Over the last few months I have slowly been tweaking my system. I use the E-MU 0404 PCI sound card which has built in hardware EQ though the interface has a learning curve (it's not graphical!) I don't how "good" this EQ is, perhaps there is better out there, but I've been using it extensively.
I calibrated my system based on the assumption that listening to a sine sweep from 20-20000hz, it should sound "perfect". In other words, my perceived volume should be the same throughout the entire sweep - there should be no (or minimal, life isn't perfect) points where the volume seems to raise or drop. Of all the anomalies I found, it turned out to be about 95% spikes and only 5% dips.
I had noticed from the beginning, starting with my HD 600 and moving onto a DT880, 600 ohm, there was always many things about the sound that did not sound lifelike. Rather than blame any potential problem with the music itself, I listened to the sweep on my Beyers. Without any EQ at all it sounds abysmal. There were points where the volume spiked so loudly it hurt my ears. This occured with the HD600 as well but sounded less balanced to my ears so I used the DT880 as a starting point.
(I also used pink noise to assist in making sure something was evenly leveled)
Right now I have *counts* ... 19 EQ inserts in the main mix of patchmix, as well as one for right channel only, and one for left channel only. (To do this I make foobar route each channel to it's own ASIO panel in the sound card. I know, I'm a genius) Now that every frequency is included and no frequency is emphasized more than the other (for all intents and purposes)... anything well-recorded (hell, anything decently recorded) sounds incredibly life like. I hear crazy details that I never heard before because they were over-shadowed by some other freq blasting in my ear or whatever other problem I corrected.
Is this true to human hearing? I don't know. I've read a lot about this but I've also heard many people say if you EQ to "flat" it will sound very wrong. There's a lot of ways to interpret this. FLAT as determined by a microphone and computer system is very different from calibration using my ears and my ears only.... in a way I've fixed everything wrong with human hearing. I somewhat believe this is the what a direct line into your brain might sound like, bypassing the ear. Of course it's not perfect. I already have my system balanced +23% to the RIGHT to compensate for my right ear not hearing as well. This has caused certain freqs (like 6500 hz) to create a very loud resonance in the right cup, so I must lower the volume of that frequency on that side to compensate. What happens is that my perceived volume of frequencies in this range is actually a combination of the ACTUAL frequency, plus the resonating frequenc(ies). This is not ideal by any means but the negative effects are not too bad, as it sounds pretty damn good anyway.
And you know what... the sound isn't bright. It isn't muddy. It isn't sterile and it isn't clinical or dead. It's alive, it's dynamic, there's just the right amount of everything. EQ is not wrong, it's not a bad word. Flat is not a bad word. Flat is not "boring"!! If this sound like this is wrong, then baby, I'm wrong.
