Hi-Rez Audio Distinguished in Blind Testing

Spectral analysis of a live blues band recording made by John Atkinson, showing content up to 40kHz, from "What's Going On Up There?"

With a few exceptions, audiophiles have long advocated high-rez music formats, believing that music should be recorded and presented in the highest fidelity possible, for our pleasure and posterity.

And yet a few people in our community, and many more outside it, have long maintained that CD is good enough—that, indeed, as a general principle, music at CD resolution is indistinguishable from high-rez music. And until recently, scientific evidence for the audibility of high-rez music, while not negligible, was thin.

It was not for lack of trying. Quite a few articles published in the relevant journals sought evidence of the audibility of higher bit depths and sampling rates. The results were mixed. In 2016, Joshua Reiss of the Queen Mary University of London published a meta-analysis combining results from 18 such studies, involving more than 400 participants and 12,500 trials, concluding that the experiments "showed a small but statistically significant ability of test subjects to discriminate high resolution content, and this effect increased dramatically when test subjects received extensive training." Some, though, thought the studies they chose were cherry-picked and so found the evidence weak (footnote 1).

At the October 2019 Audio Engineering Society convention in New York—just concluded as I write—Yuki Fukuda and Shunsuke Ishimitsu, both of Hiroshima City University, presented results that show quite clearly that listeners can distinguish sounds encoded and reproduced at different sampling frequencies. Their trials differed from the previous ones in one important way: Instead of exposing test subjects to music at different resolutions, they used test tones.

Specifically, they used two forms of spectrally "flat" signals: white noise and (Gaussian) impulse signals. White noise and impulse signals both have broadband content; they differ from each other in that in an impulse, all the frequency components are correlated in time, whereas for white noise the phase of the various frequency components is random.

The Japan Electronics and Information Technology Industries Association defines the "CD format" as having sampling frequencies up to 48kHz and a bit depth of 16. Anything higher, in bit depth or sampling frequency, is considered high resolution. For these experiments, the two researchers used a bit depth of 16 in all trials so that their lowest sampling rate data would be classified as CD-rez, the others as high-rez.

Fukuda and Ishimitsu employed Gaussian impulse and white noise test signals at 48kHz, 96kHz, and 192kHz. The test had seven subjects, all young: The average age was right at 22. ABX comparisons were made between 48kHz and 96kHz signals, 48kHz and 192kHz signals, and 96kHz and 192kHz signals. In a different round of testing, they applied a different methodology: MUSHRA, for Multiple Stimuli with Hidden Reference and Anchor. This test, too, involved seven subjects. Two women participated in the ABX round and one woman in the MUSHRA round. For each methodology, both headphones and loudspeakers were employed.

The setups were modest. Loudspeakers were Eclipse TD-M by Japanese manufacturer Fujitsu Ten, a single-driver—hence intrinsically time-aligned—powered desktop loudspeaker. The Eclipse, which is out of production but last sold in the US for $1300/pair, accepts digital data via USB and Wi-Fi, but in these experiments appeared to be used via the analog input: a 3.5mm stereo jack. The researchers employed a Fostex HP-A4BL D/A converter, which retails for $600 and is widely available at a street price under $500.

A fast-roll-off linear-phase filter was employed to minimize aliasing distortion. Headphones were the Sennheiser HD 650. Tests were carried out in an anechoic chamber.

Applying the binomial test—the common p<0.05 assessment (footnote 2)—all but one of the ABX tests yielded decisive positive results, with p values well below 0.05. The most successful tests were those with loudspeakers and the Gaussian impulse; in those experiments, p was less than 0.0001 in all comparisons. The results of the MUSHRA study, evaluated using a two-way ANOVA analysis, were more tenuous but still supported a conclusion that "there is a possibility of a discrimination between Hi-Res and non-Hi-Res audio data."

In a brief email, Fukuda, the corresponding author on the paper, told me that next they intend to study whether people can distinguish among bit depths of 16, 24, and 32 bits at a 48kHz sampling frequency.

Beyond that, it would be good to see the test repeated with older listeners—with, presumably, less acute high-frequency hearing—and with a variety of loudspeakers, to determine which loudspeaker characteristics are most important for hearing such differences; the implications could be very important for loudspeaker design. Next, experiments could be carried out in a regular room, to determine whether such differences are audible in a domestic environment or whether the anechoic chamber is essential. I'd love to see the results of interviews with the test subjects, to hear their subjective listening impressions: What, precisely, did they hear from the high-rez? What are the subjective characteristics of high-rez test signals, relative to their lower-rez counterparts?

Finally, subjects trained using these test signals could then be exposed to carefully chosen music samples—maybe music with wood block or other percussion, resembling those Gaussian impulse signals but with more going on—and then gradually moving on to other kinds of music. Only then will we be in a position to identify the subjective characteristics of high-resolution music in a way that should satisfy objectivists.


Footnote 1: See John Atkinson's discussion of the Reiss analysis and the subsequent comments.

Footnote 2: A p<0.05 means that the possibility of the result occurring by chance is less than 5%.

