Captions for Videos and Live Events

Focus Area Five
Captions for Live Broadcast (Part One)

Image: Line of cameras at an event


In this section, we will become familiar with captions for live broadcast television including news programming, sporting events, weather forecasts, and live events. Accurate and high quality captioning of live broadcast is important for people who rely on captioning to understand the content.

Developing an Understanding

Live captioning is the conversion of speech to text in real time. Live captions are generated by a person who may be in the same room, in another city, another province or even in another part of the world. As the live captioner listens to the audio of the broadcast, they type the words they hear using a steno or chorded keyboard. Some live captioners wear special headsets and microphones to listen to the audio and re-speak those words into a speech recognition software that is specifically trained to their voice (Ai-Media, 2017). Specialized software known as an encoder adds captions to the broadcast television signal transforming the spoken audio of a broadcast into words on the screen. Unlike video or offline captions, live captions are written in all capital letters as there is no time to shift between uppercase to lowercase letters.

A live captioner has to be fast and accurate. Since they cannot transcribe words until they are spoken, live captions usually lag a few seconds behind the audio. And, since the live captions are being created as the program is being broadcast, there is no time to proofread them for mistakes. This means that there can be spelling and grammar mistakes, especially with names and locations that are hard to spell or when the action is moving very quickly. Live captioners are often prepared with a list of names and difficult words prior to the broadcast, although this is not always possible (National Captioning Institute, n.d.).

Deepening Your Understanding:

The Canadian Radio-television and Telecommunications Commission (CRTC) has determined that people who rely on closed captioning for their live programming should have access to the highest possible quality captions. All Canadian broadcasters are required to follow established quality standards for English language live programming (CRTC, 2012). These standards were established in July 2012 and were amended in November 2016. These are the current standards:

Lag time: The time between when the viewer hears the audio and views the live captions on the screen must not be longer that 6 seconds (CRTC, 2016).

Speed of the captions: The captions must be word-for-word, regardless of the age of the target audience and only edited when technical limitations or time and space restrictions will not allow for all of the spoken words at the appropriate presentation rate. This rate is defined as 120-130 words per minute for children’s programming (CRTC, 2016).

Onscreen information: For all captions, both live and pre-recorded, the caption must be in a place that avoids covering up the action, visual elements, graphics or other words needed to understand the message. While captions are usually at the bottom of the screen, they may be placed elsewhere. In the image on the right, you can see that the live caption is placed at the top of the screen so it doesn’t hide the action (CRTC, 2016).

Screenshot of a hockey game with caption at the top of the screen

Image: Screenshot of a hockey game with caption at the top of the screen.

If it is impossible to present the caption without covering the graphics on the screen, the caption must take precedence as seen in the right hand image below (CRTC, 2016).

Screenshot of a baseball game with caption at the bottom of the screen.

Image: Screenshot of a baseball game with caption at the bottom of the screen.

Accuracy rate: The accuracy rate for pre-recorded programs is 100% and currently the accuracy rate for live programming is 95%. Regulations in this area are in the process of being re-evaluated based on industry and consumer input. This accuracy rate is monitored by the CRTC and every month all Canadian broadcasters must calculate the accuracy rate for two live programs and submit a report. Every two years, broadcasters must submit a full report that describes the efforts they have made in order to improve the accuracy rates for in-house and out-sourced captioning (CRTC, 2016).

In 2016, the CRTC called for comments on the 2016 Working Group’s proposal opens new window to amend and update the English-language closed captioning quality standards relating to the accuracy rate for live programming. Specifically, the Commission sought comments on the following:

  • Approving and adopting the Canadian NER Evaluation Guidelines that were included with the 2016 Working Group’s final submission;
  • Issuing amended English-language closed captioning mandatory quality standards for live programming, which would replace those set out in Broadcasting Regulatory Policy 2016-435 opens new window, reading as follows:

Accuracy rate for live programming, beginning September 1, 2019: Broadcasters must reach an accuracy level of 98% based on the “NER model”. The NER model is based on the Number of words; Edition errors (where a captioner chooses to paraphrase rather than use verbatim speech, and Recognition errors, where the captioner or the software delivers a word that is wrong, misspelled or garbled (CRTC, 2019).

Rebroadcast: If a program was initially aired live and is being rebroadcast or shown online, any errors in captioning must be corrected (CRTC, 2016).