Go to a TJ Maxx or Marshall's and you might see plenty of these Philips SHE3580 series IEMs for $7 or $8. I bought mine back in May and I discovered that my EQ almost cancels the response they measured at goldenears.net:
http://translate.google.com/translate?sl=ko&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=goldenears.net. Click on GE Review and you'll see the list of IEMs that they measured.
Without EQ, I find them pretty unlistenable, with overwhelming bass and piercing highs. To me, it sounds excellent when I copy the frequency balance of my speakers (in well-treated rooms, sometimes with digital room correction on top of that, sometimes with a house curve). The headset SHE3575 is $10, sounds pretty much the same, but doesn't have the nice stiff cable:
The correction needs only a few bands and can easily be implemented in Rockbox. The Equalizer iOS app actually needs more work, even though it has seven bands. The exact locations of the spikes depend of the particular geometry of my ears. Mine seem to match up more with the dummy head at goldenears (and less with the one that Tyll uses at innerfidelity.com).
There's a lot of heavily-discounted IEMs and headphones at Ross, Marshalls, and TJ Maxx. I've made it sort of a fun sport to pick up the cheapest open-box pair to hear how horrible it is, and to see what EQ it needs. There are a few gems out there.
For example, the Panasonic HJE120 Ergofit, to me, sounds better without EQ than the SHE3580. It's still a bit harsh, though. It doesn't have much isolation, which is sometimes useful. It trades isolation for comfort and ease of insertion. It's practically an earbud with a nozzle. I bought mine at Ross for $7. At one time it was $1 at Amazon, but it's usually $5-$6.
Over the past few months, I've picked up clearance MEElec, Dynex, Sony, Coby, Philips, Skullcandy, and iFrogz IEMs, as well as JVC Marshmallows, Clear Colors, RipTidz, Gumy+ IEMs, Ear Cushions. I've managed to match the color of the IEM to my outfit.
Using precise EQ to get these closer to neutrality has always resulted in better sound. Phase and time-domain response aren't independent from the magnitude response. I found that after corrective EQ, the imaging sounds more precise and focused, especially with binaural recordings. I wouldn't be surprised if the transient response also measures better.
I found it easier to correct the responses of cheap IEMs than those of cheap headphones. Those have so many ups and downs in the response, especially above 2 kHz. It's because you also hear the effect of the pinna.
I can share more graphs, if anyone is interested. But maybe that's the topic of another thread--EQ'ing cheapo IEMs for neutrality (or something closer to it). I'll have to order me some MP8320 sometime.