Perceived loudness of concurrent tones hinges on frequency separation via three distinct mechanisms

Evidence

The formation of the loudness percept on the frequency axis can be investigated with different psychoacoustic tests. Figure 10.11 shows a result where the level of a 1-kHz tone was adjusted to match the perceived loudness with the loudness generated by a pair of concurrent tones. When the frequency of a pair of sinusoids differed less than a few hertz, the level of the reference tone was raised by 6 dB from the level of the non-summed tones. This effect can be explained by assuming that hearing reacts to the instantaneous maximum levels in the input and, due to beating, the maximum corresponds to a direct sum of the amplitudes (6 dB equals double the amplitude). When the frequency difference is from 20 Hz to 160 Hz, the adjusted level of the reference tone is only 3 dB higher than the non-summed test tones, which corresponds to a doubling of the power of the sound. The hearing can then no longer follow the maxima of beating, but only the average of the signal, corresponding to the sum of the powers. At frequency differences above 160 Hz, the tones appear in different critical bands, and the adjusted level of the reference tone approaches 10 dB, which corresponds to the sum of the loudnesses (Fastl and Zwicker, 2007). — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1145

Categorization

• DEFINE: Explains how beating-driven amplitude summing, power summing, and critical-band separation each dominate loudness perception in a different frequency regime.
• ADDENDUM: Connects to acoustic perception mechanics underlying soundscape and loudness models.