COMMENTS
rt66indierock's picture

Jim, I think using an anechoic chamber it might be possible for well trained ears to hear differences but take Jon Whitledge's Magic Bus to a dry lake where the background noise is 17 dB and I start to have doubts about the differences being audible (Jon does this). In my office at night with the HVAC and my computer off, 23-27 dB of background noise, I doubt whether highly trained listeners can tell the difference. During the day, 36 to 42 dB of background noise in my office probably means it is not possible to hear the difference.

One way to look at this is if people cared about hi-res, there would be a market to stream it and download files. There isn't any real evidence a market exists for hi-res. It's been eighteen plus years of trying to push it with little to no results.

er1c's picture

Silly me, trusting my own direct experience. I thought I liked 24 bit better! Now I know better than to trust my own ears

rt66indierock's picture

I don’t know about silly but if you are a longtime reader of Stereophile you should been a little suspicious when “There’s Life Above 20k” in 2000 was published. And you should have gotten steadily more suspicious year after year.

I am arguing when I make a recording you can’t reliably tell the difference between the 16/44.1 and the 24/96 versions.

I do a little test at the T.H.E Show. Let’s meet up and see if you can trust your ears. Or you could pass Mark Waldrep’s current test of his hi-res recordings.

But be warned, I introduced my girlfriend to a quite a few people at the Los Angeles & Orange County Audio Society Annual Gala this year with “this person trusted their ears.” Those people failed my test at T.H.E. Show 2019.

er1c's picture

Being an Audiophile is about art and beauty and music and spirit and lifting ourselves up

PS- Greta has a posse

supamark's picture

You didn't read the article, did you? They were testing sample rates (so - 48kHz vs 96kHz vs 192kHz) at a single bit depth. It even states in the article that their next experiment will be about bit depth (16 vs 24 vs 32 bits) while keeping the sampling rate the same (48kHz).

next time read the article, THEN troll

1/10 wouldn't read again.

rt66indierock's picture

It is no secret I enjoy trolling Jim Austin. But I didn’t here. Too bad the best stuff was deleted from Audio Asylum from 2017.

But my initial comment was if you repeated the test in the quietest natural place I can easily visit (Southern California dry lakes) and other quiet places like my office at night, I doubt you would hear a difference in the tones.

My next response was a direct answer and challenge to er1c’s comments not a comment on the article.

Finally, I’ve read the paper Jim is referring to. There is an interesting battle going on behind the scenes trying to justify high resolution audio but don’t hold your breath reading about it in Stereophile.

drblank's picture

Those aren't good for listening since one can't sit through listening to a pair of speakers long enough to perform an actual listening test.

A proper listening room would have a flat response curve, but it would have TC60 reverberation times that people could listen for extended periods of times.

But yes, the room does impact one's ability to hear any differences, its just unfortunate that most people don't have proper listening environment, but some people can tell the difference even without having the best conditions..

jimtavegia's picture

I can achieve a nearly -80db noise floor in my home studio. In my software. the metering will tell me what the noise floor is when my mics are open when I practice singing. If I use a large diaphragm mic they always pick up more of the "room noise" than any of my hand-held mics. This is especially true of my Rode NT-1A's as they only have 5db of self noise, incredibly quiet. If there is noise it is not the mic, but the room and the mic preamps. Most vocal mics pick up less room noise, but also have self noise that is 17db or higher, some up to 23 db of self noise. Also helps if you some form of AC line noise filter on your power supplies to your gear. Even cheap ones are better than none.

The real issue of this noise is at what frequency does it appear? If you have noise that is at or below 30hz and your speakers or headphones can't reproduce it, it really doesn't matter. You probably can't hear it, yet it will show up on your metering in your recording software...it might even be as high as -50 or -60db. You can find out at what frequency this noise is by looking at the FFT of the wave form and it will tell you frequency of the noise and then you can decide whether it matters or not. Most of this noise is not as broad in spectrum as we hear from analogue tape from cassettes or R2Rs. Most of this sits at -60db or higher and really limits the dynamic range of recordings, mostly the long reverb tails that fall into this high noise floor.

I still recall JA1 talking about a recording session he had in a large hall and trying to find the source of some ambient room noise, only to find it was the HVAC system on the roof of an adjoining building . Not much to be done about that. Most of my location recording work, this noise is generally HVAC related and once turned off I can get a low noise floor, other than the folks as part of the audience.

You must find the frequencies of the ambient noise in the room by recording it and then viewing it by FFT. Then you can decide if it matters or not. Depending upon how resolving your audio system is will determine if it is problematic or not. In your home your refrigerator and HVAC will be the biggest noise source, other than children playing , door bell, phone ringing, or the dog barking at a bad time.

eriks's picture

One thing I've noticed over the last 10 years of listening:

A DAC's performance at Redbook varies a great deal.

Some DACs perform a lot better with 44/16 than others, and have a very small step up, if any, to 96/24. The Mytek Brooklyn is a good example as are some Schiit models.

This gap has all but disappeared of late. With the ARC DAC 8 for instance (which I owned concurrently with the Brooklyn) you could certainly hear a difference. Was this the clocks? Was it the rolloff filter? I don't know.

