[KeithEmo]

   
If you want to know whether there is a point to DSD, or rather want to disprove supremacy of DSD, you don't really have to compare it to anything at all.
All you need to do is to show whether the current system of digital capture and playback is transparent, which maverickronin's set-up lets you do. Whether the source is a DSD output, tape, vinyl, or directly from a pair of mic pre-amps in a concert venue, if you can't differentiate it from the PCM AD/DA loop, then that's that. You can't improve on transparency. One kind of homeopathic remedy doesn't work any better than another kind of homeopathic remedy.

 
I agree - if you can't differentiate the "original" from the version that's been through "the PCM loop" then you have indeed proven that PCM is "effectively perfect" at reproducing whatever your original source is. However, you seem to be missing a huge piece of the equation, which is that DSD is not "a reference source"; it's simply another way of doing capture and playback - like PCM. All that test will prove is whether PCM is capable of perfectly reproducing a particular source - which, in that case, is simply a DSD copy of a real source.
 
So, yes, if you can't tell the difference between a signal played directly from a DSD source, and one that has then been recorded and played back using PCM, then you have indeed demonstrated that PCM is "perfect enough" to make a perfect copy of a DSD source... or you could say that you've proven that PCM is "at least as good as DSD"... but that is precisely all you've proven.
 
(My point here is that, even if you find out that you CAN hear the difference between the "DSD original" and the "PCM copy", you still have NOT proven that DSD is better; you've simply proven that PCM isn't perfect. And, for the record, I promise you that you WILL be able to measure the difference between a PCM copy and a DSD original, AND between a DSD copy and a PCM original; AND between either a PCM or DSD copy and an analog original; the only question in my mind is whether that differences will be audible.)
 
The problem is that there are a whole bunch of other things we sort of need to know....
 
What if DSD significantly alters the original in a bad way?
If that's the case, then all we're testing is whether PCM causes enough additional damage to be noticeable.
 
What if DSD and PCM both alter the signal a little bit - equally? (Which is what I would expect based on the science.)
If that's the case, then both are equally good as recording formats, but neither is better, and one should avoid converting between them (and so suffering the flaws of both).
 
If you want to prove "the supremacy of DSD", then you need to prove that it reproduces something better than PCM.
Which means that you need to quantify the reproduction quality of BOTH - and then compare them.
 
Another thing you're ignoring is that there is a "downside" to DSD; it is a nuisance to record and edit, and PCM is already the de-facto standard.
So, if it turns out that both are "equally perfect", then, in terms of justification "DSD loses on points".
In order to justify its existence as a REPLACEMENT for PCM, DSD has to demonstrate that it is actually BETTER in some way.
 
 
 

 

[KeithEmo]

   
If a PCM A/D->D/A loop (or DSD->PCM->DSD digital conversion) cannot be distinguished from the analog (or DSD) original in extensive blind testing, then that already strongly suggests that PCM is indeed good enough for audibly transparent reproduction. What would be the point of DSD making less than already zero audible difference to the sound, then, considering the practical disadvantages (such as more difficult and computationally expensive DSP) compared to PCM, and the less efficient usage of a given bit rate ? DSD64 requires 4 times as much space as Red Book, which could spent on more channels - a real, audible improvement - and still better than CD quality sample rate and resolution.

 
I agree entirely.... and I suspect that will turn out to be the case.
 
However, my point is that, even if you CAN hear a difference, that won't prove that DSD is superior... only that PCM isn't perfect.
(I don't want to let the DSD fans "set this up" such that, if there is an audible difference, it seems to automatically prove that DSD is not only different but better.)

 

[KeithEmo]

   
Which one was accurately doing the d/a conversion? They can't all be performing to spec if they all sound different.
 
It's much more likely that high end DACs are jury rigged to alter the sound than midrange ones. Try comparing to the line out from an iPod or a decent blu-ray player and see which one matches those. Those are most likely to be accurate. I have an Oppo HA-1 and iPods and iPhones and Macs and an Oppo BDP-103D and a low end Sony blu-ray player... all of those sound exactly the same. If I bought something that wasn't audibly transparent with all of them, I would start worrying and would want to know which component was performing out of spec.

 
You are incorrect - but only because you're considering a specific list of specs (which, in the case of DACs, don't usually include all of the actually relevant ones).
 
