• Transcription

    Transcription services

    Text transcription of audio or video material.

Transcriptions

KERN Global Language Services offers a professional transcription service to capture German or foreign-language audio or video material in writing. A transcription is the writing down of spoken texts and situations from audio and video recordings, which are available as audio CDs, DVDs, and MP3 files or other file formats, such as Wave. The variety of texts which need to be recorded in writing ranges from video conferences, telephone interviews as a data collection strategy in social research and interviews for films and television productions to series of lectures.

Experience has shown that an hour of audio material gives approximately 30 to 40 standard pages of typed text. The time required for transcribing audio material can be measured by a factor of five to nine, depending on the quality of the recording, the speed of the speech, and the degree of accuracy required for the transcription to be adapted for its purpose. The relatively high level of time and effort required for a transcription is due to dialects of the speakers and background noises, which make it necessary to repeatedly play back the sentences. A transcription of the audio material that contains neither gaps nor errors, and in which even small details of the utterances are recorded, including when speakers' utterances overlap requires the highest level of concentration.

General information on transcription

Experience has shown that an hour of audio material gives approximately 20 to 30 standard pages of typed text. The time required for transcribing audio material can be measured by a factor of five to nine, depending on the quality of the recording, the speed of the speech, and the degree of accuracy adapted to the purpose of the transcription. The relatively high level of time and effort required for a transcription is due to dialects of the speakers and background noises, which make it necessary to repeatedly play back the sentences. A transcription of the audio material that contains neither gaps nor errors, and in which even small details of the utterances are recorded, including when speakers' utterances overlap, requires the highest level of concentration.

What are the special features of a transcription?

Appropriate transcription guidelines must be taken into consideration, depending on the purpose of the transcription. Unless the client requests otherwise, the material shall be transcribed in words, not phonetically and in detail, i.e. not in summarised form. Dialectal sounds and sounds such as “erm” and “um” etc. are generally not transcribed. Language and punctuation keep closely to written language. Long pauses are marked by ellipses and passages of texts that were particularly emphasised are underlined; interjections from other people, laughter, or sighs are noted in brackets. Changes in speakers are indicated by blank lines and, if applicable, an anonymous designation (Speaker 1, Speaker 2). Other features of an interview situation, such as entering or leaving a room or a telephone ringing may also be of interest and can be noted upon request.

After the text has been recorded in writing, a spell check must be carried out and the text must be formatted according to the client's specifications if this is requested. After the completion of the transcription, there is also the option of having the text translated.  

We require the following information from you:

  • In what format is the audio material available?
  • How many minutes/hours of audio material are there?
  • What kind of audio material is it (interviews, sections of presentations, etc.)?
  • What is the purpose of the transcription?
  • Do you have any particular formatting requests or transcription rules?
  • What file format is required for the transcribed text?
  • Do you require a translation after the transcription?

Appropriate transcription guidelines must be taken into consideration, depending on the purpose of the transcription. Unless the client requests otherwise, the material shall be transcribed in words, not phonetically and in detail, i.e. not in summarised form. Dialectal features and sounds such as er, um, etc., are generally not transcribed. Language and punctuation keep closely to written language. Long pauses are marked by ellipses and passages of texts that were particularly emphasised are underlined; interjections from other people, laughter, or sighs are noted in brackets. Changes in speakers are indicated by blank lines and, if applicable, an anonymous designation (Speaker 1, Speaker 2). Other features of an interview situation, such as entering or leaving a room or a telephone ringing may also be of interest and can be noted upon request.

After the text has been recorded in writing, a spell check must be carried out and the text must be formatted according to the client's specifications if this is requested. After the completion of the transcription, there is also the option of having the text translated.  

In order for us to give you the best possible support with transcribing your audio or visual material, your KERN project manager needs the following information:
  • In what format is the audio material available?
  • How many minutes/hours of audio material are there?
  • What kind of audio material is it (interviews, sections of presentations, etc.)?
  • What is the purpose of the transcription?
  • Do you have any particular formatting requests or transcription rules?
  • What file format is required for the transcribed text?
  • Do you require a translation after the transcription?

You are interested in a transcription? Then please get in touch – we're here to help!

Automatic speech recognition (ASR): Machine transcription

Recording, post-editing and publishing online: The creation of media content has never been as simple as today. Apart from being used in media production, audio and video content plays an increasingly important role particularly in business communication. However, the new possibilities also pose new challenges to companies. Often, they need to obtain a quick overview of what has been said to enable them to make the correct decisions.

With the machine transcription from KERN, we can create a raw transcript in just a few clicks. This subsequent human editing is even more reliable - it offers up to 99% accuracy! Our experienced language experts listen intently for you and transfer the content, where necessary, into all the desired target languages.

What are the requirements for automatic transcription?

Particularly when the audio quality is good, the number of interferences is low and the speakers speak without an accent, ASR can produce good results. This means that the transcription of professionally produced media segments has more chance of success that the transcription of an audio file of a speech recorded on a mobile phone where the speaker is speaking in their second language.

When is human post-processing needed?

Human post-processing is advantageous for transcriptions that are to be translated or that are to be used for subtitles. Alongside the required foreign language, the subtitles must include the time code so that they appear at the right moment. The length of the text is of critical importance here and the highest level of precision is required during translation. The translators and transcribers of KERN Global Language Services are trained for this and have many years of experience in the area of subtitling!

We will take the special elements of your transcription into account and we will be pleased to advise you which services are appropriate for your project! Simply contact us more more information.