Using schlieren photography to visualize sound propagation

I’m a photography hobbyist. Quite a newbie so far! I’ve come across the method of schlieren photography some time ago. As for my hobbies, I am mainly interested in sound and acoustics. I am currently tasked with improving the acoustics in a club. One major part is to figure out where sound gets reflected, and find dispersion patterns. It’s fairly difficult without visualizing it somehow and among other things I have started looking at schlieren photography. I know it can visualize minute changes in sound pressure, and apparently there’s been some research into it:

Schlieren Photography of Sound Waves,
F. D. Carlson, K. C. Clark and J. C. Eisenstein,
Electro‚ÄźAcoustic Laboratory, Harvard University, Cambridge, Massachusetts

J. Acoust. Soc. Am. 17 , 101 (1945)

(I haven’t been able to locate this paper thus far)

I would like to create a map of the whole floor with the schlieren technique (is this even possible? feasinble?), showing wave dispersion patterns.

The question is: what’s the minimal budget for performing schlieren photography in this setting? I understand the mirror in itself is very expensive. I would likely be photographing from 3-4 meters away, and would like to cover a lot of ground (maybe 40×10 meters), so maybe the mirror doesn’t have to be that perfect. I would like to see how the sound waves change at around 1.80 meters above ground (that’s where our ears are just about). I understand I could take multiple photos and stitch them together. What is the best way to go about performing this project inexpensively (if that’s at all possible)? I understand a parabolic mirror of 15 cm diameter is about $100 online, but I wonder if photographing from further away would allow me to use a less exact mirror, and maybe make one out of pieces of flat mirror. A 15 cm mirror would mean a huge amount of photos and stitching.

I assume I could do stitching because the sound waves come from a generator and always disperse in the same way. I could rig a trigger circuit so that the successive photos always happen in the same exact moment when a specific point in the test signal is being played back – therefore always capturing the same moment in the dispersion pattern, which evolves every time as the same sound is being played back over and over, but will always be the same e.g. 10.0000 seconds into the sound.

I would almost certainly approach a local photography specialist (even if just because I don’t have the right gear for this), but first need to know: does this make even remote sense, and is this possible without spending tens of thousands on a huge mirror?

P.S. I know you can make room simulations – they don’t work for me as the room is very complex. I also know of this long-exposure technique: but it’s very time-consuming.


P.S. maybe one of the other members could add the “schlieren” tag to this question.

2 responses to Using schlieren photography to visualize sound propagation

Man, this is definitely a serious overkill.

First of all, you need a good quality laser AND good quality mirrors. I mean laser-quality mirrors, to have predictable results.

Second of all, you cannot simply “stitch” resulting images, as only standing waves will show a time invariant pattern, all others will be time varying (varying with phase from source, phases from reflection).

The easiest would be: you rent a good quality microphone (one that is lab characterized) then you do a sine wave frequency sweep on your speaker setup (or multisine – faster but needs more postprocess) meanwhile positioning the microphone in space, and then you will be able to draw a 3 dimensional attenuation and phase characteristic map. You will have to be systematic, draw a cross matrix on the floor, set the microphone to increasing heights, etc.

In general, you will want to go for double the highest frequency you use, then calculate the air wavelength (consider the room temperature!) and that will be the necessary grid resolution to have any meaningful data. You have more time – make a smaller grid.

Now, why is this a complete waste of time? Because a club will have people of varying amount, they will move, etc. They themselves represent attenuation, phase distortion, reflection, etc. Unless they are sitting like rag dolls, how are you going to address that? You can do your measurements with people in the club, but WOW! Another nice challenge! Unless you are redesigning the Metropolitan Opera, I suggest not being extremely quantitative about this.

Your best bet is probably trial-error. Use movable sound absorbers, sound reflectors, move the speakers, and find a good setup.

It seems like the biggest problem would likely be using a strong enough sound to produce a detectable distortion. It should be detectable near a speaker cone perhaps, but sound dissipates and looses power exponentially. To cover an entire room with sufficient magnitude of distortion to be meaningful would likely be catastrophically loud.

The Wikipedia article on the subject has a clip of the distortion from a gun being fired and as you can see, even there it disipates rather quickly (and there is also a lot of air movement in addition to the sound waves there).

Your best bet would be to use a high speed camera and a very sudden sound and see where the waves go, but I’m not sure that you would be able to track it more than a few feet before it fades too much and doesn’t cause enough of a distortion to be detected.

Leave a Reply