Transcription rules and types

Transcription guidelines

Transcription guidelines help to ensure the finished transcription accurately represents the spoken word and is clear and understandable for readers. Here are some general transcription rules:

1. Use a consistent spelling system throughout the document, such as either US or UK English.
2. Human speech doesn’t only consist of spoken words. There are sometimes other sounds that add to the meaning. Transcription can represent all the sounds of speech as accurately as possible. This means sometimes using symbols to represent sounds and also taking into account dialects or accents. Phonetic symbols can be used to represent sounds that don't have a direct equivalent in written language. These symbols and the rules of phonetic transcription can be found in the International Phonetic Alphabet (IPA).
3. Punctuation marks should be used to indicate the speaker's intonation and to clarify meaning.
4. Transcription should reflect the speaker's stress and intonation patterns – using symbols such as an apostrophe (') to indicate primary stress, or a caret (^) to indicate a rising intonation.

Different types of transcription

There are various types of transcription, each with different transcription rules and conventions. Phonetic, orthographic and intelligent transcription are three types of transcribing spoken language into written form. Some transcription types are particularly suited to certain applications:

1. Phonetic transcription is a process using a set of symbols from the International Phonetic Alphabet (IPA). The rules of phonetic transcription allow the capture of the pronunciation of words. Phonetic transcription is often used in language learning applications and speech therapy. For example: The word ’cat’ would be transcribed as /kæt/, with each symbol representing a specific sound.
2. Orthographic transcription follows verbatim transcription rules. It uses standard spelling conventions and captures every word and sound. It's commonly used in situations where the purpose of the transcription is to provide an exact written record of a conversation or speech. Applications include journalism, legal proceedings and business meetings.
3. Intelligent transcription is a more modern method of transcription that uses artificial intelligence (AI) to transcribe spoken language. It’s particularly suited to automatically transcribing large volumes of audio or video content quickly.

Why are transcriptions used?

Transcriptions are useful for many purposes, including:

1. Accessibility: Transcriptions improve accessibility for people with hearing impairments, cognitive processing difficulties or for those who require more time to absorb information.
2. Legal purposes: Legal proceedings, such as court hearings, depositions and interviews, are transcribed so that a permanent record is available of conversations that can be used in official records and evidence.
3. Media: Transcriptions are used to create subtitling, captioning or dubbing. This is particularly important in multilingual media.
4. Business: Transcriptions are used for transcribing meetings, interviews and presentations to keep a formal record of information and any decisions taken.
5. Education: Transcriptions are used for transcribing lectures, webinars and other learning content to create documents that can be referred to in future studies.

How are transcriptions created?

Human transcribers

Professional human transcribers listen to an audio or video recording and transcribe it into written text. If multilingual transcription is required, they may also translate between languages. Here’s the process that a professional transcriber follows to create a transcription:

1. The transcriber listens carefully to the audio, paying attention to the speaker's tone, pace and the emotion of the content.
2. As the transcriber listens, they type what they hear either using word processing or specialised transcription software.
3. The transcriber ensures that the written text is accurate and grammatically correct.
4. If it’s part of the client’s brief, the transcriber can add timestamps to indicate when each sentence or paragraph was spoken.
5. If there are multiple speakers in a recording, the transcriber can indicate on the transcription when the speaker changes.
6. After the transcription is complete, it's given a final edit and proofread to ensure accuracy and consistency.

AI transcribers

Artificial intelligence (AI) transcription software is now available and it’s rapidly becoming more reliable. Here's how an AI transcriber works:

1. The AI transcriber app receives an audio or video file as input.
2. The AI transcriber uses automatic speech recognition (ASR) technology to convert the audio content into written text.
3. The AI transcriber uses natural language processing (NLP) to interpret the transcribed text. NLP helps to identify and correct any grammatical errors, spelling mistakes and inconsistencies.
4. The final output is a text file of the audio content. Some AI transcribers can also create transcripts with timestamps and speaker identification.

AI transcribers continuously learn as they process more data, and this means that they improve over time. However, it's important to remember that AI transcribers can make mistakes, particularly when dealing with different accents, background noise or low quality recordings. This means that it’s always important for a human transcriber to check the AI’s results and to edit the final transcript as necessary.

Can transcriptions be referenced?

Transcriptions can be referenced using systems like Harvard, APA (American Psychological Association) and MLA (Modern Language Association) to provide proper citations for sources included in the transcription.

Using referencing systems like Harvard, APA or MLA is important in some forms of transcription, such as legal or academic, because it ensures sources in the text are properly acknowledged.

Can AI reference transcription?

While AI transcription software can transcribe speech to text, it cannot yet automatically generate reliable citations and listings using the standard referencing systems. However, with AI transcription capabilities developing rapidly, this will undoubtedly become possible in time.

Currently, if accurate referencing is required, it must be carried out by a human transcriber.

Final thoughts

The general rules that apply to creating transcriptions are simple. However, applying different systems and using different types of transcription requires specialist knowledge.

Our multilingual transcription services offer fast, accurate transcriptions, with translations available in over 170 languages, and we offer all transcription types. We can also reference transcriptions if required.

If you’d like to talk to our team about transcription, or have a project you’d like to discuss, fill in this form and we’ll be in touch.

Download our free transcription template

Download templates for both detailed transcription and standard transcription. You can use the formats and examples in your own working document.