But of this I am certain: The revolution in Redbook performance has happened recently and the importance of HiRez files has greatly diminished for me.

doak's picture

Yawnnnnn.....

volvic's picture

I can only speak from my experiences, yes, early 16 bit wasn't very good, got better in the 90's and as I have said on these very pages, it is good enough when played through a computer setup and not a CD player. Having said that I just recently got into SACD having purchased a Pioneer player and an Arcam FMJ DV139 player because it dawned on me, I have quite a few SACD's, including quite a few of the Esoteric ones. The rule of thumb is if the recording was properly done, high-rez benefits the listening experience, even to my 54 year old ears. If the recording is poor like the Karajan's 80's Tchaikovsky symphonies then not even Esoteric's engineers can breathe new life into it. The ESoteric SACD is no different from my vinyl version and just slightly fuller in high-rez than the 16 bit. But if you weren't comparing them you really wouldn't be missing much. The Esoteric SACD of Miles Davis' Sketches of Spain shines, whereas those 90's Columbia CD reissues were horrid. The vinyl can be good too, but my pressings from the 90's include a Canadian pressing and a European one both are ok but not standouts. In short, high rez does benefit musical enjoyment and the differences can be heard, but they are not as clear cut on every recording and the advances in 16 bit recording technology has definitely closed the gap. But I will continue to purchase SACD's as some of the high resolution results can be quite impressive. One man's opinion.

Anton's picture

I'm glad they are investigating what and how we hear things, I certainly prefer my (Gaussian) impulse signals played back in hi rez.

It will be interesting to see where in the audioband these things are most audible.

I have never been fooled by MP3, but have missed the ball with some hi rez vs. Redbook playback.

This is such a fun hobby, much like wine: seeing what we can elucidate about a subjective experience under blind conditions is great fun, to me.

I think Volvic summed it up pretty well!

Jason Victor Serinus's picture

On my Rossini DAC, which upsamples to 24/384, I still hear major differences between 16/44.1 and hi-rez. The soundstage is wider, more extended from top to bottom, and most definitely deeper. Spaces between the notes are more profound, colors more saturated, and images weightier.

I insist on reviewing hi-rez whenever it is available. In fact, if it's available for download and streaming, but the label will only supply 16/44.1, I review something else. I want to review the best that's available to Stereophile's readers.

eriks's picture

You wrote:

On my Rossini DAC, which upsamples to 24/384, I still hear major differences between 16/44.1 and hi-rez

So, how do you ascribe responsibility for this? How do you say it is the fault of the sampled data, vs. the DAC not performing as well as it should with 44/16?

For decades we have said it must be the sample rate, and the war has raged between the theorists and the consumerists. My ears used to tell me that Hi Rez performed better. At least with DACs I can afford, I've heard that gap disappear within hte span of a handful of years. Did they get better with 44/16? Are they not that good at Hi Rez? Do clocks and filters built into cheap chips now perform in a better league?

I don't know, but since you have access, I encourage you to listen to the gap itself with something like a Mytek and compare it to your Rossini.

Jason Victor Serinus's picture

No time to do it now, but I have both a Mytek Brooklyn Plus and Manhattan II here. Which do you prefer for the test? It's gonna take a while - reviews come first - but I'll try.

eriks's picture

Only because it's the one I own. :) I've not been able to audition the Manhattan.

Another DAC which I've heard this from however was a Schiit. Again, great Redbook, and slim improvement if any on hi Rez.

supamark's picture

only because it's a fairer comparison. Also, if Mytek ever releases a blue one... rofl

I think it'd be even more interesting if you ever get a topline R-2R ladder DAC (or even Herb's Schiit Yggdrasil) to compare to the Manhattan and Rossini. I personally find delta/sigma DACs to sound "squishy" and smeared compared to Schiit's multi-bit DACs (including my own recordings from back in the day, with which I'm very familiar). I've never had a chance to listen to a dCS DAC, but I'm curious since it's yet another conversion method and likely sounds unique.

fbailiey's picture

Did I read that correctly - there were only 7 test subjects, all students? This is the best you can do to support your headline "Hi-Rez Audio Distinguished in Blind Testing"? I like HD music, but having to dredge up a report like this to make your point is not likely to convince many non-believers.

John Atkinson's picture
fbailiey wrote:
Did I read that correctly - there were only 7 test subjects, all students? This is the best you can do to support your headline "Hi-Rez Audio Distinguished in Blind Testing"?

I don't understand your objection. In the text it says that "The most successful tests were those with loudspeakers and the Gaussian impulse; in those experiments, p was less than 0.0001 in all comparisons." That means that the probability of the identification being due to chance was just 0.01%. That is pretty convincing evidence, I would have thought.

John Atkinson
Technical Editor, Stereophile

fbailiey's picture

I'm no scientist, but what I take from this is that hi-rez sounds different using test signals with college kids, but is less useful for music, which therefore implies that HD is overkill for music. I'm a hi-rez supporter, but this is not gong to help with the general public to put a message like this out there IMO. That message being that you have to be a college kid listening to test signals to tell the difference.

