Transcripts

What a transcript is

A transcript is the full written text of a recording, produced so that a person can read what was said without listening to it. The version sold here is edited: the words are kept, the ums, false starts and repeated half-sentences are removed, speakers are named, and a timestamp marks each change of subject.

It is not the automatic text a video platform or a meeting tool spits out. Those are a starting point, and they are wrong in exactly the places that matter: names, product terms, numbers and anything said quickly. It is also not a subtitle file, which is timed to the second for display in the picture. A transcript is for reading, quoting and reusing.

When you need one

You need a transcript when a recording has to be searchable, quotable, or accessible to someone who cannot or will not listen. That includes every podcast episode you publish, any webinar you send to people who missed it, recorded customer interviews you intend to quote in a case study, training videos that staff need to refer back to, and any recording that will become written material later. A transcript is also the fastest way for a manager to check what was said in an hour without spending an hour.

You do not need one for a video that will only ever be watched with captions on, if the caption file already exists; the caption text can be exported. And if what you want is the article rather than the record, buy podcast to article or video to article, which include the transcript as a working file.

What goes in it

Every transcript is produced the same way. The recording goes through transcription software first, because that is fast and cheap. Then a person listens to the whole recording against the text and fixes it, which is where the value is. The delivered document has:

• Speaker names. Each change of speaker marked with the person's name or role, agreed with you first so that titles and spellings are right.
• Clean verbatim editing. Filler removed, false starts removed, sentences that trail off tidied so they end. Nothing added, nothing reordered. If a speaker said it, it is there.
• Timestamps. Minutes and seconds at every change of subject, and at any point you are likely to want to find again.
• A spelling list. Names, products, places and technical terms checked against your website or a list you send, so the document does not carry a misspelt client name into everything made from it.
• Marked uncertainties. The rare word that cannot be made out is flagged with the time, not guessed. You can listen to that second and settle it.

A full verbatim version, with every hesitation kept, is available for legal or research use and is said in the quote if you ask for it.

What you get

The transcript. One document per recording, in an editable format and as a PDF. Roughly two to three pages per ten minutes of speech, so a one-hour episode runs to twelve to eighteen pages.

A summary page. The first page of the document: who spoke, what it covered in five to eight lines, and the timestamped list of subjects. This is the part most readers use.

A plain text version. The same text with no formatting, for pasting into a website, a notes tool or a scheduling tool.

A worked example

An illustration, not a client. A residential mortgage broker records a fifty-minute conversation with a real estate lawyer about what goes wrong at closing. The automatic transcript from the recording tool renders the lawyer's firm name three different ways, turns a provincial land-transfer term into a word that does not exist, and loses two of the four dollar figures the lawyer quotes.

The edited transcript fixes all of that, names both speakers, and marks eleven subject changes with timestamps. The summary page lists the seven closing-day problems the lawyer described. The broker sends the PDF to clients who ask what happens on closing day, quotes three paragraphs of it on a page about closing costs, and uses the summary page as the show notes for the episode. The recording was already made; the transcript is what let it be used four times.

How it runs

1. A half-hour call. What you have recorded, how many hours, and what the transcripts are for. No charge.
2. A fixed price in writing. Per recording or per hour of audio, agreed before anything starts.
3. The files. You send the recordings and a list of names and terms that must be spelt correctly.
4. The work. Three to five working days for a single recording under an hour. A backlog of episodes is scheduled in batches.
5. Delivery. The transcript, the summary page and the plain text version, with any flagged words listed for you to settle.

What it costs

Transcripts are quoted as a fixed price in CAD after the half-hour call. The price depends on the total length of the recordings, the sound quality, the number of speakers, and whether you want clean verbatim or full verbatim. A standing arrangement for every episode of a show is priced per episode and stated in writing. The pricing page explains how fixed pricing works across every service.

What happens next

A transcript is the raw material for everything else in this group. The usual next step is show notes for an episode, or the full repurposing set that turns one recording into an article, posts and a newsletter section. If the recording is a customer interview, the transcript becomes a case study. And if the video itself needs to be accessible, the transcript is the source for a subtitle file that is timed rather than merely readable.