Frequency response, signal-to-noise ratio, THD, IMD, and all the other "normal specs" are all what we call "steady state" measurements; they are measured using a continuous sine wave. As it turns out, this works very well for analog devices, because the types of errors introduced by flawed analog circuitry tend to show up when you test them this way. (Transient Intermodulation Distortion was the first type of distortion that was noticed in analog circuitry which "doesn't show up" when you test a circuit this way - and it turns out that you can detect it from the other measurements if you know how to interpret them.) 
 
In contrast, because they work on entirely different principles, digital circuits like ADCs and DACs tend to produce other types of errors - some of which ONLY show up when you use transients or other specific types of test signals (other than the sine waves used for the "standard measurements"). I can show you two different DACs, both of which have "normal specs" that are so good that it's a virtual certainty that the flaws present are inaudible (THD < 0.003%; frequency response flat to a small fraction of a dB; S/N ratio of 130 dB or so), yet they still do sound different. And, yes, if you look at their outputs on an oscilloscope, using a transient-type test signal, you will see obvious differences in the output waveform.
 
This is NOT some magical difference that people imagine is there; it's real and you can see it very easily on an oscilloscope - the only question is whether it is audible or not. And, yes, if you look at the output of your iPod reproducing a signal that includes transients (and not just steady state sine waves), you will find that what comes out is considerably different than what goes in... it's nowhere near "perfect". And, yes, it is also true that some vendors do deliberately "voice" their "high-end" products to sound distinctly different than everybody else (and those differences are easily measured).
 
(If you want to argue about which one is "right", the reality is that none of them is precisely correct, and the errors, while small - are different. If the digital oversampling filter is properly designed, then all of the energy is there, in the correct proportions, which gives you a near-perfect frequency response, but some of that energy is shifted slightly away from the precisely correct time. If you were to look on a spectrum display, the signal would look perfect; however, if you look at the actual waveform on an oscilloscope, it will be different... and the different filters do sound slightly different with certain content. Note that we're talking about slight differences here, to the point where they may only be audible with certain recordings, and with certain speakers or headphones - if the recording or the output device isn't especially accurate at reproducing transients to begin with, then you aren't likely to hear differences between them.)
 
However, since you mentioned it, I do have to ask you a question....
 
If, as you say, you really hear NO difference between your Oppo HA-1 and your iPod, then WHY did you spend $1200 for the HA-1?
 
 
 

 

[Steve Eddy]

I have auditioned this system about two years ago - and it is SUPERB - near field monitor.

How can it be a "SUPERB" near field monitor with an offset tweeter like that? An offset tweeter is anathema to listening near field. For near field, you really want a point source. Looks like he put all his attention into the enclosure but took his eye off the ball on some equally important details.

se

 

[Steve Eddy]

Frequency response, signal-to-noise ratio, THD, IMD, and all the other "normal specs" are all what we call "steady state" measurements; they are measured using a continuous sine wave. As it turns out, this works very well for analog devices, because the types of errors introduced by flawed analog circuitry tend to show up when you test them this way. (Transient Intermodulation Distortion was the first type of distortion that was noticed in analog circuitry which "doesn't show up" when you test a circuit this way - and it turns out that you can detect it from the other measurements if you know how to interpret them.) 

You really need to brush up on modern test and measurement capabilities. And then of course there's also difference testing which uses actual music signals as its inputs.

se

 

[KeithEmo]

  Type of filtering is the core of the Sygnalist HQ player http://www.signalyst.com/consumer.html . You can choose pre-ringing, post ringing - etc - according to the known defects of CDs that were recorded on known equipment producing these undesirable effets - and software will add exactly the opposite pre or post- ringing, effectively restoring the desired response. It takes one hell of a computer to do it right ( IIRC - 12GB RAM is minimum requirement ) - actually, they offer a music-only computer, optimized to cover for any setting that may actually tax that same computer to the max.
 
There is one fly in that ointment - you have to know on which machines your CDs have been made - and which "correction" filters are to be applied. About as useful and widespread as various non-exactly-RIAA-equalization curves for vinyl - some are known, most aren't.
 
Enter the simplicity of the DSD512 - NO filtering required anymore, no guesstimating which equipment has been used to produce your file. At least in digital domain - analog sections will still be the limiting factor and will ultimately set the SQ. Digital can be dirt cheap - analog most definitely not - if it wants to be really good. For fact - not in mfr's brochures or internet only.

 
I played with HQPlayer for a while; even on my Quad Core Dell desktop machine some of the longer filters were unable to run without problems.... but it is a cool idea.
 
However, I do disagree with a few of the claims you attribute to it....
 
