Subjective listening effort: Influence of background noise direction and speaker’s gender
In today's society, hearing and understanding of speech is essential. In many situations the target speech is masked by some noise which leads to a decrease in intelligibility. The amount of decrease depends e.g. on the direction of the noise source (spatial release of masking) and the voice pitch of the conversation partner. However, interfering noise does not only affect intelligibility but also listening effort. Therefore, we investigated the influence of these factors on the subjectively perceived listening effort.
In addition to speech intelligibility, listening effort is an important outcome measure for comparing hearing aid settings. Using the Adaptive Categorical Listening Effort Scaling method ACALES (Krueger et al., 2017) the individual listening effort was determined for various directions of the noise source. Using this method the individual SNR range from "no effort" to "extreme effort" was determined for 20 young normal hearing subjects. The listening effort measurements were performed with following conditions: S0N0 (speech (S) and noise (N) from 0°), S0N90, S0N135, and S0N180 with a male speaker. Condition S0N0 was also measured with a female speaker (all measurements applying sentences of the male and female Oldenburg sentence test (OLSA), Wagener et al., 1999, Wagener and Brand, 2005, Wagener et al., 2014).
The listening effort decreased with increasing spatial separation between target and noise up to 135°, but increases again at 180°. Similar effects have been reported in the literature on speech intelligibility measurements with OLSA. However, speech intelligibility seems to improve with increasing degrees only up to the maximum at 90°. Subsequently, speech intelligibility decreases again (Beutelmann and Brand, 2006).
In addition, the difference in subjective listening effort was evaluated with a male versus a female speaker, resulting in lower listening effort for the female speaker. In addition, the speech intelligibility with OLSA was determined for a female and male speaker. Reference values show different SRT values for the OLSA with a male speaker (-8.4 dB SNR) and a female speaker (-10 dB SNR, Wagener et al., 2014). This difference seems to be smaller for listening effort.