This FAQ addresses common questions about captions for Kaltura entries.
Sometimes the terms "captions" and "transcripts" are used interchangeably, but there are meaningful differences, particularly within the context of media management. While both contain a textual representation of words spoken during a video, here's where they diverge:
Kaltura generates its transcripts based on the captions. So if you edit the captions, the changes will be reflected in the transcript. As of August 2, 2024, the transcript can be found within the player by clicking its icon when viewing a video in My Media, a Media Gallery, or a MediaSpace channel. Most embedded videos have the transcript widget as well. The nice thing about Kaltura's transcript is that it will highlight the text that's currently being spoken, and you can click a line to jump to that part in the video.
Kaltura allows you to edit your captions by either using their web-based caption editor, or by downloading the captions and editing them in a text editor on your computer. You can also download the transcript, edit it locally, and then use Kaltura's "machine alignment" feature to create a caption file based on the edited transcript.
Below are links to relevant documentation:
You aren't automatically able to request human-made captions with Kaltura, but you can work with a vendor on your own.
Because you can upload captions yourself, you can use captioning services that you employ on your own, such as Rev.com, Cielo24, or 3Play. Services like these usually cost between $1 and $2 per minute of media. You may want to talk to your department about this cost if you think it necessary.
Typically the way captioning services work is as follows:
Do note that we can work with departments if you want to try and develop a workflow to request human captions and bill your department accordingly. Reach out to kaltura@ucsd.edu if you want to learn more.
Machine captions are sufficient, but consider taking the time to edit them for accuracy.
All videos added to "My Media" after 13 June 2020 will automatically have machine captions requested and added to the video. However, machine captions are only about 70-80% accurate for colloquial speech. With potentially 1 out of every 3 words being inaccurate, we strongly encourage everyone to consider editing their captions for accuracy.
There's no doubt that editing captions is time-consuming, particularly for longer videos. With that in mind, we encourage you to collaborate with TAs and student workers for assistance. Note that you can add another user as a collaborator on a video (specifically a "co-editor") and they'll be able to edit your video's captions. You can learn more about adding collaborators in our tutorial on sharing videos.
If a student with a disability requires captioning as an accommodation, captions will be provided in real time by a certified captionist or you may be asked to provide a complete and accurate captioned transcription of the video. Please contact the Office for Students with Disabilities at osd@ucsd.edu if you have further questions about disability accommodations.
It varies, but could be up to 2 times the length of the video.
As you might imagine, the time it takes for your machine captions to be created depends on the length of your media. The formula that dictates how long it will take machine captions to be generated is roughly as follows (according to Kaltura):
Do note that this refers to Kaltura's captions (which are themselves generated by a subcontractor called Verbit). If your media entry is the result of our Kaltura-Zoom integration, Kaltura is pulling Zoom's captions over and NOT requesting captions from Verbit. There is no formal stated turnaround time for Zoom captions to make their way to Kaltura.
Though it's ultimately a matter of preference, we find it easier to edit captions on your desktop.
Kaltura's web-based editor has a few shortcomings, particularly when it comes to "find and replace" functionality. It's also arguably a little easier to navigate a pure text document than the text within individual caption boxes in the web editor. Admittedly, there are a few additional steps involved in editing locally, insofar as you have to download and upload.
The other thing that's worth mentioning is that there's a possibility your Canvas/Kaltura credentials will expire while you're editing using the web editor. It's why we strongly recommend you frequently click "save" if you're working on your captions in the web-based editor, but it's also why we lean towards recommending downloading them and just editing them in a text editor.
The approach that we would strongly recommend is to edit the transcript instead of the captions. (In an FAQ above we discuss the difference between the two.) Since the transcript is pure text without any time codes, it's arguably easier to make edits than in a caption file where you're constantly having to maneuver around the time codes in order to make changes to the text.
Yes. You just need to make them a "co-editor" on the video.
Follow our instructions on adding collaborators to a video to add a user as a "co-editor" on your video. The video will then show up that user's "My Media," after which time they can follow the instructions on this page to edit your captions. Note that that user will also be able to perform a variety of other edits to your video, such as renaming it, changing the description, changing the thumbnail, etc. They'll also be able to download your video.
If you click "save" in the Kaltura video editor, your old captions will be cut at the timecodes you identify. If you click "save a copy" after editing, your "old" captions will be present in the new entry (cut appropriately based on your edits) AND a new set of machine captions will be requested.
If you make cuts to a video, the captions will be cut at the timecodes you identify. BUT, every new entry created in Kaltura automatically has machine captions requested. This means: