Mani, I'm just getting ready to post a topic over at CA on the "chain" from ADC through DAC.  Sending those waves through the aether to my brain again, are you?   

A quote from that Spectral Bulletin really appeals to me:


[Emphasis mine.]

There's so little attention paid to this topic.  It seems to me most people must take lack of knowledge as a barrier to inquiry, whereas to me it's a spur.

I'm going to rush in where even fools fear to tread (that is, into a discussion I'm vastly underqualified to be a part of), because of my great interest in the topic of filtering in particular and of good, accurate music reproduction in general.  (I don't know how widely such things are shared among different countries/cultures - the first phrase above is a reference to an aphorism, "Fools rush in where angels fear to tread.")

- Of course the goal is the most accurate music reproduction, and one of the best ways to tell if your system is accurate is to listen to lots of different things - instruments, types of music, etc. - and see if the system sounds better with certain types of music than others.  If it does, then it is imparting its own sound, and thus deviates from accurately reproducing only what it is given.

- Unfortunately we don't have some of the things demanded by the idealized mathematical model of Shannon/Nyquist, e.g., infinite time.  Therefore any system that samples music into digits and turns those digits back into music will be less than perfect.  There will be trade-offs involving phase shift, pre- and post-ringing, and undoubtedly other stuff I've never heard of.

- Given the premise that the filtering involved in the system won't be perfect, one must choose (or provide choices) between imperfect filters.  It seems to me this must inevitably be, to however (hopefully) small an extent, a matter of taste.  Phase shifts may drive one person crazy, pre-ringing another, and each makes the choices that minimize what bothers him or her most.

- When making (or providing) choices between imperfect filters, one should strive to guard against choosing the euphonic at the cost of the accurate.  But I imagine it is damned hard to tell which is which.  Does it sound good because less phase shift makes it more accurate, or is the touch of extra post-ringing making things "harmlessly" euphonic?

Mighty interesting stuff.

Enh - he's way too fond of calling good perfect, as in "A CD player has no measurable jitter."  Would be news to many folks, I'm sure.  And there's the same old same old about how 16 bits and 44.1kHz is enough. 

The explanation re 44.1kHz is wonderful, as in something to wonder at: We don't need higher than 44.1 because there's 8x sample rate conversion in the DAC.  Gosh, it must be far better to use a low sample rate at ADC then convert it to an 8x higher rate, than to use a higher sample rate and convert it at a lower multiple or not at all.

Hah, makes me want to plan to go to a hi-fi show as part of a European vacation, which would probably be a good way to get my wife to think of divorcing me.  :-)

Whenever you see this, Peter, if you care to reply, I'd be interested in knowing what your attitude is to the possibility of exhibiting at shows in the US.

I think it was the engineers at General Electric who used to have a saying: "The difficult we do immediately; the impossible takes a little longer."   


The Dragonfly is USB1 (apparently GR/AQ didn't want to make it necessary to download drivers for Windows), so perhaps the bandwidth limitation is a bit more justified in this particular instance.  Doesn't make things any easier for you, though.

Was looking forward to -4, then saw the note about PA not working for "needs 24" DACs, of which I believe mine (Audioquest Dragonfly) is one.

Oh well - back to anticipation.

When you say this will be resolved in a next version, Peter, you mean -5 rather than z-8, right (I hope)?

Hi, Peter.  Still on the initial -7, which does give a bit of a low level "crack" when stopping.  Nothing awful, but of course if I can eliminate it that would be good. 

Am I correct in thinking that we must wait for the next version of -7 in order to have a version with the crack fixes, the PA improvements (though I feel that PA is working nicely for me at this point) and usable with 24/96?  (I have a few recordings at this resolution that I'd rather not do without.)

Thanks!

So the 3770 will almost certainly run hotter than the 3930 (unless you reduce the speed of the 3770 - probably not what you want).  Peter can therefore afford to run the cooling fans in his rig reasonably slowly, making it nice and quiet.  And as Peter says, the 3930 has more cores/thread capability.  (Peter, are you really going to be able to take advantage of, for example, 8-10 threads in XXHE?)

