This is Your Brain on CD Resolution Sound: Read The Study

Why does CD sound not only annoy many people, but also cause them to not want to listen or to listen in the background while doing something else?

This is a known phenomenon that CD resolution apologists simply ignore, preferring to claim that it's some kind of "learned response" by people who have wrongly conditioned themselves to not like CDs.

Their chief argument is that we can't hear beyond 20kHz and most of us over the age of 50 are luck to hear beyond 15kHz so what's the problem?

Please download the study and you will read a good explanation for why CD resolution sound, chopped at 20kHz, disengages the brain

Just as I was about to post this story I was sent a link to a story by former New York Times tech editor David Pogue who's posted yet another anti-PONO, anti-high resolution digital audio story that's based on an absurd "experiment" using a Radio Shack switch box and unsuspecting test subjects.

Read it:

Neil Young's PonoPlayer: The Emperor Has No Clothes.

I wrote Pogue and told him:

You haven’t a clue David. You not only don’t have clothes, you haven’t a brain.

You are the one peddling B.S.

Please read the study.

So please read the study and then look at what Pogue has posted.

Werd's picture
zzcorey's picture

Clearly the readership on this site does not feel this way, just email Mr. Pogue and call it a day. This is just going to cause some people to get angry, I'm sure you could find tons of misinformation by computer reviewers who think they know audio. Waste of time.

zzcorey's picture

im not talking about the study, thats cool, just the pogue postings here and on facebook

rpali's picture

I don't would make more sense to me if it were the brick wall filtering, but the suggestion that the simple absence of information about 22 khz makes music more fatiguing doesn't make complete sense to me. The thing is, although vinyl can indeed reproduce sound well above 20 khz, does it do so in any meaningful way? And more importantly, *did* it do so back when we all listened to albums, from beginning to end, far more often than we did today? Say, with 70s technology?

Also, I've never heard anyone say that open reel recordings, or cassette tapes, for that matter, caused the same fatigue, and certainly cassettes has a much lower high frequency response than even CDs.

I'm not arguing with the study, but I'm not sure all the pieces fit very well when using it to explain that CDs are so fatiguing.

MonetsChemist's picture

This is pretty clear "We pay special attention to the fact that FRS is accompanied by an intensification of the pleasure with which the sound is
perceived" p.3557. How does that not make sense? Are you saying your experimental results are different?

rpali's picture

What doesn't make sense is why you're arguing with things I didn't say. I have no issue at all with the study.

Bromo33333's picture

The physiological reaction of the brain to the full range vs truncated sound says 2 things:

1. The brain stays lit up for longer and more intensely with the full range sound being present.
2. Without the full range, the brain loses attention within a song or two.

The author is implying that your attention lapses with the truncated sound, and is engaged with the full range. Where things get controversial is when polled for subjective impressions, the results of the polling don't always follow that proposition.

(And this is the point where the hoards of debunkers get a little confused - it is clear we sense and react to the high frequency that our ears don't seem to hear consciously, so their claim of "you can't hear anything above 20kHz" isn't supported by the facts. It is clear we are perceiving it, and we react to music with it in it differently than we do the music without.)

You have to read additional papers (and get into the weird world of Cognitive Science a little) to make sense of it.

But also in another silo (Academic disciplines usually remain in their speciality and generally don't collaborate with those in other fields) in the AES, there is at least a little research that shows the pre- and post- ringing of 44.1kHz filter reconstruction creates audible artefacts when put through real-world speakers (interactions between the nonlinearities of the speakers + the ringing of the filters) - and that could be interpreted as "hashness" and there are a number of mitigating factors - use of apodizing filters can send the signal to the trailing edgeof transients, lending help to the attack at the expense of the decay, and the use of a slow roll off filter that starts rolling off early to keep the overall ringing minimized.

So, it COULD be that there is something with the high frequency components that you both sense and it helps hold your attention, and it's absence will cause you mind to wander and not be sucked into the performance.

ANOTHER plausible possibility is that the nature of the brickwall filter creates arefacts that are aharmonious that you might not like listening to.

Could be something else entirely, too. Or Both.

I think the point is, that we spend far too much of our time playing the "yes you do/no you don't" dance when it comes to audio, rather than use the science at our fingertips - new OBJECTIVE measurements to learn a little and then figure out how to get even better sound.

MonetsChemist's picture

The next time someone says this in the context of hi-res music, it will be LOVELY to point them at this paper and say "no, just a human being".

Thanks so much Mr. Fremer!

abelb1's picture

Blind tests are useless for music. I can tell the difference between high res audio and CD quality consistently but only when I’m listening for enjoyment. I’m relaxed, perhaps I’ve had a couple of drinks and I’m listening for the sake of enjoyment. Occasionally there will be this moment when the hair stands up on the back of my neck and the music sounds so good my thoughts are something along the line of “this is the greatest thing in the world”. This has only happens with high res audio or vinyl and over years of listening I’ve found it to be very consistent. I can hear the trolls limp minds rattling out the words “subjective bias”, bah. The last thing I’m thinking about when enjoying the music is what resolution it’s in. When purchasing downloads I opt for the high resolution format when available but I have a large amount of CD quality lossless in there too and I don’t fuss about which is which when I’m lobbing albums at my playlist, nor do I remember which is which most of the time. When I have one of those moments I may go back and look more closely at the album files and lo and behold…

Where do these people come off trashing high quality audio, and products made by passionate advocates when clearly they don’t care either way what format their music is presented in. These products aren’t for them, but that doesn’t stop them having opinions about what other people should and shouldn’t be listening to. How about they shut-up and leave the conversation with the people who do care.

Or at the very least, shut-up and let the market decide.

J. Carter's picture

I was just posting a comment on that article when I decided to link to your Gizmodo article and low and behold I see this article you posted.

David has been responding to many of the detractors that are posting, we shall see if he responds to me.

This is what I posted there:

"So a couple of things.

First I don't care what your audio engineer friend tells you the Radio Shack A/B box IS going to affect the sound.

Second $100 headphones are a far cry from a "nice" pair of headphones and are not what I would call a typical pair of "nice" headphones people would buy in this day and age. Not when the most popular headphones on the planet (Beats) sell for $200+.

Third, if there is any possibility that the iPhone was playing the tracks louder than the Pono or vice versa the test is not valid. People automatically pick the louder music, it's instinctual.

Fourth as another commenter posted Blind AB tests are a flawed way to truly test audio. I don't think there is a great way to do it honestly.

Fifth those studies you elude to and link to have been debunked by many people. You want more info on that, I encourage you to read this and this

I appreciate your efforts and the story is interesting however flawed it may be. As someone who owns a $1000 pair of headphones (Sennheiser HD700), several thousand hi resolution files purchased/downloaded from Pono and other sites and a hi resolution portable player comparably priced to the Pono (Fiio X5) I can tell you I can usually tell the difference between MP3/AAC files and hi resolution. Much of it is very dependent on the quality of the recording/mastering of the music."

Archimago's picture

And I don't think anyone should be too impressed quite yet without replication of the Oohashi data (from 2000 - 15 years ago!).

I've also looked at the 2002 report and their most recent 2014 paper. Makes it even more confusing, Michael.

Look, bottom line is that anyone thinking the difference is massive like the artists coming out of the Neil Young car really has no evidence to support such an audible difference based on research. They'll just end up disappointed when comparing to a side-by-side 16/44 downsample using the hi-res source. That's just the reality. I've been thinking about your post earlier this week as well as Michael Lavorgna's. My feeling is that even if there were identifiable qualitative differences, it's not going to matter for most recordings due to the poor quality we're getting these days. This includes the Pono stuff I looked at last week.

Summarized my thoughts here for consideration:

As I said in the post, I'd certainly be very happy to have good 24/96+ digital audio, but it's mainly out of a desire to get the "best possible" version of the recording (assuming the mastering is good) rather than high expectation of hearing a significant difference.

Bromo33333's picture

Above a certain level, it is a game of "subtle but significant" to those who are aficionados.

I find the use of EEG and fMRI to be incredible tools to know exactly how things are working. If you look at flanking research, you will find some pretty interesting things. One thing is that if you hear tapping - in sync with your heart, or the music, or something else, may predisose you to like r not like music. Even if the tapping is your own toe. fMRI scans included with this.

Also the look of color, or things unrelated to sound that is familiar or not.

But the results have been repeated, but they are repeated in the context of learning other or new things. The Physiological response is well understood at this point - it is the "what does it mean" that is the controversial part and the part that isn't repeatable.

Yovra's picture strikes me that it must be far easier to write an article stating that audiophiles are bunch of misguided fools than prove the opposite....
The best 'blind test' I experienced was with my girlfriend when I was randomly playing Beatles tracks off cd's and LP's and halfway "Drive My Car" she said from the kitchen: "this must be from vinyl, isn't it?". No; it wasn't the dust and scratches, but even far from the 'ideal' listening position the LP was a nicer and more musical experience.

isaacrivera's picture

Of course the same people who scream SCIENCE! when presented with empirical evidence are the ones who now gonna make empirical arguments against the study. Clearly the inaudible HFCs are not silent to the brain. Now I have to get an amp that has higher HFC response... sigh!

Mark UK's picture


So none of this stuff matters as we won't hear it whatever sample rate we use.

It doesn't seem to have occurred anywhere, on this and other sites. Not even to the 'experts'.

isaacrivera's picture

The methodology section explains it well. Not only D/A conversion, but speaker frequency were addressed. Read it, it won't hurt.

Journeyman's picture

So all DACS have a cut off at 20Khz? Some sources please or articles to support such a claim.I'm always curious about such claims.
Also define what you mean by "Cut Off".
I'll give a DAC and its specs behold the Foxtex HP-A8C
Now read the output section and explain about that 20Khz cut-off that all DACs have. If you are talking about filters or generated noise please do explain.
This stuff actually matters to people who enjoy audio, be it the nay sayers or the supporters of High-Res audio.
It matters because without discussion, audio quality will stay at a status quo, the same goes for mastering.
I'm one of those guys that doesn't love the hype around Pono and those super High-res files the player reads, specially because most people won't actually use headphones High-res capable.

Toptip's picture

That the brain, as shown on EEGs, responds to HFC, even if the subject cannot identify which version of the music he is listening to, says little on its own. I suspect if you nuked a subject, his brain would also show some "unheard" activity. The only relevant suggestion is that he finds music with HFC content more "pleasant." But that part of the study is a bit woolly. Did they know which version they were listening to? Is the "pleasantness" statistically relelavant? (i.e., far higher than in 50% of the cases? The study is silent on those matters).

In any case, I am biased because the study quotes a "Prof." M. O. Hawksford and everything I have read of him was a bit rubbish-y, prof or not.

Finally do real world LPs contain any HFC, whether or not the system can reproduce it?

Michael Fremer's picture
Up to 40kHz is possible but certainly above 20kHz and below 20Hz.
rpali's picture

Possible, but the masters didn't hold any information about 20kHz when LPs were the primary way to listen to music...and no one said anything about how those LPs were difficult to listen to. Similarly, there were no claims about any comsumer tape format or even radio, so why single out CDs?

Michael Fremer's picture
Not so. There's plenty of spectral information showing well above 20kHz on many records.....if I could embed an image here I'd show you...
rpali's picture

Yea, *now* there is, though I wonder about there being 'plenty.' Still, there wasn't any before the CD took over. There wasn't any on the masters, so how could there be on the LPs? No one said anything about LPs causing listener fatigue back then, so how can the same lack be causing fatigue with CDs now? Sure there may be other reasons but a lack of high frequency signal isn't one of them.

jusbe's picture

Do you have any evidence for that? For claiming that before CD, analogue (or even digital) masters were arbitrarily truncated to 20kHz?

rpali's picture

No, I have no evidence for such a claim, because I never made such a claim.

jusbe's picture

"Still, there wasn't any before the CD took over. There wasn't any on the masters, so how could there be on the LPs?"

Perhaps you might consider not repeating such statements then, if you don't have evidence to substantiate them? There are truly interesting discussions and learning going on around digital audio, measurement, auditory research and similar perceptual studies. I would encourage you to read some of them, such as the two presented here in this blog post.

rpali's picture

I said nothing about any arbitrary limitation. If the gear can't record it, it won't be there. The frequency traces I've seen from high-def reproductions of 70s era and earlier recordings rarely have anything in the 18 kHz region, much less over 20 kHz. Granted I'm speaking lately of mainstream works, and I wouldn't claim there are no exceptions. Do you know of released recordings from the 70s or earlier with any significant response over 20 kHz?

jusbe's picture

No. But digital brick wall filtering is a relatively recent technology. I'd be curious to see the spectra you've seen.

The issue is complex, not least since there will be several recording tools, and parts within them, which are designed to the 20-to-20 range on the assumption that nothing else is important. But so many instruments, and even the human voice, have overtones and harmonics that go way up:

rpali's picture

I certainly agree that there is information above 20 kHz. I just said, for the most part, it's entirely missing from older recordings. You somehow got the idea I claimed the cutoff was arbitrary, which I did not. My evidence is not at all systematic, just frequency traces I've seen people post about new high-res releases of older recordings in an attempt to determine whether they're worth buying. You seemed to disagree with my point about the presence of signal above 20 kHz so I wondered what recordings you've heard that I've missed and you say it's a complicated issue.

I'm not sure where you're going with all this but I'm going to sit-out the remainder of this dance.

Bromo33333's picture

Further studies have shown that "The Hypersonic Effect" as it is called is truly strange.

They have shown that the above 20kHz sound is NOT heard in the ear canal, but through your skin, hypothalamus and brain stem (isn't that weird) (2005) - and that with headphones you miss all of that HFC sound - you have to be enveloped in the soundfield to get the benefit.

Oddly enough is that in that 2005 followup (people had difficulty replicating the experiment, but the researcher found out they only used headphones) the HFC part has to hit the body, but the LFC can be on phones or in the room.

THe things that is cool, is when we get past the incessant "debunking" waste of time, is we can learn some freaky things!

It also explains the ultrasonic supertweeters when people report a "better sound" despite not being able to hear it with their ears! :-)

But these studies also show the inherent weakness of qualitative studies (such as the ones done by the Double Blind Testing crew) - and these days EEG and fMRI studies are the gold standard. The line of research has shown an indisputable (and repeated!) biological effect with HFC, especially with the LFC intact - the attention is maintained longer and indefinitely, and is only appreciated when your whole self is in the soundfield.

HiRez on headphones isn't much better than CD for engagement.

Here is the follow up study:

jusbe's picture

Fascinating reading, both research articles. Thanks for the updated one.

JGonzo's picture

There's a fascinating intersection to the writing of various folks like David Pogue - it's both highly reactionary ("Back in my day, we only had red book CD and WE LIKED IT! UPHILL! BOTH WAYS!") and simultaneously, strangely anti-luddite ("it's IMPOSSIBLE that vinyl could sound better - the march of progress is inevitable!"). Mostly, I would attribute the character (hell, existence) of these articles to the need to drive eyeballs to whatever it is that they are writing.

In the case at hand, Pono is the hot new thing, and one that is getting a reasonable large amount of press. As such, there's a pretty strong incentive for the so-called debunkers to fire up the mill for a fresh addition of grist. They have a readership/bunch of fanboys and girls and, as such, demonstrate something approaching a sense of obligation to tell the so-called smart people that they're wrong, and the wrong ones that they've been right all along.

Here, the argument is almost impossible to refute, in large part because Pogue's validity criteria are self-affirming:

Debunker: "it's not possible for anything to sound better than CD, because: science."

Audiophile: "decades of experience demonstrate that nearly any format other than CD sounds better - vinyl, SACD, DVD-Audio, hi-rez downloads"

Debunker: "but SCIENCE"

Audiophile: "observational data IS science"

Debunker (fingers in ears): S-C-I-E-N-C-E!!!!!!!!!!

There's no arguing with people who have a stronger desire to be FEEL right than to actually BE correct. As such, I would encourage Mr. Fremer to dedicate his time to penning more of his excellent writing that helps us extract so much pleasure from our vinyl collections. Besides, nothing makes internet trolls more unhappy than being ignored.

Ortofan's picture

...the 20kHz upper bandwidth limit represents a problem with CD sound. Were that the case, there would a similar sense of "disengagement" when listening to FM radio - which is limited to 15kHz. I've listened to hundreds of Met Opera, BSO Tanglewood and Prairie Home Companion broadcasts on FM and never once felt "disengaged" because of the somewhat limited bandwidth.

Mark UK's picture

Anything though a DAC,
FM typically cut off at 15KHz
DACS typically cut off at 20KHZ.
CD, 44.1, 96, 192, DSD, whatever. It all gets chopped off at 20KHz with almost all DACs. And it is NOT the, no matter how hgh you put it.

And VERYONE overlloks that vital point.

So all this high-res stuff is complete BS until they make DACs with a wider frequency response. Which nobody does at present.

Michael Fremer's picture
like LP, slow, graceful roll-off.
Ortofan's picture

...the 19kHz FM stereo pilot tone while still maintaining flat response to 15kHz.

kozy814's picture

I would rather not pay $400 for a digital music player. Now, if this player and it’s specially sampled catalog of songs were to make me “feel” the way I do when I play my LP collection, I could be warmed up to the fact that this might be something to look into. All science aside, the emotional connection I have with analog music is the real deal. I almost never sit down in my listening space to play a CD. The ritual of playing a nice sounding LP (Aja for example) is one I strive to share. I invite people to sit in the “sweet” spot to hear the magic. And newbies walk away knowing there’s something to the hype. I can understand why a blind “taste” test would generate mixed results – These are tests with earbuds (even the best models are not great for listening) in a restricted environment. When the purest pleasure of music resides in the unexplained subconscious, I can see how this type of controlled test could debunk a device like Pono. This product is bound to take a pounding from all the tech-weenies that will cite science as the main reason that hi-res sound is more a myth that reality. But the same folks that can’t hear any difference in LPs versus CDs will probably never search for a reason to change their own minds. That being said, when the Pono becomes a failed experiment, I may shuck out my $50 for a used one to see what all the fuss was about.

Bromo33333's picture

Thank goodness SOMEONE decided to use Cognitive science techniques to show what was going on! This will cut through the BS and reveal that conscious subjective perception isn't all of what's going on.

If the AES and the various debunkers are truly scientific, they will review their methods, and adjust their opinions. IF.

ravenacustic's picture

Even on systems costing a half million $$$ I almost immediately go into la la land when being played digital and that includes hi rez folks. As much as I WANT to enjoy the music, that is what happens to me all the time. Ortofan has a valid point regarding FM and its frequency limitations vs digital's limitations. No, there is something else going on with digital that at least in some people is a serious shortfall.

Interestingly, I have a very science oriented friend who will defend the cd. College educated, engineering guy who will be happy to haul out his latest vitamin claiming to have all kinds of health benefits. Of course snake oil in an unregulated environment but he believes it. But he won't believe what so many are saying about digital. Go figure.

barrysconspiracyworld's picture

While the study is certainly interesting, there are much more relevant factors within the normal audible range. The engineering limitations of implementing the CD standard are quite significant. The standard itself can come under worthy scrutiny from a theoretical standpoint alone (especially linear 16 bits of resolution).