Jason Victor Serinus's picture

is not going to be convinced one way or the other by a study published by Hiroshima City University, just as it not going to be swayed one way or the other by ceaseless debates about / condemnations of MQA on audio sites. If they hear a difference, and it doesn't cost them an arm and a leg to do so, they will change their listening habits. Now that we have Amazon Music HD, which can potentially reach far more people than Qobuz or Tidal, the shift could happen sooner rather than later.

rt66indierock's picture

You would be surprised how easy is to convince true believers of hi-res recordings the general public can't hear a difference. So far it's been pretty easy to convince the general public MQA doesn't sound better. Funny thing is there is interest Dolby Atmos in the small part of the general public that cares about sound quality. The question is will it deliver better sound to the general public?

supamark's picture

recorded live for surround/Atmos (or with *really good* DSP reverb) it would be simple to demonstrate the superiority but difficult to get many to shell out the cash (and living space) for a proper set up, especially Atmos due to the height speakers. Surround really blows away 2 channel stereo for acoustic music recorded in a real space, for other music... meh. I mostly listen to non-acoustic stuff and nightclubs sound like crap, I don't want that in my living room.

If I want to hear that wrap around acoustic sound I can go see it live and get better immersion than any surround set-up could ever provide (I spent a couple years recording classical music professionally - always two mic stereo). and yeah, recording engineers are generally at least as picky about sound as audiophiles, and spend a lot more time listening - I'd trust a good engineer's ears over any reviewer's any day. It's mostly mastering engineers and record labels that are responsible for the horribly compressed/clipped sound of most non-vinyl releases today. Vinyl only escapes because it can't take the level lol. If they ever find a way to make vinyl super loud with no dynamics they'll do it.

It's like wine tasting, with training there are of people who can pick out the vinyard and year of a wine while it all tastes like grape juice with vodka mixed in to me. I'm kinda like that with audio (i.e. the crunches of a corn vs. potato chip sound distinctly different to me), and the synesthesia helps too lol (it's common among recording engineers).

jimtavegia's picture

Will never care about this issue, ever. Streaming at 256 or 320 kbps is enough for them even if it is 1/4 the data of redbook CD Quality. I believe this is the reason that so many $399 turntables are sold to many happy customers. Even at this level an LP will sound better than an MP3.

I doubt that most of the computer listeners are not even using any type of USB DAC for their audio listening, or even a decent set of cans. Even if they bought a DragonFly or a usb dac like the Focusrite Scarlette or Steinberg UR-22 which are about $160, I would believe they cared about sound quality. Trying to convert the masses is most difficult. The NY Times and WSJ don't help much as Mr. Fremer has pointed out.

supamark's picture

would like to see a larger sample size, and like Jim said more variety in age in test subjects. Also, speakers that have multiple drivers (even the LS50 if they want a point source) like in the real world. I do like that they're sticking to one aspect (bit depth, sample rate) at a time.

Mellotronix's picture

The researchers applied appropriate analysis for this study. From the information presented, the results appear to be valid and reliable; however, I would have to access the full paper to see how the sample was selected and what population is represented. Normally, you select subjects from a well defined population--for example, individuals between 18 and 25 who listen to popular music for at least 10 hours per week...Random selection and random assignment to levels of the variables are critical in order to generalize the findings to the larger population. Additionally, researchers can only claim that when it comes to white noise and impulse signals, there is a significant difference. But, good work on the part of the researchers to advance our knowledge in this area. No single study is going to answer all of the questions, but this will fill in a small piece of the bigger puzzle.

jimtavegia's picture

If you want to check this out yourself, my build 12 version of Sony SoundForge Audio Studio, across the bottom, has not only the spectrum display, the scope you can use to insure you have a stereo signal, and it has to the lower right a full graphic EQ which will show you the frequency spread of your track. Instead of guessing that you have HF detail, now you can verify it. Magix now owns the Sony line of software.

I have used this and Sony CD Architect for burning discs, with great success.

jimtavegia's picture

It can handle file formats up to 24/192.

pbarach's picture

There are so many examples of studies in medicine and psychology with striking findings that could not be replicated. A commonly cited example is the study by Bem that purported to show the existence of psi (psychic powers): https://theness.com/neurologicablog/index.php/follow-up-on-bems-psi-research/

The statistical power of the study is relatively low given the number of participants (N = 7) and in spite of the low p value. Personally I often hear differences in favor of hi-res, but as a Boulder-model-trained psychologist, I'd want to see some replications. And whether this study's findings can be generalized to music is unknown, and I like JA's suggestions for further research.

jimtavegia's picture

I have 4 Tascam SDHC card recorders, all of them can record 2496 as wave or (bwf) broadcast wave files. My DR-680MKII can do two tracks of 24/192 and 6 tracks of 24/96. I love 24/96 and find it to be very smooth and analogue like for those into vinyl. I was, and still am to a degree, but as I have given my sons 3 of my TTs and have only kept my Dual 502 with a Rega 202 arm installed. I am about the cull through all my vinyl over the holidays and get rid of the lps that don't sound very good. Too many I am afraid. I still have 4 SACD players that are also DVD players and I can burn 2496 discs on DVD-Rs and play them in all of those players. THAT was the original promise of digital, now fulfilled in my mind. Many albums I've bought as HD downloads have been burned to DVD-Rs.

I had two lps from a fav piano/vocalist from the 1980's, and the lps sounded so bad to me I had to put them away as my middle son will get them and many others shortly.

