Articulation Inc
Articulation Index Equation
Home | Interactive AI Software | Articulation Index History | Articulation Index Equation | Consulting | About
Fletcher and Galt AI Equation

The AI is a metric whose value indicates the intelligibility of speech transmitted over a communications path.  The AI ranges from 0 to 1, where 0 indicates speech is not audible, and 1 indicates perfect transmission.

Whenever you calculate the AI using  Interactive AI™,  you will also see values for the four variables V, E, F, and H.  The four variables explain why speech intelligibility is poor, and will guide you on how to increase intelligibility.

AI = V × E × F × H
The AI metric is the product of the four variables V, E, F, and H.  Like the AI, each factor takes on values ranging from 0 to 1, where 0 indicates no contribution and 1 is optimal for speech intelligibility.  The factors are calculated using step-by-step instructions provided by Fletcher and is rather complex.  The calculation requires as input the composite noise spectrum (from all sources), the composite frequency-gain characteristic, and the speech intensity level.  Each factor is tied to an attribute of the input audio signal and can be viewed as the perceptual correlate of that attribute.  The four factors (V, E, F, and H) are useful for identifying the cause of poor speech intelligibility and fixing the problem.
V:  Volume factor

V indicates effective gain or volume.  V depends on the level of the received speech above the threshold for speech.  V indicates speech audibility.  Individual speech sounds are inaudible when V = 0 and and speech above detection threshold by various amounts values are greater than 0.  The factor V is associated with the speech-to-noise ratio and is perceived speech loudness.  Increase the system gain to increase V.

E:  Too loud
E is associated with the intensity level produced when speech is louder than normal conversation.  Speech is too loud when E is less than 1, in which case speech sound recognition is decreased.  The system's gain must be lowered to increase E.
F:  Spectral balance
F is associated with the frequency response shape and is perceived as "balance."  F is equal to 1 when the frequency-gain chracteristics is flat and may decrease with sloping or irregular frequency responses.  Flatten or smooth the frequency response to increase F.
H:  Intermodulation distortion, other distortions
H is associated with the percept of noisiness introduced by ear-generated intermodulation distortion.  H equals 1 when there is no external noise.  H decreases when speech peak and noise levels are both high and of similar intensity.  Speech sound recognition decreases when H is less than 1.  Decrease the gain to increase H.


Copyright © 2013 Articulation Incorporated