[A recommendation for calibration and choice of masking noises used in speech audiometry].

Tests to evaluate speech recognition ability against background noise have still not been adequately standardized. This is due in part to the lack of a calibration procedure for the various noises used in testing. In the present study, a psychoacoustic method for the calibration of fluctuating background noises was evaluated. Three types of noise were used: babble noise, stationary noise and fluctuating speech simulating noise in accordance with (CCITT Rec. G 227). Following calibration, the masking effect of each type of noise was evaluated. Results indicated that stationary and fluctuating speech simulating noise had similar effect on speech understanding ability. The steepness of the intensity-intelligibility function was significantly greater using babble noise as compared with stationary or fluctuating speech simulating noise. Babble noise was well accepted by our subjects. However, since its measurement can be calibrated, stationary speech simulating noise may also be used in speech audiometry.