Different approaches to audio transcription
What is audio transcription?
Audio transcription involves recording the spoken word (either from digital format, such as a video or audio file, or spoken live) in written form. Audio transcription is often performed to evaluate the content of a piece of audio, for example in court cases or in a disciplinary hearing. Alternatively, it can also be a requirement when the content needs to be processed further, for example in the case of subtitles or voiceovers for video translation. Audio transcription can be performed on audio in any language (multilingual transcription) and then a translation performed if necessary for all parties to understand it.
Different types of transcription
The approach taken for transcription projects will usually depend on the final use of the transcription. The most frequently used different approaches are listed below:
Phonetic or Verbatim transcription
In this case, everything spoken is transcribed unabridged and true to the original. The resulting transcript therefore also includes gap fillers such as “Uh” or “Hm”, but also replicates slang, accent and dialect as spoken (for example “dunno” instead of “don’t know”), as well as repetitions, slips of the tongue or stuttering. This level of detail can make the transcription quite difficult to read.
Simplified or edited audio transcription
With this approach, everything that is spoken is transcribed, but the transcription is cleaned up for improved readability. Interjections such as “uh” or slips of the tongue are removed. Slang is transcribed into a standard register of English, improving the readability of the text.
Audio transcription for the deaf and hard of hearing
In this scenario, para-linguistic information, such as background noises that are necessary for understanding what is happening in an audio or video file, are also included in the transcript. Examples include additional cues such as “applause”, “dramatic music”, “dog barking”.
Audio transcription for subtitles or voiceover
Transcription for use as subtitles or as a script for voiceover needs a very specific approach. Not only are time stamps added to each line, but the number of lines and characters per line must also be specified in advance. The transcript therefore becomes a template for the subtitle file that is burned into the video or the voiceover that is applied. Moreover, during the transcription process, the spoken word may have to be shortened to fit within the specified framework. For subtitles, if the transcript is not edited and the subtitles are too long, there will not enough time to comfortably read the on-screen text. For voiceovers, if the transcript is too long, the voiceover will need to be read too quickly, making it sound unnatural.
When transcription is performed with a view to creating multilingual subtitles or voiceovers, the transcript also serves as a template for the translation, and translators must also take into account the character limits per line for the same reasons as above.
Quicker audio transcription using technology
Audio transcription software can shorten the transcription process significantly, leading to better lead times and reduced costs. Intelligent speech recognition software automatically creates a first draft of the transcript so that the draft only needs to be revised by a professional transcriber. To do this, the linguist works in a portal where the video or audio runs on a screen in parallel with the transcription. Timestamps are also inserted automatically, which is incredibly useful for video work (subtitling and voiceover). In short, good transcription software (used by transcription professionals as part of the process) can save time and money, whilst still delivering the same quality output.
Before starting the project, discuss with your account manager what format the audio is in (e.g., MP4, WAV, AVI) and what file format the transcription should be delivered in. Depending on the intended use, in some cases a transcript in the form of a simple Word file is sufficient. In other cases, such as with video translations, an export file such as SRT or VTT is required, or the transcript can be added directly onto the video as subtitles or a voiceover recorded.
Whatever your transcription requirement, we’d be delighted to discuss the specifics of your project with you. Please get in touch at [email protected] or on 0161 850 0060.