My old ears can't really tell much, if any, difference from 24/96 and 24/192, but I can say that I love all of my SACDs. I think that 24/192 needs a super stable clock and don't have one to use at that format. I can surely hear the difference when I record 16/44.1 vs. 24/96. Even my 72 year old ears can do that.

JBLMVBC's picture

Reminds me of people who claim they can't differentiate superlative wines from wines sold by the gallon... Often, this is a cheap excuse for ignorance and lack of curiosity: it is much easier and less expensive to experience some of the best wines in the world than some of the best cars, stereos in the world.
Back to sound, Hi Rez is like 45 RPM versus 33 RPM. If you can't notice the difference -dynamics, clarity of tone, resolution of instruments, stage depth, micro-information- then no need to spend $1,000s in audiophile products...

Archimago's picture

Should this factor in the experimental design immediately stop us in our tracks and consider what relevance this has with actual, real, high quality, music?

There are examples with test signals where we can show things like "absolute phase" being important (I talked about this in my blog June 2019) yet the moment you apply this to music, the differences become of questionable significance.

So too with "hi-res" vs. CD.

Sure, for an album I simply adore, recorded/produced/mixed/mastered immaculately, I'll happily buy the 24/96 or whatever... But those albums are rather few and far between for me. And even then, if someone has to justify audibility based on test signals like this with young people, average age of 22, most readers I think will not (and should not!) be impressed.

Graham Luke's picture

...you've gone and done it now, Jim...

downunderman's picture

As a test environment this appears to come very (blindingly) close to cracking it - one addition could be for the subjects to go through a hearing test beforehand to establish the state (frequency range) of their hearing.

jeffhenning's picture

Here has been my experience:

• Recording at 24 bits as opposed to 16 bits offers much more benefit than going from 44.1/48kHhz than double those sampling rates (256 times the resolution as opposed to just 2)

• Not that it's an issue anymore, but if you do have only 16 bit multitracks to mix from, mixing them into 24 or 32 bit masters will sound better than a cumulative 16 bit mix... though not drastically

• Obviously, 96kHz and 24 or, better, 32 bit recording is preferred. I'm very dubious that a higher sample rate is really worthwhile. Having done some experimenting on it myself, I'm not finding sample rates above 96kHz to be worth the data space or processor overhead.

AJ's picture

The papers actual title LOL.
After 30 yrs of floundering to provide "proof" in favor of "Hi Rez", we have this desperation Hail Mary.
Will read in more detail, but at first glance, the signs of nonsense are all there.
Btw, are audiophiles who flatly reject blind testing suddenly going to tout *this* blind test too, like other "good" ones? Stay tuned...

Also Jim, Joshua Reiss of LANDR - a for profit hi res service, gave a particular paper the highest "trained" weighting in his "Meta".
"Further Investigations of the Audibility of Digital Audio Filters in a High-Fidelity Playback System" Jackson, Stuart et al 2016
THERE IS NO SUCH PAPER in the AES library, because it FAILED review.
Ever seen Reiss et al mention that since?
Me neither ;-)

AJ
Soundfield Audio

gwybob's picture

The Japan Electronics and Information Technology Industries Association defines the "CD format" as having sampling frequencies up to 48kHz and a bit depth of 16. Anything higher, in bit depth or sampling frequency, is considered high resolution.

gwybob's picture

The Japan Electronics and Information Technology Industries Association defines the "CD format" as having sampling frequencies up to 48kHz and a bit depth of 16. Anything higher, in bit depth or sampling frequency, is considered high resolution.

David Harper's picture

It seems to me that attempting to discern between CD and hi-res by listening to dynamic cone drivers in a wooden box is like attempting to discern between standard and HD video by looking at a 360 line CRT tube TV. Maybe a difference would by discernable with electrostat speakers,or high-quality electrostatic headphones, but not with the low resolution high distortion sound inherent in all wooden box speakers.

rt66indierock's picture

You need better facts. From Hi-fi News & Record Review February 2015 review of the vintage AR-7. Distortion at 90db SPL/1 meter is 100 Hz 1.6%, 1kHz 0.3% and 10 kHz 0.5%.

My AR-4x's measure slightly better and I've owned them 35 years. They on their fourth crossover and have been repaired as necessary and restored by noted AR expert the late Carl Richard shortly before his death.

I admire your desperation to find anything to support the notion that high resolution sounds better. Keep looking I'll talk you off whatever ledge you find yourself on when you fail.

Bogolu Haranath's picture

As an additional note ....... Hi-Fi News measurements show, Revel Performa F228Be 'wooden box' speakers have distortion levels of 0.6% at 100 Hz, 0.1% at 1 kHz and 0.1% at 10 kHz :-) .........

