And now let’s move on to the most interesting test. It is the test of the headphone amplifier under load. The measurement method was adjusted to a new recording sound card, which was an ASUS Xonar Essence STX. This card’s line input is designed to record peak to peak signal up to 5.65V, but most headphones are unbearably loud at such a voltage. To perform the test under more real-life conditions, but also to comply with the signal level requirements of RightMark Audio Analyzer, I had used to increase the sensitivity settings in the Creative X-Fi Elite Pro mixer. The ASUS card’s driver does not offer this opportunity, so I just used RMAA’s own ability to increase the volume of the recorded signal within 9dB which had been used to measure at an even lower volume. Thus, the system was calibrated for a volume of -7dB before the test, which agrees with the volume I had used in my two previous sound cards reviews, yet the new results cannot be directly compared with those published earlier. Particularly, the Xonar STX did much better in this test whereas the C.E.C. HD53R-80 did somewhat worse than they had done in my previous test sessions.
I test amplifiers under load using a homemade splitter that connects the amplifier’s output with the headphones and the sound card’s input. To minimize noise, the splitter is made from screened cable, which results in a considerable growth of crosstalk under such load as a pair of Grado SR 325i headphones (32Ohm impedance), so you should only use the Stereo Crosstalk results for comparisons within this test session only. I also tested the C.E.C. amplifier with a different cable connected into the duplicate connector of the right output.
The summary table with the results of three amplifiers indicates that the integrated amplifier of the Auzen Forte ensures the lowest level of harmonic distortions. Moreover, it delivers the best result among all devices we have tested in our labs. The dynamic range is obviously limited by the quality of the Auzen Forte’s own line input and the problem with noise when measured via a different sound card. Anyway, 90dB is quite a good result, too. The only drawback of the Auzen Forte’s amplifier is its non-zero output resistance, which leads to insufficient electric damping of the dynamic heads at low frequencies. This explains the poor results of the intermodulation distortion test in which a high-amplitude 60Hz tone is used.
On the other hand, the ASUS Essence’s integrated amplifier is not any better. It is the C.E.C. that has the lowest output resistance which can be observed in the distortions at low frequencies as well as in the influence of the nonuniformity of headphone impedance on the amplifier’s output voltage.
Amplifiers gain-frequency characteristic under heavy load
The graphs above make it clear that the Essence STX amplifier has a somewhat larger frequency range but the Auzen and C.E.C. amplifiers cannot be blamed as their frequency range at -0.5dB is 15 to 40,000Hz.
Auzen X-Fi Forte 7.1 nonlinear distortions spectrum
The harmonic distortion spectrums of the Auzen amplifier do not differ much when measured via the card’s own line input or via the Xonar Essence STX. The second and third harmonics are no higher than -100dB. The higher-order harmonics fall down quickly, and the most dissonant seventh harmonic is almost completely lost in noise. That’s an excellent result!
Asus Xonar Essence STX nonlinear distortions spectrum
The ASUS amplifier might be called excellent, too, if the second and third harmonics at a load of 32 Ohms were somewhat lower. The higher-order harmonics fall down very quickly here.
C.E.C. HD53R-80 nonlinear distortions spectrum
It is the C.E.C. amplifier that boasts the best linearity. Its fourth harmonic is already hard to see. In fact, the only unpleasant thing is the rather high third harmonic. However, in the previous test of the same amplifier with the same headphones but with a Creative X-Fi Elite Pro and with a different mainboard, the third harmonic was not visible at all in the amplifier’s right channel. So, I am not going to blame the amplifier. It is excellent even with a load of 32 Ohms.
I also measured the card’s line output using its own line input with every opamp and was surprised to see that the measurements were indicative of the difference in quality and character of distortions. It is not big but the measured characteristics of the Auzen Prelude did not depend at all on the opamp employed. It is the AD826 that delivered the lowest THD. The LME49720 and OPA2132 had the highest THD. The others were in between with the same result of 0.0009% but with different shape of distortions. Perhaps this is just a measurement inaccuracy, but the results were repeatable.
Modern digital-to-analog converters, including the AK4396, offer a choice of several oversampling and digital filtering modes. The sharp roll-off mode ensures better damping of high-frequency noise and is the most popular, but the slow roll-off filter handles the signal more carefully, which may result in a more natural sound. Judging by the smooth decline of the output signal amplitude at high frequencies and the sound improvements with the high-speed opamps AD826 and LM6172, the Auzen Forte uses the slow roll-off mode. These modes are selected on the software level with the DAC configuration register, so the setting may change with driver updates, making the card sound more like the Auzen Prelude.
And finally, I have to tell you about a queer bug in the first version of the Auzen Forte driver. It is going to provokes some trouble for people who use the hibernate feature in Windows XP (I did not test the card in Windows Vista). The problem is that the front output and headphone amplifier begin to sound very dull after your computer awakes from hibernation. The rear outputs, serviced by another DAC, are not affected. This reminds de-emphasis, so I guess the AK4396 is not initialized properly and enables de-emphasis on leaving the hibernation state. This is odd because the AK4396 works with the X-Fi chip without any problems on the Auzen Prelude.