Schlieren Photography
W. Beaty         Dec 1999

Years ago I stumbled across an article by Dr. Winston Kock called SEEING SOUND (see amazon ref in links below.) In the early 1960s Dr. Kock worked out a Moire-pattern method for displaying sound waves by sweeping a microphone back and forth in the region near a sound emitter. I recently realized that there is another possibility for making sound waves visible in realtime.

[Lens focuses light onto edge of razor, some light passes, spreads out, and falls on screen.]
In a "Schlieren" optical system, we can make small variations in the density of air become visible by using a point-source illuminator, a razor blade, and large lens (or a large telescope or searchlight mirror.) The "Schlieren" setup is also called "the Foucault Mirror Test" because it can be used to test the curvature of homemade telescope mirrors. The razor blade is adjusted so that it partially blocks the light... the shadow of the "test object" is cast upon the viewing screen... and if some heated air near the "test object" should slightly deflect the light, then the light will either miss the razor blade completely, or it will be blocked. As a result, any regions of the air which deflect the light will cause dark and light patterns to appear on the screen. An image of the moving air will appear on the screen! A warm hand will cast a hand-shaped shadow, but it will be surrounded by a plume of rising warm air which resembles flames.

A "Schlieren Camera" can photograph the warm air rising above any hot object. This type of camera is commonly used by researchers to photograph shock waves in hypersonic wind tunnels. But think: a shock wave is a kind of sound wave. Perhaps a Schlieren system can photograph other sorts of sound waves? But sound waves MOVE, they move at around one foot per millisecond, so in order to capture them, we'd have to set up some sort of high speed flash photography. Maybe not: if we could flash the light source in synch with the sound waves, then the sound waves should become "stroboscopically frozen", and if these density-waves in the air are dense enough, the shadows of the waves should be visible on the Schlieren Camera screen!

A superbright LED would be a good pointsource illuminator if it was much brighter than normal. However, Charles Yost has discovered that LEDs work fine in this system if the frosted screen in the above diagram is replaced with a television camera with the camera's lens removed. The camera is positioned close to the razor blade, so that the image of the "test object" is very small yet partially covers the CCD sensor within the camera. Simply view the result on a television screen.

Other possibilities: this article mentions two other simpler "Schlieren" setups. In one, a bright pointsource is placed on a camera lens, facing outwards, and the camera is aimed at a distant retroreflective screen. The light from the LED normally bounces back and hits the LED, so the camera sees a dim field. Any density patterns will bend the light so it comes from the wrong place, then returns to shine in the lens. A second technique: a large wall is painted with a fine vertical grating, a camera is aimed at this wall, and an opaque photo of the grating is inserted in the camera at the location of the grating's real image. If the test object is at a different distance, the camera can focus on this object and blur the grating. Yet any deflections in the light will make more/less light hit the opaque grating, causing brightness changes.

I have not yet tried the above idea. Will it work? Maybe not, because sound waves are typically 100,000 times less intense than 1 ATM of pressure. Sound waves are density waves, but the changes in density might be too small to photograph. (Photos of bursting balloons apparently work OK though.)

To guarantee that the variation in pressure is maximized, we must use VERY LOUD sound. Also, we must use sound of fairly high frequency, since low-freq waves will be too big to fit into the viewing field. Sound moves at about 1100 feet per second, therefore a one-inch sound wave would be 13,000 cycles per second. A low frequency could be used if the sound consisted of brief pulses. A 100 microsecond pulse would give a sound wave that's a little under an inch thick. If the same 100uS pulse was applied to the LED as a 1A current, the viewing field should be quite bright. We should pulse the LED in synch with the camera frame-rate in order to avoid visible "beats" caused by the video scan.

Another idea: use a resonant chamber. That way the sound could be extremely intense, yet the required wattage would be fairly small. Build a clear plexiglas chamber and drive it at high frequency with a signal generator, an audio amp, and a small piezo-tweeter bought from Radio Shack. Or even drive it with narrow, high-power pulses, so the loudspeaker reinforces a sheet of high pressure rather than a long wide sine wave.

Dr. Winston Kock's "VISIBLE SOUND"

Wiston Kock used a small microphone attached to a long motorized arm which slowly swept out a raster-scan pattern. Adjacent to the microphone was a small neon pilot light which was driven by an amplifier connected to the microphone. Loud sound received by the microphone would light up the pilot light. The whole setup was photographed in darkness with a long exposure. The mechanical arm swept out its pattern, and the resulting photograph depicted the sound as bright patterns. Kock also managed to photograph the sound WAVES by adding a reference signal from a stationary microphone to the signal from the scanned microphone. This acted to stroboscopically "freeze" the pattern of waves. The resulting photograph showed a bullseye pattern radiating from any small source of constant-frequency sound.

Around 1982 I tried to build a real-time version of the above system. I attached a row of microphones to an old vinyl record album, with each microphone feeding an op-amp chip which drove an adjacent LED. When the whole assembly was spun at about 500 RPM, a glowing band of light appeared. The sound from a tiny loudspeaker would create a bright region in the band of light. When I played a constant tone into TWO tiny loudspeakers (headphones, actually), the dark bands of an interference pattern appeared in the band of light from the spinning LEDs.

A few years ago the Exploratorium science museum built a large version of a "Winston Kock Scanner".

Created and maintained by Bill Beaty. Mail me at: .
View My Stats