Especially at 3, 6 and 9 kHz, there are sometimes greater deviations from professional plots – but for a general, rough comparison between various in-ears and a rough idea of how they sound, the results are sufficient, and in the mids and lows, they are even (very) accurate. If the treble or any other frequency band deviates from my perception, I'll note that below the graph.
You are probably also interested in what I am using to measure the in-ears, so I'll list my gear:
- First and foremost, I am using my Acer Windows 7 64 Bit Home Premium Laptop with Intel i7 quad-core processor, 16 GB of RAM and many other cool features - but that doesn't really matter, as about any computer is able to do the measurement stuff. For recording the plots, I use ARTA as software.
- The most important thing when doing the measurements with multi-driver in-ears that have a varying impedance response over their frequency range is having a source with as low as possible output impedance, hence I'm using the LH Labs Geek Out IEM 100 for that purpose.
It features a ruler-flat frequency response without and with complex load from 20 Hz to 20 kHz, which is important for getting a good result.
- The microphone/measurement coupler I am using is the Vibro Labs Veritas, which I calibrated to mimic a real IEC 711 coupler's response with applied diffuse-field compensation target.
- The interface I use for capturing the microphone input is actually quite simple - it is a cheap StarTech.com ICUSBAUDIO2D. Its microphone input section sufficient, as it has got loads of amplification power to achieve a strong input level, and most importantly, its input frequency response is flat from 20 Hz (-0.2 dB) to 20 kHz (+/- 0 dB)
Please note that my older measurements (that have the same compensation and calibration as the newer ones though) have a 5 dB-scaling on the Y-axis whereas the newer ones have got a 10 dB-scaling on the Y-axis. Please keep this in mind when comparing older with newer graphs.