A different kind of typing
Stenography is a method of writing in which a trained person presses several keys together to enter whole sounds, syllables or words in one movement, instead of typing letters one at a time. Applied to live captioning, it lets a skilled writer follow natural speech and produce readable text within moments.
Stenographic captioning is often what people mean by CART, short for communication access realtime translation, although that term is used in some places to cover other methods too. In courtrooms, the same skill produces official records, and many captioners have a background there.
The machine and the strokes
A stenographic keyboard looks quite different from an ordinary one. It has far fewer keys, which are pressed in combinations known as chords. Each chord, called a stroke, represents a set of sounds. Rather than spelling a word, the writer records how it sounds, often in the same stroke as other sounds, and words are built from one or more strokes.
The system is largely phonetic. The writer is not spelling “through” letter by letter; they write the sounds, and the software supplies the spelling. Common words and phrases can be reduced to a single stroke, which is a major part of why speeds can be so high.
From strokes to text
The keyboard sends the strokes to computer software that includes a dictionary. The dictionary maps each stroke or sequence of strokes to words. The software then displays the result as ordinary text, usually with capitalization, punctuation and spacing applied by the writer’s strokes or by built-in rules.
This is where preparation comes in. A stroke that does not exist in the dictionary shows up as raw steno notation or as a guess, which readers find confusing. So a captioner who expects unusual names or technical words will add them to their dictionary before the session, assigning each a quick stroke. This is the stenographic equivalent of providing custom vocabulary to an automatic tool.
Some captioners also use conflict handling for words that sound the same, setting dictionary entries so that a particular stroke yields the form likely to be right in context. Mistakes of this kind, such as the wrong homophone, still occur in live work.
Training and skill
Learning to write steno takes sustained practice over a long period. Writers build up a vocabulary of strokes, practice at increasing speeds, and learn to handle numbers, names and unfamiliar words. The details of training and certification differ between countries and organizations, so it is best to ask a provider about their writers’ qualifications rather than assume one standard.
Because the skill is specialized, availability can be limited in some regions and for some languages. Steno systems exist for a number of languages, but coverage is uneven.
What it is good at
- Very fast speech. Stenography is generally regarded as capable of keeping up with rapid talk, with relatively little need to condense.
- Close to word-for-word text. Writers can often capture most of what is said, which suits meetings, classes and formal sessions where detail matters.
- Complex subject matter with preparation. With a prepared dictionary and a briefing, writers can handle technical content well.
- Judgment about speakers and sounds. A person can note speaker changes and relevant non-speech information, and cope with overlapping talk better than software typically does.
What to keep in mind
Delay. The writer still has to hear and process speech, and the translation software adds a moment more. Text appears a little after the words are spoken. How long depends on the writer, the setup and the connection, and it is rarely zero.
Errors. Unfamiliar words, ambiguous sounds and fast changes of topic produce mistakes. When a stroke isn’t in the dictionary, the result can look odd until it is corrected.
Fatigue. Sustained live writing is demanding, so long events normally use more than one writer, taking turns.
Cost and booking. Stenographic services are generally priced above automatic options, and skilled writers may need to be booked well in advance.
Audio dependence. Writers need to hear clearly. They can work in the room or remotely through an audio feed, but a poor feed hampers them as it does any method.
How it fits with other methods
Stenography is one of several ways to produce live text. Respeaking uses recognition software driven by a trained speaker. Automatic recognition removes the human from the loop. A single event might use more than one: for example, a stenographic writer on the main stage and automatic text in smaller rooms. Each approach has different trade-offs in cost, delay and accuracy.
Working with a steno captioner
Some practical steps help:
- Share materials early. Slides, scripts, speaker lists and a list of terms let the writer build their dictionary in advance.
- Provide a clean audio feed. A direct line from the mixing desk, not a room microphone, is best.
- Brief the speakers. Steady pace, one voice at a time, and repeated audience questions all help.
- Plan for breaks. Agree how a long session will be covered.
- Offer a contact. Someone should be reachable if the writer needs to ask for a spelling or report a problem.
- Ask about transcripts. Many writers can supply the session text afterward. Because it was produced live, it will usually need cleaning before it is published, as with any live transcript.
Requirements for access to events and meetings differ by place. Where obligations may apply, a qualified professional or the relevant authority is the right source. A captioning method on its own does not guarantee compliance or perfect text.