Speech Intelligibility

In general terms, the speech intelligibility of a speech transmission system depends on the characteristics of the system (e.g. a room, or sound system in a room), and the absolute sound levels of speech and background noise at the receiver.

IRIS calculates two speech intelligibility parameters:

  • Speech Transmission Index (STI)
  • Useful-to-Detrimental ratios (U)

Speech Transmission Index

IRIS calculates the full STI following the indirect method as described in IEC 60268-16 Ed. 4 (2011) [1]. Three STI metrics are presented: STI Male, STI Female and STIPA (IR). The implementation was validated with the technique by Cabrera et al. [2].

Useful-to-Detrimental Ratio

Useful-to-detrimental ratios, proposed by Bradley [3], can be determined from the early-to-total sound energy ratio and the speech-to-noise ratio. IRIS calculates U50 and U80 ratios in octave bands using:

\[\mathrm{U}_{t} = 10\log\left[\frac{\mathrm{D}_{t}}{1 - \mathrm{D}_{t} + n/s}\right]\]

where D is the early-to-total sound energy ratio with integration time t (50 or 80 ms), and n/s is the noise-to-speech ratio. Without the n/s term, U is the same as C (Clarity).

[1]International Standard IEC 60268-16 Ed. 4.0, Sound system equipment - Part 16: Objective rating of speech Intelligibility by speech transmission index (2011).
[2]D. Cabrera, D. Lee, G. Leembruggen, D. Jimenez, Increasing robustness in the calculation of the speech transmission index from impulse responses, J. Build. Acoust. 21 (3) (2014) 181-198.
[3]J. Bradley, Predictors of speech intelligibility in rooms, J. Acoust. Soc. Am. 80 (3) (1986) 837-845.