About the offer
The Department of Psychology has purchased a powerful computer to be used for transcribing audio and video files. The computer is placed in a security cabinet and is handled by Lars Kindberg.
The transcription program is OpenAI’s Whisper, a free program which can transcribe 82 different languages and also translate these languages into English. Whisper can be run locally on the computer and does not need to be connected to a network. The transcription program can transcribe files in mp3, mp4, mpeg, mpga, m4a, wav and webm formats.
After each run, when all the files have been transcribed, the computer is “reinstalled” by writing an image of the computer’s hard disk, as it was before the run, back onto the hard disk and encrypting the hard disk to ensure that no trace of the files remains on the computer. The only thing that is saved from the computer to the USB stick before the computer is restored is a text file containing information about the number of files that were transcribed, how long the total time of the transcribed material was, and the time taken to transcribe it. The computer is not connected to any network during the process, it only connects to the Internet when it needs to be updated.
Don’t expect the transcription to be perfect. The program “hears” errors, repeats and sometimes even makes things up. But it still provides a good foundation to start from.
Routines for transcribing with Whisper
- Contact Lars Kindberg, lars [dot] kindberg [at] psy [dot] lu [dot] se, to book an appointment for transcription.
- Upload your files to a USB stick. All files to be transcribed should be in the same folder on the USB stick and no other files should be in that folder. Do not use the letter å, ä or ö when naming the files. The USB stick can be encrypted, but then it must be encrypted using the software Bitlocker so that Whisper can transcribe the files directly onto the USB stick.
- Lars helps you complete the transcription. When the program has transcribed a file, the content is saved in the same folder as the file is in on the USB stick and the content is saved as five files in different formats. One file is in json format, which is a standardized format for moving data between different systems, three files are different variants of time-stamped text files and one is a plain text file (txt).
