Creating Coherent Sound Beams, Easily

Lasers work by emitting light that is “coherent” in that it doesn’t spread out in a disorganized way like light from most sources does. This makes extremely focused beams possible that can do things like measure the distance from the Earth to the Moon. This behavior isn’t just limited to electromagnetic waves, though. [Gigs] via [CodeParade] was able to build a device that produces a tightly focused sound wave, essentially building an audio laser.

Curiously enough, the device does not emit sound in the frequency range of human hearing. It uses a set of ultrasound speakers which emit a “carrier wave” in the ultrasound frequency. However, with a relatively simple circuit a second signal in the audible frequency range is modulated on top of it, much the same way that an AM radio broadcast has a carrier wave with an amplitude modulated signal on top of it. With this device, though, the air itself acts in a nonlinear way and demodulates the signal, producing the modulated signal as audible sounds.

There are some interesting effects of using this device. First, it is extremely directional, so in order to hear sound from the device you would need to be standing directly in front of it. However, once the ultrasound beam hits a solid object, the wave is instantly demodulated and reflected from the object, making it sound like that object is making the sounds and not the device. It’s obvious that this effect is hard to experience via video, but it’s interesting enough that we’d like to have one of our own to try out. It’s not the only time that sound waves and electromagnetic waves have paired up in interesting ways, either.

Thanks to [Setvir] for the tip!

51 thoughts on “Creating Coherent Sound Beams, Easily

    1. That is effectively what the guy is doing here, by using a whole array of transducers that all are in parallel (and as such all in phase). This is the reason the effect is so localized. In theory, if you managed to feed each transducer its own signal with settable phase, you could make the beam steerable as well.

    2. In theory there is no reason it wouldn’t work. Essentially you could have a phased array speaker, able to direct sound in different directions easily. The harder part is controlling the phases of high bandwidth signals (as the phase of each frequency component would need to be controlled separately), so a variable delay (as opposed to phase) would work better on such systems.

      1. Yamaha’s old soundbars did this, during setup it would run through each “beam” individually, you could feel/hear the audio sweep across the room as each beam was activated. It was a really strange sensation, customers always got a kick out of it.

    3. It is done all of the time in a number of fields. It is a phased array.

      For example, many ultrasonic inspection devices (the industrial version of the medical imaging device used to look at foetuses in situ) beam steer with phased array techniques. Real useful when mapping thicknesses or localizing flaws in a thicker test piece, compared to a standard probe, and reduces the need for mechanical scan. Also useful when imaging the interior of a test piece.

      Also used in geologic work (though at a different scale).

      Phased array microphones are also quite handy. Much easier to implement these days using a DSP than when I was in school and needed to use delay-line techniques (I learned a lot and it worked to a point)

      A nice paper that covers the math (medical imaging application, but substantially the same math) is https://scholar.dickinson.edu/cgi/viewcontent.cgi?article=1202&context=student_honors

    4. This is how some modern professional speaker systems work.
      Some beam form in only one direction, some others attempt to work in 2d.
      The trick with sound is that the long wavelengths involve dictate a very large source to effectively steer low frequency energy.

      In the concert sound world EAW does this with their ‘Adaptive’ line of loudspeakers.
      There are also a few companies taking the concept even further such as Holoplot with wave field synthesis.

      1. This is beyond active beam-steering though as used by EAW, Renkus Heniz and others, which I believe work fully in the audible spectrum, right? This is similar to what AudioSpotlight does: AM modulate audio onto a beam-formed ultrasonic carrier. Hard surfaces (like listeners heads) act as the detector; the carrier ‘falls off’, and you’re left with a the fairly low bandwidth audio frequency content.
        Holoplot is freakin’ bonkers. I got my ears on a demo back in December at MSG and was blown away. Walked a hundred feet up the isles away from the speaker with virtually no fall-off. Then I walked 20 feet along a row and I was out of the beam. They described the technique as virtually placing a point-source an infinite distance behind the speaker array, and you’re left with near a planar wavefront. Many of them, to steer as you please. The demo really blew us away.

