Obviously national labs and acoustic companies have invested millions in SPL calibration and we can't hope to compete at that level. Still, I'm trying to think of some easily reproducible "thing" that could get within 2-3 dB or so. Ideas that come to mind are a certain size and weight paperback book being tipped over onto a flat surface. A specific ball bearing dropped onto a granite surface plate. It has to be something that a person could reproduce using nothing more than a verbal description and uses commonly available materials. An impulse is fine, since we have sound cards and can look at the waveform, though the frequency should be "easy" for available microphones. This is a brain teaser so if your first thought is "impossible", keep thinking. There's the thread on spark gaps, but that's only doable by those with some electronics skills and is better for frequency response rather than absolute level.
mass-produced items come to mind, in order to get the same initial conditions.
Drew
- a golf ball and a 3mm piece of plate glass (like from a framed picture), dropped from 1 meter high
- a guitar string 6, high E, plucked mechanically, like in a harpsichord
- two wood boards clapped together, like the boards used in orchestra. assuming identical dimensions, this may be limited by differing species of wood
- somehow using water as a standard mass
- a plastic recorder, but how to deliver the same air impulse into the mouthpiece
Drew
Good ideas. I'm thinking this has to start with some reproducible energy, so the things based on mass and gravity seem like a good place to start. Trying to think of what anybody could go out and buy, with the same dimensions, mass etc. is hard. IMO, vernier calipers, postal scales and such, are all fair game, being inexpensive and commonly available, so simple fabrication is OK. I'd love something that could make a tone. Easy to get a close frequency, like a bending beam on the edge of a desk, but no good way to control amplitude. I also thought of water, some specific volume poured over a specific time, but not sure how to get a consistent sound out of it. Maybe a standard whistle of some simple construction, but I don't know how the average person can control airflow. A mass in a tube will do it, but not everybody can easily construct that. Keep 'em coming!
You can rent a calibrator for a day for $20. Cheap accuracy.
https://www.sgsgalson.com/equipments/quest-qc10-calibrator/
https://www.sgsgalson.com/equipments/quest-qc10-calibrator/
- two vessels, one with water, one with air, connected with "standard" tube
- air vessel is connected to "standard" recorder by "standard" tube
- water is gravity fed into air vessel, pushing air through tube into recorder
still, I imagine the air to be so elastic that there may be random "oscillations" in air pressure.
A 10" pan filled with 4" water, brought to a full boil. It's going to make some noise - but will it be the same amount of radiated sound power anywhere on the planet this is done? A mic placed a meter above the water surface...
A wood lathe faceplate spun by an electric motor with an AC line frequency dependent synchronous RPM. Put some tines made from 4 - 6" bolts occupying all the holes in the faceplate. When spun, it's going to make some noise...should be the same amount everywhere, say, for an 8" plate spun at 1800, 1500 RPM with 16 tine occupied holes arranged in the plate in the usual way.
From the ones I see available, an adapter from the motor shaft to the plate threading would be needed. Looks like an extra $20 for that part... I've no idea what RPM such a contraption could sustain safely.
A wood lathe faceplate spun by an electric motor with an AC line frequency dependent synchronous RPM. Put some tines made from 4 - 6" bolts occupying all the holes in the faceplate. When spun, it's going to make some noise...should be the same amount everywhere, say, for an 8" plate spun at 1800, 1500 RPM with 16 tine occupied holes arranged in the plate in the usual way.
From the ones I see available, an adapter from the motor shaft to the plate threading would be needed. Looks like an extra $20 for that part... I've no idea what RPM such a contraption could sustain safely.
iPhone with SPLnFFT is very accurate, and probably easier to get hold of than a piece of material of any given size
There may be accurate Android phones with an app too
Brian
There may be accurate Android phones with an app too
Brian
I once picked up a microphone calibrator at a fleamarket; old thing in a wood box that looked cool. All that was inside was an oscillator and a compression driver. One would think even the less expensive CDs when driven by a miniscule, known signal level would put out a consistent db spl into a mic, via a sealed coupling.
I made a guitar talk box out of the driver from the fleamarket one. The nicely machined stainless adapters for the "University Sound" driver - with integral rubber seals - fit to the hose nicely.
I made a guitar talk box out of the driver from the fleamarket one. The nicely machined stainless adapters for the "University Sound" driver - with integral rubber seals - fit to the hose nicely.
I have calibrators and much more than that (the miracle of surplus) but am looking for a way to get close without any real resources. More good ideas above and now I have to think about the pros and cons of each. Thanks!
I used to work in a calibration lab. There was a handy microphone calibration tool, a cavity filled by lead shots, that were dropped on a diaphragm. You had to turn it over, and it generated a shot noise for about 1 minute.
I found it:
https://www.diyaudio.com/community/...ise-source-for-spl-meters.37656/#post-5452902
I found it:
https://www.diyaudio.com/community/...ise-source-for-spl-meters.37656/#post-5452902
Last edited:
There is a difference between a calibrator and a standard. A calibrator for this use is often an electrical signal source and a transducer. It requires calibration, usually an annual one.
A standard does not require calibration. It uses physical properties that do not change. The usual acoustic standard is a pistonphone. It is a motor driven piston moving a fixed distance at a fixed rate inside a tight fitting cylinder. You do have to use a barometer to check the air pressure to insure accuracy. That is actually something you could build from off the shelf parts! Of course you would need another standard to calibrate it.
