Captioning

Posted April 22nd, 2020 at 1:27 pm.

Captioning involves converting the audio content of an event (e.g., lecture, webinar or performance) to text that is displayed and synchronized with visual content. For a live, in-person event, captions are often displayed on a large screen near the podium or stage. For an online event, captions are displayed at the bottom of the window (closed captions) or in a separate window.  Captioning is different from transcription, which involves producing a textual record of audio content for access after the event.

Captions can be produced in one of two ways:

  1. Human-generated captions: Professionally trained human listeners type captions in real time. These captions are also called CART (Communication Access Real-time Translation) and can approach 99% accuracy as long as the captioner can hear speakers clearly. CART services must be scheduled in advance; current lead times are 2-3 weeks.
  2. Automatic speech recognition (ASR): Speech-to-text computer algorithms generate the captions. ASR captions can be produced quickly and inexpensively, but they often require significant human correction/editing.


 

Captioning Live Online Events

College policy requires you state that reasonable accommodations are available on request in all advertising for events.

  • If a participant has requested captioning OR you know that disabled individuals who require this accommodation are participating, then you should arrange human-generated captioning (CART) for the event.
  • If no one has requested captions, then human-generated captions are not necessary. You can add ASR captioning if you would like.

ASR captions can help viewers without specific disabilities compensate for background noise, assist viewers with comprehension, and enable viewing without the sound. They do not reach the level of accuracy and simultaneity needed for disability accommodations, but they are better than nothing and can be used as a last resort. (Note: If you are recording an ASR-captioned event, you will need to edit the captions for accuracy before posting to a college website.)  

Captioning options for College-provided platforms:

  • Currently, Zoom only supports human-generated captioning. It does not have a built-in ASR captioning, but the college is trialing third-party ASR captioning services that might fill this need.
  • Microsoft Teams currently only supports ASR captioning.

 

Human Captioning for Zoom Events

Providing real-time, human-generated closed-captioning for Zoom Meetings and Webinars hosts is similar to the process of providing CART captions for in-person events. The captioner joins the Zoom event as a participant (Meeting) or panelist (Webinar), and either type captions directly into Zoom or types them into a special captioning platform which feeds them into the Zoom using an “API token” that is unique to that meeting.

Step 1. Schedule a Captioning Service

CART services need to be scheduled well in advance of an event — most vendors now require 2-3 weeks lead time due to increased demand.

  • Access Services will help instructors schedule human captioning for academic Zoom meetings for students with documented disabilities.
  • Other hosts should contact CART serves directly:
Step 2. Exchange Information Before the Event

The captioning service will typically need to know:

  • Meeting date, time, and duration
  • Title of the event
  • The Zoom meeting link and any passcodes the captioner will need to join as a participant (Meeting) or panelist (Webinar).
  • The Zoom host’s name
  • A contact number or email address in case of last-minute technical or logistical difficulties
  • Information that could help improve captioning accuracy, such as a list of proper names or technical vocabulary that might be referenced during the event

The Zoom host should make sure they know:

  • The captioner’s name (so they can be sure to admit that person!)
  • A back-up phone number or email address in case of last minute technical or logistical difficulties
  • Whether the captioner needs to be assigned the captioner role during the meeting so they can type directly into Zoom or whether the host needs to copy an API token for them
  • If the service needs the host to copy the API token — which can only be done after the meeting is started — how should the host share it with the captioning service (e.g. paste it into the chat window, email to specific contact, paste into a web form)
3. Enable Captioning During the Event

(See Zoom’s support article for full details and a tutorial video.)

Typically the host starts and captioner joins the meeting few minutes early to get set up. Once the captioner has joined, the host clicks on the Closed Captioning button in the meeting controls bar. (Co-hosts will not see this button. If you are the host and do not see it, click the … to display hidden options.)

  • If the captioner will be typing directly into Zoom, click Assign participant to type, then choose the captioner from the Participant list.
  • If the captioning service needs an API, click Copy the API token, then share the API token with the captioning service using your pre-agreed method (e.g., paste in meeting chat, send in email, paste into web form). When the caption feed begins a Subtitles available notice will pop up; click on it and choose Show subtitles to display them for all.

 

ASR Captioning for Microsoft Teams Meetings

  • Any Bryn Mawr participant can turn on ASR captioning for a Teams meeting hosted through Bryn Mawr’s Teams license.
  • However, non-Bryn Mawr participants, including Haverford participants, will not be able to view the ASR captions.

 

Captioning Recorded Video

Captions are not only essential for viewers with hearing impairments, but they can also help all viewers compensate for background noise or poor audio quality, assist with comprehension, and/or enable viewing without sound.

  • ASR-generated captions are better than nothing, but they typically do not meet the 99% level of accuracy required for ADA compliance without human editing.
  • If you are captioning a video as disability accommodation OR will be using the video in a course or on a College website, you must either edit ASR-generated captions to ensure they are accurate or request human generated captions.
  • For questions about captioning of Tri-Co library resources (DVDs or streaming video), please email library@brynmawr.edu.

 

Captioning in Panopto

Panopto is the College’s primary platform for lecture capture and hosting streaming video libraries.

  • Panopto automatically adds ASR captions to videos that you create in or upload to Panopto, if they do not already have captions. (ASR captions do NOT overwrite existing captions.)
  • Creators can:

Access Services will work with instructors and LITS to set up third-party captioning for course-related videos to meet the needs of students with documented disabilities.

If you would like to have a third-party service caption Panopto videos for other reasons, please email help@brynmawr.edu for more information.


Recordings Made in Zoom

If you did not arrange for the Zoom event to be live-captioned, you can upload it to Panopto and produce captions there before sharing the recording.

If you did arrange for the Zoom event to be live-captioned, then Zoom will save captions as a caption (.srt) file. If you chose to Record to the cloud, that caption file will be uploaded to Panopto when the recording is uploaded there (you can also edit them as usual). Viewers will be able to view the captions during playback on both platforms.

 

Filed under: Accessibility,Panopto,Teams,Zoom Tags: by Ellen Farr

Comments are closed.