DIY Ultrasonic Microphone?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

hkung

Member
Joined
Jun 27, 2024
Messages
15
Location
SE
Hey guys!

New to the forum but have been reading for quite a while, so first of all thanks for being awesome both by sharing knowledge and maintaining a nice community.
It's super inspirational and has given alot of confidence into try building something for myself.

Background:
I work as a sound designer and composer in the film and gaming industry. Recording for this purpose usually entails alot of non-traditional recording and out of the box approach. Something in particular that is useful is ultrasonic recordings done with 192 kHz or more samplerate, which utilizes the fact that some mics (most mics, but some way better than others) record above audible hearing. Basically the idea is that you pitch these frequencies down to hearing range while still adhering to the nyquist-shannon theorem, aka create clean sounding source material allthough it's pitched down as much as 3-4 octaves. Add some effects / sampling magic and get sofa king creative results.

Thread topic:
How would you go about building a good sounding DIY ultrasonic microphone, and what would be the most important things to think about in your view?

Example could be:
What capsule and why?
What in the circuit should one pay special attention to?
Have you done any experimentation on your own in this scenario that you'd like to share?

Already existing mics:
Examples of mics I've had good results with are: SDC: Micbooster/LOM mics with Primo capsules, Sennheiser 8000 series (goes to about 60kHz). LDC: Ehrlund-M (goes to about 87kHz) and obviously the Sanken CO-100k which goes to +100kHz and is heavily marketed for purposes like this. There's also a bunch of measurement mics and what not that goes to 300+ kHz, but in my experience dont necessarily sound good for sound design.

It would be interesting to hear your thoughts, and hopefully excite you to ponder a more out of the box scenario.

Cheers
 
I use this Knowles 0410 ultrasonic mic capsule from Micboosters as the basis for the ultrasonic mic I use for recording bat echolocation ultrasound.
I'm using it with my Zoom F3 recorder , which has a 192KHz sampling facility....
It also uses a 32 bit floating point format which - as you no longer need to worry about gain settings - makes things lot easier when recording in the dark! :)

The mic capsule is interfaced to the phantom powered balanced XLR input of the F3 using the attached schematic.....
Quite pleased with the results, although the capsule is not really in the same league as the Sennheiser or Sanken mics you describe... ..

It seems that there are alternative specialised ultrasonic capsules available, although they tend to be more suitable for serious scientific work.
And at the price some of them are listed at, you might be better off taking another look at Sennheiser or Sanken.....1641.MEMs.mic.preamp.schematic.no notes.jpg
 
Sure!

Please care for your volume, this is literally making noise with mics. Put together some quick & dirty (and noisy) experiments just to demonstrate. I'd say they're quite good examples of things you can do cool stuff with, as they contain a wide range of low and high frequencies from the get go so they dont necessarily need downpitching to turn into interesting materials. Anyway...

All follow the same structure: Raw recording followed by downpitched octaves followed by the same last octave with added creative FX.

1. "RubberD2" 0, -12, -24, -36, -36 + FX. Result... maybe some Wall-E / R2D2 robot voice. Link:
2. "Metal Scrapes" 0, -12,-24, -24 + FX . Sci-Fi / Mechanoid alarm perhaps? Link:
3. "Cello Rips", 0, -12, -24, -24 + FX. Lets call these some spaceship passbys. Link:
this last one is also with micbooster mics rogs

(I suspect the noise in the pitchdowns of the last one are either internal from the preamp or from the computer fan. Note on is also that artefacts like this arent usually bad when sound designing, usually they make quite cool flavour when you distort and effect it together with whatever you're recording)
 
Last edited:
Here's a short recording of a Noctule bat made with the mic I described above ...slowed down 20 times so you can hear it (the original is centred around 30KHz).
The quieter chirps in the background are being made at around 55KHz by soprano pipistrelle bats .... (the 'pips' are always about - or so it seems ! :) )
View attachment Noctule slowed down 20 times.mp3
 
I use this Knowles 0410 ultrasonic mic capsule from Micboosters as the basis for the ultrasonic mic I use for recording bat echolocation ultrasound.
I'm using it with my Zoom F3 recorder , which has a 192KHz sampling facility....
It also uses a 32 bit floating point format which - as you no longer need to worry about gain settings - makes things lot easier when recording in the dark! :)

The mic capsule is interfaced to the phantom powered balanced XLR input of the F3 using the attached schematic.....
Quite pleased with the results, although the capsule is not really in the same league as the Sennheiser or Sanken mics you describe... ..

It seems that there are alternative specialised ultrasonic capsules available, although they tend to be more suitable for serious scientific work.
And at the price some of them are listed at, you might be better off taking another look at Sennheiser or Sanken.....View attachment 133031
Ah, that's cool! I've seen these, but haven't tried them out yet :) Thanks for sharing. Do you have any recordings to share by any chance?
That is basically the same process as these recordings. Zoom F3 and F8 with 192kHz and 32bit float. Such a handy feature for field recordists
 
Here's a short recording of a Noctule bat made with the mic I described above ...slowed down 20 times so you can hear it (the original is centred around 30KHz).
The quieter chirps in the background are being made at around 55KHz by soprano pipistrelle bats .... (the 'pips' are always about - or so it seems ! :) )
Beat me to it! I think you might've posted an .XML file instead of an audio file though?

