This report will be (relatively) brief, for two reasons. First, I've covered in earlier MAN articles the underlying philosophy that Hugh Dean applies to the design of his amplifiers, including most recently his revised flagship amplifier, the Maya (Issue 565, September 2022), and the story of his much modified floorstanding speakers (Issue 580, December 2023).

Second, I could not stay to the end of Hugh's presentation, as a local arborist was scheduled to appear at my place at 7 AM the next morning. He was to be there to help m deal with the consequences of the ferocious storm of 20 February. I was far from alone in having been affected by that short and savage event, losing mains power (and therefore, hot showers) for five days and, with even more ominous possibilities, the pressurised sewerage that deals with gray and black water from places such as mine up in the hills.

Hugh's experience in designing, building and selling audiophile amplifiers goes back nearly 30 years under the Aspen Amplifier moniker. The story starts with a 28-watt valve hybrid amplifier (the Glass Harmony ) from 1996, then four years later the solid-state AKSA 55, followed by the Lifeforce, Soraya, NAKSA, Titan (thee model demonstrated to the club in August 2022) and, most recently, the Maya. The Maya is a 200-watt MOSFET Class AB amplifier. It first saw the light of day in 2015, and the one demonstrated last month (and in the December DIY meeting) was its fourth iteration. Cost is $6,800 or $3,200 if you buy the individual modules and assemble it yourself. The three-way, floorstanding speakers I won't describe, as I covered them in the January issue of MAN.

The amplifier was fed by high-resolution digital files, via a HP laptop using JRiver software and an SMSL D6S DAC, which uses the high-end ES9029 chip from ESS. The DAC costs only $370-400 online, depending on supplier, and at that meagre price it again makes you wonder how other companies using the same chip can justify charging 10, maybe 20 times, as much for essentially the same device. For those interested in such things, the DAC also supports MQA, DSD and DoP64. It has balanced and normal (i.e. RCA) inputs and outputs. Via USB it will handle files up to 768 Hz at 32 bit depth. And this provides a neat segue into the central theme of Hugh's presentation (and to my report): a focus on high-resolution files as the musical food for the set up. Unlike many earlier evenings, the demonstration made routine use of high-resolution files and Hugh was explicit in his reasoning for using such sources rather than, say, the 44.1 kHz 16 bit files standard to CDs.

Patricia Barber, a singer beloved by many audiophiles for her superb recordings as well as her sensitive interpretation of jazz classics, started the night with Hugh playing her track "Too rich for my blood". He pointed out it was a 96 kHz 24 bit file - and that such resolution comes only with the expense of large data files, in this case, 332 MB. Whereas such large files would once have caused computers and their hard drives much indigestion, perhaps even a bout of choking, they are no longer a problem for modern large-capacity data storage systems. Nevertheless, 96 kHz 24 bit is comparatively miserly in its storage requirements compared with some other high-resolution formats: 1 minute of music at 96 kHz 24 bits requires 33 MB; a doubling of the sampling rate to 192 kHz demands twice as much space, and in between is DSD64 at 41 MB per minute. But dwarfing all these is DXD (352.8 kHz 32 bit), which requires 121 MB per minute. Ouch: the 45 minutes of The Doors' L.A. Woman at that resolution would require 5.5 GB of storage. We are not taking 51/4 inch floppy discs with a storage capacity of 360 kB any more (more on this later). Note that the humble CD requires only 10 MB of storage for a minute's worth of musical pleasure (or pain).

I see that "Too rich for my blood" came off Patricia Barber's Cafe Blue album, which was released in 1994. The year of release made me wonder whether it had been recorded natively at 96 kHz 24 bit or whether the original had been merely up-sampled by the distributor in order that the appropriate 'high-resolution' indicator was illuminated in the DAC when it was replayed. This type of digital trickery - dare I call it dishonesty? - is far from rare. The audiophile magazine Hi-fi News and Record Review routinely publishes Paul Miller's laboratory analyses of hi-res downloads and in many cases, even of contemporary releases, the supposed resolution bears no relationship to the actual (native) resolution of the recording. Too often the putative 'hi-res' file is merely an up-sample from a far lower resolution original, and in the worst cases the files are inept compilations (within the one track!) of multiple resolutions and, presumably, bit depths. In the January 2024 issue, for example, the analysis of Joe La Barbera Quntent's World Travellers indicated that most tracks were genuine 96 kHz recordings but Track 1 was an up-sample of a 48 kHz original. Similarly, Stephen Isserlis' A Golden Cello Decade 1878-1888, while listed as 192 kHz, has parts that are upsampled from 96 kHz originals (April 2023 issue).

