Audio and video
transcription

Accurate. Clear. According to their function.

Audio and video transcriptions have many functions. They are the basis for subtitling. They are used to archive media content. They are used for the scientific recording of interviews and experiments. That's why we transcribe your media with their function in mind, from good readability to the rules of Dresing & Pehl.

Transcription - Symbolic image: Pegasus leaps from one medium to another over a mountain valley

Functionally optimised

Why human transcriptions?

Software such as Adobe Premiere and online platforms such as Happy Scribe can auto-generate transcriptions. That’s why we at Beyond the Blur use these software solutions ourselves. However, we use them only as a rough starting point for a clean revision by a human. Even though audio-to-text and video-to-text systems are improving, they are still very error-prone. Unfortunately, this is pre­cise­ly where it counts the most. In corporate videos, company, product and personal names are routinely misspelled. In scientific discussions and interviews, the systems fail precisely because of the spe­cial­ised vocabulary.

In addition, these software solutions do not optimise for the function of a particular transcription. Above all, transcriptions for subtitling should be easy to read and occasionally contain explanations for the hearing impaired. Scientific transcriptions should be precise and accurate. Explanatory com­ments may be necessary as well. The transcription of an interview for psychological purposes, for example, must often contain information about the behaviour of the interviewee. Does he laugh? Does he cry? Does he seem insecure? Professional transcriptions therefore require human revision with these functions in mind. Your advantage:

  • Repetitions, fillers etc. can be deleted
  • Transcriptions for the hearing impaired possible
  • Transcriptions according to Dresing & Pehl possible
  • Special requirements can be taken into account

How much do our transcriptions cost?

Fair prices. Excellent results.

  • Price by audio and video length

    The price refers to projects with good audio quality and without pronounced dialects. It includes a correction loop. If the audio quality is poor or if you require an extended transcription for the hearing impaired or a transcription according to Dresing and Pehl, the price per minute will be higher.

    €3/ min.+€30basic fee
    Request a quote now

      Advantages

    • Software-supported, but with ‘Humans in the Loop’
    • More accurate and flexible than fully automated software transcriptions
    • Specialised vocabulary and names of persons, institutions, brands, etc. are checked separately
    • Express deliveries possible
    • Reliable budget planning possible
    • Offers can be compared easily

    • Our standards. Scientific standards.

    Transcriptions according to
    Dresing and Pehl

    There is no doubt that a lot is at stake in marketing and the entertainment industry. But let’s not kid ourselves: The scientific standards of research are much higher. This is also the case with tran­scri­ptions. Because they are used in qualitative analysis and therefore require at least a basic, general understanding of this analysis, the creation of transcriptions for research purposes is more difficult to automate.

    A look at the rules that Thorsten Dresing and Thorsten Pehl discuss in their book illustrates why such transcriptions are significantly more time-consuming:

    Semantic transcription

    • Verbatim transcription, no phonetic or summarised presentation.
    • Adapting word slurs to written German.
    • Translation of dialects into High German, if possible.
    • Colloquial particles (‘gell, ne’) are recorded.
    • Smoothing or omission of stuttering and broken words.
    • Cancel character ‘/’ for incomplete half sentences.
    • Punctuation that promotes readability; focus on units of meaning.
    • No transcription of reception signals (e.g. ‘hm’) as long as they do not interrupt the flow of speech.
    • Marking of pauses from approx. 3 seconds with ‘(...)’.
    • Mark emphasised words with VERSALS.
    • Separate paragraphs for each speaker contribution, with time stamps.
    • Emotional expressions in brackets (e.g. ‘(laughs)’).
    • Incomprehensible passages are labelled with ‘(unv.)’, if necessary with the cause (e.g. ‘microphone noise’).

    Extended semantic transcription

    • More detailed labelling of pauses, filler sounds and stresses.
    • Marking of speaker overlaps with ‘//’.
    • Precise depiction of hesitation and word-finding inhibitions.

    Standardised spelling

    • Standardised representation of particles and filler sounds (e.g. ‘hm’, ‘ähm’).
    • Write out units of measurement and symbols.
    • Literal spelling of quotations and idioms.
    • Number rule: Small numbers (up to twelve) as words, larger numbers as digits.

    General guidelines

    • Clearly define the purpose and objectives of the transcription before starting.
    • Orientation towards readability vs. depth of detail depending on the research objective.
    • Use of transcription programmes to increase efficiency.
    • Consider the required time: 5 to 10 times the interview duration for simple systems, up to 60 hours per hour of material for complex systems.

    The good news is that although scientific transcriptions cannot be fully automated, there are systems that have specialised in this and speed up the creation of such transcriptions compared to con­ven­tional systems. One example is the transcription software f4, which we can also use as a support for such transcriptions. However, even with f4, it is not possible to do it completely without a post-editing and revisions.