HiFiMan HM-601 review and test
I'm new on this forum. It looks like a pretty interesting place. I thought I'd contribute some information, perhaps someone will find it interesting.
I'm an engineer and a musician. Recently I bought the HiFiMan HM-601 to use with my relatively high impedance headphones. I thought it had some interesting features, and, truth be told, the whole audiophile thing had an effect on me, although I'm usually very skeptical.
I bought it online and thought some music sounded very good on it, and some sounded kind of strange. I'm not going to say anything about the design. The battery life is poor, user interface very buggy, but I'm not going to dwell on that, either. It's supposed to be about the sound. Again, I didn't think that everything sounded right on it, but these things can be pretty subtle and difficult to judge.
Then I found this forum and an extremely long thread about the HM-801, which is now closed. In essence, the HM-801 has a noticeable roll-off at high frequencies, and somewhere in that thread Fang (the designer, as far as I understood) explained that it sounded less digital that way. I thought it was a pretty strange claim, and the most likely explanation not involving black magic was that it had some digital artifacts, which the roll-off helped to hide. I decided to test my HM-601 against other devices, using RMAA.
HM-601, firmware v0.04
Sansa Clip+, firmware 1.02.15A
iPod 5.5, Rockbox 2.7.1
Sorry if iPods are not allowed on this forum.
Test ADC: M-Audio FW-410.
Test software: RMAA 6.2.3
The test signal was recorded as a 44.1 kHz, 16-bit wav file for the iPod and the HM-601, and converted as a 256 kbps mp3 for the Sansa, which was running native firmware and did not seem to support wav.
The frequency response of the HM-601 is very similar to that of HM-801. The other devices exhibit a flat response, as expected.
Perhaps, this roll-off sounds "less digital". The question is why. I think the plot below illustrates the problem:
These are the spectra of the 1-kHz sinusoidal signal, which is played by the DUTs at approximately the same level. It's the part of the test protocol that is supposed to measure the THD. Note the 2nd harmonic of the Sansa, by the way. Most noticeable of course are the peaks at every 100 Hz in the HM-601 output signal, which are clearly not harmonics but digital artifacts.
Here are the same data plotted from 0 to approximately 3.5 kHz.
The HM-601 has similar problems at 10 kHz, although the pattern is a little more complicated and the spacing is not 100 Hz. I think these artifacts are probably much more audible at higher fundamental frequencies, to use the term loosely.
I'm not sure where these artifacts are coming from; most likely, poorly implemented resampling.
All of the tests above were done at 44.1 kHz, 16 bit mode for the RMAA ADC input, but the results are similar at other sampling frequencies and bit depths. Output levels were approximately matched between iPod and HiFiMan. RMAA has its share of problems, apparently, but I'm reasonably sure that these results accurately reflect the DUT performance.
I was already thinking of returning this magical audiophile player because of numerous other flaws, these results did not make me want to keep it.
It took several emails to figure out the address to send it to, but I got the refund very quickly. Just as I was returning it, they came out with a new minor firmware release, 0.05, which I did not try.
Last edited by sergivs; 03-18-2011 at 08:49 AM. Reason: typos
|Thread Tools||Search this Thread|
All times are GMT -5. The time now is 12:00 PM.