Gizmodo dude is clearly the Audio Ninny of the Year, but there are numerous more powerful and relevant arguments against his drivel, which he delivered with Churchlady-like self rightiousness.

Mark UK's picture

I see they still carry on as if I never pointed this out.

Ferquency response. What else? Not filters or whatever, just frequency response. Same as you see on a specification of a speaker or an amplifier.

You don't need a reference or a 'white paper' from 'experts' These guys didn't think of it either in their 'white paper'.

JUST READ ANY DAC MANUFACTURERS LITERATURE. the frequency response will be there somewhere. Likely the same place as it tells you the sample rates it can handle.

A typical DAC frequency response is 20Hz to 20KHz. That's your lot. Allow say, another 2KHz for rolloff (and I'm being generous).

So no matter whether on say a 192 file or a DSD file, and REGARDLESS of musical information on the disk at say 40 KHz, and the filter at say 50KHz, NONE of it is going to come though. GOLDEN EARS OR NOT. It's worse than a vinyl cartridge.

So lets see-
Amplifier - 10 Hz to 100 KHz.
Speaker - 20 Hz to 30 KHz
DAC 20 Hz to 20 KHz.

Golden ears, DSD, 192 KHz file. whatever.


Journeyman's picture

Mark something must be wrong with your observations and I say this with all my respect.
I'm one of those guys that doesn't believe in audio voodoo but the frequency response of good DAC these days goes way beyond 20kHz.
Look at this link from Archimago:
I wont change your opinion I'm sure but at least Archimago's site is always nice to visit.
Best wishes.

Michael Fremer's picture
First of all we don't use the term "Golden Ear" around here so shove it. Secondly you are referring to measurements of DACs typically given for 16/44.1 audio. Here's a measurement from Stereophile's review of the PS Audio DirectStream DAC:

"but the 192kHz response (blue and red traces) rolls off gently above 50kHz, reaching –3dB at 60kHz and –9dB at 80kHz."


Bromo33333's picture

Actually, most higher end DACs can output the higher sample rates. So do higher end CD players ...

Mark UK's picture

Why would a 'higher end' CD player need to output the higher sample rates when a CD is limited to 44.1?

And of course 'higher end' CD players are 'dedicated' to achieve 'the best sound quality'. None of this high street rubbish of allowing access by other sources as that would of course degrade the 'purity.

So 44.1 it is.

Bromo33333's picture

Meant to say: Most commercial Silverdisc gear can output above 20kHz (like for SACD or DVD-A), and DAC's as well. When limited to a 44.1kHz file, OF COURSE it will limit itself to 20kHz. But when fed a 96kHz file or a 192kHz file? Higher than 20kHz.

bdiament's picture

When I first started mastering for CD, in January of 1983, I found it very difficult to go home and listen to vinyl. That is, after an evening of enjoying my vinyl collection, I found it particularly difficult to endure a day of 16/44 digital.

Then I found the secret.
To make CD sound better, stop listening to vinyl. ;-}

Actually, nowadays you'd have to stop listening to properly done high res digital too, especially 4x rates like 176.4k and 192k, which to my ears, are what digital promised back in '83.

Another way might be to spend some time listening to eMPty3 files. CD sure would be a pleasure after a few of those.

I do understand that some folks say they don't hear a difference between 256k mp3 and CD and don't hear a difference between CD and 24/192. That says they find 256k mp3 and 24/192 indistinguishable. Or, one can toss 95% of the information and not lose anything --- to their ears. Personally, I believe if they're enjoying their music, more power to them. I wouldn't want to take that away. Just don't touch my 24/192.

I don't understand why folks would insist that what they hear or do not hear reflects what others hear or do not hear. More importantly, I don't understand why they would care. As I see it, if they were truly confident, they wouldn't care. Instead we see "white papers" saying 24/192 is *worse* than CD. (Wait a minute! I thought you couldn't hear any difference! And Charmin is a 'white paper" too.)

I'm reminded of the manager of a major NYC recording studio who once told me (though for the life of me, I don't know why) that women don't have orgasms. (!?!) His "evidence" was that over the years, he had been with dozens of women and on not one single occasion had he witnessed one of them having an orgasm. (!?!)

A) Was it good for you too?
B) Wait a minute. Let me check my EKG and I'll let you know.

Best regards,

Billf's picture

Great post.

PeterPani's picture

Yesterday in Vienna we had on air a very interesting comparison of the new - in all the media - highly appraised Callas Remastered box. It was compared with earlier CD's a n d vinyl. And this on the most important Austrian radio station, which is still aired in FM and received with my good old tubed Leakthrough tuner. You can imagine the conclusion and it was very clear in front of the speakers: nothing beats the vinyl. Even when playing Callas "difficult to replay" voice, digital remastering at highest level is no contender. Who is able to understand German - the program is still 7 days on the "radiothek" (logically by now only in digital format):

Erik_Rohr's picture

I love vinyl, it makes me forget the technology and enjoy the music. Digital keeps me thinking I love 192kHz, it's so hi-res!

Mark UK's picture

It's not a 199% of course - I see the one you mentioned say 80KHz.

But it is a big majority. And any DAC with 192K sample rate and a 20KHz top frequency response is inherently stupid.

Some expensive ones don't mention it as all.

Journeyman's picture

Most 24/96 DACs these days can deliver a decent frequency response, there was a time in the 80s when DACs were really bad, its was one of the reasons for audiophile bashing audio CDs besides really bad audio mastering but thats another story and I won't go there, too much content for a simple comment.
Even so I get your point about that 20Khz cut-off in a way, a DAC isn't just the digital side, one also has the analogue output & power supply stages.
Doesn't really matter the specs of the DAC chip/hub if the output has PSU noise and a bad implementation of the analogue output, bad ICs specs, etc. I enjoy electronics but building DACs is out of my league, I "know" guys who build them but they are engineers with lots of free time! XD Again best wishes to you Mark.

Journeyman's picture

If possible use the "reply under my comment" that way I can receive an update on my email and we don't spam the comment area of the site.
I'll be around.

cgh's picture

... I really am. I don't need any convincing, but this study establishes what would be called a neural correlate. Jumping to some of the conclusions I've read is just spurious, specious, and smacking of confirmation bias.

