Accessible Transcripts

Downloading and Editing Panopto Caption Files

For WCAG 2.1 AA compliance, videos must have accurate captions and/or transcripts. If you use Panopto’s auto-captions, always work from the caption file with timestamps, not from text copied out of the viewer window.

Step 1: Open the video and go to the Captions settings

  1. In Panopto, click the Settings button for that session (usually a gear icon). (If you are viewing the video, you may need to first click the three vertical dots in the top-right corner.)
  2. In the Settings panel, click the Captions tab on the left.

This is where Panopto lists all Available Captions associated with the video, including the default captions. For me, these appear as “English (United States)”.

Step 2: Download the caption file

  1. In the Captions tab, under Available Captions, expand the captions you want to edit.
  2. You should see an option to Download file (or similar) to download the caption file. Click this to send the captions to your Downloads folder.

If you edited your video in Panopto, the captions are based on the edited file; there is also an option to download the unedited captions (if you wish).

You should now have a caption file saved to your computer, probably as .txt file. It will contain your transcript with timestamp lines, for example:

1
00:00:02.000 --> 00:00:06.500
Welcome to our discussion of eigenvalues and eigenvectors.

This is the correct file to edit. It includes both the text and the timing information.

Step 3: Edit the caption file (for example, with generative AI)

Next, clean up the caption text for accuracy, punctuation, spelling of mathematical terms, and readability. You can:

  • Edit the file manually in a plain text editor (e.g. NotePad, TextEdit... I actually right-click and open in TeXShop), or
  • Use an AI tool (e.g. ChatGPT, Gemini) to clean up the transcript.

If you use AI, upload or paste the entire caption file and give very explicit instructions. I recommend manually editing any typos that appear on the first few lines, where you presumably set up the topic for the lecture. This step will help the AI understand the lecture content.

Here is a prompt you can try (and adjust as needed):

I need to fix a video transcript for a recorded lecture on the Chain Rule. It needs to be a perfect transcript matching what I say. Include filler words like um, like, and well. No latex, it needs to be plain text only. Do not edit or paraphrase anything. Please preserve the timestamps so that the transcript matches the original lecture.

You may want to also give it some notational conventions specific to your lecture, like calling your functions f and g rather than F and G.

Step 4: Copy the corrected transcript back into a .txt file

Once the transcript has been corrected (either by you or by AI), you need to put that text back into a plain .txt file so Panopto can read it.

  1. Select all of the corrected transcript text, including:
    • All of the timestamp lines (for example, 00:00:02.000 --> 00:00:06.500), and
    • All of the caption text under each timestamp.
  2. Copy the selected text.
  3. Open the original file in the downloads folder, you can simply paste the corrections there. (I recommend that you also make a copy to preserve the original.)
  4. Paste the corrected transcript into the file. Skim through it to check for strange symbols or AI hallucinations.
  5. Save the file, making sure the file type is Plain Text (.txt).

This saved .txt file is what you will upload back into Panopto in the next step.

Step 5: Replace or add the cleaned-up captions in Panopto

Panopto allows only one caption file per language for a given session.
You therefore have two options:

Option A: Delete and replace the original caption track

Before you do this, be sure you have a copy of the original saved.

  1. Return to the video’s Settings > Captions tab in Panopto.
  2. Delete the existing caption track for that language (for example, “English (United States)”).
  3. Upload your edited caption file as a new caption track for that same language. If you edited your video in Panopto (clipped and trimmed, for example), then you will want to check "Captions are based on the edited session." (In this case, I strongly recommend you proceed with Option B instead.)

This option is straightforward if you are confident the edited captions are correct.

Option B (recommended): Upload the edited file as a different language variant first to check it

If your Panopto instance allows multiple language variants (for example, US English and UK English), you can safely test before overwriting the original.

  1. Keep the original English (United States) captions for now.
  2. In the Captions tab, upload your edited file as a different language/variant, such as English (United Kingdom). If you edited your video in Panopto (clipped and trimmed, for example), then you will want to check "Captions are based on the edited session."
  3. Save, and then give it some time to process the new captions.
  4. Open the Panopto viewer for the video, turn on the new caption track (for example, “English (United Kingdom)”), and scroll through some moments to check for accuracy.
  5. Once you are satisfied, delete the original US English captions and re-upload the edited file.

Summary

  • For Panopto, always download the time-stamped caption file from Settings > Captions and edit that file.
  • When using ChatGPT or any editor, explicitly instruct it to leave timestamps and timing structure unchanged.
  • In Panopto, you can either replace the existing caption track or temporarily upload your edited file as another language variant to test it.