Media@UCSF: Machine Captioning

Video content is an integral component of education at UCSF. To improve accessibility for students, it is highly recommend that all instructional video content be captioned. As of September 2022, all new content uploaded and shared in the CLE using the Media@UCSF plug-in will be automatically captioned.

The automatic, machine captioning is surprisingly accurate, and free. All you need to do, is upload a video and wait a few minutes for the process to complete! This help article will introduce you to the machine captioning feature in Media@UCSF and address your frequently asked questions.

Captioned_video_playing_from_the_CLE.png

What is machine captioning?

Machine captioning is a method for adding captions to video content in which an artificial intelligence analyzes recorded audio and translates it to text, which can be displayed with the video. This is the most common form of captioning. "A.I." captioning and "machine" captioning are synonymous in this case. You can expect about 80-90% accuracy in normal speech... names and medical terminology not withstanding!

How long does it take?

That depends on the length and complexity of the video. Shorter videos with a single speaker are often done processing within a few minutes. Longer videos can take upwards of 30 minutes to process.

How do viewers see the captions?

The Media@UCSF video player has a standard CC (closed captioning) button in the bottom-right corner. Users can click to select English captions or toggle the captions off.

CC_feature_in_media_player.png

Can I edit the captions?

Yes you can! Follow these steps:

  1. From your My Media repository in the CLE, click to open the video.
  2. From the Actions menu, choose Edit.
  3. Click the Captions tab.
  4. Click the Edit Captions button.
  5. Click inside the text boxes provided to edit spelling and punctuation as needed.
  6. Click the Save button to apply your changes.

NOTE: If you accidentally change the timing or placement of captions, you may need to delete the caption file and re-order it. This also requires you to re-upload the source video file, however. If you need help with this process, please contact us.

Closed_captions_editor_in_the_CLE.png

Can I download the captions?

Yes, you can! Follow these steps:

  1. From your My Media repository in the CLE, click to open the video.
  2. From the Actions menu, choose Edit.
  3. Click the Captions tab.
  4. Click the Download button.

The save location will vary depending on your browser settings. The file will be in the standard .SRT format, which can be viewed and edited with any text editor, such as TextEdit (Mac), WordPad (PC), or Word. 

Captions_tab_-_download_button.png

For some videos, I prefer to use Otter. How do I choose which caption file is used?

It is possible to upload a caption file from another system, such as Otter. When more than one caption file is attached to a video, viewers will see multiple options in the CC menu.

After you have successfully added the additional .SRT caption file to the video, we recommend that you disable the other, machine language caption file to prevent confusion among viewers:

  1. From your My Media repository in the CLE, click to open the video.
  2. From the Actions menu, choose Edit.
  3. Click the Captions tab.
  4. Click the “Don’t Show on Player” button next the appropriate captions file.

Hide_machine_captions_button.png

Are other languages available?

No, not at this time. Only English is available with this free and automatic captioning service.

I have a transcript that was created outside of the CLE, can I use that for captions?

Yes, you can use the Alignment feature if you have a document that was created manually or by another system, and if it does not have timecode for aligning the captions to the video. Media@UCSF will attempt to do that for you.

  1. From your My Media repository in the CLE, click to open the video.
  2. From the Actions menu, choose Caption and Enrich.
  3. Under the “Order Captions & Enrichment Services” group, Select Alignment from the Features drop-down menu.
  4. Click the “Upload txt” button to locate your file.
  5. Click the Submit button.
  6. If necessary, use the edit captions feature to fine-tune the captions.

NOTE: The uploaded transcript file must have the .TXT file extension. Transcript files are often basic text or word processor files. Captions files, often in .SRT format, contain the text along with timecode, so the video player can display the text at the right moment in time. 

Alignment_option_within_Caption_and_Enrich_group.png

What about all of the videos I had uploaded before Sept 2022?

All of that content can be easily captioned as well, but it will not happen automatically. The owner of the video(s) can “order” captions from their My Media repository. It just takes a few clicks, and multiple videos can be ordered in a batch:

  1. Open your My Media repository.
  2. Scroll down, and use the keyword search or filters to identify any number of videos.
  3. Click the selection box next to each video, or click the box at the very top to select all videos that are visible.
  4. From the Actions menu, select Caption & Enrich.
  5. Accept the default options, and click the Submit button.

The captions will begin processing for all videos selected, and there is nothing more than you need to do! When the processing is complete, viewers will immediately have access to the captions.

NOTE: If you select a batch of videos and any of them have already been processed for machine captioning, you will get an error message, "Service was already requested for this video." In this situation, you must re-select videos and omit those which already have machine captions enabled. If you need assistance with this process, please contact us.

Bulk_Caption_and_Enrich.png

Have more questions? Contact us