Public Sector Body Accessibility Regulations (PSBAR) require that audio and video recordings made available from 23 September 2020 must meet accessibility standards - one way to achieve this is to add captions to pre-recorded content.
All new video and audio recordings in Re:View (Panopto) are automatically captioned by the system using Automatic Speech Recognition (ASR). However, the accuracy can vary depending on a number of factors. A few changes to your audio set up could make your captions more accurate and save time when editing the recordings afterwards.
We spoke to Sacha Goodwin, AV Support Manager at the University, and asked for his top tips about how to ensure ASR captions accurately reflect what is being spoken.
Use a good-quality microphone
Use either an external microphone or headset plugged into your device as opposed to your machine’s built-in microphone. This enables cleaner audio of your voice and rejects background noise – your voice recording will be clearer, and background noise will be reduced, thus helping the transcription. If the session is interactive, use headphones rather than speakers in addition to a good-quality microphone. Apple users may consider using EarPods or AirPods which contain a built-in microphone.
Use the Panopto Desktop Recorder or Panopto Capture for pre-recorded content
Some staff use Zoom meetings to pre-record video content because of their familiarity with the platform. Instead, use the Panopto Desktop Recorder or Panopto Capture to pre-record content. Here you’ll need to adjust your audio gain settings before you record as it’s not automatic – aim to be at around the top of the green section but not hitting the red.
When presenting, speak clearly and at a pace that is neither too fast nor too slow. Concentrate on correct pronunciation of technical vocabulary, as these words are often transcribed incorrectly by ASR.
Microsoft PowerPoint online has a built-in Presenter Coach that allows you to practice your presentation. The built-in Artificial Intelligence will give you tips about the pace of your presentation, tone of voice and any ‘ums’ and ‘ahs’ you may be repeating.
Adjust your Zoom audio settings
Within Zoom you can adjust its audio settings to improve the recording quality:
- Test your microphone and adjust your volume level.
- Suppress background noise - once you have a quiet environment, turn off the ‘Auto’ setting. Set this to Low if you have a good microphone.
- Music and professional – enable this setting if you have a good-quality microphone and are wearing headphones (not using your PC’s in-built speakers). This can improve the quality of audio and intelligibility of your transcriptions.
- Although there is currently no legal requirement to provide live captions during webinars, these can be enabled by the host once a Zoom meeting has been started. Do a test recording.
- In a test recording, practice any technical vocabulary that may not be picked up clearly by ASR. When you watch and listen back to the test recording, is the audio clear? And how accurate are the captions that have been generated?
Look out for upcoming functionality in Re:View (Panopto)
Panopto have announced upcoming functionality to their platform:
- Improved ASR accuracy and dictionary updates so that as many words as possible are recognised.
- Faster, easier caption clean-up. Low confidence words will be highlighted, and a ‘find-and-replace' feature will be introduced, as in Microsoft Word.
We will update you when these features are rolled out more widely.
SubZero Headset Mic – these are great if you are recording only as they are visually unobtrusive and are good at rejecting background noise. You can combine them with earbuds or normal headphones if you are recording an interactive session.
Contact the University of Bath IT shop for equipment needs.