[Jensigner]

You can really notice the E12s uneven frequency response, with a slight grade at ~ 1 kHz. Measurements are with 33 ohm loads and with line-in level ~ 1.0Vp and gain adjusted to get ~ 2.5Vpeak headphone output:
  http://www.jensign.com/FiiOE17/RMAA/CompareE17E12/Compare_E17_E12.htm

 

[Evshrug]

Interesting charts.
What happens if you use the FiiO E17 with the line-out adapter to E12? I question the quality of the Creative X-Fi external DAC, and it seems like an unnecessary extra variable. I haven't tested my E12 to make measured graphs, but I am surprised at the frequency graph because it's different from any other E12 graph I've seen.

 

[Jensigner]

Quote:

Interesting charts.
What happens if you use the FiiO E17 with the line-out adapter to E12? I question the quality of the Creative X-Fi external DAC, and it seems like an unnecessary extra variable. I haven't tested my E12 to make measured graphs, but I am surprised at the frequency graph because it's different from any other E12 graph I've seen.

I haven't tried that but the L7 line-out adapter for the E7/E17 probably won't work with the E12 since the E12 doesn't have a DAC.  That Creative sound card has a very good line-in A/D which I have characterized and used extensively (although the line-out of that sound card limits the noise floor a bit). I doubt anyone (for higher efficiency phones below 50 ohm) would hear any difference though (I can't hear any difference of any type with my Senn HD598 50 phones).
You can see a bit of that "ramp" in response of the E12 in the graph with this review:
  http://www.inearmatters.net/2013/03/impression-fiio-e12-mont-blanc-way-to.html
and also it is evident (not the db scale used) in FiiO's own specs here:
  http://fiio.net/UploadFiles/main/Images/2013/1/20130128152834.jpg
FiiO apparently changed that bass EQ detail in later models so that might also affect the FR.

 

[Jensigner]

A more detailed frequency response plots comparing E17 and E12: 
http://www.jensign.com/FiiOE17/BodePlots
Note that the low frequency peak with Bass enabled for the E12 has been moved to higher frequency (~ 50Hz) compared to earlier version.

 

[stv014]

Quote:

Interesting charts.
What happens if you use the FiiO E17 with the line-out adapter to E12? I question the quality of the Creative X-Fi external DAC, and it seems like an unnecessary extra variable. I haven't tested my E12 to make measured graphs, but I am surprised at the frequency graph because it's different from any other E12 graph I've seen.

 
Why do you think the FiiO would have better DAC performance ? From the results it looks like the opposite is true. Contrary to popular belief, sound cards tend to have fairly well performing converters.
 
The frequency response graph is consistent with Tyll Hertsens' measurements at InnerFidelity. The crosstalk for the E17 is not, but it cannot be reliably measured with a typical single ended sound card loopback because of grounding issues. Although the InnerFidelity graphs also show impossibly good crosstalk for the E12 (maybe because a low impedance load was not connected ?).

 

[Jensigner]

The main ADC on the X-Fi Elite Pro sound card I use is the AK5394 (123 dB 192kHz/24bit) which is very good for a sound card.
At any rate, using a simple loop-back (sound card line-out to line-in) this is my measurement capability (at least based on RMAA which i realize overstates the Xtalk):
   http://www.jensign.com/XFiRMAA
so the measurements for the FiiO devices under 33 ohm loads are pretty close to my measurement floor:
   http://www.jensign.com/FiiOE17/RMAA/CompareE17E12/Compare_E17_E12.htm
 
Crosstalk measurements as already mentioned are influenced by ground loops (Creative actually discuss this in relation to using their sound card for RMAA measurements). Also the exact loading and output signal level are very important (which often is not mentioned). Even accounting for the RMAA inaccuracy of crosstalk, still with 33 ohm loads my results under the specified signal level show the crosstalk to be quite good (perhaps more or less consistent with innerfidelity results)? 
 
For total noise measurements (not using RMAA) the total input noise of the X-Fi line-in in 40kHz noise BW is ~ 10uV ... not great.  I use an extremely low-noise M-Audio preamp to raise my total noise measurement capability to ~ 0.6uV.