I realize this makes me unpopular here, but objectively all this study says is that the brain starts consuming some glucose when exposed to ultrasonic auditory stimuli. Take a chess grand master and pit them against a chess novice. Record their respective fMRI. The latter's will light up like a fcuking Christmas tree, the former not so much. The reason? Pattern recognition. To say that the latter is better at chess is erroneous. My point is that 1) the brain lights up like a christmas tree at all kind of stimuli; 2) we know that these "neural correlates" lie to us all the time, and this is when we know precisely controlled cause and effect; 3) to say that measured brain activity with ultrasonic input means that this is why CD suck is a leap. There are too many other levers in CD playback in addition to the cutoff.

Bromo33333's picture

One of the main ways that people trying understand the brain works with an fMRI is to note which part of the brain lights up with what kind of stimulus. A good study usually will go a little different than "lights up like a Christmas Tree" - usually noting what parts of the brain lights up, noting if the situation is familiar (say Chess) and unfamiliar (say Cards Against Humanity). Once could do the same with a piece of music, too.

Given the studies, you can conclude that the brain activity was sustained for longer with the HFC than without. You can also make conclusions about what part of the brain has lit up - the intensity as well. The author makes his claim, but the puzzling piece is that when given subjective surveys in later studies, the subjects expressed opinion did not agree with the fMRI. This means there is more to the story to find out in the realm of CogSci.

For you and me? Clearly the above-20kHz spectrum is able to be perceived by us in some manner, and likely has an effect on our perception of the music we're listening to. Of course, this is what many audiophiles have been contending, and the subject of a lot of controversy in those Double Blind Tests.

And what I find encouraging about the studies is that fMRI and EEG measurements are objective measurements, and always more reliable than anything subjective (generally at the end of a traditional DBT there is a survey where they ask people if they can tell the difference - and then use the results as their proof - which is at its core a subjective qualitative measure, and inferior to quantitative data.)

Using the DBT + the DBT+EEG and even the fMRI you can make a few statements supported by facts:

1. We can sense above 20kHz. (Ear-based hearing it well understood, but later studies show we hear also with our skin and brainstem!)
2. The above 20kHz portion of sound engages our brain if it is with the below 20kHz part - and the engagement is longer than with just the below 20kHz portion.
3. Looking at the body of work, this engagement may be associated with increased pleasure. But is also may not be - the results from the body of work shows this is not clear - the subjective appreciation is mixed (but so is music, too).
4. This level of engagement is not reliably measured with the questionnaires given in the DBT favored by a lot of debunkers. I think it is safe to say it is entirely missed.
5. The DBT crew might be better served by adopting better measurement methods with an emphasis on the quantitative rather than qualitative.
6. It is clear there is a lot we don't yet understand about how we perceive sound and react to it.

And, most importantly, we ignore and dismiss our own senses and states of mind at our own peril. Empirical observation is the first step in a deeper understanding.

Especially the "I know you have a piece of paper with equations telling me that CD is everything we can sense or need, but I seem to like LP and high Rez music better - my mind doesn't wander, and I get caught up with it easier" - that's a sign that there might be more going on than meets the eye.

J.D.'s picture

On the pono/ayre sound, lossless. Uncompressed. As-is.

On the apple side, AAC compression-format. Psychoacoustic algorithm modifies the soundfile, to conceal data reduction.

People in Pogue's "test" -who, let's face it, are neither a random sample or an expert panel-- preferred the reconstituted files, the processed material. Referred to as "perceptual coding" this uses psychoacoustic models to reduce file size and conceal the loss.

People like processed everything.
Especially when it is constructed to cater to them. Mp3, equalization, bass-enhancement, botox, collagen, youtube-channels, facebook-feeds, 'those-who-purchased-this-also-bought' sidebars on amazon. Whatever it is in chicken mcnuggets. Hydrolized corn syrup. Processed stuff -- designed conspicuously and entirely created for Popular Appeal, for immediate consumption.

Wears off over time, most of that. The "appeal" part, anyway.

Point being that Pogue's article is a hatchet job not only trying to villainize pono/ayre but the whole conception of high resolution sound. Fraudulently using "processed" files to compare to unprocessed.

Analogico's picture

I'm a music and LP lover but also a tech geek so this topic really pick my interest, One thing that I find amusing about the arguments held on both sides of the audio table is that they frequently reduce to absolute terms (i.e. this sucks, this doesn't) what should always get qualified using a systemic view.

Music reproduction is the result of a system, if the recording, the first link in the reproduction chain, is poor to begin with then who cares if the file is an MP3 or DSD? Let's assume that the Pono player, a self-contained system itself, is somehow built to play "perfectly" Hi-Res files, what can be said about the interaction between the Pono player and the headphones used? This gets even more complex on an LP-fronted system.

Regarding frequency hearing... in his book "This is your brain on music" Daniel Levitin assures that most adults over 40, me included, won't be able to hear anything above 15khz as in, for example, the last seconds in The Beatles' "A day in a Life". Maybe I'm not able to hear it all but that doesn't hold me from enjoying those mono LP reissues way more than the stereo ones.

Bromo33333's picture

Read the paper. You might surprise yourself. You may not be able to directly hear above 15kHz, but you might still be able to sense it.

I have a Pono. It's pretty nice - for $400 you really do get a first rate portable player. The rest is a sad little article by a journalist trying to be the "Audio Shock Jock" by insulting his readers and Neil Young. There is a special circle of Hell for people who try to profit by hurting others.

Engelsstaub's picture

I own, love, and collect vinyl LPs. I also have CDs, "Hi-res" burned to DVD-As, and even cassettes.

I just like music. These sort of debates do nothing to further my enjoyment, but they rather only serve to distract me from what's important.

christopherf's picture


This has to be the most unbalanced article; perhaps you will write a rebuttal in the form of a letter to the editor. Bob Clearmountain and Bob Ludwig are referenced in the piece; it seems such an old argument; I thought each format has its' own strengths but this article is very patronizing; it says what vinyl lovers are enjoying are "distortions"; the tone is patronizing; you can like vinyl but don't say it is better than CD; indeed!

Here is the link to the piece:

I am amazed that you have not grown weary of addressing this with the "experts" in this case the engineers; I regret that most people who read this article will take it as gospel.

bdiament's picture

It depends on which engineers you speak with. As I've always said, if you ask three engineers a question, you'll get at least four different opinions. (Five of which may be wrong.)

I remember in the earliest days of digital, when a college said to me "Just look at the frequency response" to which I responded "Just listen to the music!"

What is "good", "better", or "best" depends entirely on what one is seeking.
I know lots of engineers to whom, like me, CD is the "cassette" of digital. (And some other formats are the 8-track.)

