The Holo May states it could do DSD 512 in itās manual, you are right. Does @sridavid can maybe tell us more about it if he still got this unit?
The May manual does say DSD512 over DoP, though I wonder if that is exclusive to Windows, or Windows and Linux. There are a number of DACs that will do DSD1024 over native DSD to Windows, or Windows and Linux, but are limited to DSD256 over DoP with MacOS. Iām supposing itās a matter of driver/kernel support on Windows and Linux that is lacking in MacOS Core Audio.
In the same post, sridavid says he has no problem with DoP512 on a Mac with a Denafrips Ares, so that would answer affirmatively regarding DSD512 on MacOS, and be a hopeful sign regarding DSD1024 on Windows/Linux.
I want to see the results of any listening tests using multiple subjects (subjective or double-blind) that will provide definitive evidence that modulating PCM to DSD512 or DSD1024 provides a tangible level of auditory/sensorial perception in contextual harmonic and dynamic energy; in direct comparison to modulating PCM to DSD128 or DSD256 as auditioned through the average audiophile playback systemā¦
Personally speaking, I donāt want my DAC running at the red-line of its performanceā¦ I want the headroom of running at cruising speed.
I donāt think that people who prefer listening to PCM modulated to DSD512 or DSD1024 do this via a set-up which is comparable to the āaverageā audiophile playback system.
So in this point you might be right.
Matt
Iād even like to see a multi-subject study (subjective or double-blind) done on one of the most sophisticated playback systems in existence, that can provide definitive evidence of tangibly perceptible, contextual harmonic and dynamic energy in modulating DXD to DSD512 or DSD1024 as directly compared to DXD modulated to DSD128 or DSD256ā¦
High End listening is not about scientific studies. It is the subjectivity of an individual who prefers listening to the highest DSD rates in comparison to lower rates and finds pleasure in it. If you go over to WBF you will find a lot of these people. It is similar to drinking coffee on high end espresso machines or wine tasting.
Matt
Thatās okay and I understand the nature of that sort of cognitive-biasā¦ However, when these sorts of things are used as marketing points, then I want to see evidence of the rationale for the existence and applicationā¦ I enjoy a fine wineā¦ if I get a chance to indulgeā¦ however, Iāll need a consensus or a personal experience, before I buy into the hypeā¦ Now my feelings are different when a native DSD recording is modulated to DSD512, but the catalog of these products is not wide enoughā¦ so the limited DSD512 products become āshowpiecesā for the technology, and are far from mainstream.
We all here on this and other forums are far from mainstream. Some people do offline upsampling of ordinary files in order to create new ones in the GB area which need a lot of storage space: A toast to PGGB, a heady brew of math and magic - Software - Audiophile Style
This month I will visit High End Munich which is also far from mainstream. It is just interesting to listen to these audio systems, to get new impressions and to compare to my own set-up.
Matt
Yes, this is also my opinion. I canāt make such a test, maybe never, I donāt have such a DAC, donāt know anybody who has and also a place (store?) when can test. Anyway, in a public place canāt make such tests because: 1. Need to have a very good special mood 2. Proper files 3 Proper software 4 At least one hour 5 To not be disturbed by anything/anybody. Itās not like mp3 vs flac. Until that I think itās (just) marketing. Itās a consume society. They need to sell and sell and sell.
Plus I donāt have an absolute hearing and perfect perception. So even if it is a difference maybe I will not notice. So why to waste my money? Just because others, maybe paid to say something or with a better perception than mine or not such pro think itās better? To be in ātrendā?
And then there will be the tedious remasters of the remasters of the remasters!
Iād like to see the multi-subject study (subjective or double-blind) comparison of PCM and DXD up-sampled by PGGB, directly compared to PCM and DXD modulated to DSD128 and DSD256ā¦
When I speak of āmainstreamā it is in the context of the high-end audiophile experience.
I am not fond of offline upsampling like PGGB and mentioned it only because I think it is even more esoteric and outside āmainstreamā than on-the-fly upsampling to higher rate DSD.
I also mentioned that I donāt think that you get any consensus about SQ with āscientific studiesā, but YMMV.
Matt
Well, without personal experience, a consensus derived from a variety of folks with various levels of hearing acuity and listening skills, brings us closer to a reality less fettered by cognitive-biasā¦
Maybe, but that doesnāt matter much for your personal listening behavior. In the end you choose the software and hardware components of your audiophile system and listen to the music you like. Even when a study claims that there is no difference but you hear one it might be very important for you. Further, studies are like fashion, some change every season, good luckā¦
Matt
I agree with you in the subjective experienceā¦ in a true double-blind study, there is little room for cognitive-bias and agenda to creep into the resultsā¦
Thereās real science on human hearing thatās informative, and can also be a lot of fun if itās the kind of thing that interests you (which it does me).
