Skip to main content

Canadian Writers and Poets on Tape

A guide to library and archival resources.

Guidelines for Transcription


Transcription is the translation of the spoken word into the written word, which is not always as straightforward as it may at first appear. It involves a number of choices at each stage of transcription, so this Guide is designed to help you make these decisions as painlessly and consistently as possible. Consistency is key!

The Guide is not intended to be exhaustive, but aims to suggest techniques for handling the most common situations that arise in transcription. If you discover a situation not covered in the Guide, let us know so we can find a solution and share it with the others!

The guidelines presented here are largely based on the book Transcription Techniques for the Spoken Word by Willow Roberts Powers, which you may wish to consult for further details.

Speaking Element

Recommended Technique/Notation


Identifying speakers


  • Use initials to identify speakers, followed by a colon.



  • Use standard spellings, as opposed to phonetic spelling, even if words seem to be mispronounced. The main goal of the transcription is not to reproduce pronunciation—the original tape is better for that purpose!
  • Transcribe common contractions (can’t, won’t, don’t) as speakers say them; slang words such as “y’know” or varieties of yes (“yeah,” “yup,” “mm-hmm”) are also to be transcribed as they are spoken.
  • People sometimes create words or use very local terms for which standard dictionaries might not be very helpful. Spell them out as you hear them.
  • Never use [sic]. The speech of interviewees is what it is




  • Applying punctuation to the spoken word is not always obvious and requires a certain


amount of interpretation. Keep punctuation

basic and let the idea of ease of reading and comprehension be your guide. Use basic punctuation—commas, periods and dashes—as follows:

  • Commas­­­­—use to indicate very slight pauses between parts of a sentence
  • Periods—mark the ends of sentences. Knowing that sentences express a complete thought and include a subject and a predicate, use your judgement to place periods appropriately.





  • Pauses occur frequently and are a normal part of speech. A brief pause can indicate sentence structure, in which case you can represent it with a comma.
  • Other brief pauses occur amidst repetitions, broken sentences and false starts. See below for ways to handle these situations.
  • For pauses immediately following the interviewer’s question, as in the case when the interviewee is collecting their thoughts or even hesitating to answer, you can write [pause] using square brackets.
  • You can also write [pause] in cases of prolonged pauses elsewhere in the interview.



False starts and incomplete sentences


  • Speech does not always take the form of full sentences. Speakers start, stop and start again.
  • Transcribe all false starts, partial sentences and repeated words using em dashes between them. The difference between false starts and incomplete sentences can be difficult to detect, but both use the same notation (em dash)


Faltering speech


  • Faltering speech refers to occasions when you


hear a speaker’s voice trail away so that the


sentence remains incomplete

  • Faltering speech is not the same as thoughtful speech when a speaker chooses their words carefully or simply speaks slowly.
  • Identify faltering speech using ellipsis points




GB: I don’t know how I felt. I just . . . I just . . .

Interviewer: Yes?


Overlapping speech


  • When two people are speaking at the same time, enclose the overlapping speech in flourished brackets or braces { }. Place a bracket at the point that joint speech begins and another where it ends. Space the overlapping speech on separate lines so that the overlapping parts start together:



M.R.:                                       {XXXXXXXXXXXX}



Assent and dissent sounds


  • If a speaker makes an assent sound (uh-huh, mm-hmm) or dissent sound (uh-uh, uhn-uhn) that interrupts or overlaps with the other speaker, you can insert it in parentheses with the initials of the speaker (MR: uh-huh).





  • Laughter is usually recorded as a context note in square brackets. Use [laughs] if one person is laughing and [laughter] if more than one person laughs.
  • Laughter does not always accompany humorous remarks, which might instead be communicated by tone of voice. In these cases, you can use exclamation points.
  • However, humour and irony are sometimes


hard to detect, so if you are uncertain you can enclose a comment in square brackets such as [speaker may intend a joke] or [?humour].



Other nonverbal sounds


  • Unlike laughter which conveys the emotional tone of the recorded event and should be transcribed, other nonverbal sounds might simply be a meaningless nonverbal act. An example is if someone simply clears their throat or coughs. Still, record nonverbal sounds using square brackets [coughs] or [clears throat].


Filler words


  • Filler words refer to words or phrases people insert into their speech, often unconsciously. Examples: “You know,” “so” or “like.”
  • Record filler words even though they can be frustrating for the reader and may be edited out later.


Reported speech


  • If the speaker is quoting the words of another precisely (and not just giving their version of what the other said), place the reported speech in quotation marks.
  • Note: quotation marks may also be used to introduce ironic words or invented expressions, but only do so if the words are somehow marked in this way by the speaker.



Inaudible words or sections


  • If after multiple attempts you are still unable to understand what is being said, indicate this by writing [inaudible words]. If possible, give the number of missing words [inaudible 3 words].
  • If you guess what is being said, insert your transcription preceded by a question mark [?recoiled in horror].




Poetry reading



  • Canadian Poets on Tape includes poetry readings by the poets. Use MLA style “blockquote” formatting.


Foreign language


  • Speakers may include some non-English words or phrases. If you are familiar with the language (or word or phrase), record it and consult a dictionary for accuracy.
  • If you are not familiar with the language, indicate the use of non-English words with square brackets [words in French] where they appear. In your own notes, record the time on the digital recording where the words appear.




  • When a speaker talks for a long time, this creates a large block of text in the transcript. Use the natural pauses speakers use before starting a new thread of ideas to start a new paragraph. You can also use “context-oriented chunks” to break text into segments focused on a specific topic.