Best regards,

audiof001's picture

Always the entertainer, Pogue was one of my favorite tech reviewers... I say WAS. He's lost all credibility with me now. Done.

marmaduke's picture

She blinded him with science!

AnalogJ's picture

Most of the audio engineers I know listen for audible frequencies and separation of instruments in the soundscape. There's a store in Waltham, Massachusetts called Goodwin's High End. They sell very expensively priced equipment. I once went in looking for a VPI HW-19jr turntable, a table at that time retailing for about $1000 with tonearm w/o cartridge. They said they carried it for parents to buy their kids for their dorm room. But I digress.

All the sales consultants are audio engineers. They listen for detail and soundstaging. I listen for musicality, musical flow. They will almost always go for "accuracy" over what sounds like music. I remember clearly hearing nuance and micro-dynamics in one integrated amplifier and stilted flow in another. The sales guy liked the other because he could hear more instrumental separation in spite of it sounding more sterile.

Anyway, David Pogue has responded in his post by reconsulting the Radio Shack salesperson, who assured him that copper is copper is copper, and the A/B switch used was just copper (No accounting for the connectors or anything else). When you want to consult someone on high fidelity, Radio Shack is the perfect place to find an expert. I posted that he could find someone from Stereophile, or someone at McIntosh in Binghamton NY, or consult Bob Ludwig at Gateway, or Bernie Grundman. There are a myriad of people who truly deal with the high end.

How a guy like David Pogue could lack the intellectual scientific curiosity for addressing this issue is befuddling.

Mark UK's picture

A whole lot of '192' DACs cut off at around 20KHz. We don't all buy ones costing several thousand dollars.
Why not? I've got money coming out of my ears. But it's pointless. Unlike amplifier etc. DAC technology moves fast. A five thousand dollar one now will be bettered by a thousand dollar one next year.

You don't like being disagreed with do you? I see why the guy you are whingeing about would not publish your comments.

I also see you removed your "BETTER YET BUGGER OFF" from your post above. Who are you? Just some self-appointed so-called 'expert' who makes a living by keeping the audio pot boiling.

Bromo33333's picture

There are a slew of $500 DACs that put out more than 20kHz.

And, friend, a good DAC today will be good in 10 years time. It is a question of priority, and based upon the studies performed and additional research afterwards (it is fascinating if you are so inclined) - you will get more out of your hirez music if you go that route.

If you don't, and don't want to, or "can't" (which I doubt) ... not sure it matters - it is a personal choice and says nothing about the science of it all.

Mark UK's picture

Personally I don't, for the reasons I stated. But my other stuff is quite expensive.

Hearing? It is not a matter of telling others what they can't hear.
But 50 years of hearing testing says adult human beings can't hear above 20 KHz and the opinions of a few 'audiophiles' aren't going to change that. And above 16 KHz it has to be loud. 19 KHz or so the required volume starts to cause permanent hearing damage.

Nyquist? It's a theorEM, not a TheorY. So it's true, just the same as Pythagoras' triangles are true. Given the conditions we are all aware of, (frequency no higher than the half the sample rate) the reconstruction is perfect. It is NOT a step function as often drawn for the convenience of non-mathematians.

Bromo33333's picture

It's actually "The Nyquist Sample Rate" and it is both correct that if you sample at twice the rate of the highest frequency you hope to reproduce, it is mathematically possible to reconstruct that waveform.

It is also extremely amateurish engineering to set the anti-aliasing filter right near that 2*f point because it quickly becomes expensive, difficult and unimplementable - with all kinds of distortion (ringing artefacts) and so on. In the 30 odd years since Redbook, recording studios have had to go to very high sample and bit rates and perform DSP gymnastics in order to have good sounding CD's and decent performing equipment. The Redbook 44.1kHz rate was set there, not because they made a mistake, but because of the nature of the tech they were working with. Had they waited 6 months, they could have chosen 48kHz which would have made that Brickwall filter easier, and waited 12-24 months, then 96kHz would have been possible, and you could have a really easy - no-ring filter to nail it.

The positive side of the nasty standard that is Redbook, is that it forced us to get really good at DSP to begin to overcome some of the flaws (but unfortunately the pre and post ring is with us!)

Michael Fremer's picture
Well since this is my site Mark, built for me by people who have faith in my abilities, I am not a "self-appointed" expert. They have appointed me one. Yes I removed "bugger off" because I thought better of it after one of your stupid, angry posts about how DACs don't go beyond 20kHz.
Mark UK's picture

Try Journeymans's post to see how it should be done.

"We don't use the 'term 'Golden Ear' around here".

No, a quite old 'reviewer' wouldn't. But you think you are. Your calling the other guy who wrote something you disagreed with an "Outsider" proves that.

Mark UK's picture

I find vinyl more 'engaging' than digital at ANY sample rate. it keeps my attention more. But I can't live with background noise.

But then my '192' DAC only goes up to 20KHz (as do many others) whatever the sample rate. So even with 'high quality' ears and 12 files anything above that won't be coming out of the speakers anyway.

What to I buy? The music I like of course. At the highest sample rate available. What's a few dollars more?

Can I hear the difference. Of course not, for the reason above and also because I'm a human being. But next year a 1000 dollar DAC may be better than the one I have now.

And my ears may improve with age, like some around here appear to do. Why not mine too?

soundman45's picture

This is one argument I wish would go aways. They are two different things. One is not better than the other they are just different ways of capuring an electrical value.and in order for digital to sound exactly like analog you gotta have a hell of alot of information in those numbers.

Mark UK's picture

As a person with 40 professional years in the 'big' computer business (we do know - we've been doing computers since 1943 and they wouldn't work at all, both then and now if it wasn't figured out before they started). The requirements are -

1) 24 bit sample depth. (16 should be ok but some eccentrics would consider it marginal.)

2) Human hearing top frequency. 20 KHz

3) Nyquist.

That's it.

Bromo33333's picture

The human ear is about 20kHz.
The high frequency effect is a strange one - and it appears the body senses the sounds our ear can't hear (and to be fair, the stuff it can, as well, but less well).
Nyquist tells you on paper the minimum sampling rate you need to reconstruct the waveform at that rate or below. But it is a poor spec (you should know this as an engineer, I assume you are) that can't be implemented properly. Given the nature of the anti-aliasing filter that has to be so darn sharp (they should have given the poor engineers that had to try to implement the spec writer's product a break and made the sampling rate 70-80kHz or higher, actually, to have a slow roll off filter since with sharp ones, you start having phase shift issues about an octave below the knee in a high order filter!)