First, ordinary DACs ALWAYS oversample, and so ALWAYS apply oversampling filters. In order to avoid this entirely, you would need to use a non-oversampling DAC (which have other "issues"). (Signalist used to recommend buying or building a NOS DAC to use with HQPlayer if you wanted to avoid this.) However, by upsampling a 44k signal to 192k in software, then feeding that 192k signal to the DAC, you are allowing the DAC to operate in a mode where it can hopefully use gentler filters that have less effect on the audible signal.
 
Second, while it's possible to correct for certain types of production flaws, some of the damage or alteration caused by an aggressive filter is irreparable. So, while they may be able to "improve" content recorded using certain ADCs with certain filters, they cannot "make badly damaged content perfect again". It's more like sharpening a fuzzy picture; if you do it well, you can often improve things, sometimes dramatically, but it isn't "perfect".
 
Third, your statement about "no filtering" being required for DSD512 is incorrect. The fact is that, because of the very high sampling frequency, the filtering required is at such a high frequency that it can be very gentle, and so should be totally innocuous in terms of damaging the audible portion of the frequency spectrum - but it isn't "not there". You also seem to be under a slight misapprehension about how difficult it is to design a DAC. A current high-quality DAC chip costs somewhere between about $2 and about $20; and it's already the supporting circuitry - things like the power supply and analog sections - that take the most design effort. In short, producing a DSD512 DAC isn't going to be especially cheaper or easier than the current designs. Of course, the huge downside to DSD512 is that, since none of the current studio mastering equipment or playback equipment supports DSD512, that would mean replacing everything.
 
The other thing you need to do is to maintain perspective between the various "levels" of the different formats. DSD64 is approximately equivalent to 24/96 PCM, and DSD128 is pretty close to 24/192 PCM. Therefore, if you want to reasonably compare DSD512, you should be comparing it to 24/384 PCM, or perhaps 24/768 PCM - which offer similar frequency response - and require similar bandwidth and storage space. (In other words, DSD512 is NOT some magically easy way to get wonderful sound quality with very little effort. It's just another alternative, which has its own strengths and weaknesses, and which is going to have to justify itself before anyone will bother implementing it.)
 
Without belaboring the point, look around a bit and read the various discussions about why SACD was a resounding commercial failure - because most of those reasons apply even more to DSD512. (Instead of comparing CDs to SACDs, use the same types of comparison between 24/384k PCM and DSD512.) DSD is difficult to record, difficult to edit, takes up at least as much space as equivalent quality PCM, and hasn't shown any benefits sufficiently compelling to overcome those obstacles.

 

[analogsurviver]

How can it be a "SUPERB" near field monitor with an offset tweeter like that? An offset tweeter is anathema to listening near field. For near field, you really want a point source. Looks like he put all his attention into the enclosure but took his eye off the ball on some equally important details.

se

Sorry, this speaker IS superb - and its LED system for pointing the speakers EXACTLY right is extremely narrow - if you can't see it, BOTH from the listening position, you will not get the correct reproduction.
 
Please read again trough it all, watch the videos - better yet, audition the system yourself.  It is one of the least coloured speakers ever made.
 
Now - to be fair - how many nearfield monitors you know that actually are point source? Outside Tannoy ?
 
That is why I would love to see the driver from this ( one can see that they worked much harder than usual on the cabinet, curved and mirror-image asymetrical -  measurements for this speaker are superb, but I have yet to see and hear it in flesh )
 

 
packaged into enclosure like that 
 

With the hole for the tweeter faired over - - if not actually made specifically for the Technics coaxial driver.
 
The result should be really, REALLY good - specially if supporting amplifier would, again, ensure the pairing of the left and right channel acoustic output not to exceed 0.5 dB - as in all "Eggs".
 
I DO love point sources, from the diminutive Visaton FRS-8 full range 80 mm driver ( mounted in spheres if required...) trough Technics SB-RX50 (which is the predecessor from the current reincarnation, mid 80s, normal "box" ) - and use them on regular basis. 
 
A coax in egg shaped resonance and diffraction free enclosure, positioned atop some "woofer" that also eschews diffraction as much as humanly possible (or some narrow stands + subwoofer ) , should be capable of producing sound of great precision - one that has not been achieved yet.
 
I have auditioned the original sE Egg 150 - and that was the most uncoloured thing I ever heard in pro audio, offset tweeter or no offset tweeter. Imagine what described above could achieve...

 

[Ruben123]

I like this debate, but more important is if the differences are audible and then how much. Or is it placebo. Or does "audiophile" gear maybe have some sort of tweaked DAC to actually REALLY sound different to let audiophiles hear the difference between those and "consumer grade", correctly adjusted DACs?

