From Sunday 27 September 2020, videos uploaded to Kaltura MediaSpace, via the web interface or from within Moodle, will have computer-generated captions added automatically.
These automated captions are about 70-85% accurate (Kaltura, 2018). The accuracy of the machine-generated captions will depend on variables such as:
After Kaltura generates the captions, you should use the in-built caption editor to edit and correct them before you publish the video.
This caption file allows captions to be displayed on your video, and it also produces a transcript from the text for viewers to download.
Pending requests
Kaltura Reach takes around 30 minutes to generate captions, depending on the size of your original clip.
Remember to make your captions visible when you have finished editing them.
Kaltura Reach captions are machine generated. They are less accurate if the clip contains specialist vocabulary, strong accents or if the sound quality is poor.
A new screen will load with the Captions Editor.
You can watch the the clip with the captions, and press pause each time you want to edit a caption.
Alternatively, you can scroll through the captions without watching the clip. If you need to check the video soundtrack, select the caption. The video will jump to that point and then select play.
Select the text of any caption to start editing.
Use the Kaltura's Replace function to correct any repeated errors.
You can add the name of the speaker to selected captions.
From the top right of the caption editor page:
Once you have completed editing your caption file and saved it, return to your video on MediaSpace.
If you have more than one caption file attached to a clip, you can choose which file will play as default by selecting the tick alongside the relevant captioning file
You can change the name of the file as it appears when users select the CC button on the video player.
To edit the name of the caption file, select the edit (pencil) icon.
To delete the caption file, select the delete (x) icon.
You can download your caption file. Select the download icon under the Captions tab. You can open the file as a text file in Notepad. When viewed as a TXT file, you will see the text of the captions and the timecodes.
You can save the file as an SRT file. The SRT format is a standard caption format that can be uploaded to other platforms.
Select the screen icon to make the captions unavailable without deleting the file. When viewed in the player, the captions will be unavailable.
You can upload a captions file to your video by either downloading a file from another video or creating one yourself.
You may wish to do this if you have previously-corrected your captions on an older video, but have had to make a small edit to it (without changing the timecodes) and subsequently uploaded a new copy of the video.
We recommend saving your caption files as either .SRT or .VTT files.
If you have downloaded your caption file from another MediaSpace video, it will be a .SRT file by default.
Ensure you hide or delete any other captions files on the video to avoid confusion for your viewers.
You can request caption correction for MediaSpace videos on the IT Self Service Portal.
You will need to provide the following information to help the team process your request.
Automated speech recognition can be inaccurate. Captions or transcripts of your recordings may contain words incorrectly transcribed as profanities e.g. swear words or inappropriate words. Alternatively, you might find that the captions or transcripts on your recordings are censoring words with a row of asterisks which means that students relying on captions are missing key information.
LEaD are gathering information on the impact of profanities and censorship in recordings. If you find any examples, please raise a ticket with us via the LEaD General Request Form on IT Self Service Portal.
Please include as much information as possible, including system used - Lecture Capture, Teams, Zoom, links to the recordings on Moodle and timecodes.