Case Study: Flexagon Capital Solution

With this new case study, we want to present the transcription project commissioned by the British company Flexagon Capital Solutions LLP.

 

Flexagon Capital Solutions LLP (“Flexagon”) is an investment manager MIFID firm that combines transaction structuring, credit origination, and asset allocation capabilities with a strong risk management discipline.

For private investment deals, Flexagon has an established granular network of contacts in the territories of interest, so to source first-hand speciality and niche opportunities.

For asset allocation, Flexagon employs robust techniques for building a pool of investments that match the risk/reward objectives of its client.

Flexagon can offer its structuring/execution and portfolio management services to professional and institutional clients and fund management companies.

Transcription service

At Translations Universe we provide a professional audio/video transcription service. By relying on native-speaking industry professionals, we report in writing, audio or visual files. Depending on the goal of the requested transcription, we have selected two main categories:

Soft transcription

The transcription produced in digital format reconstructing the dynamics of the text while maintaining the meaning of the conversations of the source file. In this way, portions of the original source could differ, as in the case of exclamations, meaningless sentences, truncated sentences, sounds or words in the background. We aim to transcribe the meaning of the given audio/video file.

Literal transcription

Literal transcription: the transcription produced in a digital format not affecting the audio of the file. All parts of the speech will be transcribed, including truncated sentences, exclamations, interjections, ideophones. This type of transcription is required if the contents to be transcribed are of a legal nature or in any other case if requested by the client.

The files are generally returned in the same format as the source of the commissioned text unless special requests from customers. Our translators work on editable digital texts.

For the audio or video parts to be transcribed to be uncoiled, it is necessary that the files provided by the clients are audible to return a transcription quality close to perfection.

We are able to translate audio or video media files, this requires the transcription and translation of the media files, thus, it will apply the combined rates for the transcription and translation service. The customer can make a specific request for the type of transcription.

The project

The client chose Translations Universe because of its need to transcribe a complex project, both for its legal-commercial nature and for the amount of work to be developed. For the transcription of this project a team of 3 experts, two transcribers, and a project manager with the task of splitting the files to be transcribed and coordinating the staff.

The main difficulty, besides having a very short delivery time, was managing to deliver a work that was linear, in the terms adopted and in style. As time was short, it was necessary to proceed by creating a terminological database allowing transcribers to work with ease.

As a rule, we prefer to assign this type of work to a single competent professional so that the project does not differ in style and glossary. Yet, when the client needs to receive the documents, in a short time, it is necessary to divide the workload between several professionals. As mentioned above, glossaries allow the meanings to be conveyed by using the same terms, especially in the case of soft transcription where it is preferable to enhance the meaning of the transcribed moment.

In this specific project, the client requested a Literal transcription, so that every element played in the audio files would be clearly stated in writing. This type of transcription is preferred in case the audio extracted from a conference, meeting or speech in general needs to be reported in full written form. Usually, this type of procedure is used in the case that the final piece of work is used for legal, medical or fundamental importance purposes.

Moreover, an extra level of difficulty is due to the participation of several interlocutors within the same discourse. In this case, the customer produced an audio file in which eight members of the company participated, each with their own phonetic tone, dialectal inclination, and style in the spoken language. The file format was not a video format and thus we conducted the transcription by using our transcription software able to recognize the vocal timbre of the speakers detecting the wavelength of the sound played.

We transcribed two files, for a total duration of 360 minutes. It takes about 10 minutes to transcribe 1 minute of an audio/video file to get an idea of the work done. The fee charged per word was 3.80€/word. The service, in this case, was carried out in a mono-language form. We are able to offer a combined transcription/translation service that provides two distinct but competing levels of elaboration to develop a contextualized and linear work, as well as localized for the phase in which it will be consumed.

Resume

There are several ways to approach an audio/video transcription project, one of them is the one we have described. The approach varies according to the client's request and the expertise of the file that may determine the choice of a Soft or Literal transcription.

Price per minute of transcription varies according to language and area of expertise and can be estimated on the Pricing page.

To shorten delivery times we may rely on the cooperation of several professionals, coordinated by a manager and the creation of a glossary that provides linearity to the entire project.

In this project, more than 45,000 words have been transcribed, equal to a minute of 380 minutes of recording. There were 8 speakers participating in the speech and they were identified thanks to sound wavelength detection software.

Files in mp3 format have been delivered to the client in Docx format, presenting a matching that allowed to identify the content of the reproduced speech and the agent speaker at that particular time.