Article

Verbatim vs Clean Verbatim Transcription: Which Do You Need?

Verbatim transcription records every word and sound exactly as spoken, including fillers, false starts and pauses. Clean verbatim keeps every meaningful word but removes hesitations and repetitions without paraphrasing. The right choice depends on use: evidence and discourse analysis need verbatim, while meetings, media and most research interviews suit clean verbatim.

Person wearing headphones working at a computer by a window

Key takeaways

  1. Verbatim transcription keeps every word and sound as spoken, including fillers, false starts, stutters, pauses and meaningful non-speech sounds.
  2. Clean verbatim transcription removes fillers, stutters and false starts but keeps the speaker's own words and meaning, without paraphrasing or summarising.
  3. Evidential recordings and conversation or discourse analysis usually need full verbatim, while business meetings, media and most thematic research suit clean verbatim.
  4. A verbatim transcript can always be cleaned later, but a clean verbatim transcript cannot recover the hesitations it left out.
  5. When a transcript will be translated, agree how fillers and broken sentences should be handled in the target language before work starts.

Anyone ordering a transcript is asked the same question sooner or later: verbatim or clean verbatim? The answer changes what the transcriber writes down, how long the job takes, and whether the transcript is fit for its purpose, whether that is a witness interview, a research focus group or a board meeting.

This guide explains both styles in plain terms, compares them line by line, and shows which one suits legal, academic and business recordings. It also covers what happens to each style when the transcript is translated, and how to brief a transcriber so you get the style you actually need. For the wider picture, from file preparation to translation and subtitling, see our guide to transcription and translation of audio and video.

What is verbatim transcription?

Verbatim transcription (sometimes called true or full verbatim) is a word-for-word written record of a recording that keeps everything the speakers say, exactly as they say it. Nothing is tidied up, even when the speech is ungrammatical or hesitant.

A verbatim transcript normally includes:

  • Fillers such as "um", "er", "you know" and "like".
  • False starts, where a speaker begins a sentence and abandons it ("I went to the – no, we drove to the office").
  • Repetitions and stutters ("I, I, I didn't").
  • Non-speech sounds that carry meaning, such as [laughs], [sighs], [coughs] or [door slams].
  • Pauses, interruptions and overlapping speech, marked with agreed symbols or tags.

The W3C's accessibility guidance uses the same idea when it describes a good transcript of a press conference as "a verbatim record of everything the speakers say", with each speaker identified and significant sounds such as "applause, laughter, questions from the audience" noted. That guidance supports WCAG 2.2 Success Criterion 1.2.1, a Level A requirement for audio-only and video-only content.

What is clean verbatim transcription?

Clean verbatim transcription (also called intelligent verbatim) records every meaningful word a speaker says but removes the noise of natural speech. Fillers, stutters, false starts and unhelpful repetitions are left out, while the speaker's own words, word order and meaning stay unchanged.

Clean verbatim is not paraphrasing. The transcriber does not rewrite sentences, improve the vocabulary or summarise. Light corrections are limited to things such as removing a repeated word or a filler that adds nothing.

The closest public model of clean verbatim in the UK is the parliamentary record. The Northern Ireland Assembly's guide to its Official Report (Hansard) describes the report as:

"substantially the verbatim report, with repetitions and redundancies omitted and with obvious mistakes corrected."

According to the same guide, that definition is based on the one in Erskine May and was adopted in 1907 by the Select Committee on Parliamentary Debates at Westminster. More than a century later, it is still a good one-line description of what clients usually mean by clean verbatim.

Verbatim vs clean verbatim: what is the difference?

The difference between verbatim and clean verbatim transcription is what happens to the hesitations, repetitions and non-verbal sounds of real speech. The table compares the two styles feature by feature.

FeatureVerbatimClean verbatim
Fillers ("um", "er", "you know")KeptRemoved unless they carry meaning
False starts and self-correctionsKeptRemoved; the completed thought is kept
Stutters and repeated wordsKeptRemoved
Grammar errors made by the speakerKeptUsually kept; only obvious slips are corrected if agreed
Laughter, sighs, crying, background eventsTagged in square bracketsTagged only when relevant to meaning
Pauses and overlapping speechMarkedUsually not marked
ReadabilityHarder to read; closer to the audioEasy to read; closer to written prose
Typical usesEvidence, investigations, discourse and conversation analysisBusiness meetings, most research interviews, media, content reuse

Both styles should label speakers consistently, flag inaudible passages, and avoid any summarising. Those basics do not change with the style.

Is edited transcription the same as clean verbatim?

Edited transcription is not the same as clean verbatim: an edited transcript goes further and rewrites the speech into correct, readable written English. Sentences may be restructured, grammar fixed and rambling passages condensed.

