Transcription Guidelines for Professionals

By: Danielle Chazen
man sitting in front of his workstation looking at a book

Serving as an effective stenographer or professional transcriptionist requires both skill and speed. In addition to personal preferences for how to tackle projects efficiently, there are transcription guidelines and transcription rules to be aware of when producing materials.

For example, there are many transcription nuances to consider. Is there a correct way to handle transcribing numbers? Do you know how to timestamp transcriptions in media? Transcription details, use cases and requirements are seemingly endless. These include meeting legal requirements to ensure transcripts are admissible in court, writing a transcript of an interview for media purposes and transcribing verbal statements from courses for students who are hearing impaired. We’ve assembled high-level transcription guidelines with general tips to help professionals succeed. Many of these tips may seem intuitive, but they’re worth noting.

Transcription Rules and Guidelines

Do: Research difficult words

With the quick-paced nature of this profession, it’s common to not catch or understand certain words you are transcribing the first time around. If you encounter difficult words, context should never be guessed. If you encounter a speech or a term that is unclear, try to resolve it with research. You may also hear the term clearer in another location in the audio. If it’s still unclear, it’s good practice to use an [inaudible] tag. Sometimes, you may need to do a research search, such as looking for terms on Wikipedia, Google or Bing.

Do: Use appropriate regional spellings

It’s important to use the appropriate language variation for your location and for the purpose of the transcription. Words spelled differently in the US vs. the UK should be spelled to match the spelling of that region.

Do: Make transcripts verbatim if required

A verbatim transcript includes all dialogue spoken, word for word, including fillers, false starts, incorrect sentences, slang words, stutters and repetitions. When approaching a Verbatim job, you must ensure that everything the speaker utters is written. For stutters (I- I was going), use a dash. For fillers (about the, um, presentation), use a comma. For example, here’s the difference between a clean read vs. verbatim: Clean read: “I was about to tell you about the presentation that we saw yesterday.” Verbatim: Um, like, I- I was about to tell you, you know, about the presentation that, um, we saw yesterday.”

person typing on a laptop on a table with notebook and pen, mobile phone, and an orange fruit - transcription guidelines

Don’t: Paraphrase or change grammar

Don’t paraphrase or change grammar even if a speaker makes mistakes. Use the original speaker words. Don’t alter speaker words even if they’re incorrect grammatically. Don’t add or remove any words or information regardless or how relevant or irrelevant they seem. For example, a single word changed can make or break a case when considering transcriptions in legal settings.

Do: Note interrupting noises

It’s rare for transcripts to not include sound events that interrupt the flow of a lecture, deposition or dialogue. Note any background noise when it occurs with brackets and a short description explaining the sound. A note should also be made for silence when necessary. Try to keep these descriptors at one-to-three words. Examples include: [laughter], [applause], [phone ringing] and [music] for noise, or [silence], [cuts off] or an ellipsis: “I think… John is the one who should be responsible for that payment.” when you encounter silence.

Do: Take into account transcript structure requirements

Clients, depending on whether they are academic, legal, media or others, often have specific requirements for the format of their transcripts. From breaking the transcript into paragraphs to how to use spaces to items to keep an eye on for grammar, it’s best to gain knowledge of these requirements ahead of time. Technology-based transcription tools can also help to make customization processes less manual. For example, media entities have specific requirements around timestamps or SMPTE timecodes and legal agencies may request speaker differentiation to be presented in specific ways and in custom templates.

Do: Consider using technology, including Artificial Intelligence

Many technologies have been developed to help speed up transcription services. Technology can help to ensure both accuracy and speed, so you don’t have to turn away projects and can expand your services. The most advanced method of legal transcription is AI software, or speech-to-text transcription software. This software utilizes artificial intelligence and machine learning to process audio or video and produce a transcript, often in real time.

AI-based software allows for an efficient, affordable and accurate transcription process. Automatic speech recognition machines often self-learn to ‘get smarter’ and recognize more terms to provide more accuracy with each use. Technology can also help to detect difficult terminology or accents and differentiate between speakers accurately in instances when humans may struggle more to detect these differences. Then, transcriptionists can allocate time to fact checking and editing the technology’s work and get time back to work on additional projects.

Therefore, automating the process with technology and AI-based transcription software can not only save transcriptionists time, but it can help them reduce operating costs and make more money with the ability to handle more work.