Re-Thinking Dialogue Intelligibility: Why High SNR Still Fails Viewers

Dialogue intelligibility remains the top complaint in broadcast audio — and it’s getting worse. A recent Xperi survey reports that 84% of consumers have difficulty understanding dialogue in TV shows and films. Surprisingly, this issue extends well beyond those with hearing impairments and is affecting general audiences at scale.

So what's going wrong?

This talk dives into the disconnect between traditional intelligibility metrics and actual viewer comprehension. Despite high speech-to-background ratios and near-perfect word recognition scores, comprehension often breaks down — pointing to the limitations of current assessment models.

Researchers at the University of Salford’s Acoustics Research Centre, funded by Xperi, are investigating this phenomenon using biometric and comprehension-based testing. The goal? To develop next-generation metrics that go beyond SNR and word recognition to capture the real-world effectiveness of dialogue enhancement in broadcast contexts.

We’ll discuss:
·       Why intelligibility ≠ comprehension in today’s media environments

·       The shortcomings of traditional intelligibility measures

·       How biometric and cognitive data are informing new approaches

·       Implications for designing and evaluating speech enhancement algorithms

If you're working on audio post, codec design, loudness standards, or broadcast tech, this session will give you fresh insights — and a new perspective — for solving one of the industry’s most persistent problems.