In this post, Mario shares an update on the work he carried out as part of the EAD project. His research focuses on how changes in reverberation can be better understood and used to support accessible film sound for visually impaired audiences.
As part of the project’s mission to move beyond traditional Audio Description, the study we performed investigated how reverberation, which is connected to the acoustic characteristics of a space and how sounds are modified when emitted in a certain environment,, can function as a cue to communicate space, depth, or changes in scene through sound alone. For example, a shift from a dry, close-sounding voice (that is, a voice that isn’t strongly impacted by the acoustics of a space) to one with more reverb might signal a character moving from a small room into a large hall. To apply this effectively in sound design, it is essential to determine how much a reverberation setting (such as the length of the reverb or the balance between the direct voice and its reflections) needs to change before listeners can reliably perceive a difference.
This study addressed that question by measuring the Just Noticeable Difference (JND) in reverberation time. JND refers to the smallest detectable difference, whereas Reverberation Time (also sometimes referred to as RT60) is the time it takes for sound to decay by 60dB after a sound source has stopped. The measurements were done under varying direct-to-reverberant ratio (DRR) conditions, meaning how much of the original voice we hear compared to its reflections off surrounding surfaces. The goal was to define perceptual thresholds that sound designers can use to guide their decisions when applying reverb for spatial clarity and immersion in accessible audio.
To conduct the test, we selected speech samples and processed them to change the reverberation settings. Each sentence was rendered at three DRR levels (low, medium, high) and across a range of reverberation times, from 0.2 to 1 second.
The stimuli were presented in a listening test where participants evaluated whether changes in reverberation time were perceptible. In this very early stage of the work, we worked with sighted participants to establish a solid working protocol for the listening test. Our next phase will be a study with visually impaired participants to determine whether there are any significant differences experienced.
By identifying such thresholds, the study is an initial step towards making more precise decisions when mixing EAD soundtracks, as well as having the potential to impact audio-only media and audio-visual media productions more generally. The results support practitioners in making reverb choices that are not only aesthetically appropriate but also perceptually effective.
These findings were presented at Tonmeistertagung 2025 in Germany this past November (2025). The results showed that listeners needed reverberation time changes of approximately 0.26 to 0.33 seconds to notice a difference, depending on how prominent the reverb was in the mix. When the reverb was more subtle (a drier, more direct sound), much larger changes were needed for listeners to perceive them. This means that for sound designers small reverb adjustments are unlikely to be noticed, while more substantial changes are required to effectively signal spatial transitions such as a character moving from one room to another.
Future work will look into whether the same results are found when working with visually impaired people.