For about a year now I've been testing rental earbuds (ie earbuds that come back from rentals) by using SMAART to produce a rough impedance curve of each bud to look for variations from standard. Now I'm just musing about ways to make it a quicker and more portable setup.
Here's how it works: I set up SMAART to show the usual transfer function. Don't really care about phase, just magnitude. The generator produces pink noise, which goes through a ~30 ohm resistor (I've forgotten the exact value) to the earbud. The test signal is taken from the junction between resistor and earbud. Reference signal is the pink noise directly from the interface output, as usual. With no earbud connected the output is just pink through a small resistor, so the transfer function is a flat line. With an earbud connected the impedance of the earbud drags the signal at the resistor/earbud junction down to a lower level, dependent on the impedance at each frequency.
Rational Acoustics has a paper on using this basic approach to measure loudspeaker impedance curves (https://www.rationalacoustics.com/files/Impedance_Tech-Note.pdf). I've found that with tiny earbuds you can use pink noise instead of a swept sine wave and still get very consistent curves. You can even put your finger over the earbud hole and see a change in the curve in real time with many earbuds.
The point is of course to see if there's anything wrong with a pair of earbuds without having to listen to each one - we're a rental house and we move a lot of IEM systems, so a quick test is important. I don't know if the impedance curve is actually very accurate, but it *is* very consistent across the same model. So if one of the drivers in a dual-driver earbud is bad the curve will be significantly different.
So the hardware for this test is pretty simple - a resistor. Seems like it would be possible to write (or adapt) a simple iPhone app to send out pink noise and then show a transfer function from the microphone input. This might complicate the hardware the trivial extent of needing a DC-blocking capacitor to stop the DC voltage meant to power an external mic, but that's probably about it. Should even be possible to do the test with a quiet enough pink noise that the earbuds wouldn't even need to be out of the ears, say if you needed to do a quick test during a show. A purpose-written app could have reference curves for various popular models, user storage slots, etc.
Here's how it works: I set up SMAART to show the usual transfer function. Don't really care about phase, just magnitude. The generator produces pink noise, which goes through a ~30 ohm resistor (I've forgotten the exact value) to the earbud. The test signal is taken from the junction between resistor and earbud. Reference signal is the pink noise directly from the interface output, as usual. With no earbud connected the output is just pink through a small resistor, so the transfer function is a flat line. With an earbud connected the impedance of the earbud drags the signal at the resistor/earbud junction down to a lower level, dependent on the impedance at each frequency.
Rational Acoustics has a paper on using this basic approach to measure loudspeaker impedance curves (https://www.rationalacoustics.com/files/Impedance_Tech-Note.pdf). I've found that with tiny earbuds you can use pink noise instead of a swept sine wave and still get very consistent curves. You can even put your finger over the earbud hole and see a change in the curve in real time with many earbuds.
The point is of course to see if there's anything wrong with a pair of earbuds without having to listen to each one - we're a rental house and we move a lot of IEM systems, so a quick test is important. I don't know if the impedance curve is actually very accurate, but it *is* very consistent across the same model. So if one of the drivers in a dual-driver earbud is bad the curve will be significantly different.
So the hardware for this test is pretty simple - a resistor. Seems like it would be possible to write (or adapt) a simple iPhone app to send out pink noise and then show a transfer function from the microphone input. This might complicate the hardware the trivial extent of needing a DC-blocking capacitor to stop the DC voltage meant to power an external mic, but that's probably about it. Should even be possible to do the test with a quiet enough pink noise that the earbuds wouldn't even need to be out of the ears, say if you needed to do a quick test during a show. A purpose-written app could have reference curves for various popular models, user storage slots, etc.