which
likely to happen any time soon, if ever. I've written on this subject
at length in my article Voice
CART, but briefly: Automatic
speech recognition is not currently a substitute for human
transcription, because computers are unable to use context or meaning
to distinguish between similar words and phrases and are not able to
recognize or correct errors, leading to faulty output. The best
automatic speech recognition boasts that it's 80% to 90% accurate, but
that means that, at best, one out of ten words will be wrong or
missing, which results in a semantic accuracy rate that's often far
lower than 90%, depending on which word it is. Companies claim that
the software is steadily improving, and indeed it has been getting
better over the last two decades or so -- but asymptotically, meaning
that its theoretical ceiling of improvement is still significantly
lower than that required for trustworthy output. Without the ability
to understand what is being said, a speech engine will always fall
short of a human transcriber.
Can you provide CART for people with
A voice writer is someone who uses speech recognition technology but
also introduces that vital human error checking mechanism into the
equation. Voice writers train their software to their own specific
voice pattern, which improves the translation rate considerably, and
they go beyond simply repeating words into a microphone, developing
methods to pronounce similar words differently and using a variety of
software macros to further increase accuracy. All of this requires
years of training and significant natural talent, equal to or greater
than what's demanded of realtime stenographers. Consequently, truly
verbatim realtime voice writers are rare, and usually charge about the
same as stenographic CART providers. Because people assume that voice
writing must be easier than CART, there are many improperly trained
voice writers in the marketplace, so it's vital to ask for a
real-world demonstration before hiring anyone who purports to be able
to provide realtime, regardless of input technique. I offer free
one-hour consultations by appointment, and I'm happy to demonstrate my
services at any time, either on my own or in head-to-head competition
with other providers.
Many of my clients have both hearing and vision loss, and I work hard to provide
appropriate CART accommodations for people with various types of vision loss. For
instance, one of my clients with macular
degeneration needed magnified
text, so I increased the font size of my CART output and put it on a
small tablet computer so that she could hold it just a few inches from
her eyes. Another client had Usher
syndrome, which meant that his
field of vision was narrow, but his central field acuity was pretty
good. In his case magnification would have been counterproductive,
because the bigger letters would have meant fewer words at a time
within his useful visual field. A third client didn't disclose his
specific vision issues, but mentioned that he had trouble reading
information on the classroom's PowerPoint screen. Most students prefer
that I don't transcribe onscreen information, but in his case it was
very helpful to have all of the big screen text available on the CART
display together with the professor's lecture. I'm happy to make
accommodations for all forms of vision loss, and to work with my
clients on font size, shape, and color so that their reading
experience is as clear and comfortable as possible. So far I haven't
had any CART requests from Braille users, but I'm very interested in
the possibility of outputting to a refreshable Braille display, and
I'd welcome any information on the subject.
Why do you stop writing when your
client leaves the room?
My job is to provide equal accommodations for my Deaf or hard of hearing
clients. I transcribe whatever would be within their range of hearing
if they did not have hearing loss. If they would be able to hear it,
they should be able to see it. Conversely, if they show up late, leave the room,
or fall asleep, I stop writing. I am not a personal assistant or
notetaker. People who use my services get equal access, no more and no less.
What do the various NCRA
The National Court Reporters
Association offers a number of realtime certification options for
CART providers, captioners, and court reporters. They include:
There are also several non-realtime certifications, though they
are all focused on court reporting rather than CART or captioning.
- The CCP, Certified
CART Provider. This consists of a five minute skills test of
literary material at 180 words per minute, given in unedited realtime,
plus a written knowledge test on various
subjects, including realtime writing techniques, remote
CART technology, spelling and punctuation, Deaf culture, and client
- The CBC, Certified
Broadcast Captioner. This requires the same skills test as the
CCP, but a different written knowledge test, which covers closed caption
encoding technology, broadcast preparation, quality control,
troubleshooting, and caption placement.
- The CRR, Certified
Realtime Reporter, which is a two-voice realtime skills test at 200 WPM, at 96% accuracy, but does not include a written knowledge test. However, when I attained the certification, its skills test was identical to that of the CCP and CBC skills tests; I was given grandfathered CRR status when the test was changed to 200 WPM.
I currently hold all of the NCRA stenographic certifications: Certified CART Provider, Registered Diplomate Reporter, Registered Merit Reporter, Certified Broadcast Captioner, Certified Realtime Reporter, and Registered Professional Reporter. The RDR supersedes the RMR and RPR, so it's the only one I tend to use when listing my credentials. Likewise, when I attained them, the CRR, CBC, and CCP all shared the same skills test and were only differentiated by written tests, so I tend only to list the CCP in most contexts, as it's most relevant to what I do in my day-to-day profession.
The RPR, Registered
Professional Reporter. This is the baseline certification
level for court reporters. It includes three five-minute edited
(non-realtime) skills tests: Literary at 180 words per minute, Jury
Charge at 200 words per minute, and Testimonial at 225 words per
minute. After the 15 minutes of dictation are complete, test takers
are given 75 minutes to turn their steno notes into a polished
transcript, which they then turn in to the grading committee. There is
also a written knowledge test on legal subjects.
The RMR, Registered
Merit Reporter. This consists of three five-minute edited
(non-realtime) skills tests: Literary at 200 words per minute, Jury
Charge at 240 words per minute, and Testimonial at 260 words per
minute, identical to the RPR in everything but speed of dictation.
The RDR, Registered
Diplomate Reporter. This consists only of a written
knowledge test, on advanced legal subjects. There is no skills test.
Where can I learn more about