David Harper's picture

your post is completely incoherent.The point of my post was that hi-res is, for the most part, worthless. And your AR box speakers cannot possibly have the speed, transparency, detail or resolution of a good electrostatic, no matter what you think you know.

rt66indierock's picture

You said “Maybe a difference would by discernable with electrostat speakers,or high-quality electrostatic headphones.” So, let’s review, my modified AR-4x will play my reference albums and recordings will electrostatics? I’ve listened to many and the answer is no. The trouble starts with The Eagles “Already Gone”, continues and by the time I get to a recording I made of a Derring Goodtime Midnight Special (resonator banjo) I’m pretty much done with electrostatic speakers. And don’t get me started with what happens using my DAW. There is a reason silk dome tweeters will be in my new studio monitors.

hrboucher's picture

I’m with Audio Note. A Redbook CD played through one of their non-oversampling DACs is an equal or better musical experience for me than the same recording in Hi-Rez. Note I said “musical” experience, not “audio” experience. De gustibus non disputandum.

dcrowe's picture

I scored 5 out of 6 on the test at https://www.npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality, I used an iPad and Sennheiser HD800 headphones. I am 71. And I listen to high resolution audio on TIDAL. However, the differences are small and it took switching back and forth between the test recordings many times to identify the best fidelity. There are audible differences, even though deciding which is higher fidelity is less obvious. I am in the process of assembling a system with the Raal SR1a driven by Benchmark DAC3 and AHB2 from my iMac Pro to listen to TIDAL hi-rez streaming. This level of fidelity is beyond the basic requirement for enjoyable listening, but it is a luxury I value to improve it beyond the threshold.

rt66indierock's picture

There are some interesting rumors about Tidal. I would have a backup plan.

The last time I conducted a similar test, a John Atkinson recording of the Portland State Chamber Choir. The AAC version won with Maggies and Rogue Audio Receiver over CD and Hi-Res.

I've had the Benchmark stuff in my office, very transparent and sounded great with Joseph Audio Pulsars and my AR's.

Dr AIX's picture

The NPR and virtually all other ABX testing of hi-res vs. CD res fail because the source materials are not bona fide hi-res. The commercially released tracks used were standard-resolution to start with.

dougspeterson's picture

These tests typically use random people. The question is what does this have to do with me or you? I have observed over 40 years as an audiophile, my sensitivity to hi resolution audio increases with exposure to fine equipment, 1000s recordings, live music, i.e. experience.

The superior time base stability of CDs over vinyl I heard immediately after years of LPs on DD turntable and RR and cassettetape.

Early Telarc CDs were criticized for steely highs which I did not hear. A decade later the same disks sounded like listening thru a metal fog, the original Stereophile reviewers perception was now mine.

Hires now sounds like liquid music.

rt66indierock's picture

I hope you don't mind me challenging your expertise and powers of observation.

I've met a lot of audiophiles the last four years at the Southern California audio shows and RMAF.

There isn't much expertise in the show goers. And a simple test causes me wonder about the hearing ability of a disturbing number of show attendees at the last couple of T.H.E. Shows.

Dr AIX's picture

Several of my readers linked me to this article. I find it interesting but seriously flawed as it bears no relation to the real world of music and listening. Seven young listeners, listening to noise and gaussian impulse bursts, in an anechoic chamber should convince us that high sample rates are perceptible?

As a strong advocate for real hi-res music (I spent millions of dollars and 20 years producing and releasing them on AIX Records), I have recently modified my views on the perceptibility of hi-res over CD "Redbook" audio. I invite everyone that is curious about this issue to participate in the Hi-Res Audio Challenge II at RealHD-Audio.com. Feel free to join over 600 other participants in a real world survey using real hi-res music tracks on your own systems. You can sign up at the RealHD-Audio blog site.

The reality is that virtually all hi-res audio was not recorded using hi-res equipment. The files we cherish are transfers of standard-resolution audio to higher sampling rates and longer word lengths. The fidelity of the original master isn't magically enhanced to the new specifications.

David Harper's picture

I know far less about all this than you do but I can state that I own a number of 24/192khz bluray audio music discs. And a much larger number of redbook CD's. My system is a Schiit Vidar amp,Oppo BRP, and magnepan LRS speakers. Having listened extensively to all of my source material for several years (over and over) it has become apparent to me that there is no difference in sq between the CD's and the blurays that can be reliably attributed to the disc file format. Some of the hi-res discs sound good, some sound mediocre, and some sound terrible. Same for the CD's. Some of the redbook CD's sound better than the hi-res discs.Some sound worse. The difference seems to be in the original recording quality not the file format. I'm aware there are many variables involved in this comparison that I don't know about (such as your aforementioned source file format for the hi-res discs) What is clear to me is that file format is no indicator of sound quality. And if it is both common knowledge and common sense that speakers have distortion several orders of magnitude greater than electronics, maybe it makes sense to forget file format since the resolution of our speakers would mask entirely any difference in sq due to file format.

Dr AIX's picture

You're absolutely correct. The fidelity of the original master as established by the engineer, producer, and mastering team are far more important than the ultimate delivery format. However, it is extremely difficult — virtually impossible — to enhance the fidelity after the original production. Releasing an old master at 192 kHz/24-bits or processed with lossy MQA is a fruitless exercise and meant to enhance profits not fidelity.

