Audio and Video accessibility
Pre-recorded video or audio content published on or after 23 September 2020 must be accessible.
The 2018 Digital Accessibility Regulations require that, from 23 September 2020, all ‘time-based media’ (video and audio) must either provide a transcript or captioning or both (video only). Media published before that date is exempt.
Updated guidance from the Government Digital Service and Jisc has stated that:“Captions (and transcripts) that are accurate are compliant. Captions that have not been checked and that contain mistakes are not.”
How we will meet the regulations
The University has agreed that we will provide automatic transcripts, i.e. machine-generated transcripts, for all video recorded via services that Information Services provides, primarily Zoom, Echo 360 and Microsoft Stream. The quality of these transcripts is claimed to be between 75 - 90% accurate, although strong anecdotal evidence suggests it is less than that. Improvement of transcripts needs to be addressed locally.
To try to manage student expectations, information on the accuracy of auto transcripts has been provided to students.
Providing transcripts and or captions of recordings can have a positive impact on learning for all students by:
- Making recordings more understandable for students
- Enhancing the learning experience for students whose native language is not English
- Improving flexibility of learning - captions allow recordings to be viewed in an environment with no or poor-quality audio.
- Providing searchable recordings – some transcripts allow users to search for keywords and jump to the place in the video which will start playing where that word appears.
- A Transcript is the text of the spoken words and non-speech audio information in the recording
- Captions appear on the screen as the video is being played and are the text of what is said and a description of any sounds that are important to the understanding of the content. Captions can be open (always there) or closed (toggle on/off)
- An audio description is a form of narration used to provide information surrounding key visual elements in a media work, typically placed during natural pauses in the audio, and sometimes during dialogue if deemed necessary
Accurate (corrected) transcripts should be prioritised at School/College/Service level and recordings considered under 3 suggested categories (required, recommended, and requested).
Suggested categories of recordings:
- Required – Must have accurate edited captions and transcript available for one or more disabled students. Externally facing high visibility marketing and promotional material should also be in this category.
- Recommended – Some editing of the auto transcript, provision being ‘good enough'. Staff should spend no more than 15 minutes (as a minimum) per hour of recording correcting transcripts in this category.
- Requested – Automated transcripts. Students can request edited transcripts.
This model will need to evolve based on experience and feedback.
Before you start
- If the media is replicating or summarising existing text-based content (e.g., a script) and it contains no more information than the text then it can simply be labelled as such and does not require a transcript.
- Live video less than 14 days old is exempt from the legislation. 14 days is the maximum amount of time given, and if a recording of a live event is made available after the event, then a transcript and/or captioning must be provided as soon as possible.
- Transcripts and/or captioning means not only having written text of what is being said, but also a description of anything that is happening on the audio file, for example, music playing or wind blowing through the trees. In most cases, this may not be an issue, but in some instructional videos it will be.
- If a document is used in the video (PowerPoint, Word, or similar alternatives), then an accessible version of that document must also be provided.
Considerations for recording
- The impact accurate transcripts and or captions will have on student learning.
- The time and resources available to edit automated transcripts and include captions. Staff should spend no more than 15 minutes correcting transcripts unless there is a requirement for accurate transcripts.
- The delivery model of the course the recording belongs to.
- The likelihood of the recording being used in future years.
- The length of the recording.
Staff should spend no more than 15 minutes per hour of recorded video (as a minimum) correcting transcripts unless there is a requirement for a fully corrected transcript (see Prioritisation)
Choosing a recording tool
Decide which tool you will use to make your recording. The University supported systems that will provide auto-generated transcripts are:
- Echo360 – is available in many lecture theatres and can be used on a personal device. Auto-generated transcripts are generated for any recording that is published to a class
- Zoom – University of Glasgow licensed. Provides auto-generated transcripts generated and provided to students along with the recording
- Microsoft Teams – Recordings automatically are uploaded to Stream. Auto-generated transcripts are created alongside the video. Students access the video via office 365
There are many other tools that can be used to create recordings that will generate auto-transcripts and allow editing. Staff are free to use these tools to generate accessible resources though support will be more limited.
|Record video remotely||Record video in teaching spaces||Use on mobile||Edit video||Add to Moodle||Live classes and seminars online||Automated transcription|
|Microsoft Teams||Record live meetings||No||Yes||No||No||Yes||Yes (if recording)|
|Microsoft Stream||Yes (up to 15 minutes)||No||No||No||Yes||No||Yes|
The accuracy of auto-generated transcripts and therefore their usefulness will vary depending on several factors including audio quality, pace and volume of the presenter, and the number of subject-specific technical terms included in the presentation. Improving the quality of audio recording will improve auto-transcript generation.
- Use a good quality microphone to improve the accuracy of the auto-generated transcripts. (Link to further advice and loan equipment)
- Consider the positioning of the microphone e.g., is it close enough, is it rubbing against your clothes, etc.
- Speak at a steady pace. This will give the best results from auto transcription. Many students will appreciate this too.
- Consider background noise. Try to record somewhere where background noise is as low as possible.
- Practice! Anecdotal evidence strongly suggests that practicing using your preferred tools, or the Microsoft translate plugin for PowerPoint makes a significant difference to the quality of transcripts. Are there key words the auto transcription always gets wrong?
- Consider creating a script where practical. You would not need to provide a transcript for captions as the script would be the alternative format.
Break each session down into chunks more suited to an audience sitting at their computer, e.g., 10–15-minute segments. This makes working with transcripts easier and is advised pedagogically.
Provide a link where possible to a subject-specific online dictionary (e.g., using the University’s online reading lists via Moodle), as well as an additional glossary of any subject-specific/technical terms that are used in recordings.
3rd party audio/video content
- If the audio/video is not owned or paid for by yourself, or the University, then it is 3rd party content which you do not have control over and is exempt from the legislation around providing a transcript and/or captions but under the Equality Act you have to make reasonable adjustments to make sure that it is accessible.
- If you have created audio/video yourself then you own the copyright and you are responsible under both the 2018 digital accessibility regulations and the 2010 Equality Act to make sure that the video is accessible (see sections above on making accessible recordings)
- If there is a performance video (such as from the Box of Broadcasts) which the University subscribes to (i.e., pays for) then you can only alter the video for the service of a disabled person – you cannot distribute this altered accessible video to all your students, as this would breach performance copyright law