    5. First impression is that the description is wrong in so many ways. More details needed. In particular I’m dubious about a “non-linear” behavior of air. I know you can get detection in the ear based on non-linear response from the bones and other parts of the inner ear, but the air itself? Wouldn’t this mean losing energy at a very high rate as the signal propagates? (I won’t go into the problems with calling this coherent.)

      However, this is the basis of some stage tricks that are very clever and never published.

    6. “…‘beam-forming’ with sound waves just like you do with antennas, does anyone know if this is a thing?”

      Sure… this isn’t the same tech as in the article… though is another method called SASER:
      https://en.wikipedia.org/wiki/Sound_amplification_by_stimulated_emission_of_radiation

      And another method that reminds me more of microwave or optical collimating methods that uses a Lens:
      https://www.physlab.org/story/acoustic-lens/

      The first method I ever experienced, albeit large was with the Whisper Dishes (a.k.a. Accoustic Mirrors) @ the NM Museum of Space History:
      https://en.wikipedia.org/wiki/Acoustic_mirror
      https://en.wikipedia.org/wiki/New_Mexico_Museum_of_Space_History

      There are also other methods that even the military eludes to being dangerous outside of thresholds. There are also references to “sonic” bullets:
      https://drive.google.com/file/d/0B3kLL6AnKjj5eVgyMnVhbjNGMEU/view

      However, like other dangerous threshold devices akin like with “sonic bullets” potential… the devices can be used for health improvement if used critically carefully safely for health improvement for the longest life cycle intent:
      https://www.scientificamerican.com/article/acoustic-lens-turns-sound/
      https://phys.org/news/2014-01-acoustic-lens-tunable-bullets-ultrasound.html

      Figure music alone has the potential to stimulate, depress and disorient as an example of the effects of sound. Figure beam forming can localize the effect even more.

  1. Years ago I read a story about a similar device, but in the normal audio frequency range.
    They had lots of fun by going to a library and whispering messages into the ears of completely unsuspecting strangers from 20m distance.

  2. “essentially building an audio laser”
    Really? A SASER = Sound Amplification by Stimultaed Emission of Radiation?
    In the case of an antenna system nobody talsk about “essentially building a radio laser”. Or a RASER. I can’t find the principle of “stimultaed emission” in the case present.

    1. Yeah, this. There’s a (steerable maybe) coherent (maybe) wavefront there, but no real evidence of amplification as a result of the physics beyond wavefront summing. Now if there were a resonator in the system somewhere…

    2. I think coherence is like line width in a spectrum. This thing puts out a kind of “blue sound” with a frequency modulated high frequency. (I’m surprised the ultrasonics are high enough to do FM well at all.) The video says FM and HaD says AM for some reason.

      For example the coherence length of white light is about 3 wavelengths. The 40 kHz has a wavelength of about 28 feet, so several wavelengths is a pretty good distance. On the other hand, FM with 10 kHz onto 40 kHz will have a bandwidth probably less than two times the modulation width because … Bessel functions. Anyway, wild guess on a coherence length is more than 3 and less than 9 wavelengths.

      The Wiki is strange and lacking in details, especially about some of the behavior. Like saying the sound is always being demodulated in the “beam” but can only be heard in one direction. The video shows some cool phenomena, and I have a sneaking suspicion that it does not work like the Wiki says. For one thing, non-linear response means no superposition (and really bad distortion), which is the basis of the math of wave phenomena. I’ll make a small wager that this audible result is from side-bands from the interface between the beam and any reflecting or absorbing material and that the audible sound does not exist in the open air. Maybe it could be detected with Schlieren optics. (Or to be audible maybe the changing frequency of the carrier “shoves” air forwards and back as the number of wavelength is a given length change, in which case the sound will be dependent on the rate of change of frequency and the 40 kHz would be better than higher frequencies.
      For example the coherence length of white light is about 3 wavelengths. The 40 kHz has a wavelength of about 28 feet, so several wavelengths is a pretty good distance. On the other hand, FM with 10 kHz onto 40 kHz will have a bandwidth probably less than two times the modulation width because … Bessel functions. Anyway, wild guess on a coherence length is more than 3 and less than 9 wavelengths.