Edited transcripts suit published interviews, articles based on podcasts, or minutes where readability matters more than the speaker's exact phrasing. They are unsuitable for anything where the precise words may later be scrutinised, because the edited text is partly the editor's language rather than the speaker's. A summary is a further step away again: it records what was discussed, not what was said.

The right transcription style depends on how the words will be used. As a rule, the more likely it is that someone will argue about exactly what was said, the closer to full verbatim the transcript should be.

Police interviews, witness accounts, investigation interviews and recorded calls that may be used in proceedings normally need full verbatim. Hesitations, self-corrections and pauses can matter when a court assesses a witness or a suspect's account. In England and Wales, section 139 of the Criminal Justice Act 2003 allows a witness in criminal proceedings to refresh their memory from a transcript of an earlier sound recording, which is one reason accuracy matters. Our guide to transcription for legal evidence and police interviews covers this in detail.

Research interviews and focus groups

For research, the analytic method decides the style. Conversation analysis and discourse analysis need full verbatim with detailed conventions for pauses and overlaps, while thematic analysis often works well with clean verbatim. The UK Data Service stresses "consistent use of transcription conventions, standard templates" and clear speaker labelling. See transcribing and translating research interviews for anonymisation and data protection.

Business, marketing and media recordings

Meetings, webinars, podcasts, training videos and marketing interviews almost always need clean verbatim. Readers want the content, not the hesitations, and clean text is far easier to reuse for minutes, articles, captions or translation. If the end product is captions or subtitles, read subtitles vs transcripts vs captions first, because subtitles are condensed further still.

How does the transcription style affect translation?

The transcription style affects translation because fillers, false starts and hesitations do not transfer neatly between languages. A clean verbatim transcript translates into clear, natural text; a full verbatim transcript forces the translator to decide how to render each "um" and broken sentence.

When a verbatim transcript is translated, agree in advance how the translator should handle:

  • fillers and hesitation sounds (keep them as tags such as [hesitates], or render them with natural equivalents in the target language);
  • ungrammatical or incomplete sentences (mirror the break in the translation rather than completing the thought);
  • words spoken in a third language or in dialect (leave them in the original with a translator's note).

For foreign-language recordings, you may want different styles at each stage, for example a full verbatim source-language transcript for the file and a clean verbatim English translation for readers. Our article on foreign-language transcription and translation explains the workflow options.

How do you brief a transcriber on the style you need?

A transcription brief should state the style in writing and give examples, because "verbatim" means slightly different things to different providers. Use this checklist when you order:

  1. Name the style: full verbatim, clean verbatim or edited, and say what the transcript will be used for.
  2. List what to keep or remove: fillers, false starts, laughter, pauses, background noise.
  3. Set the speaker labels: names, roles (Interviewer, Participant 1) or codes for anonymised data.
  4. Choose timestamps: none, at each speaker change, or at fixed intervals such as every two minutes.
  5. Agree tags for problems, such as [inaudible 00:14:32] or [crosstalk], with a timecode so passages can be checked.
  6. Send a spelling list of names, places, products and technical terms.
  7. Ask for a short sample page before the full job starts, especially for long or multi-file projects.

Common mistakes when choosing a transcription style

Most transcription problems come from a style that was never specified or was specified too loosely. These are the mistakes we see most often:

  • Ordering clean verbatim for evidence. A transcript that later has to be used in a dispute cannot recover the hesitations it left out. When in doubt, order verbatim; it can be cleaned later, but not the other way round.
  • Ordering verbatim for content reuse. Full verbatim makes minutes, articles and subtitles slower to produce and harder to read.
  • Mixing styles across a project. If two transcribers work on one research study with different conventions, the transcripts cannot be compared reliably.
  • Accepting raw automatic transcripts as "verbatim". Speech recognition drops fillers unpredictably and mishears names; its output always needs checking against the audio.
  • Forgetting the translation stage. Decide the style for the translated version as well as the source transcript.

How TranslationWorks can help with verbatim and clean verbatim transcription

TranslationWorks is a UK-based translation agency that transcribes audio and video recordings and can translate the transcripts into 150+ languages, working with native-speaking linguists. TranslationWorks is ISO 17100 (translation services) certified, ISO 9001 (quality management) and ISO 27001 (information security) registered and Cyber Essentials certified, which matters when recordings contain confidential interviews or meetings. When you ask for a quote, tell us how the transcript will be used and which style you need, whether full verbatim, clean verbatim or edited, along with any conventions for speaker labels, timestamps and tags. Find out more about our audio and video transcription services and translation services, or request a free transcription quote, which we send within 24 hours.

Not sure which transcription style your recordings need?

Tell us what the transcript is for and we will recommend verbatim, clean verbatim or edited style in your free quote.

Keep reading

More from TranslationWorks

Related reading

Get a quote Call