1

Dear Sound Design community,

we are doing research in bioacoustics and need to determine the make and specifications of microphone elements inside digital voice recoerders that were used in the past, especially with respect to their signal-to-noise ratio.

We have used the Olympus DM-620 and Olympus DS-40 digital voice recorders.

Different information requests to Olympus in the USA and Germany did not lead to any sharing. Olympus does not disclose that information, and it also commonplace for the other major manufacturers, which at best report microphone sensitivity.

Their product manuals say:

  • DM620 has a 40 Hz to 23 kHz frequency response at the highest sampling frequency and depth (48kHz 16 bit)

  • DS40 has a 50 to 19 kHz frequency response at the highest quality recording setting (XQ mode)

So we disassembled the units, and we found circular microphones with solder pads, with numbers etched onto the microphone elements, and we measured their height and diameters to find out whether we could reasonably estimate their SnR from similarly specified microphones with this Digikey data base: http://www.digikey.com/products/en/audio-products/microphones/158/

So far, the numbers etched onto the microphone elements did not yield any result. But we can guess the SnR based on similar microphones in that database, unfortunately that may not be a very clean approach.

Our questions are:

do Olympus manufacture their own microphone elements, so that specifications are entirely inaccessible?

Does it make sense to assume that microphones of a similar type (we assume it's an electret condenser), frequency response, height and diameter, will have a similar SnR?

How could we measure the SNR of these microphones?

If by any chance you simply know the SnR of these microphones, that would be the simplest answer.

Here are pictures of the microphone elements: https://i.stack.imgur.com/7LoWX.jpg Sorry that I cannot post more links due to too little reputation.

Improve this question
1
  • To measure the SNR of a condenser, attach it to the mic imput of a very HQ AD DA converter / soundcard, record silence on it, and check out how low the noise base is, if it's -80dB lower than zero, that's the SNR, Use a zoom recorder with the electret plugged in, they have a good snr electronics, around 95
    – bandybabboon
    Dec 10, 2017 at 9:34

1 Answer 1

Reset to default
0

...and we measured their height and diameters to find out whether we could reasonably estimate their SnR from similarly specced microphones

Physical dimensions do not correlate well with SNR, and so this approach will probably not yield any useful data.

do Olympus manufacture their own microphone elements, so that specifications are entirely inaccessible?

It's certainly possible. It's also possible that someone else manufactures microphone elements, solely for Olympus, or for a select group of companies, and therefore the specifications are inaccessible.

Does it make sense to assume that microphones of a similar type (we assume it's an electret condenser), frequency response, height and diameter, will have a similar SnR?

IMO, no. Mainly because SNR is not simply a function of the microphone element itself, but also the input amplifier circuitry and noise induced by the surrounding circuitry. There's a chance that similar microphones may have similar SNR performance when connected to the same input circuitry, but no guarantee.

If by any chance you simply know the SnR of these microphones, that would be the simplest answer.

The easiest way to solve this problem is measure the SNR directly. The only trouble here is that if you want to do this with a high degree of precision, you need some fairly expensive equipment. An audio analyzer (Audio Precision is a highly-respected brand) set up with the proper accessories (at the very least, you'll need a speaker and a reference microphone to calibrate out the response of the speaker, plus a chamber suitable for microphone testing). An audio analyzer will allow you to measure SNR, frequency response, and many other parameters of the entire system (the whole voice recorder) which will be much more useful than simply the performance of the microphone capsule alone.

Improve this answer
3
  • What about recording the same test sound in a really quiet place, both with the Olympus recorders and other recorders with known SNR. Then, looking at the difference in self-noise level between recorders after equalizing the sound level to the test sound. Even though crude, would that be a valid approach? We would assume the added SNR of the pre-amps to be null.
    – kdarras
    Mar 15, 2017 at 13:35
  • You will likely want to first use a reference tone to achieve equal digital levels (in order to make the gains between the recorders equivalent) and then perform the test, but it seems like that may work.
    – uint128_t
    Mar 15, 2017 at 14:45
  • Alright, to improve that test design, we would simply plug a microphone with known SNR into the same Olympus recorders, so we can avoid the possible bias of different pre-amps. I edited the post to make it more general and would upvote your answer if you could integrate that alternative more practical solution so that it is useful to more readers.
    – kdarras
    Mar 16, 2017 at 9:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.