We often do not hear speech or announcement clearly at churches, gymnasiums, stadiums or factories. This happens because the original sound is interfered by its reflected sound and noise.

The human ears judged the easiness of speech listening in the past. Such judgment is ambiguous and prone to be influenced by individual variations among listeners and the state of voice.

The Speech Transmission Index (STI) provides the sound transmission with an objective value, focusing the physical phenomenon of sound mixing.

In short, the easiness or difficulty of speech listening can be assessed by the resemblance of the original sound wave and transmitted sound wave to ears.

The STI values range from 0 to 1. If the transmitted sound to ears includes much more reflected sound than the original, in other words, the transmitted sound badly differs the original, the STI value decreases toward 0. (poor intelligibility)
Contrarily the transmitted sound resembles the original sound, containing less reflected sound, the STI value increases up to 1. (good intelligibility)

The STI value is generally required more than 0.6 for human speech to be intelligible.

Let’s try listening to some actual sounds.