The curve from Parts Express looks like someone forgot smooth the measurement. However the small wiggles are probably not the microphone. Typically the low frequency response does not need to be measured since its quite flat to below 20 Hz (unless your electronics have a high pass filter). The HF response from CSL above looks like a pressure response without the correction for free field on axis. Or the Dayton data is at 90 degrees from axis. However the Dayton mike (and its close clones) seem to start rolling off around 10 KHz and may be down as much as 4 dB at 20 KHz. Good accurate measurements of microphones is not easy. The industry standard doesn't work for the ECM capsules (electrostatic actuator) so you are a little out in left field. if using the substitution method you need to position the alternative microphones very very precisely and anechoic conditions with optimum windowing to remove reflections. At higher frequencies it gets really challenging since the wavelengths are so short and a millimeter becomes a significant part of a wave.
The curves are presumably as measured, so you are seeing raw data basically. The small fluctuations are small, +/- 0.5dB, and possibly partly artifacts of the measurement method (high performance anechoic chambers are massive, small ones don't perform so well). I guess you could smooth the data if you wanted.
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.