First, we have to realize that we hear quite differently from each other, not only in matters of acuity like high frequencies and low volumes, but very fundamental things such as whether a passage is ascending or descending in pitch or what notes are playing in which channel. See https://deutsch.ucsd.edu/psychology/pages.php?i=206 and https://deutsch.ucsd.edu/psychology/pages.php?i=208 . Thus the idea of arriving at objective consensus via listening tests by multiple people doesnāt necessarily work, because the minority isnāt wrong, theyāre actually hearing something different.
The other very problematic thing with A/B listening tests, blinded or not, is that echoic memory (our memory for sounds, other than things like a loved oneās voice that youāve heard many times - thatās pattern recognition, which weāll get to) has been confirmed by multiple experiments over decades to last about 4 seconds. And the very worst circumstance in which to try to retain echoic memory is when hearing a second series of notes right after the first one youāre trying to remember - in other words, exactly the classic A/B test. See https://deutsch.ucsd.edu/psychology/pages.php?i=209 . If we canāt even recall what note was played after 4 seconds, how well do you think we retain the kinds of subtle differences between DSD128, 256, 512, and 1024?
So then how do we recognize a loved oneās voice over a bad phone connection, a melody transposed to a different key, etc.? Pattern recognition. We are absolute fiends for pattern recognition, much better at things like facial and voice recognition than even the best current AIs. Our brains love to do pattern recognition so much that we often impose patterns where there are none, and so we have optical and audible illusions.
How do we get from the impossible task of trying to retain echoic memory for more than 4 seconds to the very natural and easy recognition of a pattern? Practice/training, hundreds, perhaps thousands, of hours of it. (Thus the famous ā10,000 hours to become an expert.ā) So a musician might have a good idea of whether a system provided an accurate rendering of an instrument theyāre intimately familiar with. A jazz musician I know once discerned that a 7 foot rather than a 9 foot piano was being used on a recording heād never heard before, for example.
Now I donāt have such abilities. So what I do is check where my DAC or similar ones measure best, which happens to be at DSD256 or DSD512. Whether I can really hear the difference I donāt know, but why not start out at the best place possible?
I have been a student of psychoacoustics for decades and the psychology of perception/hearing neurology and cognitive-bias on these thingsā¦
This trueā¦ however they complete the spectrum of potential influences that may bias the studyā¦
Iāll give you my experience in the 90ās as a subject for Dr. Floyd Tooleās and Sean Oliveās double-blind study apparatus at Harman International / JBLā¦
Floyd had devised a listener curation test which included identifying fixed frequencies at different levels and harmonic distortion covering various spectra and level as well as dynamics sensitivityā¦ It took about 30 minutes to complete and we were not given any clues or multiple choicesā¦ I aced it to his amazement, and was placed in the center-chair position on the qualified listening panel which included other audio professionals/audiophiles.
We were evaluating speakers that were placed on a mechanical carousel-type exchanger, which was randomized automatically, completely unknown to the testers all gain adjustments were done automaticallyā¦ This is done completely in the darkā¦ all music material was specifically chosen for specific attributes or lack thereof, and this is randomizedā¦ We were only given a label for each test event 1A, 2B, 3C, 4D, etc, and we were to choose which event we preferred, the array of speakers and note it on a formā¦ Even after the test, we were not given our individual results in comparison and we were only given the brand/model of the speaker(s) as they ranked in preferenceā¦ These speakers ranged in price from inexpensive to ultra expensiveā¦ I donāt remember today what group of speakers were being evaluated in this test, I think the most expensive was B&W.
We all come to the listening experience with biases, only through a process of filtering these biases in a qualified test/study will we get closer to a ātruthāā¦ What we prefer, is what we prefer and this is what counts in the final assessment.
Yes, Iāve been very interested in the Toole/Olive studies.
In at least one study, they had groups of ātrainedā and āuntrainedā listeners. Even the āuntrainedā listeners were asked to read material online that indicated speakers which had flattest frequency response across the audible range were superior. The ātrainedā listeners were given further instruction along the same lines. Unsurprisingly, both untrained and trained listeners indicated a preference for the speakers that had flattest frequency response across the audible range.
Now a preference for flatter frequency response might seem perfectly unobjectionable. But hereās something interesting: The speaker brand Iāve had for decades, Vandersteen, uses time-aligned and phase-aligned drivers in pursuit of better imaging. In order to do this, Vandersteen crossovers are linear phase. In order to be linear phase, the crossovers must be āfirst order,ā and first order filters donāt cut much. This causes an unavoidable bump in frequency response at the crossover point. Of course Vandersteen designs their speakers to minimize such bumps as much as possible, but theyāre there, more so than with speakers explicitly designed for flattest frequency response.
Now whether from initial preference, long listening experience, or both, I love the imaging of Vandersteens and donāt hear any objectionable frequency response humps. Does this make me wrong and the folks in Oliveās and Tooleās studies correct? Or are we just hearing differently? 
There is such a thing as āinherent biasāā¦
It was interestingā¦ I identified ālobesā at the cross-over points on every speaker we listened toā¦ none were āpoint sourceā as it turned-out and Floyd, Sean and I spent the rest of the afternoon discussing this over lunch. The Toole studies are hard to swallow for a lot of audiophilesā¦ 