Friday, September 08, 2017
Color, not angle
Article in Nature on bat echolocation caught my eye.
Claims that bats run into smooth surfaces like windows because the smooth surface reflects sound only at one angle, like light from a mirror. Natural surfaces reflect diffusely in all directions, so the bat can always pick up the reflection.
Immediate problem with the angle idea:
Bats emit sound broadly, not in a focused laser beam. No matter where the wall is, or how it's angled, the bat will pick up a direct perpendicular reflection from some part of the wall. (If the wall is almost behind him, he could fail to hear it; but he's not flying backwards, so it doesn't matter.) Secondarily, the sound reflected from the wall also spreads as it travels, so the echo has a better chance of reaching the bat. My quick animation doesn't show this echo-spread.
I'd bet the problem is in frequency, not angle. Bats emit chirps or sweeps that run through a wide range of frequencies quickly. The bat picks up the response as a "temporal painting" of acoustic colors. Natural surfaces have characteristic absorptions and resonances in sound frequencies, just as they have characteristic absorptions and reflections in light colors.
Plate glass doesn't have a natural pattern of resonances and absorptions. It reflects all the sounds in a bat's range equally, so it appears as colorless sky in the "temporal painting".
Not coincidentally, our auditory system also includes a tremendous set of upsweep and downsweep sensors. We don't consciously use them for echolocation, but blind people learn to do it. Some blind people use a tapping cane to generate transient outputs, others do it the bat way with tongue chirps.
Normally we use our sweep detectors for speech, to read the changes in vowel formant frequency and the intensity changes of syllables.
= = = = =
Later thought: You could test angle vs frequency with a wall made of thin sheet metal, almost foil, just thick enough to hold a straight surface. Glue a collage of absorbing foams, tree bark, silicon glop, etc, to the BACK side of the steel. The front surface would then be optically straight and uniform but acoustically colorful. Leave half of the wall unglued. Do the bats run into the unglued part, or the entire wall?
Bats emit sound broadly, not in a focused laser beam. No matter where the wall is, or how it's angled, the bat will pick up a direct perpendicular reflection from some part of the wall. (If the wall is almost behind him, he could fail to hear it; but he's not flying backwards, so it doesn't matter.) Secondarily, the sound reflected from the wall also spreads as it travels, so the echo has a better chance of reaching the bat. My quick animation doesn't show this echo-spread.
I'd bet the problem is in frequency, not angle. Bats emit chirps or sweeps that run through a wide range of frequencies quickly. The bat picks up the response as a "temporal painting" of acoustic colors. Natural surfaces have characteristic absorptions and resonances in sound frequencies, just as they have characteristic absorptions and reflections in light colors.
Plate glass doesn't have a natural pattern of resonances and absorptions. It reflects all the sounds in a bat's range equally, so it appears as colorless sky in the "temporal painting".
Not coincidentally, our auditory system also includes a tremendous set of upsweep and downsweep sensors. We don't consciously use them for echolocation, but blind people learn to do it. Some blind people use a tapping cane to generate transient outputs, others do it the bat way with tongue chirps.
Normally we use our sweep detectors for speech, to read the changes in vowel formant frequency and the intensity changes of syllables.
= = = = =
Later thought: You could test angle vs frequency with a wall made of thin sheet metal, almost foil, just thick enough to hold a straight surface. Glue a collage of absorbing foams, tree bark, silicon glop, etc, to the BACK side of the steel. The front surface would then be optically straight and uniform but acoustically colorful. Leave half of the wall unglued. Do the bats run into the unglued part, or the entire wall?Labels: Grand Blueprint
¶ 2:30 AM
