Captions for audio or video media make your media more accessible. Machine captions are generated by computer-based speech recognition and can be created for media in MediaSpace. They are approximately 75% accurate depending on content complexity, and should only be used when you review and edit them to increase their accuracy. Inaccurate captions can cause disruption and confusion for the user. Allow 3.5 times the duration of the media to review, edit, and improve their accuracy. Since editing to improve accuracy takes time we recommend you use this tool with shorter recordings.
If you are teaching a student who would benefit from captions, the McBurney Disability Resource Center
offers captioning services for students with disabilities. If you have longer or more complex content, request captions reviewed for accuracy by professional caption editors through our recommended vendors
. To increase the initial accuracy of machine captions we recommend that you:
- Use a dedicated microphone (not your built-in laptop microphone) when recording your screencast or webcam if possible.
- Avoid a noisy recording environment. For example if you have people talking in the background, or a fan, or air conditioning on, it will add noise to the background your recording which will make auto-detection of what you are saying more difficult.
- Launch Chrome or Firefox, go to Kaltura MediaSpace, and login: mediaspace.wisc.edu. If you need more information on how to log in to MediaSpace please see Kaltura - How to Log in to Kaltura MediaSpace (UW-Madison) .
- Click on your name or the person silhouette icon in the upper right corner and click My Media:
- Search for or scroll until you find the media you want to add captions to.
- Click on the thumbnail of the media item you want to add captions to.
- Click the Actions drop-down menu and select Caption & Enrich:
- The Order Captions & Enrichment Services options will display underneath your media item. If you have previously requested machine captions for the media item, it will display here and you can edit them as described in Kaltura - How to edit captions in Kaltura MediaSpace . If you would like to order machine captions and none have been previously ordered, select your options. There are drop-down menu options for
- Service - select Machine
- Source Media Language - select the appropriate language. Note that while Arabic, Dutch, English, French, German, Hebrew, Hindi, Italian, Japanese, Korean, Mandarin Chinese, Portugese, Russian, Spanish, Swedish, and Turkish are all available we have not had the opportunity to test the machine captioning accuracy of languages other than English.
- Feature - select Captions
- When you're done selecting your options click the Submit button.
At this time you can only request machine captions for a video once so only do this after you are sure you are done making any changes to your video.
- The Existing Requests screen will update to show your machine captions pending request order. You should also receive an email notification that indicates a new caption file was added when the machine captions are done processing.
It usually takes a minimum of 5-10 minutes for machine captions to be processed even for short videos. For longer videos it generally takes about twice the length of the video. Please be aware that during times of heavy usage (like in 2020 during the COVID-19 situation) processing times can lengthen to hours. Keep an eye out for the email notification indicating your machine captions are complete.
- If you refresh the window you will need to re-click on Actions and select Caption & Enrich to check when the machine captions are completed. If they are still in process it displays "Processing." Once they are done you'll see the Existing requests will indicate the captions are completed:
- Captions will not be enabled until you display them. See the instructions below to display captions.