• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

DAC filters?

Sounds plausible, due to the variable impedance of DT880, lows might be boosted by a high output impedanceamp, balancing the peaks in treble. Despite the more natural presentation created by a bass boost however, the peaks in highs will still be there and still will be fatiguing in my experience.

View attachment 374934
I have no problem with the treble on my DT-880 - 600's, and also have DT-990 - 600's I like with this amp too, which have even more treble. I have found a tube buffer to be the answer to alleviate fatigue for me. I now use linear phase fast roll-off for the DAC filter.
 
Last edited:
I have no problem with the treble on my DT-880 - 600's, and also have DT-990 - 600's I like with this amp too, which have even more treble. I have found a tube buffer to be the answer to alleviate fatigue for me. I now use linear phase fast roll-off for the DAC filter.
What I hear you are saying effectively is that you don't mind +4dB at 5-8KHz range, but it's the 0.5dB at 20KHz that was bothering you.
 
I disagree with comments to the effect that there is no difference between DAC filters. The fact that some people can't hear it doesn't mean that it isn't there or that those who can won't find it bothersome. Everyone's hearing is potentially different.

I have always been able to hear higher frequencies than most people. When I was a kid, I could hear dog whistles (even the one at the end of Sargent Pepper on the original LP and the early CD release that included it.) I lost some hearing due to injury, when I was in my 20s, and more in my old age. All in the high range, but not continuous or consistant in the latter case (it varies a lot from day to day) which appears to be hereditary.

That said, I have a Loxjie D30 and D40. They both have the same DAC chip (but the D40, for which I used the balanced outputs, has two of them). On the D40, I find that Filter 1 generally has the sound that I like the best. It's the most clear and analytical to me, but it does seem to have boosted the upper treble range and this can be shrill or have a ringing effect on some recordings. So, I have started playing with Filter 2 as well and trying the different Sound Color settings on both of them. (I read something about Filter 3 quite a while ago that convinced me that I shouldn't use it, but I don't remember the reason.)

So far, to my ears, the settings that lessen the shrillness kill too much of the high range or greatly reduce the volume of the mid-range (which sounds even worse). It's hard for me tell, though because (as previously stated) what I hear varies from recording to recording and from day to day. It would help if anyone who knows exactly what each filter and sound color does (or is supposed to do) would explain that here, for us to use as a starting point.

Thanks.
 
I disagree with comments to the effect that there is no difference between DAC filters.
I'm not sure people are saying that.

Some of the slower filters start to roll off the response from around 10Khz, and that is quite likely to be audible in an ABX test for people with good hearing above that frequency. It's not going to be a huge "night and day" difference though.

However, that type of filter is essentially broken by design. If you are comparing two devices with filters with fast rolloff above 20IHz and good attenuation (typical of the default filter used on competent DAC designs) then you are exceedingly unlikely (to the point of impossibility) to hear the difference.
 
Which test signals would make the differences between the filter options maximally detectable to human ears?
 
Which test signals would make the differences between the filter options maximally detectable to human ears?
White Noise (doesn't need to be high-pass filtered) is good for spotting the general spectral change, especially when comparing a NOS or "slow rolloff" filter vs. "sharp rolloff".

Single-sample pulses can also be used and might expose any ringing (to young and trained ears only, IMHO). The spectral effect can be heard as well, because a pulse's spectrum is white, too. The pulse will appear to be a bit softer.

A further possibility are steady-state sine very close to fs/2. These signals are hard to use because one must still be able to hear a 22kHz sine clearly as a tone. With non-perfect filters one can hear a beating, the volume of the sine swells up and down. Say, you are 10Hz below fs/2, then with an imperfect filter you will get an image at fs/2 + 10Hz with almost the same level. These two closely spaced tones produce a 20Hz beat frequency, that is, an apparent fs/2 sine is pulsating up and down in volume 20 times per second, which sounds rough and gritty.
This could well be (a part of) the mechanism behind the notion that filterless NOS DACs tend to emphasize transients despite the lower steady-state volume in the top octave... and that is not exactly intuitive.
 
I have had 3 Loxjie DACs: the D30 (which I recently gave away), the D40 (the original, not the D40 "Pro"), and the D60. To my ears, they all had the same basic sonic characteristics, except that each successive version sounded a little less shrill in the higher frequencies and had more of what I would describe as ambiance overall (like an ambiance, or reverb and delay effect that you would apply to a track in a DAW).

Unfortunately, some music still sounds, for lack of a better word, bad -- regardless of the filters or sound color used. My view of the filters is that they can alter some of the sound to slightly mitigate some of the objectional characteristics of a given recording. However, this comes at the expense of either worsening other aspects of that recording or worsening the sound of other tracks. The conclusion I have drawn is that filters are not a good way to address problems that really should be equalized and, if an equalizer were added to the signal chain, I would still have to adjust it from track to track. The "sound colors" on the other hand, seem to simply add varying types and degrees of ambiance to the mix and, like filters, enhance some characteristics and diminish others.