 

[limpidglitch]

   
So, yes, if you can't tell the difference between a signal played directly from a DSD source, and one that has then been recorded and played back using PCM, then you have indeed demonstrated that PCM is "perfect enough" to make a perfect copy of a DSD source... or you could say that you've proven that PCM is "at least as good as DSD"... but that is precisely all you've proven.

 
No, you've demonstrated that PCM is audibly perfect, and DSD is at best equally as good. As such the debate is rather pointless really. 
 
 

   
(My point here is that, even if you find out that you CAN hear the difference between the "DSD original" and the "PCM copy", you still have NOT proven that DSD is better; you've simply proven that PCM isn't perfect. And, for the record, I promise you that you WILL be able to measure the difference between a PCM copy and a DSD original, AND between a DSD copy and a PCM original; AND between either a PCM or DSD copy and an analog original; the only question in my mind is whether that differences will be audible.)

 
Let's worry about that after the PCM DBT fails.
Until then we can discuss what kind of weapons we want to use to fight the coming zombie apocalypse. 

 

[Steve Eddy]

Sorry, this speaker IS superb - and its LED system for pointing the speakers EXACTLY right is extremely narrow - if you can't see it, BOTH from the listening position, you will not get the correct reproduction.

That's fine. I don't like listening with my head in a vice to keep it perfectly positioned in order to avoid the lobing effects that are inevitable with an offset tweeter design like that.

Please read again trough it all, watch the videos - better yet, audition the system yourself.  It is one of the least coloured speakers ever made.

Will you provide the vice?

Now - to be fair - how many nearfield monitors you know that actually are point source? Outside Tannoy ?

Not many. Which is why I can't help shaking my head at the plethora of so-called near field monitors with offset tweeters. It's basically a contradiction of terms.

That is why I would love to see the driver from this ( one can see that they worked much harder than usual on the cabinet, curved and mirror-image asymetrical -  measurements for this speaker are superb, but I have yet to see and hear it in flesh )

That would certainly be a step in the right direction. But I'm sorry, I just can't take seriously anything that calls itself a near field monitor but uses an offset tweeter.

I have auditioned the original sE Egg 150 - and that was the most uncoloured thing I ever heard in pro audio, offset tweeter or no offset tweeter. Imagine what described above could achieve...

You're aware that cabinet diffraction can be (and has been) effectively dealt with using equalization networks, yes?

se

 

[analogsurviver]

   
I played with HQPlayer for a while; even on my Quad Core Dell desktop machine some of the longer filters were unable to run without problems.... but it is a cool idea.
 
However, I do disagree with a few of the claims you attribute to it....
 
First, ordinary DACs ALWAYS oversample, and so ALWAYS apply oversampling filters. In order to avoid this entirely, you would need to use a non-oversampling DAC (which have other "issues"). (Signalist used to recommend buying or building a NOS DAC to use with HQPlayer if you wanted to avoid this.) However, by upsampling a 44k signal to 192k in software, then feeding that 192k signal to the DAC, you are allowing the DAC to operate in a mode where it can hopefully use gentler filters that have less effect on the audible signal.
 
Second, while it's possible to correct for certain types of production flaws, some of the damage or alteration caused by an aggressive filter is irreparable. So, while they may be able to "improve" content recorded using certain ADCs with certain filters, they cannot "make badly damaged content perfect again". It's more like sharpening a fuzzy picture; if you do it well, you can often improve things, sometimes dramatically, but it isn't "perfect".
 
Third, your statement about "no filtering" being required for DSD512 is incorrect. The fact is that, because of the very high sampling frequency, the filtering required is at such a high frequency that it can be very gentle, and so should be totally innocuous in terms of damaging the audible portion of the frequency spectrum - but it isn't "not there". You also seem to be under a slight misapprehension about how difficult it is to design a DAC. A current high-quality DAC chip costs somewhere between about $2 and about $20; and it's already the supporting circuitry - things like the power supply and analog sections - that take the most design effort. In short, producing a DSD512 DAC isn't going to be especially cheaper or easier than the current designs. Of course, the huge downside to DSD512 is that, since none of the current studio mastering equipment or playback equipment supports DSD512, that would mean replacing everything.
 