C_Hoefer's picture

David Harper, Dr. AIX, the arguments you give, for it being pointless to digitize old analog master tapes (or whatever-generation tapes) at 24/96 or 24/192 instead of at Red Book, are flawed. First of all the fact that loudspeakers have relatively high distortion is irrelevant, unless the distortion is so bad that the speakers are no good for music listening in the first place. Not all forms of distortion are the same, nor are they perceived by the ear/brain system in the same way. We are able to hear, say, 0.3% IM distortion from an amp, through speakers with 1% distortion at some frequency range; the former distortion is not rendered inaudible by the latter. Ditto for the kinds of distortion introduced by digitizing a signal at Red Book. I grant you that if the digitization is done with skill and care, a 16/44.1 digitization of a master tape, played back through a high-quality DAC, can sound fabulous, and it will be hard for most listeners, in most circumstances, to spot the difference in A/B comparisons. Whether musical enjoyment would be the same under prolonged, relaxed listening conditions is a different matter, and more relevant to music enjoyment by consumers. This is of course hard to test rigorously.

Here's a way to see that the argument of "do as little harm as possible to the analog original" is not wrong, no matter when the analog original was recorded. If the limitations of the original tape (noise floor, dynamic range and frequency range) made the additional distortions of digitization irrelevant, then you would think that one could run the signal through as many loops of A/D - D/A conversion as you wish without lowering the fidelity of the copies. But we all know this is false. After a certain number of A/D-D/A loops at Red Book, the sound quality will be noticeably harmed. After a higher number of loops at 24/96, it will be harmed also. This shows that every trip through digitization degrades the original to some extent, and does do more at Red Book than at hi-res. Whether the degradation done by only one loop is audible and significant is a further question; but there is some harm, and more at Red Book than at hi-res, therefore the argument of the hi-res advocates at HDTracks that they want to give you a file that stays as close to the original master as possible is not wrong, nor is it pure marketing BS.

Again: all that said, whether the difference makes a perceivable difference to the consumer, and affects how much they enjoy the music they purchase or stream, is a different question, and a hard one to answer. Personally, I feel I can usually hear the difference when I stream Qobuz at 16/44.1 vs. when I stream it at 24/96. Not always, and sometimes I worry that I am comparing apples to oranges because the hi-res file may have been "remastered".

I look forward to trying your Hi-Res Audio Challenge II, haven't had time to start yet.

John Atkinson's picture
C_Hoefer wrote:
Here's a way to see that the argument of "do as little harm as possible to the analog original" is not wrong, no matter when the analog original was recorded. If the limitations of the original tape (noise floor, dynamic range and frequency range) made the additional distortions of digitization irrelevant, then you would think that one could run the signal through as many loops of A/D - D/A conversion as you wish without lowering the fidelity of the copies. But we all know this is false.

In my 2011 Richard Heyser Memorial lecture to the Audio Engineering Society, I mentioned Peter Craven's experience in digitizing 78s - see www.stereophile.com/content/2011-richard-c-heyser-memorial-lecture-where-did-negative-frequencies-go-measuring-sound-qua.

"In his keynote address at the London AES Conference in 2007 . . . Peter Craven discussed the improvement in sound quality of a digital transfer of a 78rpm disc of a live electrical recording of an aria from Puccini's La Bohème when the sample rate was increased from 44.1 to 192kHz. Even 16-bit PCM is overkill for the 1926 recording's limited dynamic range, and though the original's bandwidth was surprisingly wide, given its vintage, 44.1kHz sampling would be more than enough to capture everything in the music, according to conventional information theory."

Peter noted that "With such a recording there is more to the sound than only the music. Specifically, there is the surface noise of the original shellac disc. The improvement in sound quality resulting from the use of a high-sampling-rate transfer involved this noise appearing to float more free of the music; with lower sample rates, it sounded more integrated into the music, and thus degraded it more."

Peter postulated that the reason listeners respond positively to higher sample rates and greater bit depths is that these better preserve the cues that aid listeners in the creation of internal models of what they perceive. If you going to archive analog recordings by digitizing them, therefore, then why not do so at the highest practical sample rate and bit depth? All it has cost you is some additional hard-drive space, the cost of which is negligible these days.

John Atkinson
Technical Editor, Stereophile

rt66indierock's picture

John, I'm going to need some folks to verify Peter's claim. Preferably tests in Asia, the Americas and Europe.

You seem to be grasping at straws to justify high resolution formats.

AJ's picture

I've searched the AES library and could not find any paper by Peter you refer to, where he detected "improvement" with high sample rates LP noise.
John, could you please provide a link?
Thanks and happy new year.

AJ

Soundfield Audio

Jim Austin's picture

AJ, I don't see how you could have missed that JA wrote that Craven delivered a keynote address; he didn't claim that Craven delivered a paper. It took me two minutes to confirm 1. that Craven delivered the keynote; 2. that Craven was at the time very interested in 78s (it's mentioned in his conference bio), and 3. That his keynote was called "Resolution: Who Needs It?" I was also able to confirm that JA shared a panel with Craven, and that Mark Waldrep also presented at that meeting, on DVD-A surround sound, so perhaps he remembers the content of Peter Craven's lecture.

