Transcription is a procedure that allows to turn speech into writing.

Words are carried away by the wind. We all know that the ephemeral nature of speech can be inconvenient when it comes to proving that such and such words were ever spoken. Recording speech on a durable medium (such as an audio file) allows sound to be preserved and reproduced.

However, if the goal is to analyse, edit, encode, process, study, or search for oral sequences in a sound recording, some form of transcription is needed.

Speech can be transcribed in different ways, depending on the levels of information that you’re interested in reflecting. We’re going to show you some of the most common forms of orthographic transcription. We don’t deal here with other systems of representation such as phonetic transcription, in which different symbols of the common alphabet are used.

Literal transcript with semantic coding

The form of orthographic transcription that’s most faithful to the original sentence consists of reproducing the complete sequences, word by word, noting them with coded labels that provide information related to the following aspects:

  • Structural aspects such as utterances, including markup for moments when when several people speak at the same time;
  • Identification of the speaker for each utterance;
  • Errors, repetitions, delays, hesitations, unintelligible fragments, rectifications, interruptions, etc.
  • Semi-lexical vowel elements (for example, pss, shhh) and non-vocal elements (gestures) with a communicative function;
  • Non-communicative sound phenomena (e.g. closing a door, footsteps, music).

This type of transcription is used when there is a need to represent in writing even the smallest detail of speech captured in the sound recording, generally for the purpose of developing a corpus for the study of the spoken language.

Court transcription

The form of transcription required in the forensic field is a literal non-coded orthographic transcription, although it may contain explanatory annotations related to the identification of the speaker or to sounds that may be considered relevant. For example, barely intelligible passages or noises relevant to the conversation are often highlighted.

The transcription of judicial proceedings or sound recordings for forensic purposes can be presented in different formats, depending on the client’s specifications.

  • Simple certified transcription: certified court transcript, duly signed and stamped, structured by utterances, identifying the speaker;
  • Certified transcription with time codes: a certified court transcript where each line is preceded by a time code that indicates the minute, second and tenth of a second when that particular sequence begins and ends;
  • Synchronisation of text and audio on a still picture with subtitles: further to the certified transcription, we can deliver the audio with subtitles on a still image;
  • Synchronisation of text, audio and video with subtitles: if a video is transcribed, this option is perfect for following and analysing the subtitled conversation, even in a court of law.

Although in Spain the concept of sworn transcription doesn’t exist, a lawyer or linguist can certifify the quality of a transcription. At textinnova we certify the quality of our transcripts and guarantee absolute privacy throughout the process.

In addition to transcribing the content of the recording, in some cases it needs to be translated. The translated transcript will only be officially valid if it’s a sworn translation. At textinnova we take care of the entire process.

Do you need a transcript for judicial purposes? Contact us. We’ll deliver the certified transcript and, if applicable, duly translated into the language you need.


Editorial transcription

When what matters is the essence of the message rather than word-for-word accuracy, superfluous elements and some features of speech are removed in order to obtain a more polished transcription. Repetitions are eliminated, hesitations or truncated sentences are not transcribed, pronunciation defects are omitted and grammatical errors are corrected.

The proceedings of Congress are a good example of this type of transcript. But it’s not the only one. Interviews, lectures, and panel discussions are typically published after an edited transcription process.

At textinnova we provide editorial transcription services in various languages for magazines and other publications.

Do you need a transcript for publication? Tell us about your case and we’ll advise you.

Post-edited machine transcription

Based on voice recognition technology, systems have been developed that are capable of automatically transcribing with some reliability, provided the voice is intelligible. This type of transcription is useful for obtaining a quick draft of the content of a recording (for example, an interview) in order to extract excerpts or to edit the transcript for publication.

This technique can also be useful if you want to transcribe your own recorded voice, whether simple voice notes or longer talks, such as classes or lectures.

Machine transcription is not perfect. If the recording quality is poor, if there are background noises that make understanding difficult and if the system has not yet been trained in recognising the voice in question, the resulting transcript may (and often does) contain errors.

We offer you a refined service of post-edited machine transcription, so the result is exact, or edited according to various criteria, for publication purposes. Do you need to transcribe a recording quickly and you want to do it using a reliable transcription system for publication? Send us your query and we’ll give you an estimate that suits your needs.