Matters seem to have improved markedly over recent years, as a couple of reviews I have from a few years back show how commonplace was such up-sampling before downloading. A January 2019 analysis in Hifi News and Record Review of Connie Han's Crime Zone, for example, concluded that "while this download will certainly register as '96kHz' [sic] on your rendering software/ DAC, tracks 3 and 5 are evidently upsampled from 44.1kHz [sic]". Ditto for a February 2018 analysis of Gregory Porter's Nat "King" Cole & Me: "So forget the LED on your DAC or proud boast of '96kHz' [sic] on its display, for this file is obviously an upsample" (emphasis in original).

A peer-reviewed overview of the history of high -resolution recording (Melchior 2019) notes that 44.1 kHz was the near-standard protocol in 1991, but by later that decade 96 kHz 24 bit had started to be used. Given Patricia Barber's audiophile credentials, it's not unreasonable to believe that she did, in fact, record Cafe Blue at a higher-than-CD resolution. If so, she would have been one of the first to use the technology.

The usual web-based sources (e.g. Discogs) provided no information on this matter, but I did see that when released Cafe Blue was available also as a gold-plated HDCD. That it was an HDCD suggests that the original was recorded at the CD standard of 44.1 kHz and 16 bit. Certainly in 1994 there would have been no way to make commercially available a 96 kHz 24 bit recording. SACD and that sadly failed technology DVD-A came out only in 2000, so six years earlier CD-based and CD-compatible formats such as HDCD would have been utterly dominant. Food for thought perhaps?

Track 2 was a 384 kHz 24 bit track by Johnomo Preludio, a Japanese guitarist. Track 3 was an old faithful, Jennifer Warnes' "The well", but played back at a resolution of 96 kHz 24 bit. This track came off her album The Well, which was released in 2001, and so I can be pretty much sure that it was not recorded at the played -back resolution. Maybe it was an up-sample of a lower-resolution original too?

Track 4: "Time", of Hans Zimmer's film score for Inception, but arranged by Alan Walker. I found this track peculiarly spatial, which is odd considering that it's entirely electronic and cannot have had any spatial component introduced by having been recorded in, say, a real-life recording studio using a microphone set-up that maximises spatial information (e.g. a binaural recording). Odd indeed, but by no means unpleasant.

Track 5 : Adele's "Skyfall", which as Hugh pointed out was in some ways a channelling of Shirley Bassey. He's quite correct: close your eyes and Adele has a very Shirley Bassey sort of voice on this recording. The song was used in the soundtrack of the 2012 James Bond film of the same name; the incomparable Shirley Bassey recorded the theme songs for Goldfinger (1964), Diamonds are Forever (1971) and Moonraker (1979). It's perhaps odd that some of these are among the better James Bond films (e.g. Goldfinger and Skyfall , may be not so much the spectacularly silly and comical Moonraker) and this leads me to posit that the quality of the title track is related in some way to the quality of the rest of the James Bond film.

On reflection, I'd venture an even better hypothesis - the quality of James Bond films is correlated 1:1 with the quality of motor vehicles that feature in them. Goldfinger had not one but four Aston Martin DB5s, without a doubt one of the most glorious motor vehicles ever made. Not to mention it features also a Rolls Royce Phantom III and, for the plebs, a Ford Mustang. The beyond-gorgeous DB5 returned for Skyfall, assisted by Moneypenny's Land Rover Defender. Diamonds are Forever had to make do with a Triumph Stag (nice in its own, unreliable, way), although things were livened up by the Cadillac Funeral Coach that provided the hearse. By contrast, Moonraker had nothing better than a pathetic AMC Concord station wagon. My case is proved beyond all reasonable doubt: the better the cars in James Bond films, the better the movie.

Track 6: "Tempus fugit", written by Bud Powell in 1950 and recorded many, many times since by different jazz groups, including in 1979 by Charlie Parker tribute band Supersax, on the Dynamite!! album. At this stage, the monthly meeting halted for the mid-demonstration break and I had to leave in order to be in tip-top form so I could wrestle manfully with the fallen trees the next day.

