Note: The creation of this article on testing Audio Control was human-based, with the assistance on artificial intelligence.
Explanation of the success criteria
WCAG 1.4.2 Audio Control is a Level A conformance level Success Criterion. It ensures that when audio plays automatically for more than 3 seconds on a web page, users must be able to either pause or stop it, or adjust its volume separately from the system’s overall volume.
There was a period time in Internet history where website frequently used background audio, typically music, to set a mood while visiting the site. Thankfully, in my opinion, this implementation has dwindled in usage.
Who does this benefit?
- People who use screen readers can listen without interference from other sounds. This is especially helpful for users who are hard of hearing or whose screen readers rely on system volume, making it impossible to lower the audio without also lowering the screen reader.
- It also helps individuals who have trouble concentrating on visual content, such as text, when background audio is playing.
Testing via Automated testing
Automated tools can quickly scan for <audio> and <video> elements with autoplay attributes, flagging potential issues across many pages. They’re great at spotting code patterns, like missing controls or known autoplay configurations.
But they fall short on key points: they can’t tell if audio actually plays, how long it plays, or if it starts automatically via scripts or delayed events. They also can’t judge if audio interferes with screen readers or user focus. Browser behavior, JavaScript-triggered audio, and other nuances often slip past automation, making manual review essential.
Testing via Artificial Intelligence (AI)
AI can rapidly scan many pages or apps to spot autoplay audio, especially when it’s embedded in standard code like <audio autoplay>. It also detects complex triggers, such as JavaScript-injected media or autoplay inside modals, that simpler tools might miss. AI-based tools can continuously monitor for regressions, helping maintain compliance in fast-changing environments. Advanced AI can even flag media lacking volume, pause, or stop controls.
However, AI often can’t reliably tell if audio plays longer than 3 seconds without simulating user interaction. Audio triggered by user actions, complex scripts, or third-party content may go unnoticed without human review. AI also can’t judge how audio affects user experience or screen readers and may miss custom or obfuscated autoplay implementations unless specially trained.
Testing via Manual testing
Manual testing shows how real users experience auto-playing audio and whether controls are easy to find and use, something automated tools often miss. It verifies if audio plays longer than 3 seconds and checks that pause, stop, or volume controls work correctly, are properly labeled, and accessible via keyboard or assistive tech. It also ensures consistent behavior across browsers and devices.
However, manual testing can be time-consuming, especially for large sites or frequent updates. Judging if controls are truly discoverable or independent of system volume can be subjective without clear guidelines. It’s not ideal for quickly scanning many pages, since some auto-playing audio may be hidden or triggered only by interaction, requiring thorough, careful testing.
Which approach is best?
No single approach for testing 1.4.2 Audio Control is perfect. However, using the strengths of each approach in combination can have a positive effect.
The best approach to testing WCAG 1.4.2 Audio Control combines automated, AI-based, and manual methods: automated tools quickly flag pages with audio elements, while AI helps identify complex patterns like delayed or interaction-triggered audio. However, manual testing is essential to confirm whether audio plays automatically for over 3 seconds and to verify that pause, stop, or independent volume controls are present, accessible, and easy to use across devices and assistive technologies, ensuring a thorough, user-centered evaluation that no single method can achieve alone.