Transforming Transcriptions

Ian Noah's picture

There has been some debate in oral history circles about the merit of full transcriptions over key word summaries. Summaries may certainly be quicker, demand limited resources, and prove useful for search engines, indexing and initial research purposes. However, there can be little doubt that - when it is possible and financially viable -  a full textual rendition of an entire interview recording fundamentally and exponentially enhances the value of the recordings by providing a substantial number of additional and important benefits:-

  • Increasing the efficiency of research by providing quick access to relevant words, phrases or sections of a recording without having to spend time trawling through the recording.
  • Providing clarity for researchers where recordings - due to sound levels, accents or interference - may render the sound difficult to decipher.
  • Making content available in a more accessible format to the hearing impaired and to the various text to speech devices they may use.
  • Making content available to those without the equipment to play back audio or video recordings or in situations in which this may not be practicable.
  • Creating a readily available transcript as the basis for future documents or publications.

Until very recently, this range of advantages was out of the reach of most documentarians (production teams, oral historians, researchers, etc) as the costs of a transcription (with most professional agencies charging between £50 to over £100 per hour of recording) or the time involved in doing it oneself, were substantial and frequently prohibitive. Digital technology gave us early speech to text software in the form of dictation software, which unfortunately proved of limited value to transcribers as it could only recognise one key voice, and even the required a lot of hours of training). Fortunately, advances in artificial intelligence and speech recognition technology are now able to deliver dedicated transcription software; automated programs which can recognise any voice and thus provide a transcript of a conversation even between several speakers.

We know that digitising of recordings reduces wear and tear on original audio and video tapes, and provides versatility, flexibility and efficiency for archive storage, indexing and searching as well as access and distribution. Now, we can also add the value of a fully automated, digital transcription into the mix of benefits.