The six tracks, along with Hugh's introduction to them, did get me thinking about the veracity of supposedly 'hi-res' downloads or streamings. Paul Miller's analysis in Hi-fi News and Record Review indicates that what we are actually getting might be, in too many cases, quite different from what we think we are getting. Does it matter? Well, from one perspective, Yes, as what's on the label of the can does not accurately indicate what's inside it. On the other, No, as long as the sound quality is acceptable. And here there is a complete lack of agreement among audiophiles. There is only one way to reliably test for audible differences or preferences among tracks of different resolution, and that is via repeated double-blind tests. Forget one-off listenings, in which the listener knows beforehand the resolution of what he's/she's listening to. Humans are too subject to confirmation bias, placebo effects and a million other sensory and mental flaws, let alone inadvertent feedback from the tester that provide complicating clues to the listener, to make this anything other than a flawed and uninformative test.

We think we are unbiased listeners? Garbage: we are all biased, some to greater extents than others, but there is no such creature as the unbiased human. Experimental scientists in all disciplines have known this for decades, and the double-blind test is an accepted gold-standard for all scientific and medical experimentation. (Multiple lines of evidence suffices for other sorts of scientific or medical investigation where experimentation is impractical or unethical, such as in epidemiological studies.) Indeed, the Hydrogenaudio website specifically excludes any supposed audio comparisons that do not conform to the double-blind protocol. As my statistics lecturer told us all when I did a 3rd year course in biometrics many decades ago, "Anecdotes are not data".

When double-blind testing is done with highresolution audio files, the results are often minimally better than random, in some cases, no different at from what would be expected by chance alone, that is, by guessing (e.g. see Monty 2012; Fisher 2015). But there is also some evidence that the better trained and more experienced listener can reliably tell differences, if the latter are consistent, even if they are nevertheless slight (e.g. Reiss 2016). And the hearing of differences across various audio resolutions will depend to a large degree on the resolving power of the playback equipment. If you compared tracks of various sampling frequency and bit depths via your phone and ear buds, you are unlikely to detect any differences at all. And to complicate things even more, it must be rare for the same master to be used for different audio formats or playback resolutions, and this renders many putative 'comparisons' highly fraught.

Of course, higher sampling rates and bit depths have two very great advantages over the 44.1 kHz 16 bit standard: (1) they allow better headroom and noise levels, which creates room for the inevitable losses in sound quality that occur down the recording and production chain until the product is finally released to the consumer; and (2) with higher sampling frequencies the job of the anti-aliasing filter in the DAC is made much easier, and we don't have to endure the brick-wall filters that were necessitated by a 44.1 kHz original. Mind you, oversampling by your DAC or CD player achieves the same result. Whether the higher frequencies of reproduction are audibly important is debatable, but many audiophiles reckon the reproduction of the >20 kHz components do make a significant difference with high-resolution recordings, especially with regard to the sensation of 'airiness'.

A final comment in defence of high-resolution recordings is that we must remember the 44.1 kHz 16 bit standard used for CDs (and their relations, such as HDCD) dates from the late 1970s. Sony and Philips started to collaborate on digital audio in 1979 and the Red Book, which describes the CD standard, was published in 1980. The first CD players came out in late 1982 or early 1983. So the technology is 40+ years old. It was a brilliant technology, especially the very, very clever CIRC approach developed for error correction.

But it's a 1980s technology, dating from the time when powerful home computers were merely a dream, when storage for such devices that did exist often took place on cassette tapes, and when the first floppy discs (so named because they were floppy) stored an amazing 360 kB of data. I remember the first domestic/ laboratory IBM desktop computers of the mid- 1980s: the XT and, for speed freaks, the AT. The XT came with an astonishing 128 kB (not MB!) of RAM and a 10 MB hard drive. As I've music at the CD standard. Ye gods, I did the statistics for my BSc(Hons) research on a IBM mainframe in 1979, with punched cards, that had to be run overnight it was so slow! Now that was purgatory.

So yep, there is lots of reasons to move with the times, and things digital have changed markedly over the intervening four decades. Highresolution files are likely to be one such advance, but we must be very aware that what's sold to us as 'hi-res' might not be as 'hi-res' as we are being told.

To conclude: thanks Hugh for yet another fabulous presentation.

Paul Boon