So, you get phase shift (which is compensatable provided you make the recorded product with a lot more bits and a lot more frequency).

You also get pre- and post- ringing in the filter response. Itself is prety low level, but when it gets mixed with the nonlinearities of the speaker and other elements in the chain - you will get audible products - and the only thing you can really do, is increase the sample rate to make the filter better.

The issue with digital is not now nor has it ever been the theory. It has been the implementation and the box the spec throws a designer into.

SO - given this engineer is weighing in, and needs a little mercy from the impossible task, let me humbly submit the following:

1) 24bits or more (to allow your production people some slack to play around)
2) Human hearing tops out at 20kHz, but sensory perception seems to go higher, to say, 50-100kHz.
3) Need for low level nonlinear artifacts, so take Your Nyquist and double it. This will spare that poor Engineer from a drinking problem and allow him or her to enjoy his or her work again.

Michael Fremer's picture
Filters ring, filters produce phase shift, filters are the big problem. They are audible. They suck. Insisting on keeping them so close to the range of human hearing is just plain stupid. Obstinate and stupid.

Beyond that, if you've convinced yourself that the reconstruction is "perfect" fine. Enjoy. Something about it doesn't sit well with my brain.

Thus this website.

Mark UK's picture

I haven't convinced myself of anything. You don't have to convince yourself of a theorem. Just have to read and understand it. It's not subjective, it's fact. You don't have to 'like' Pythagoras' triangles.

I simply suggested 96 might be better for precisely the reasons you state.

So why ignore that part of it?

Yes, within the limits it states. So we move the limits.

Nyquist is Nyquist. A theorem not a theory. It reconstructs perfectly provided the 'rules' are obeyed. It has nothing to do with any 'best' sample rate, that's a 'variable' you plug in to it and off it goes. Just tells you what it has to be to achieve the conditions you want.

Which is AGAIN why I suggested 96 might be better as Nyquist will again reconstruct it perfectly but the frequencies it can reconstruct perfectly will be higher.

Simple and clear. As it was in my earlier post.

bdiament's picture

I would submit that any DAC with 192k on its spec sheet in the sample rate section but 20 kHz at the top of its frequency response is not going to do 192k very well. Perhaps this is one reason why so many folks don't hear a difference between 192k and 44.1k. I've heard many DACs spec'd for 192k that actually sound *worse* at this rate than they do at the easier 2x rates like 96k and 88.2k. (This, I attribute to clocking that is not up to the task and analog stages that are not performing at the wider bandwidths.) Apparently it is easier to put the number on a spec sheet that to actually design to the potential of the format.

My favorite unit specs its frequency response as 1.8 Hz - 64.7 kHz. It is also an ADC and via its mic preamps, the frequency response is, admittedly, not as good. It is only 2.9 Hz - 64.7 kHz.

By the way, one of the most pronounced benefits I hear in 4x digital rates, when they are properly done, is in the bass, which sounds more like bass in real life than I've heard from *any* format.

Another by the way: 96 dB is the *signal-to-noise* ratio of 16 bit and is all too often confused with its dynamic range. Listen to low level sounds at 16-bit and hear the bleached harmonics and truncated spatial cues. In my opinion, 16-bit has a real world dynamic range that would be happy to approach 10 dB. Okay, 20 dB if you don't mind it getting a bit raggedy. (It seems a lot of folks don't hear this. It also seems that a lot of other folks do.)

Ultimately, the recording itself makes a much larger difference than anything that follows it. I've long said that in my experience, 90-95% or more of a recording's ultimate sonic quality has already been determined by the time the signals are leaving the microphones. Everything after the microphones--the mic cables, A-D converters (or analog tape and machine), mains power, recording format, mixing (if any), mastering, and delivery format--only determines how much of that original recording we get to hear.

All just my perspective of course.

Best regards,

Mark UK's picture

How will I know if my equipment is any good? All we have is the file.

So you, the guy from Sony, the guy from CBS etc. come to my house.

I play you each a couple of your own recordings.

You all say "Yes, that's exactly how I wanted it to be".

I can give up 'agonizing' and 'obsessing' at that point and go and cut the grass :)

junker's picture

I'm sure there is something to the study, but it is my feeling that is has very little to do with frequency response. Why because my vintage Altec horns don't go much above 16kHz and they are some of the best speakers I've ever had - or heard - especially with the external Duelund-Mundorf crossovers I've built for them. I mean they sound really damn nice and musical!

Phase issues related to digital filtering? Sure. Pre-ring? Sure. 16-bit quantization noise/error? No doubt. A side effect from the use of dither? I dunno. The catch-all boogey man, jitter? Jitterbugging my brain for sure. Reconstruction anomalies not represented by some idealized Shannon-Nyquist theory and maths? Absolutely no frigging idea.

One thing though is that hi-res and DSD are better than 16/44.1 which is better again than even 320AAC. Now my main problem is watching the %^&* out for these remasters. I keep thinking I'm getting newer and better, and they are just compressed and limited to hell. Reminds me of how TV commercials were before Congress stepped in to make that illegal. They must be doing one single mastering effort that gets elected and approved by the studios for iTunes, CD, and hi-res releases. They all need to sound good in the car, and with earbuds, right?!

If we can get some decent 24 bit content > 48kHz with good mastering - less is more - we might be some real great progress here. A lot of the SACDs just seemed to be mastered very well particularly if they were just archival transfers from the master tapes. This remastering stuff right now has stood out to me as being Public Enemy #1 even before the format wars. Thanks for at least fighting the good fight Michael against the nihilists.

Michael Fremer's picture
That is why we name names....
Mark UK's picture

Remastering is just some big-headed ego-filled jerk saying -

"I'm a big deal up to date record producer. I can do it much better than the original one did"

Ego, that's all.

I've just re-written Shakespeare's 'As you like it'. Shall I send you a copy?

junker's picture

Hah! No thanks!!! ;)

Michael Fremer's picture
Your ego is even larger. If the mastering suite's electronics are improved it's very possible that the "re-master" will sound better than the original or a previous re-master. If the original mastering was dynamically compressed or limited because of the limited playback abilities of turntables during the 1960s it's very possible that a re-master can be produced without compression. If an original was mastered on Altec A-7s it's very possible that a re-master using a modern loudspeakers will produce a more pleasing result played back on a modern loudspeaker.
junker's picture

I can produce a litany based on DR Loudness vs pub date is we want to go that way but I'll start with
Ian Cooper at London’s Metropolis Studios on Morning Glory by Oasis:

And while we are naming names let's throw in Rick Rubin and Vlado Meller:

Howver, I think it has less to do with the mastering engineers than what their customers (studios and producers) are requesting stylistically - it goes to 11!!!