The other thing you need to do is to maintain perspective between the various "levels" of the different formats. DSD64 is approximately equivalent to 24/96 PCM, and DSD128 is pretty close to 24/192 PCM. Therefore, if you want to reasonably compare DSD512, you should be comparing it to 24/384 PCM, or perhaps 24/768 PCM - which offer similar frequency response - and require similar bandwidth and storage space. (In other words, DSD512 is NOT some magically easy way to get wonderful sound quality with very little effort. It's just another alternative, which has its own strengths and weaknesses, and which is going to have to justify itself before anyone will bother implementing it.)
 
Without belaboring the point, look around a bit and read the various discussions about why SACD was a resounding commercial failure - because most of those reasons apply even more to DSD512. (Instead of comparing CDs to SACDs, use the same types of comparison between 24/384k PCM and DSD512.) DSD is difficult to record, difficult to edit, takes up at least as much space as equivalent quality PCM, and hasn't shown any benefits sufficiently compelling to overcome those obstacles.

I agree on most of points you made. Of course, comparisons among "similar size file" or "levels" of DSD and PCM should be considered.
 
No thing is magically better than the other - the reason why DSD is superior to PCM will FOREVER be in the pulse response - unless PCM is made many times the size of the DSD file in order to actually start approaching the pulse capability of the any chosen level of DSD. The negative side of DSD vs PCM is of course noise rising above 20 kHz - which gets reasonably low enough by DSD512.
 
I agree - it requires the change of everything. And in this bush the rabbit that would like to proclaim "perfect sound forever" is really hiding - trying to prolong the CD redbook to as much as possible. There is not a single device (except maybe mic cables ) in a 48/24 capable studio that would not represent a significant bottleneck in say DSD256 environment - if not preventing doing recording at all.
 
DSD512 capable DACs exist - and start at approx 500-600 $ (depends where you live ). DSD512 commercial recorder, at least not to my knowledge, does not exist - yet. There is DSD256 recorder available - from Merging (Horus and Hapi) and from Weiss (IIRC). I will have to look into it a bit more if any other has also appeared.
 
I can not afford anything above DSD128 at the moment - yet if and when I decide to bite the bullet, it will be for DSD 512. Listening to the samples I posted a few posts back is convincing enough - if DSD256 can be that better compared to DSD128, I have no doubt that DSD512 would be even better - a little, audible on the best of equipment, if implemented correctly. I sure prefer DSD256 or even DSD128 over DSD512 that "stutters" for any reason.
 
There are a few DSD512 test tracks available - arrived at by upconverting DSD256. DACs that can play them are available - now and for reasonable price. If money is no object, there is always http://www.enjoythemusic.com/superioraudio/equipment/0814/gryphon_audio_designs_kalliope_dac.htm  I *guess* I will never even see it in flesh - but it does exist.

 

[KeithEmo]

You really need to brush up on modern test and measurement capabilities. And then of course there's also difference testing which uses actual music signals as its inputs.

se

 
I assure you I'm quite up to date on what can be measured - and on what usually is. Our AP test sets here at Emotiva do an excellent job of testing transient performance - amongst a lot of other things. However, I'm sort of missing your point here. The comment was made that "all good DACs measure and sound the same" - to which I replied that they only all measure the same if you only use the basic steady state tests.... and it doesn't appear that you disagree with me on that point.
 
Doing "difference testing" with a music source is not that difficult - but evaluating the results is... which is why most "basic tests" use a sine wave as their source. (In the old days, it was very easy to simply null out the original sine wave and then measure the total of anything that was left. On modern equipment, it's also easy to look at the spectrum of the output, and also easy to measure and calculate things based on the knowledge that anything other than the line that represents the pure sine wave doesn't belong there. With music as a source, you have to either analyze the full complex spectrum of the music, then do the same for the output, then figure out what types of errors are causing which differences, or you have to look at the two waveforms and determine the differences that way.... which you can do by nulling them against each other....  at which point you still have to analyze your difference/error waveform and figure out what the differences actually mean. Unfortunately, some differences in waveform that are visibly obvious are still inaudible, while some highly audibly forms of distortion are barely visible on the waveform.)
 
Back to DACs, though.... with an decent oversampling DAC, you will end up with a flat frequency response because the overall energy response is very accurate, and - at least with a sine wave - you will also end up with very low THD, yet you will end up with visible differences in transient or irregular waveforms due to ringing. (And which tests they will show up on, and which they won't, should be rather obvious once you understand how the tests work and what they're testing.) However, whether specific amounts and types of ringing are audible at all, and, if so, whether certain types "sound better", still seems to be a matter of debate, although the currently accepted "wisdom" is that post-ringing is more innocuous than pre-ringing with most content (which is why Dolby's latest encoder includes an option to upsample while "reducing pre-ringing at the expense of adding more post-ringing" as a way to "improve sound quality using post-processing"), and it also seems likely that certain individuals simply find one or the other more or less audibly annoying.

