Real-time, semi-real-time and deferred transcription

We interact daily with real-time transcriptions: think about when you use your smartphone’s voice assistant, which helps when you are unable to type out your message or to simply save time. Speech-to-text is essentially the spoken word converted mechanically into written text before it is sent. This widespread real-time transcription takes place automatically, supported by the basic functions of any modern smartphone, without any real end processing.

Professional real-time transcription

Real-time transcription performed at a professional level is where transcribers use an initial audio file (WAV, AIFF, MP3, FLAC, ALAC, M4A, OGG, WMA, AAC, etc.) to transfer what is said at a certain time (usually in the context of work, business, training or consultation) into text so that the information can be used almost immediately, for instance to have a record of a meeting for minuting.

In a legal or corporate setting, the transcription is sent to a notary or corporate affairs manager, or to an in-house legal assistant who drafts the meeting minutes for the board of directors, the board of statutory auditors, an internal control committee, risk and control committee, audit committee, ESG committee or any other board committee.

Given the importance and intended use of the document produced by the professional, there is no margin for error.

Real-time transcription by CMT

Unlike speech-to-text automatic transcription systems, the real-time transcription service developed by CMT is the result of work by a team of transcription professionals, who successfully decipher every single word and grasp even the most imperceptible nuances of a speech, which is then transferred manually into a final text. The result achieved in writing is identical to the words spoken.

This type of live transcription faithfully puts down in black and white everything that is said. Different levels of precision are used based on the customer’s requirements or the type of document to be used or archived. Verbatim transcription is when we need to transfer speech to text word for word. The transcriber must therefore type out every word, expression and voice present in the audio being transcribed.

For a flawless service, CMT’s expert transcribers prepare by studying the reference sector’s language (specific terminology, expressions, technical terms, etc.), any foreign language words that might be spoken by the participants, the names of the speakers, particular acronyms and so on.

Transcribing directly on site or remotely

Real-time transcription can take place on site, i.e. directly from the customer’s offices where the event is being held, or remotely. The result is the same: regardless of whether the team of transcribers is on site using proprietary software or manages the transcriptions remotely, partial versions are delivered for live viewing during the event. Afterwards, CMT gets to work to provide the final version as quickly as possible (having agreed on delivery turnarounds in advance with the customer).

The full transcription is formatted and made available on a shared document (Google Docs, Microsoft OneDrive, Dropbox, etc.) or in Word and sent to the customer in the chosen manner (hand delivered USB, via email, uploaded to an SFTP server, a repository or the secure areas of CMT or the customer, where available).

Edited transcription for a more “usable” result

Customers can decide whether to use the text as is, provided directly by the transcribers, i.e. with a transcription of everything that is said, or they can request an edited version. The edited transcription can be “cleaned up” by removing everything considered superfluous for the sake of comprehension and fluid communication, such as repetitions, stutters, self-corrections, interjections, false starts, etc. The text can also be perfected in terms of grammar and syntax, without misrepresenting the content. This editing service is additional and is agreed as a supplementary task, over and above regular real-time transcription.

Video and audio synced transcription

With an audio or video script, we may need to listen back to (or rewatch in the case of video) the point when a concept was expressed to check its accuracy or even just to experience it again. CMT can embed the audio or video in question into a web page, either on its own site or the customer’s web space, so that the transcription can be consulted easily. The transcription is synced in time with the video or audio.

Placing your cursor over a specific point in the reproduction (of the audio or video), sends you to the corresponding point of the script. Likewise, by selecting a word in the script, it is possible to proceed automatically to the corresponding part of the audio or video. The transcription is shown at the same pace as the video.

Guaranteed security and privacy at every stage

In addition to precision and quality of service, CMT guarantees privacy and maximum protection of sensitive data using the best security technologies, software and hardware on the market, in full compliance with GDPR.

Transcriptions and delivery times

Transcription services are distinguished based on the delivery time required by the customer.

  • Real-time transcription involves the continuous transmission of the work carried out by transcribers. At any given event, transcribers send partial versions of the speeches (from the beginning to that moment in time), so that the participants can read what has been said live and word for word.
    It is used to remember specifically what has been said by the speakers, to see parts missed due to distraction or misunderstanding, to recall useful details, note down important points and lastly to respond promptly to questions, requests or opinions expressed by the participants.
  • Semi-real-time transcriptions require slightly more time to send the final text to the customer. Professional transcribers carry out transcriptions at the same pace as the event, but the transcribed text is delivered a few hours after.
  • Otherwise, for deferred transcription, delivery times are established for transcribing the audio files, from 8 working hours after receiving the files.

In all cases listed the text is provided in full. Naturally, customers can ask for an additional editing service to refine the content or make it usable for purposes other than plain transcription.

Our transcription service proves extremely useful in the following cases:

  • Transcriptions of conference proceedings
  • Transcriptions for meeting minutes (AGMs)
  • Transcriptions for board meeting minutes
  • Transcriptions for press conferences or sport and journalism interviews
  • Transcriptions for strategic meetings
  • Transcriptions for board committees
  • Transcriptions for ESG committees
  • Transcriptions of company meetings
  • Transcriptions of interrogations
  • Transcriptions for meetings with fellow tenants
  • Transcriptions of university lectures and training courses