I tested all the options on both DACs over Thanksgiving weekend. I am not going to speculate on the technical aspects of what is causing this or what the best solution is. I'll leave that to those with more technical experience and better ears than me. I'm simply going to describe what I heard and what I did to minimalize what I perceived to be the most objectional aspects of the sample selection I was listening to -- namely, the second number ("Now / Later / Soon") on the Broadway cast album of "A Little Night Music". I have an LP of the first pressing, which I bought when the original production was running in the 1970s. I also have the first release of the CD, which I believe came out in the 1980s. I compared both to the streaming version on Amazon Music Unlimited as of two days ago. This is what I heard:

On the original LP, the first thing I noticed was a lot of ambiance, like it was recorded in a theater or a recording studio suited for opera or classical music. The orchestra sounded very lush. The strings were sweet and drowned out when the brass came in, as you would expect it to do in a live performance on Broadway. The vocal tracks appeared to have been overdubbed and mixed so that the volume of the singer pretty much matched the volume of all the instruments playing together, if not slightly louder. There was a lot of ambiance on the vocals, too. And, despite the fact that this is an old record with scratches, pops, and surface noise, if I ignored them it still sounded great. The original CD sounded the same, when played on my entry-level Denon 600NE CD player using the built-in DAC (but without the scratches, pops, and surface noise, of course).

The Amazon version was the worst (and the Spotify only marginally better) even though it was streaming at a higher bit rate than my CD player. In the number above, the volume of the strings was dramatically reduced and the brass sounded even softer, regardless of the DAC filters used. On Anne's sustained high note in "Soon", the highest violin line actually drowns out her voice. In the spoken dialog that follows, there are what sounds like loud bursts of static, which I think is actually just the breath of the actress whom I assume was standing a little too close to the microphone (or projecting her voice like she was still in the theater). It doesn't sound good on any medium. But, at least, you can tell what it is going on when you listen to the LP (and the CD when the player's internal DAC is used.) On the streaming versions, is sounds like a short circuit or a noise generator on a synth or old school drum machine.

So what did I do about this? Well, I basically tried every filter and sound color combination to try to remedy the above. On the D40, I found that I could minimize the "noise" described above by using the third filter (FL3), which I gather was intened to fix phasing issues. It diminished the noise but really killed the highs. Using Sound Color 3 helped mitigate that a bit. But it made everything else I streamed from Amazon sound worse. So, I reverted to the combination of filter and sound color that I had been using until now (the D40 forces you to use a filter and sound color): FL1 and Sound Color 4 (unless the highs were too muted, in which case I use Sound Color 3).

It was sort of the same story with the D60. Any of the phasing filters lessened the "noise" after Anne's solo, but made everything else sound muted or muffled. In the end, I resorted to my default filter, the one labeled "Linear Fast" (which may or may not actually be a linear fast filter, since some have posted that Loxjie has mislabeled some of the filters). I found that using the "Rich 1" sound color adds a touch of the ambiance missing from the streaming versions of this album, but I would advise "seasoning to taste". I have a feeling that the option to use no sound color might be better for other tracks or when streaming movies from Netflix or Amazon.

I think the bigger issue here is that some eq is needed, at least with Loxjie's DACs. And the user may have to adjust it for each selection because one curve may not work for every recording. What a pain! It makes me yearn for the days when you just put a record on the turntable and you got what you got. But everybody in the industry knew what the limitations of the medium were back then and made the best choices they could, so the listener didn't have to. Sure, we had equalizers, but we used them broadly, primarily to compensate for the poor quality of the loudspeakers that most people could afford back then. Noise reduction was handled by Dolby and dbx. (And yes, I do still have a dbx coder/decoder somewhere, but no dbx coded records -- which were few and far between even back in the day.)

Even the early days of CDs were better than a lot of what I hear on streaming versions of this same album. The DAC in my CD player isn't half bad, it just handles classical (or let's say orchestral) music better than jazz, or rock, or pop. The DAC in my Emotiva preamp does better with the latter. It was a pain to have to switch between the two, depending on the recording. But having to constantly adjust an external DAC is just as bad, if not worse. So, now I kind of regret having purchased several of them (although I really need one on my video system, where the DAC doubles a balanced preamp) instead of waiting for whatever will come next. (But then, I'm getting too old to wait for much of anything anymore!)

MORE - I've just noticed that Amazon has added another version of my test album ("A Little Night Music") to its Roku app. At first glance, it appears to be another digital rendering of the original recording, though the running time slightly differs. (And no, it's not the movie version. The movie version has another audio problem and it's called Elizabeth Taylor.) Stay tuned.

UPDATE - The new version (release date Nov 1 1998) does not seem to be a new recording, but a remix of the original release. It is the best digital version and the closest to the original LP and sounds best to me with the Linear Fast filter and Rich 1 sound color.
 
Last edited:
Back
Top Bottom