 

[lamode]

  No thing is magically better than the other - the reason why DSD is superior to PCM will FOREVER be in the pulse response - unless PCM is made many times the size of the DSD file in order to actually start approaching the pulse capability of the any chosen level of DSD. The negative side of DSD vs PCM is of course noise rising above 20 kHz - which gets reasonably low enough by DSD512.

 
This is incorrect as the impulse in both cases is a result of the low pass filter, and not the format itself. Furthermore, perfect impulse response (which neither system achieves) is meaningless as it requires infinite bandwidth which is totally unnecessary for human ears.
 
DSD does have other disadvantages such as greater distortion and noise and of course incompatibility with EQ, which is a total deal breaker, imo.

 

[analogsurviver]

That's fine. I don't like listening with my head in a vice to keep it perfectly positioned in order to avoid the lobing effects that are inevitable with and offset tweeter design like that.
Will you provide the vice?

Not many. Which is why I can't help shaking my head at the plethora of so-called near field monitors with offset tweeters. It's basically a contradiction of terms.
That would certainly be a step in the right direction. But I'm sorry, I just can't take seriously anything that calls itself a near field monitor but uses an offset tweeter.
You're aware that cabinet diffraction can be (and has been) effectively dealt with using equalization networks, yes?

se

That is humo(u)r - at its worst. 
 
You do not have a SINGLE clue what a true audio "vice" is. Compared to that, admittedly small area provided by Egg is a VAST space.
 
Enter big Audiostatic ESLs. I mean BIG (900ES - IIRC) . Imagine a LARGE listening room - some at very least 15 m long and 8 metres or so wide. At a fair - with rows and rows of seats , capacity over 100 people.
 
With ONLY the seats EXACTLY in the middle axis between the speakers occupied (and vigorously "fought for") - because anybody who sat one single seat removed from the centerline discovered in less than five seconds the distributor could well spare the trouble of bringing all those seats - except those 10 or so exactly on the centerline. 
 
It gets worse - speaker was so directive you were penalized by "outside heaven/nirvana" performance the minute you dislocated your head by more than an inch or so - even if seating 10 m away from the speakers. Bur, true, if you managed to discipline yourself (most helped themselves by supporting their head with arms...) , the award was one of the best reproduction I have ever heard. Off > one inch - catastrophy...
 
So, in this case, I rather endure quality headphones and cable - at least can move my head a bit...
 
This is an audio vice - hors categorie.
 
I am aware that cabinet diffraction can be dealt with equalization networks. 
 
The same goes for resonances within the cabinet itself. If there are no standing waves - or are very well dispersed with none being particularly large (high Q) - whatever remains can be much more easily dealt with than an regular rectangular box that has to rely solely on the damping material. Heavy damping will always have reduced dynamic range as a consequence - even if it achieves reasonably flat measured frequency response - but will be audible on well recorded material.
 
I hope you realize prevention is better than cure ?

 

[KeithEmo]

I like this debate, but more important is if the differences are audible and then how much. Or is it placebo. Or does "audiophile" gear maybe have some sort of tweaked DAC to actually REALLY sound different to let audiophiles hear the difference between those and "consumer grade", correctly adjusted DACs?

 
I agree.... and the reality is almost undoubtedly "all of the above". Some "audiophile DACs" most definitely ARE tweaked to make them sound "special", and frequently some of the filter choices on units that offer multiple filters frequently DON'T have a flat frequency response, so of course they sound different. Non-oversampling DACs are another example; a typical NOS DAC has a high-frequency roll off of about - 3 dB at 20 kHz, which makes an obvious difference in how it sounds, which in turn makes it very difficult to judge how much, if any, difference the NOS topology itself makes. I have several DACs which offer a choice of filters, and, while each filter sounds slightly different with some content, it's not as easy to say that one or the other is "better" - they're just slightly different. And Sabre DAC chips are most certainly "tweaked" to produce their characteristic flavor - which some people love and others don't.
 
Placebo effect is also a major factor with lots of audio gear.... which includes DACs.
 
Also, to put this in proper perspective, the differences between most decent DACs are FAR smaller than the differences between different speakers, or headphones, or phono cartridges. This means that they aren't likely to be significant unless you first notice them, then decide that they are specifically significant to you. They also tend to be the type of differences that you only notice with certain source material, and with certain headphones or speakers.