Speaker labels, timestamps, bracketed notes and verbatim speech, timed. Transcription is typing as the whole job, and what counts is the pace you can hold, not the pace you can hit.
Spoken language, written down verbatim, is messy prose with a layer of notation on top. Speaker labels in capitals followed by a colon; timestamps like [00:14:22]; bracketed notes for a laugh, a pause or an inaudible word; the ums and false starts a strict-verbatim brief requires you to keep. The words are ordinary; the notation is capitals, brackets, colons and digits, and it recurs every few seconds.
The text in this test comes from the Transcription & Captioning course rather than a general corpus, so it reads like a transcript: labelled turns, timestamps at intervals, bracketed conventions, and speech as people actually speak it. Type it exactly as shown, including the brackets.
Most transcriptionists come in five to fifteen per cent below their prose speed on this text, and the loss is almost all in the notation. The result marks each miss on the text; look for brackets typed as parentheses, a colon dropped after a label, and digits swapped inside a timestamp.
Sixty seconds is a sprint, and transcription is not. The rhythm reading on the result page — how even your pace was across the run — predicts your hourly output better than the peak speed does. A steady 60 beats a bursty 75 over a four-hour file.
The Transcription & Captioning course drills speaker labels, timestamps and bracketed notes separately before mixing them into long passages, and the coach builds new practice from the keys and sequences you miss most. If the brackets were the problem, the next set is dense with them. The longer timed tests on this site are the place to measure endurance.
A speaker label is a run of capitals and a colon: Shift held for a word, then a shifted punctuation key. A timestamp is two brackets and six digits with colons between them, all on the number row and the shifted layer. Until these are automatic, each one is a place where your hands leave the words and your eyes leave the audio.
Endurance is the other half. The five- and ten-minute tests on this site are the honest measure of how a shift will go, and the consistency figure across a longer run says more than any sixty-second peak.