Advice please: low frequency sound transducer

OK this is going to sound weird, but stay with me: I need a method of recording and analysing a low frequency sound in my building.

Over the last few weeks, between about 23:30 and 01:30 on certain nights, a barely audible sound starts in my flat, on the top floor of a council block. Actually, it’s less of a sound than a vibration. The sound fluctuates, and sometimes stops all together, then restarts. The best description I have for the sound is that of a diesel locomotive idling at a standstill. It’s a mechanical throbbing.

The volume is too low to be picked up by a regular dynamic microphone. So I’m thinking what I need is a transducer that can be attached directly to the walls. I plan to feed the output into my computer, and use spectrum analyser software to identify the frequency of the sound. Then record some of it, and use filtering and amplification to make the signal more audible. Then perhaps I can identify the source.

See, I told you it’s weird. Does anyone know what sort of device I should use? Internet searches come up with evesdropping devices, designed to pick up speech through walls, but this frequency is far lower than that, probably 300 Hz or lower.

Any ideas?


1 Like


1 Like

You could use a contact mic for it.
To be honest the main consideration will be having a pre-amp that has lots of clean gain - you might need as much as +70db with no audible noise floor. Unfortunately pre-amps with lots of clean gain can be quite expensive.
JRF contact mics are well thought of and quite affordable
But you would still have the issue of clean gain.
I really like Sound Devices and Metric Halo stuff for clean gain, but both are much ££££
So that’s some helpful stuff and some less helpful stuff.

It might be a Pico mic that you need, and amplify it at "line level’ through an audio interface.

I’m wondering if you weld/glue it to an aluminium plate or better brass or something it might enhance the quality of low frequency it could pick up from your wall, ceiling or floor. Different building materials through your home will conduct the sound and an ear against the surface could find the sweet spot pretty well.

I’ve not done a lot of spectral analysis using visual method, I have experimented with various frequency shift tools to move the information into hearing range by transposing the harmonic frequency upwards out of sub range. Many subharmonics go undetected by human ear and you can see information about what’s going on around your home by analysing the subharmonics, maybe you can then compare an image of before and after the noise pollution and find the clues, should you find a regular pattern you might get to the scene early and find the culprit

Cool project!

The Beatles used a bass speaker and reversed it to record electric bass guitar, I presume it would work to use a surface mounted transducer speaker in the same way to get a wider amplitude than the Pico, they are inexpensive although the Pico is bout 1/10 the cost of a transducer speaker. I’m sure the speaker in reverse would generate bass input, but if your rumble is moving a big diaphragm like that I would be scared!

look up some designs for Seismographs (Earth quake detectors) they look at a lot of low freq ranges ?

We had a heating pump that used to do something similar, almost a heterodyne / harmonic noise just in the background.


1 Like

A vibrator? :smiley:

Another suggestion: in general dynamic mics are just not all that sensitive - maybe just try a condenser mic if you have something to phantom power it? Will potentially mean you need much less gain to get it audible
They’re usually (and annoyingly) good at picking up low frequency vibrations too.

also: assuming you do get a decent recording, and you have sections where you just have the ambiance of the room and sections where you have ambiance and the noise recorded, maybe it is worth trying to phase reverse the “clean” recording and combine the two in Audacity as the phase cancellation may be good enough to give you a pretty clear recording of whatever it is and help to identify it? If that makes sense?

or, if money is no object then maybe…


Once again, I am blown away by the ingenuity, expertise and sense of humour of our Maker Space membership.

I got excited last week when by chance I thought I had identified a strong contender for the source of the noise, but more about that in a moment…

To respond to your excellent suggestions:

@Gawain the contact mic is probably the right approach, and I suppose that was what I was really asking for without knowing the correct name for it. Those jez riley mics look like a good choice too. I have a pretty good pre-amp (I think) in both a mini mixer and my audio interface but have never tried them in a high-gain situation like this. Could be a problem, as the noise is very faint.

@techfolderkites has some great ideas also, especially frequency shifting to identify the culprit. Agreed that the noise is too quiet to move a big speaker in reverse, but it might be worth a try.

@Courty has a good point that this is almost more a seismographic event because the frequency is so low. I checked our boiler pump, as well as several other sources like the lift in our building, but none coincide with the noise.

@the_duckchild also has a good approach which I had not thought of. I have been meaning to buy a decent large diaphram condenser mic for a while, and this is a good excuse! Both mixer and AI have phantom power, so no problem with that. Looking at specs, some affordable mics go down to 20 Hz so might do the trick, which I did not know. I suppose that is why they come with shock mounts. I have used the phase reversal trick to cancel noise in television production before (shoot next to a fish processing plant) , sometimes it works sometimes it doesn’t. But worth a try to isolate the real character of the sound and at least determine the length of the cycle of on/off.

@emuboy’s contribution had me laughing out loud in a crowded train carriage to the surprise of fellow passengers! When I re-read my description of the sound, I was almost senseless and people were starting to move to the other end of the carriage… . made my day. Fortunately, none of them asked me what was so funny.

But on Friday, a breakthrough… or so I thought:

A tunnel boring machine! 2.5 miles away! Throbbing noise - sounds right. Intermittent - probably. Low frequency - definitely. And there was no noise on Friday night!! Cracked it, right?

Well, maybe not. The noise was there again last night (Sunday). Have they already moved the TBM and started work in a new location? Possibly. but I doubt it really.

So the quest goes on. Thank for all for your suggestions. Stay tuned…