Page 23 - Fall2019

this 6 dB difference has at least three important implications. First, humans listening to birds communicating in a noisy environment will underestimate the effect of noise on bird communication because humans hear better than birds by 6 dB. Thus, if a human can barely hear a bird singing in the distance, a bird perched on the listener’s shoulder will not hear the distant bird at all because the sound has to be 6 dB higher to be detectable by the bird under those same condi- tions. Put another way, for a point source over a flat reflective surface, sound level decreases by the inverse square law, and a bird would have to close the distance by half (where the sound is 6 dB louder) to hear the singing bird. For a line noise source such as a highway and/or another type of surface, the decrease in sound level with distance can be somewhat greater or less than 6 dB.
The second implication of the 6 dB difference between humans and birds is whether a faint sound off in the distance (e.g., from a construction site) might cause stress in birds. If a human listener can barely hear the distant construction noise, it would be inaudible to birds at the same location as the human listener.
The third very practical implication is that human listen- ers, without using sophisticated sound-measuring acoustic equipment, can judge the range over which 2 birds might communicate in a noisy environment by using their own ears and applying the simple rule that a 6 dB difference is roughly equivalent to a doubling of distance. In other words, a bird’s threshold for detecting a distant bird is about half the distance that it is for a human detecting the same birdsong.
Finally, for a more quantitative estimate, critical ratios mea- sured in birds are a good predictor of how much masking is caused by anthropogenic noises that do not sound like white noise. In fact, critical ratios measured with pure tones and white noises perfectly predict the amount of masking from snowmobile noise and a variety of other anthropogenic noises from man-made sources (Dooling and Blumenrath, 2016).
Different Aspects of Hearing
Consider again the case of listening to speech in a noisy res- taurant. It is one thing to detect that someone is speaking in a noisy environment and quite another to be able to under- stand what is being said. In other words, hearing is obviously more than just detecting a sound 50% of the time in a hearing test, which is the classic definition of threshold and what we measure in a common hearing test.
But if the task is not just to detect whether a sound occurred but to tell whether two sounds are different, the level of the sounds must be 2-3 dB higher than that needed to detect whether a sound occurred. And sounds have to be 2-3 dB higher again to actually recognize a particular sound such as a word. These three aspects of hearing, referred to as the detec- tion, discrimination, and recognition levels of vocal signals, have been measured in the laboratory with a high degree of precision in both birds and humans. These studies have shown that the differences in SNRs required for detection versus dis- crimination versus recognition are nearly the same in birds and in humans (Dooling et al., 2009; Dooling and Leek, 2018).
As suggested above, we can add still another aspect of hearing in addition to the detection, discrimination, or recognition of a communication signal. The sound for this aspect has a level at which the sound is heard well enough to have comfortable communication. The SNR that represents a comfortable com- munication level in animals is impossible to assess because there is no way to ask an animal whether it is communicating comfortably. But in humans, the SNR required for comfortable communication is about 15 dB (Franklin et al., 2006; Frey- aldenhoven et al., 2006). Because the SNR differences between the hearing levels of detection, discrimination, and recognition are similar for birds and humans, it is possible that a comfort- able communication level also exists for birds and that level would be about the same SNR (15 dB) as it is for humans.
If so, then it is reasonable to postulate that there are four different aspects of hearing in birds that are relevant for com- munication, each of which requires a different SNR (Dooling and Blumenrath, 2016; Dooling and Leek, 2018). For humans, the distinction between these aspects of hearing is intuitive. Field researchers who rely on song playback techniques and monitor the behavioral responses of birds to determine whether a song was heard or not are familiar with a simi- lar phenomenon (Nelson and Marler, 1990). Klump (1996) described this issue as a just-noticeable difference that may be tested in the laboratory versus a just-meaningful difference between stimuli that may be measured in the field.
Working with masking results from the laboratory, it is pos- sible to estimate the theoretical maximum communication distance between two birds using the inverse square law and excess attenuation (i.e., attenuation that occurs as sound trav- els through a medium like air) for different environments such as open plain versus dense forest (Dooling and Blumen- rath, 2016; Dooling and Leek, 2018). These relationships are
Fall 2019 | Acoustics Today | 23