Jim Austin, Editor
Stereophile

AJ's picture

Hi Jim,

Yes, that's the only part I found, that he (Craven) was a conference presenter.
It seems his claim about LP noise being, umm, "improved" with high sample rates was mere conjecture, not based on any sort of scientific listening tests.
It just caught my eye because I have done blind listening tests with the audio club where we compared a live analog LP feed with a simultaneous sampled version...with LP noise, clicks etc, in real time.
I deliberately sampled at 16/44 even though higher rates were possible. But they weren't needed since no one (including devout vinylphiles) could identify the "pure" vs sampled version...again, "noise" and all.
Lots of laughter was involved.

cheers and Happy new year

AJ

Soundfield Audio

Jim Austin's picture

>>Lots of laughter was involved.

Sounds to me like an audience that had already made up its mind.

One point that many of the quasi-objective side seem to miss is that in a world (think pro audio) where subjective opinions are really the whole point, subjective opinions of well-trained people are taken seriously. As a former scientist, I rather like the idea of acknowledging that depersonalized data isn't the only thing that matters--especially in a world where the impact music makes is the ultimate test.

Jim Austin, Editor
Stereophile

AJ's picture

Strange comment from someone who wasn't there. The real answer is that we were having an enjoyable, entertaining experience, especially with the results given what many believe. Perhaps that's what you meant by made up minds?
In audio, only (oft inflated) egos are bruised, no skin or bones broken.
The club has plenty of good sports I guess.
No one left and dumped their vinyl (or "digital") either. They just left with a better understanding of (audio) fact vs fiction. And lots of smiles. Audio should be fun.

cheers

AJ
Soundfield Audio

ChrisS's picture

But there is no differentiation of "fact" from "fiction" when the test is amateur and the "lab" is someone's basement.

You poo poo the results of a university study and expect others to accept those of an "audio club" with "plenty of good sports"?

Enjoyable entertainment is one thing, but it's not science.

Jim Austin's picture
I was thinking that folks were laughing AT the hypothesis being tested. Jim Austin, Editor Stereophile
AJ's picture

That's a distinct possibility! Or they could have been laughing at my audacity as well. Who knows? ;-).
I don't think most thought they would have any difficulty given the oft absurd comparisons between "Vinyl" and "digital" aka "CD" (which of course should be easily identifiable).
Surprise!
I make no claim it was any type of absolute test or even submittable to AES, given the umm, "difficulty" in matching voltage levels to 0.1V precision with live vinyl playback (using a test record), to say the least.
As I mentioned, the laughter was because the participants were both friends and good sports. I imagine there are other devotees who may not have reacted as kindly. Despite this essentially being, listening to music....

cheers

AJ
Soundfield Audio

AJ's picture

What sort of "tests" are you doing down there Chris??

JA clarified there was no "University study" by Craven (he works for Meridian fyi), just some conjecture about LP noise.
Would have been interesting see how old guys "heard" >22k of course.

Happy new year to you and your folks upstairs

ChrisS's picture

...not a social club meeting for laughs and enjoyable entertainment.

Jim Austin's picture

>>Craven (he works for Meridian fyi)

Craven became an independent researcher in 1981. His affiliation is, or until recently was, with a consultancy called Algol Applications Ltd. According to recent Companies House filings, he is one of two directors in an equal partnership with Malcolm Law. Do you have reason to believe his status has recently changed?

Jim Austin, Editor
Stereophile

AJ's picture

I believe you are correct and I may have mistaken that because MQA was spawned from Meridian, that Craven had some involvement there too. My error.
It's unfortunate there are no details about the high sample rate vinyl playback by Craven that John references. I would have been genuinely interested in the details.
Circling back to the article Convention submittal paper 10318, I wish them good luck with a formal review, since there are a great many details missing, provenance of system used, etc, etc.

rt66indierock's picture

Peter Craven has been a shareholder in MQA Ltd since June of 2015. He was issued stock before Bob Stuart was issued his.

In Bob Talks he talks about long-term collaborators Peter Craven and Michael Gerzon. Peter is an author of MLP Lossless Compression and The MLP Lossless Compression System papers

Let's hope Algol Applications Ltd isn't still involved with Algol.

Sure looks like he was involved with Meridian.

Jim Austin's picture

I know what Craven has done and what he's been involved in. He is a longtime close collaborator of Bob Stuart, and he's a crucial partner on the development of MQA. But, as AJ has acknowledged, the claim that he works at Meridian is simply false. So, where are the "corrections"?

Jim Austin, Editor
Stereophile

rt66indierock's picture

In 1996, the late Oxford mathematician Michael Gerzon joined with audio
developer Peter Craven to patent a compression system which packs audio code without loss. Meridian Lossless Packing (MLP) was licensed to Dolby.

Read more: https://www.newscientist.com/article/mg16021554-200-longer-player/#ixzz6AMYzcnbe

Max Townshend's picture

We have been following Hi-Res Audio since Pioneer introduced the DAT recorder in 1992, which the late John Bamford was a great proponent. It was obvious that 192/24 was significantly better than 96/24 on DVD Audio IF YOU HAVE A GOOD SYSTEM.
Tony Faulkner, the renowned recording engineer, masters all at 192/24. If he has to downsample to 96/24 for Apple or whoever, he has to add extra artificial reverb to make up for the loss of fine detail, especially the venue acoustic.
It is a crime not to record everything in 192/24 and the excuse that it uses bandwidth and too much space is rubbish these days.

X