Many people, including me, really like SSDs almost everywhere.  I don't think Peter does so much, but I'd have to review more threads here to remember exactly where he thinks they may be good (if anywhere), and where they may not.  My guess is that in terms of noise, the CPU cooler fans and power supply and case fans, if any, will be the major contributors, reasonable HDDs not so much.  (I have the same CPU cooler fans in my homebuilt PC as Peter does in his rig, and they are *very* quiet.)

I'd say between the SSDs the other builder wants you to get and the CPU, you've got the best part if not all of that $1000 difference.

"Less powerful" can mean a lot of things.  Specifically what is different, as far as you know?

Edit: I see you've already asked Peter himself in another thread about the differences between this "boutique" PC and his, so if you want to continue over there rather than responding here, I won't mind.   

I myself would strongly consider both of the DACs you mention.  Have not heard either, though.

Re PC, if the "boutique" PC is expensive, there's no question I'd choose Peter's, unless you'd prefer to build it yourself.  PC parts are commodity pieces, nothing special or exotic.

Would you consider Spectral well made gear?

Some extremely well made gear has response to DC as a design/sound quality choice.  Take Peter's advice and do *not* try this without checking DC first.  You want to take a chance on frying thousands of dollars worth of equipment/speakers because you wouldn't buy a $30-$40 multimeter?  I can hardly think of a better recipe for feeling sick and stupid if you take a chance and you're wrong.

Ah.  Well then I could simply have been impatient, because I did not wait anything like 4 minutes.  *Maybe* a minute or two at most.  But from what you say below, I don't want to listen to PA without the OS minimized anyway.


Sorry, I wasn't clear enough.  I don't think changing from the 22W amp to the 100W amp had anything to do with it.  When testing with the 22W amp originally, I thought it must be muting the output, because although the XXHE UI showed everything correctly for PA, I heard no sound.  That's why I thought of trying the 100W amp, and using the N22 as a preamp. 

But now I think it may have been incorrect buffer settings.  4096 didn't work with the 22W *or* 100W amp, 512 worked with the 100W amp.  I don't want to fuss around with wires again to try the 22W amp with a buffer of 512, but my guess is that it would work.

No ticks at all with Special Mode.  With Adaptive Mode and PA engaged, no ticks, but if I halt playback by clicking the Stop button *in the middle of a song*, occasionally I get a fairly good thump, not a tick.  Nothing on a level to endanger the speakers, but, errm, interesting.

Yes, I agree it's substantially better, even on my fairly low-level office setup.  It seems to bring out all the "magic" in the songs, what the artist wanted to communicate - which I suppose translates to exactly what you are saying about realism even at low volume.  No details of intonation, phrasing, breathing are missed, so to the extent the artist controls these, you hear what he or she intended and get the full emotional and artistic impact, even without making it loud.  (Not that I object to making it loud on occasion.    )

I had not been able to get PA to work previously, so for whatever it may be worth, here is what was successful and what was not.

What was not successful:

- Buffer at 4096, SFS of 2.

- Attempts without OS minimized (is setting the OS to minimize part of the PA instructions I managed to overlook?).  If attempting without OS minimized, I would get the dialog about turning off services, and it never went away.

- Also didn't get it to work using the N22 as an integrated amp (around 20wpc, as opposed to the PSE's 100wpc), but actually I don't think this was critical.

What was successful:

- Keeping everything as close as possible to my former settings in KS Special Mode, i.e., buffer of 512, SFS=430.

- Having the OS minimized.

Barry -

It sounds a whole bunch better than Foobar to me right "out of the box," so to speak.  That was true on my laptop, and is now true on my desktop.

What's also true is that if you listen to Peter's advice, and well short of zillion-dollar anything, you can very much improve the "out of the box" sound.  (I don't know what's in the Acer, so I'm not sure how many of the improvements will actually be available to you.  Easy enough to try them and see.)

Yes, stopping the one process that the AV requires a reboot to turn off allows XXHE to shut down the remaining processes on its own.

For future reference by anyone running the ESET AV software, the particular process that requires a reboot to stop is called "HIPS" (Host-based Intrusion Prevention System).  The ESET manual page about it is currently here: http://download.eset.com/manuals/eset_ess_5_userguide_enu.pdf#page=41

Stopping it requires you to open the AV's full UI; go into Setup; under Computer/HIPS click on Enabled, choose to disable, click OK in the AV's UI window, and reboot.