Audio spotlight

It's something I've been wondering too... like maybe a parabolic reflector for it?

Fascinating idea... definitely needs a reflector as Matt said but sound waves are notoriously difficult to control, especially at low frequencies. Conversely high-frequency drivers are directional as as result of the physics of how sound propagates through air.

I would guess (and that's all it is) that you can't focus audio particularly well unless you stick to a single frequency, which is counter to what you're doing?

I seem to recall a demonstration of a flat - square shaped speaker using ultrasonic emitters that worked to make a narrow cone of sound? 

The higher frequency sound we produce the more directional it is - but to make a truly directional sound we need to limit it to a very narrow frequency and a pure sine wave. That's why bats use ultrasonic to echo-locate their prey, it's not that those frequencies are out of the normal range of the predator's hearing, it's because at 18KHz (and above) sound is increasingly directional. European bats have a range from 18 to 120Khz depending on the species.

It's hard to show this with real speakers as most of us have already lost the ability to hear that high by the time we're in our late teens. But if you can, you can set this up with an amp and a signal generator (or a microphone sensitive to H/F) as see how each frequency propagates.

Or maybe something like the military LRAD system - but a more "don't want to disturb the neighbours" instead of "cause physical pain through noise production" type of system?

A system like HyperSound HSS 3000 speakers is along the lines that might work?

https://en.wikipedia.org/wiki/Long-range_acoustic_device

LRADs are highly specialised and apart from producing ludicrous amount sound pressure, they're similarly ineffective at focusing speech (or any complex audio signal) over a range of more than a few 10s of meters. A single tone can be focused much like a laser. The problem is the relatively long wavelength of sound, especially the "speech frequencies" used in telephony. Visible light is much easier to focus because it has very short wavelengths (assuming I've remembered my physics correctly, it's been a while).

Here is another document I found regarding using ultrasonic sound used to make audible spotlight:

https://www.researchgate.net/publication/332608570_Audio_spotlight_Sound_from_ultrasound

This an interesting (if a little mathematical) discussion from Dr. Pompei.

Limitations of Loudspeaker Directivity | Audio Spotlight (holosonics.com)

And a someone on YouTube: (1) The Voice Inside Your Head - Holosonics Directional Audio - Part I - YouTube

As I've mentioned above, we can't "focus" audio frequencies because the wavelengths are so long. I don't know what frequency the actual transducers operate at, but it appears that the Dr. Pompei has developed a system to create functionally effective audio in a focused cone.

I haven't read his patents (and even then they are hard to read if you don't read patents - I can't even read my own application... damn lawyers). 

My guess (and this is all it is) is to use a pair of ultrasound drivers using beat frequencies and frequency modulation could produce a low-frequency (audio) signal at a distance. The application seems to be aimed at places like museums where the directionality is more important than quality. I would imagine this is the biggest problem designing these things because the drivers have to produce the same sound pressure level.

The idea of beat frequencies is a fascinating idea that also works with "binaural beats" where two slightly different sine waves are played into each ear. Our brains try to lock onto the difference frequency which is far lower than we can actually hear (called a "beat"). It's the low-frequency version of Dr. Pompei's system: assuming I've understood what he was describing. Here's a better explanation of beat frequencies.

Beat Frequency Formula: Definition, Formula Derivation, Solved Examples (toppr.com)

I don't know the fundamental frequencies that they're using but this is highly likely this is what they are doing because the higher the frequencies are the easier they are to focus due to shorter wavelength.

A better question is what the quality is like. The speech frequencies are (for telephony) in the 300-3KHz range but the wider the bandwidth (20Hz-20KHz) requires the transducer to be able to accurately reproduce its fundamental and cycle around the whole spectrum.

Ultrasonic transducers (typically made with cheap piezo crystals) are only efficient in a very limited range - this is the the difficulty for amateur builders. 

Murata (which makes the power convertors Matt and I used for power on the USB C microphone) also make these:

datasheet_maopn.ashx (murata.com)

I guess the real proof of the pudding (so to speak) is in the listening, but my guess is quality isn't exactly HiFi. But I can always be proven wrong.

That is something that I did not initially consider - the quality of the sound. Still - I find the idea of projecting a focused beam of sound interesting. I suppose if one wanted to have directed, focused, hi fidelity sound while watching TV, the best option is headphones (which I use all the time: AirPod Pros for something I really want the "spatial audio" effect, and bone conduction headphones for general purpose office work.)

I suppose one other factor one has to take into consideration is any harmful effect blasting ultrasound into the environment will have on creatures who can and do respond to those frequencies? Whales and navy sonar for example. 

I appreciate the time you've taken to entertain my thoughts, and the information you have provided me!

@justj Bats would be the worst effected, but your curiosity is admirable, it's been an enjoyable conversation.

I'm curious to know how they've managed to produce drivers capable of producing variable frequency ultrasound since you can to modulate it somehow or it's just ultrasound (if you catch my drift).

I've considered other modulation methods but nothing else seems to work. Class D amplifiers use PWM to get up to 80% efficiency and keeping their output transistors cool in the process. First time I heard one of those fire up I nearly needed a change of underwear.

That might seem a daft thing to say but this was the sound from a chip smaller than a postage stamp (well that an a few inductors)! Damn it was LOUD - even with the speaker in open air. Class A and B amps can't get close to that sort of performance and PWM is simple.

You can make a parabolic speaker quite easily by using 5 drivers (on a square panel). Essentially you'd want 4x drivers on each corner, these would be for producing bass frequencies, with a fifth attached in the centre of your material to produce the higher frequencies.

The material itself would most likely have to be mylar, but it plays quite nicely and I have no doubts that you wouldn't even need special tooling to get a decent curvature from your outer edges into the centre. In an ideal world, your speaker frame would be circular but the size is an issue. An oval or an obaliod rectangle may be a solid compromise.

I would hazard a guess you could also use a canvas like material, and rear channeling should be a non-issue in practical terms as your actual effective speaker size can be relatively... Huge. 

What could be fun is to get one of those hula hoop toys to use as an outer rim and make a giant "speaker". I can't speak for what it would actually sound like, but I'd imagine for outdoor listening it would be fine.

Inductors like this for the corners:

https://a.aliexpress.com/_EvZtwOs

And most likely a repurposed tweeter to manipulate the centre of the parabola

Well in for a penny I guess... 🙂

I've snagged one of them for a panel speaker design I've been wanting to try out for a couple of years!

@marcdraco if you are talking about the inductors/surface speakers they are great fun to play around with!

The very same and I may have recently ordered one from Ali to see.... well you know... 😏 😏 

Massive fun, and no mistake.