Audio3 Ltd. and Phonak launched a study to see investigate under which room conditions soundfield amplification improves speech intelligibility. Steven Dance and Lorenzo Morales of the London Southbank University collaborated in the effort. We tested Phonak's excellent DigiMaster 5000 SFA system in a reverberation chamber adjusted with acoustic treatments to achieve several specific reverberation times, while injecting noise of various levels, we then measured the speech intelligibility using STI-PA with and without the soundfield system activated.
Results produced an empirical equation that was able to predict measured speech transmission index based on: background noise, room reverberation time, and whether amplification was active or not. The correlation between the model's prediction and the measured results was significant (R2 > 0.98). We discovered that soundfield amplification does not provide benefit when background noise levels are below 40 dBA but that large benefits can be predicted and observed when background noise levels are high and when the room configuration has a lower RT. In optimal configurations we observed that adding soundfield amplification can produce an effect equivalent to reducing the environmental background noise by 7.7 dB.
Using this type of model it is possible to determine whether acoustic treatments or soundfield amplification makes more sense for a given situation. Paper has been submitted for publication.