Originally Posted by stv014
This statement is false. Power requirement is a function of sensitivity (in dB/mW), and it is the same for all three impedance versions. Indeed, here is what Tyll Hertsens measured for the DT880:
32 Ω: Power Needed for 90 dB SPL: 0.47 mW
250 Ω: Power Needed for 90 dB SPL: 0.38 mW
600 Ω: Power Needed for 90 dB SPL: 0.43 mW
These values are within a range of less than 1 dB, and it is basically just random manufacturing variation. Higher impedance needs higher voltage, but it is not an "exponential" function, the voltage required - assuming a constant sensitivity - is proportional to the square root of the impedance. So, a 600 Ω driver needs 4.33 times higher voltage than an equally efficient 32 Ω one, and only 1.55 times as high voltage as a 250 Ω one. Note that the voltage most sources can output decreases when the load impedance is very low.
As far as power is concerned, "properly driven" is when even the highest peaks of the audio signal are not clipped or significantly distorted at the preferred listening volume. There are other factors that affect sound quality, but they generally do not improve (and often even get worse) with a lower load impedance. The E11 is enough for many people who do not listen to music with a huge dynamic range, and do not like (or have already) hearing damage. When driving high impedance headphones, its performance also does not significantly deteriorate with increasing output voltage until the clipping level is reached. There is no evidence that headroom that is never used inherently improves sound quality. Otherwise, it would be best to connect headphones directly to high power speaker amplifiers, and get better sound that way. In subjective comparisons, more powerful amplifiers often sound "better" simply because of not matching the volume properly (matching the volume by ear typically errs on the side of making the device that is louder at the same position of the volume control louder; this is also why higher gain - even digital gain - is often perceived as sounding better).
Sorry, you are right. I should have written "voltage" instead of "power".
However, to be more clear, the voltage required is actually:
Voltage^2 = Power * Impedance
Voltage = Square Root of (Power * Impedance)
Since Impedance is not constant depending on the frequency, the voltage requirement may still shoot up. You are assuming both constant sensitivity and constant impedance, which is not the case all the time.
For instance, take a look at the impedance over frequency chart of the DT880 250 Ohm.
Around 100Hz (mid-bass), the impedance shoots up to well above 300 Ohm. At which point, the headphone will need to draw more voltage in order to output the same power.
Same story with DT990 250 Ohm here. In fact, much worse:
DT990 250 Ohm suffers from an impedance swing up to 350 Ohm in the mid bass region.
Lack of voltage will obviously drop bass volume significantly for the DT990. But looking at the charts, it's also obvious that both the DT880 and DT990 will lose "clarity" (high frequency) as well since impedance ramps up for higher frequencies.
Disregarding the voodoo around subjective and objective "listening", physics says you need more "volt" in general to drive the DT880 and DT990... and from my experience, the Fiio E11 is not adequate. Simple as that.
You may get away with the Fiio E11 at lower volume with less dynamic music, but when you ram into something with high dynamic range, or when you need to ramp up volume because the source is too quiet, then there's a problem.
Also, not all speaker amplifiers output high voltage out of their headphone jacks. It depends on how stuffs are connected inside as well.
Edited by Bill-P - 9/29/12 at 9:12am