I think the easiest version would be a whistle, flute or organ pipe run from a known air pressure source. A simple very old fashioned air supply was a sealed container with an output pipe and the input pipe slowly feeding water. Seems to be the modern version could be a sealed paint can fed from another much larger open one. The height of the water source would determine the pressure.
There is a difference in calibration between a continuous sound source and an impulse source.
A standard does not require calibration. It uses physical properties that do not change. The usual acoustic standard is a pistonphone. It is a motor driven piston moving a fixed distance at a fixed rate inside a tight fitting cylinder. You do have to use a barometer to check the air pressure to insure accuracy. That is actually something you could build from off the shelf parts! Of course you would need another standard to calibrate it.
I think the easiest version would be a whistle, flute or organ pipe run from a known air pressure source. A simple very old fashioned air supply was a sealed container with an output pipe and the input pipe slowly feeding water. Seems to be the modern version could be a sealed paint can fed from another much larger open one. The height of the water source would determine the pressure.
There is a difference in calibration between a continuous sound source and an impulse source.
I have been thinking in the exact same direction - but have not come up with a solution yet. My take it should not only be a level calibrator but also a standard source for FR calibration in that the contraptions frequency distribution is well known. It doesn't have to be flat, just ideally quite broad band but most importantly, reproducible - say within +/- 0,5 dB?
Lets keep thinking. My take is that it is two pieces of plates, joined together in one end like a clap. It can't be depending on a surface so I believe it must be hanged in a string (defined length and material) the clap is possible hung spread out 180 deg (like a bird coasting) - it shall then be possible to initiate the clap by letting both wings collide by its own weight, making a impulse.
Challenges:
C'mon - we can do it - it would be so very useful - I can think of a number of very useful and cool applications for this! 🙂
//
Lets keep thinking. My take is that it is two pieces of plates, joined together in one end like a clap. It can't be depending on a surface so I believe it must be hanged in a string (defined length and material) the clap is possible hung spread out 180 deg (like a bird coasting) - it shall then be possible to initiate the clap by letting both wings collide by its own weight, making a impulse.
Challenges:
- Reproducibility; the thing must not wobble as one fire it as this will probably make every fire a bit different... the trigger mechanism need to be well thought out...
- Directivity; the thing triggered must sound the same within the full 360/360 sphere (this rules out the clap perhaps and something spherical is needed - a balloon etc?)
- It need to be sufficiently loud so that its use isn't easily noise environment limited.
C'mon - we can do it - it would be so very useful - I can think of a number of very useful and cool applications for this! 🙂
//
perhaps better to use water as the standard mass for providing kinetic energy in the test, but not flowing.
- cup of water hanging on end of stick+gravity provides potential energy to drive piston (air pressure), or to drive a plectrum on a string
- cup of water hanging on end of stick+gravity provides potential energy to drive piston (air pressure), or to drive a plectrum on a string
Something with a sphere is probably in the right direction... hmm, not trivial... at it again... reproducibility is a must.
//
//
What accuracy are you after? For high accuracy its challenging. Real accuracy requires reciprocity testing. It is intrinsically accurate if you know a few physical parameters but a real pain to do. Pistonphones are the usual standard and lots more practical but not easy to make. Most mike calibrators have a little speaker like transducer. The good ones also have mikes to stabilize the driver.
One thought that may work- a standard inexpensive fan. Bruel & Kjaer have a "fan" product https://www.bksv.com/en/transducers/acoustic/sound-sources/reference-4204 I believe a fan in free 1/2 pi space (a floor) a known distance should generate approximately the same spectral energy. If you can check the RPM to be sure its at the right speed it should be pretty consistent. Something like this: https://www.amazon.com/Amazon-Basic...9Y2xpY2tSZWRpcmVjdCZkb05vdExvZ0NsaWNrPXRydWU=
If this resonates (pad pun) I would be happy to buy one and do a full calibration (I have the resources to validate the microphones). Would there be a market for these? With an FFT you could get a good idea of the response curve.
One thought that may work- a standard inexpensive fan. Bruel & Kjaer have a "fan" product https://www.bksv.com/en/transducers/acoustic/sound-sources/reference-4204 I believe a fan in free 1/2 pi space (a floor) a known distance should generate approximately the same spectral energy. If you can check the RPM to be sure its at the right speed it should be pretty consistent. Something like this: https://www.amazon.com/Amazon-Basic...9Y2xpY2tSZWRpcmVjdCZkb05vdExvZ0NsaWNrPXRydWU=
If this resonates (pad pun) I would be happy to buy one and do a full calibration (I have the resources to validate the microphones). Would there be a market for these? With an FFT you could get a good idea of the response curve.
It would be nice if it could have som longevity - like 100 years 🙂 - and be solely mechanical. I mean if you built one today and measured it and built one in 50 years and measured it - and compared measurements - they be within the spectral distribution and level spec.
Conrad - what excitation to excitation difference would you need. I'd say 0,33 dB to pick a number where I know the hearing starts to struggle to make out a difference.
//
Conrad - what excitation to excitation difference would you need. I'd say 0,33 dB to pick a number where I know the hearing starts to struggle to make out a difference.
//
History gives us the Rayleigh Disk. It is far easier to make a good sensor than a good sound maker. Recent literature explores possible errors in the Rayleigh Disk, but this is how acoustic calibration bootstrapped itself.