Edit: That works now. Sounds super clean. May I ask what algorithm / software you used to slow/timestretch it 20x?
 
Last edited:
Beat me to it! I think you might've posted an .XML file instead of an audio file though?

Edit: That works now. Sounds super clean. May I ask what algorithm / software you used to slow/timestretch it 20x?
Yes, I didn't post the original audio file correctly .. my fault!

I use a rather ancient version (v3) of Adobe Audition for my editing and 'time stretching'.
It's also useful to convert the 32 bit files onto more 'usable' 16 bit files - not everyone can use 32 bit float files yet!
The spectral display mode (sonogram) is also quite useful for analysing bat recordings.

As I say, it's pretty ancient 32 bit software, but to quote the old saying - " if it ain't broke, don't fix it!" :)
 
Very cool sounds. Much more organic than synthesis, which I like.
Briefly looked at the Sanken -- uh, some serious money. I get the DIY aspect...
 
Homebrew ultrasonic transducers (one sending, one receiving) swept in the 50-100kHz range, used as a navigational aid for blind people. Note the clip, modeled on a Shure SM58 clip, for attaching to the white stick that blind folk carry.
 

Attachments

  • bat.JPG
    bat.JPG
    1.5 MB · Views: 0
Very cool sounds. Much more organic than synthesis, which I like.
Briefly looked at the Sanken -- uh, some serious money. I get the DIY aspect...
Thank you! Yeah this approach is quite cool to get interesting results fast :D I'm a big fan of Ben Burtt, so it's sort of in the footsteps of his recording > manipulate approach just with more modern gear. In fairness many (most) times I weave stuff into synthesis as well, adds nice flavor having a recording instead of an oscillator as base.

Agree, the Sankens are a bit of an anomaly (and in my view not even the best mic for the job)
 
Yes, I didn't post the original audio file correctly .. my fault!

I use a rather ancient version (v3) of Adobe Audition for my editing and 'time stretching'.
It's also useful to convert the 32 bit files onto more 'usable' 16 bit files - not everyone can use 32 bit float files yet!
The spectral display mode (sonogram) is also quite useful for analysing bat recordings.

As I say, it's pretty ancient 32 bit software, but to quote the old saying - " if it ain't broke, don't fix it!" :)
No worries! Haha, that's pretty cool. I wonder how that algorithm is different from lets say Reaper or Ableton (which are pretty good).. I got curious while writing this so I found some threads in other places about people actually prefering Audition V3 timestretch algos to later versions. Interesting
 
Homebrew ultrasonic transducers (one sending, one receiving) swept in the 50-100kHz range, used as a navigational aid for blind people. Note the clip, modeled on a Shure SM58 clip, for attaching to the white stick that blind folk carry.
Wait, how does this work? It sweeps ultrasonically for echolocation and then warns in audible range, or?
 
There is a fairly detailed functional description here :
https://sites.aph.org/files/manuals/ksonar.pdf
At a nuts and bolts level, the sweep signal is generated by an Atmel AVR MCU implemented as a NCO (numerically controlled oscillator) driving a DAC hence to the transmit transducer. The return signal will have shifted by an amount determined by the time of flight to the object(s) ahead. This signal is mixed with the current oscillator signal via a 4 quadrant multiplier giving an audible output for the user.
 
I use this Knowles 0410 ultrasonic mic capsule from Micboosters as the basis for the ultrasonic mic I use for recording bat echolocation ultrasound.
I'm using it with my Zoom F3 recorder , which has a 192KHz sampling facility....
It also uses a 32 bit floating point format which - as you no longer need to worry about gain settings - makes things lot easier when recording in the dark! :)

The mic capsule is interfaced to the phantom powered balanced XLR input of the F3 using the attached schematic.....
Quite pleased with the results, although the capsule is not really in the same league as the Sennheiser or Sanken mics you describe... ..

It seems that there are alternative specialised ultrasonic capsules available, although they tend to be more suitable for serious scientific work.
And at the price some of them are listed at, you might be better off taking another look at Sennheiser or Sanken.....View attachment 133031
Has anyone checked the output section given by the resistors (47 k, 10k) and the capacitor 10 nF ? It turns out that this provides a cutoff near 2 kHz. Where is the ultrasonic ?
 
Has anyone checked the output section given by the resistors (47 k, 10k) and the capacitor 10 nF ? It turns out that this provides a cutoff near 2 kHz. Where is the ultrasonic ?
I think you may have misread the value of the output resistors -- they are 47R (47 Ohms) , not 47K (47,000 Ohms).
So, the -3dB cut off frequency is around 160KHz, which is high enough not to be of concern for use with a recorder sampling at 192KHz.
The value of 10K for the DC feed resistors will only attenuate the output by c. -0.04dB, not really a significant amount.
 
imho high sampling rate is only moderately important you still need a good sounding microphone. I've done this exercise with lab mics with impressive specs and the result is unusable from an artistic standpoint.
 
Back
Top