Anthropology

If you have an audio or video recording from your fieldwork that needs to be transcribed, you might save some time creating a machine-generated transcription that you can then edit. Princeton offers free tools to upload audio/visual files and create transcriptions from them. Instructions for how to do this using Kaltura are below. Instructions for Princeton-supported services will be added if they become available for all Princeton users.

Uploading A/V Files to Kaltura

• Request media upload privileges for Princeton Media Central at mediacentral@princeton.edu
• Log In to Media Central
• Click “CREATE” at the top of the page.
• Click “Upload”
• Upload Files (3GB maximum size, unlimited storage)
• The machine-generated Caption and Transcription process starts automatically. The quality will depend on the quality of the recording. The average accuracy rate is around 80%. (This audio recording generated a 92%-accuracy-rated transcript.)
• After the media file is processed, follow the following instructions to download the transcription file or edit the captions. (You may have to refresh the page to show the processing is complete.)

Note on Privacy and Confidentiality: You can upload, create transcriptions, and delete the files without having to make them public. Kaltura content is encrypted at rest and in transit. Once the content is uploaded, only the owner of the content and local Kaltura admins at Princeton have the ability to access that content unless the owner grants access to others.  Once the content is deleted, it disappears from both Kaltura and Princeton servers. Some more info on Kaltura security here.

Downloading Transcription Files in Kaltura

• Click My Media and Log in https://mediacentral.princeton.edu/my-media
• Click the Media to Edit
• Click ACTIONS below the Media displayed
• Click “Edit”
• Click Attachments
• For the .txt file, under “Actions,” Click the Download button
• Edit the text file as appropriate

Editing Captions in Kaltura

• Click My Media and Log in: https://mediacentral.princeton.edu/my-media
• Click the Media to Edit
• Click ACTIONS below the Media
• Click “Caption & Enrich”
• Under “Status”, click Editing button

If you use Zoom cloud recording for interviews, you can set Zoom to automatically produce transcriptions. See these instructions from the Zoom Help Center.

If you're a programmer, you might consider automated transcription in the programming language R using OpenAI's Whisper model. Instructions are here.

There's a program for Macs called MacWhisper that I've been using. It works pretty well, isn't free. At the time of writing, a single-user licence is €59 (approx. $69). Everything is done on your computer, so there are no privacy concerns. Transcriptions can be exported or copied and then deleted from the program. There are similar products for Windows such as Whisper Desktop (free but somewhat complicated to install) or Whisper Transcribe ($39.00/month or $240/year minimum) but I haven't tried them.

Otter.ai is a popular service offering high-quality transcriptions. Use without a subscription is very limited. [It might not be relevant, but those concerned for privacy might be interested to know that Otter is the target of a class-action lawsuit claiming it secretly records private work conversations.]

• Free version: 300 monthly transcription minutes; 30 minutes per conversation; Import and transcribe 3 audio or video files lifetime per user
• Pro version $100/year (1200 monthly transcription minutes; 90 minutes per conversation; Import and transcribe 10 audio or video files per month)
• Business version: $240/year (6000 monthly transcription minutes; 4 hours per conversation; Import and transcribe unlimited* audio or video files)