Here is fft analysis for 5.2kHz at 0, -10, -13 (dBfs)
SR = 48kHz, 24bits
Note the higher thd as the output level is increased. Level adjustment was via the dac's 32bit volume control. The load impedance was 20K ohms and a LM4562 preamp was used between the dac and the soundcard (Asus dx).
Note: the LF noise is due to a slight earth loop.
SR = 48kHz, 24bits
Note the higher thd as the output level is increased. Level adjustment was via the dac's 32bit volume control. The load impedance was 20K ohms and a LM4562 preamp was used between the dac and the soundcard (Asus dx).
Note: the LF noise is due to a slight earth loop.
Attachments
Last edited:
some more information to accompany these measurements might render them useful to someone who doesnt own the AP. as it stands i'm not sure what to make of it. doesnt look too flash at all, but without something to gauge it by its not really that legible, to me
Sorry if the screen grab is a bit low-fi.
What I was trying to demonstrate here is that the 3rd harmonic makes an appearance near maximum output level. The thd looks to be up around 0.01% at 0dBfs but reduces to < 0.005% at -10dBfs and with only a 2nd order visible.
Whether you think that this level of distortion is significant is a matter of opinion - but ftr using the same test bed setup I can't see any distortion from my power amp (at any non clipping power level) and into 8 or 4 ohm loads.
The issue is clearly analog distortion in the stages following the DAC. Not too severe, but it shouldn't be there.
The input level to the sound card is the same for all 3 dac output level tests - as it appears on the analyser. I used a class preamp between the dac and the sound card to keep the sound card input level constant for the 3 measurements. Sure, the sound card and the preamp are contributing to the total distortion but the only parameter that's being varied in all 3 tests is the level at the dac output. FWIW the preamp + sound card thd contribution is difficult to measure - being so low.
So how do you keep the input to the sound card the same with the three different DAC output levels?
I'm not trying to split hairs just trying to get a complete picture.
jan didden
hey man, that wasnt my point, sorry if I came over a bit cross hehe wasnt intentional, although I have been known to be a bit of a grump from time to time. not lofi, just without the legend or scale, not really meaningful to someone like me who isnt so familiar with the test equipment you are using. it was fairly obvious to me that something didnt look right for a 9018 dac and certainly doesnt match my experience with them and similar measurements on an AP in the development of the D1 IV stage for sabre thread, but how bad and how exactly wasnt clear. that amount of distortion at any level from a dac with the purported spec of this one that even goes to the trouble of using super capacitors in the power supply, along with a list of other flavor of the month audiophile specifications seems amiss.
monkeys like me need prettier pictures to understand stuff about stuff, thanks for the clarification
Last edited:
A DAC based on the ES9018 can be programmed to sound everything from muffled to sibilant.
With the optimal settings and proper electrical design it sounds awesome...
The gain of the preamp (between dac and sound card) was varied (+10dB, +13dB) to compensate for the -10 and -13 output levels.
Yes, it would usually be much better to notch the fundamental at the dac output and I did try doing that with a twin T notch but I found that the mkt caps in the filter produced more distortion than the sound card - whch uses LM4562 op-amps.
Points taken. The X axis is a linear frequency scale with 1kHz per division and the Y axis is 10dB per division.
Hope that helps. It's not really great test equipment - just a pc, some fft sofware and a sound card with low distortion. By no means is it a precision analyser setup but it's good enough for what's being measured here.
Interesting.
My w4s sounds completely different to every other digtal source I have. That's one of the reasons I decided to measure the distortion. It is higher than most dacs at full output - due to the zero fb I>V and buffer but those levels really don't explain the big SQ difference that exists between the W4S DAC and everything else I have. Yes, I also tried IMD tests but even the difficult 19+20kHz resulted in 3rd order below -90dBr0 which isn' t a bad result.
Well that's subjective but I'll attempt to use some audiophile magazine adjectives to describe SQ.
To my ears the difference between the W4S and my Marantz SA11 CD player, a Benchmarks DAC1 pre, and a modded MF XDACv3 is that there's a lot less 'air' and ambience - at high frequencies. Also, rather than the treble being fast and silky smooth or 'liquid' sounding it's exactly the opposite - being thick and textured almost like a valve amp. As a result it lacks resolution.
And, it's not rolled off because I did check with a sweep test and found a flat FR to 20kHz.
Perhaps it's just mine that sounds odd because the reviews have been excellent and none of what I'm hearing has been identified by any of the golden eared reviewers.
Losing air and ambience is a sign of getting sibilance IME. Some people notice the sibilance, others don't - if you don't, it tends to mask other details in the HF. I lived with digital sibilance for years as I just assumed that's how audio sounds. Then one day I changed how I was listening and it was a 'wow' moment how I'd missed it all those years. I think you can improve it with attention to grounding, and its dependent on the noisiness of your mains supply and what other kit you connect. As far as I recall the W4S DAC has no inbuilt mains filter?
- Status
- Not open for further replies.