Kaltura MediaSpace: Captioning a media file

Captioning with Kaltura

When media is uploaded to Kaltura, it is automatically captioned using automatic speech recognition software. These machine generated captions help provide a starting point for creating accessible content, however machine captions are not reliable enough to provide equitable access and should be reviewed and edited to ensure accurate representation of audio content in any media. By default the machine generated captions assume that the broadcast language is English, but video owners can request automated captions for any of the 15 other available source language. This is not a translation service, but will improve the direct transcription of the content. Translation services can be used after content is fully corrected for more useful subtitles if desired.

Accessing the Caption Editor

The caption editor can be accessed by editing the media through MediaSpace.

1. Navigate to MediaSpace.
2. Find the specific media and click the Edit icon, visually indicated as a pencil icon.
   Note: if you are using the Action menu select Edit from the drop down.
   
3. In the media editor, select the Captions tab from the navigation options.
   
4. Select the Edit Captions button to launch the Caption Editor.
   

Caption Editor Features

The caption editor allows access to all caption files associated with the media. Copies of media often have multiple versions and Zoom recordings may have associated caption files created from other services. These files can be accessed from the Captions dropdown menu if available. The editor has a fully functional search and replace option for quick navigation or bulk replacements, there is the option to add speaker tags to individual or multiple cues at once. The changes can be saved or reverted as desired, and the captions can be set to auto scroll while previewing the media to validate timing and accuracy of the current caption file.

Updating Caption Details

1. Select the desired caption file in the Captions drop down.
2. Select Edit details to set correct language, accuracy percentage, and label that will appear in the caption selector.
   
   • The default Language is set to English, this can be changed from the drop down if needed.
   • The caption fidelity of the captions can be indicated with a slider that adjusts the accuracy percentage, 100 percent is the goal, the default percentage is set from an algorithm but should be adjusted when completing new edits.
   • The label needs to be adjusted because Auto Generated is for non-edited caption files before review or manual manipulation. Delete the Auto Generated text from the label after editing the captions and select Save.
     
     

Having accurate labels can ensure that the proper file is available for viewers.

The media editor displays all caption files in the captions tab and indicates whether the caption file is displayed on the player and if it is set as the default caption file.

The checkmark in the action menu sets the default caption file when enabled on the player by a viewer, the pencil icon for the specific caption file allows for quick access to the language, accuracy, and label, the x icon deletes the file, and the download icon allows for the file to be saved locally. The final icon in the Actions column is a toggle button that allows the caption to be hidden or displayed as an option for viewers. The icon with a line through it indicates that the file is hidden if the file is visible, the player icon does not have the diagonal line across the screen .

Search and Replace Caption Content

The search option allows individuals to find specific words or phrases within the caption file. If you use the replace with option, any instance that matches the search criteria will be replaced with the new content desired.

To use the Replace with feature:

1. Type a word into the Search in Captions field to find the instances of that search within the caption file.
2. Enter a replacement work or phrase into the Replace with field.
3. Select Replace.
   

Add Speaker Tags

Speaker tags are required to help identify multiple speaker’s specific content.

To use the Add Speaker feature:

1. Select the checkbox(es) for an individual multiple caption cues that were spoken by a specific individual.
2. Type the speaker’s name in the Add Speaker field.
3. Select Add.
   

Editing Captions

When captioning content, it is important to follow captioning guidelines to help comply with ADA requirements for fully accessible content.

Auto-generated captions need to be reviewed and usually adjusted for full ADA compliance. When editing captions, it is important to follow the guidelines and best practices for captioning educational videos.  The Described and Captioned Media Program provides a comprehensive Caption Key that should be used. 

Quick captioning tips 

• Ensure Accurate Representation
  • Errorless captions of audible information is the goal 
• Use a Consistent Style
  • Caption timestamp blocks should be 1-6 seconds in duration 
  • Consist of 32 characters or less per line 
  • Only have 1-2 lines of text per screen 
  • Use proper case, punctuation, and appropriate line and screen breaks 
• Provide Clear Representation  
  • Include speaker identification, non-speech information, and important sound effects for clarity 
• Check Readability 
  • Synchronize the timestamps with the audio, captions cannot obscure or be obscured by other visual content 
• Verify Equality
  • Preserve the meaning and intention completely to provide an equitable experience  

For more information on understanding captions visit the Web Accessibility Initiative site from the World Wide Web Consortium.

Steps to Edit Captions

1. Locate the block of text containing content that should be edited.
2. Play the media clip to check for accuracy and timing. The captions are highlighted in light blue as they appear in the media.
3. Edit the block of text for the caption cue, total characters per block are displayed for reference in real-time.
   
4. Separate the caption blocks as needed following the caption guidelines so that modifiers are not split and so that each caption cue is between 1-6 seconds in duration with 64 characters or less of content on no more than 2 separate lines of text.
5. To add new caption cues, select the add caption button that can be activated for the specific location between existing caption cue blocks.
6. Select Save, a confirmation window will provide an option to select Yes to save or No to return to editing.
7. Review the captions by viewing the video with the captions enabled.