Introduction
The DUNU DN2000J is a pair of earphones with great resolution and good soundstage for an in-ear. However, after comparing its sound signature to my calibrated desktop 2.1 system and HiFiMan HE400i, I’ve noticed a prominent harshness or sibilance in the treble. So I set out to not only locate the treble peak but also to develop a very accurate EQ setting to “neutralize” the earphones.
Measurement
Target
My equalization target is the average of 47 subjects’ head-related transfer functions, a study done by Dorte Hammershøi and Henrik Møller. This is the most accurate target response according to Rin Choi’s blog.
Note: because Hammershøi and Møller’s graph doesn’t cover below 100Hz and over 16KHz, I improvised the upper treble by following the direction the graph was going before it cut off. For the bass, I used a 5dB sub-100Hz boost as expressed by Dr. Sean E. Olive’s research (more on its application on InnerFidelity). It’s to make up for the missing “subwoofer in the room vibrating your body” effect. I found it to sound quite right compared to my flat-calibrated 2.1 speakers, especially outdoors.
Notice the Compensated Frequency Response graph. It shows that the DN2000J has hollow mids, emphasized upper bass and slightly excessive warmth, and a lower treble peak for that “sharp sound” – the source of the harshness. Isn’t that an accurate description of how the DN2000J sounds to us?
Applications
- Equalizer APO config file (a graphic equalizer with variable custom bands).
- EQ convolver file (for Foobar’s Convolver plug-in and other impulse response convolvers).
- VDC file for Viper4Android‘s DDC feature (which is practically a parametric equalizer).
Click here to download them all plus the Excel spreadsheet.
Methodology
As you can see, my method of derive an equalizer setting is not eyeballing at frequency response graphs nor using “audiophile hearing” which I don’t have. I am using the calculation method because I don’t have a dummy head nor equipment to measure the DN2000J’s frequency response in my own ears. Basically what I did was model points at different frequency intervals on both DN2000J and target frequency response and calculate their difference to get the EQ values. I don’t know about you but it sounds really damn flat to me. Try my EQ setting and let me know how it sounds!