      The Wiki is strange and lacking in details, especially about some of the behavior. Like saying the sound is always being demodulated in the “beam” but can only be heard in one direction. The video shows some cool phenomena, and I have a sneaking suspicion that it does not work like the Wiki says. For one thing, non-linear response means no superposition (and really bad distortion), which is the basis of the math of wave phenomena. I’ll make a small wager that this audible result is from side-bands from the interface between the beam and any reflecting or absorbing material and that the audible sound does not exist in the open air. Maybe it could be detected with Schlieren optics. (Or to be audible maybe the changing frequency of the carrier “shoves” air forwards and back as the number of wavelength is a given length change, in which case the sound will be dependent on the rate of change of frequency and the 40 kHz would be better than higher frequencies.

      1. One of these days I am going to buy one of these modules and maybe even a sound lazer and study how they are operating if the info isn’t open source. My guess if it was… it will disappear and people will act like this isn’t real.

        Anyhow… I assumed just a simple heterodyne effect… though there may be a pulse modulation or something else effect happening to make these work. Interesting the capabilities and confusion just caused by reading into.

      2. I still need to read into the math regarding the demodulation. I’m not comprehending how that is happening based on the demonstration. Appears like the system isn’t a heterodyne or pulse train system from watching the video. Just an ultrasonic transmission carrier frequency with a audio signal modulated that is demodulated by certain materials. What are the materials characteristics causing this? I’m not visualizing clearly at the moment.

  3. I wonder how the rigidity (or lack thereof) of the perfboard holding the array affects the performance of the system. It’s probably rigid enough for ultrasonic frequencies, but I’d like to see testing done (or do it myself).

  4. So there’s a company called Ultrahaptics that does exactly this — they create an ultrasonic phased array and then use that to create the sensation of an object in the hand. The force is quite weak, but interesting application.

    1. About the same as audible frequencies… the problem with ultrasound is that you don’t hear it, so you can be exposed to a very loud environment and not realize it. That being said, these little emmitters do not pose any signifficant danger, they simply don’t have the power.

      1. I’ve wondered other than the energy level concern at a high level… about entrainment thresholds where the effect is more in regards to the lowest possible energy range required to sympathetically resonate with human body electrophysiological signals. I’ve read like with Dr Robert O Beckers work… the signals can’t be to high in power or the body rejects or adapts to.

  5. I want to see this technique used to create a single device that can hang from the ceiling of a room and bounce these signals off of walls in a full circle. This would make it easy to reproduce real 360 degree “stereo” with every sound coming from the exact location you want.

  6. Is there a GitHub page or anything like that? YouTube videos are not a great medium for carefully studying and understanding a project or technology to begin with and doubly so when the effect of the project doesn’t convey well or at all in video form (like this, or volumetric displays, or super high resolution graphics, etc.)

  7. I bought one of these from a guy on Kickstarter, wanting speakers to play only to me at work. Fun to play with, but audio quality isn’t there. The cat really hates it, even with no input.

  8. At UCSB back in the 1970’s, Glen Wade’s laser lab had a poster describing a swimming pool with a raised platform that had a stack of rocks on it. Dropping a rock from outside into the pool set a wave into motion that disturbed the floating platform, dropping more rocks into the pool. Eventually, the wave built up til it popped out of the pool. It was called a WASER. Water Amplification by Stimulated Emission of Rocks. I suspect the idea came from John Landry, but can’t confirm.

  9. Another interesting use of modulated ultrasonic beams involves aiming the beams so that they intersect at some point. Then frequency modulate one of the beams with audio of some sort. The beams interact at the intersection to reproduce the audio at that point. Can only be heard at that point.

  10. I was just wondering if I could build something like this to use in Bay Area traffic. Specifically so I could beam a message at a specific car and tell them to use their %^$^%-ing turn signals.

  11. As well as everything everyone else mentioned, the very first issue of Make: I ever read (volume 5) had an article about something that worked exactly like this: Woody Norris’s Hypersonic Sound technology, famously used in LRAD/MRAD. Article starts on page 28.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.