And we also see this pretty much universally on vinyl vs. CD remastering. Take Jack White's Lazaretto for example:

Ortofan's picture

...look at the in-room frequency response of the Wilson Alexandria XLF and MAXX3 speakers chez MF.
Doesn't look like much above 15kHz would be audible from either of those speakers.

Michael Fremer's picture
(or well-interpret graphic information): "To get a more detailed look at the Wilson's behavior in the frequency domain, I moved the microphone forward, along the line connecting the height of MF's ears to the height of the tweeter, until it was at my usual distance of 50". The black trace in fig.3 shows the Alexandria's output averaged across a 30° horizontal window centered on the tweeter. The response is generally very flat—flatter than at the listening distance—but with the lack of energy between 2 and 4kHz still apparent.

The output of the soft-dome Convergent Synergy tweeter smoothly extends at full level almost to the 30kHz limit of this graph, whereas the inverted titanium-dome tweeter used in earlier Alexandrias, as well as in the MAXX 3, peaked at the top of the audio band.

How all of these complex measurements integrate at the listening position is something a microphone measurement can only approximate.

I had two Russian guys insist to me that I had "boomy bass" in my room because of how they interpreted the measurements. I insisted I did not have "boomy bass". They insisted I did.

SO I INVITED THEM OVER. They were shocked that I would do that.

They came and brought their test CDs. When they were finished they concluded in unison:


Then I said "let me play some vinyl" and they said "no, we don't like vinyl". And I said, "I insist" so they glumly sat there as I placed the stylus on the record (a copy of "Porgy And Bess" with Maazel and the Cleveland Orchestra on U.K. Decca).

After a few minutes they said in unison "WE HAVE NEVER HEARD SOUND LIKE THAT. CAN WE GET THAT ON CD?"

I said "yes but it won't sound like that."

A short time later I got an email: "We got the CD. It does not sound like that record!"

Ortofan's picture

Take a look at the graphs in figures 6 & 7 of the XLF review.
The data appears quite clear, with at least a 15dB rollof by 20kHz starting below 15kHz. If you're suggesting that the information in the graphs is not valid because the microphone output is only "approximate", then why were they published in the first place?

What does "two Russian guys" and their concern about "boomy bass" have any relevance to frequency response at the other end of the spectrum? The wavelengths and room modes involved are entirely different.

Michael Fremer's picture
Was that they (ENGINEERS) interpreted the graphs to indicate "boomy bass". They had to hear it to know that wasn't true. At the other end of the spectrum. Same thing. What's heard taking all of the performance parameters into account and what's measured by one graph are not necessarily the same thing.They were published in the first place because that's what Stereophile does: publish in-room and semi-anechoic measurements.

Measurements tell you some things like "CDs are perfect" but listening tells you or at least it tells me, something's not right....

Werd's picture

That Radio Shack switch box would drag down the sound of both players.

Tim_Corn's picture

I have the solution that will solve this problem once and for all.

Then us vinyl guys who could care less about digital won't have to listen to a bunch of deaf digital guys argue over who's poo poo smells better!!

All said in jest of course... or maybe not... :-)

Werd's picture

I think a alot of people are missing the point. The study shows neuro awareness in a HF range. It isn't the audiblility of sound in that range. It is the high frequnecy noise that exists in the 20k+ range. The point of high resolution in sampling is to rid out that noise. The study suggest that ultra HF noise is perceived. So the goal of a 96 dac is to raise the high frequency noise above a value that 48khz can get, Thus improving exposure to fatigue. No one is going to hear a cymbal crash value at 40khz but they will perceive a 40khz noise pitch.
That is what the study when appllied to home audio tells me.

Jim Tavegia's picture

even I can hear the benefits of 2496 as increased detail and a unmistakable smoothness to the presentation and love all of my SACDs, even if they are not perfect sound forever. I own 4 SACD players and everything I record for my friends get recorded at 2496 and redbook and even they are believers even if it really doesn't matter to them. I am the only audiophile in the bunch. It is there loss not mine and I am over letting it get to me.

I do wonder what Dr. Kal's take is on this study? He is really the high rez guru at Stereophile I would think.

Werd's picture
Bromo33333's picture

The additional research after this rather important paper form 2000, shows that the High Frequency sound ins't sensed through your hears, but through your skin and other parts of your body. Likely those parts aren't losing as much acuity as the ear over time.

But also, the lower resolution files have ringing, which itself would be ignorable, but mix it with the nonlinearities of a voicecoil and it becomes audible since it creates intermodulation products with the ringing.

tnargs's picture

Please don't tell me that Analog Planet has only just now come across the Oohashi paper from 2000. Are you also aware that every attempt to replicate its finding, unless Oohashi is involved, gets the opposite result?

Science #101: An experiment(er) gets a new result that contradicts past knowledge. What next? Do scientists and interested observers everywhere leap up from their chairs saying "I knew it!"? Only on Analog Planet, perhaps. On Earth planet, they replicate the experiment and its result, before getting carried away. Classic (but far from lonely) case in point: cold fusion. Oohshi is the cold fusion of audio.

cgh's picture

So forget my arguments (and a few others) that the conclusions presented here were not supported by the paper. I failed to even notice the pub date. Can you cite the experiments / peer reviewed pubs that failed to support the 2000 conclusion?

Bromo33333's picture

Actually that is not true, there were and are papers in peer reviewed journals published even today along this thread - but it is in the field of Cognitive science - and given the nature of our academics won't get much press in that other silo occupied with the AES.

The physiological response is well documented and not in dispute. What IS currently puzzling is that the listener preference is all over the map, and that the physiological response ot the high frequencies is lost with headphones, and is only repeated when the entire body is exposed to the soundfield. In fact, you can pump the below 20kHz into headphones and just the above 20kHz into the soundfield and get an even STRONGER response than just pumping the whole signal into the room.

The interesting thing here, is that in judging scientific results, objective measures (such as an EEG or fMRI) generally should be given more weight than subjective ones (the survey after a double blind test is such an instrument. What would be cool is to augment it with an fMRI and see if the declared preferences are in contrast of the measured brain response or in contrast to it.

Also there was another rather small paper actually published in the AES that no forum warior seems to be aware of ( that measured the distortion of a 44.1kHz samples signal through speakers - the ringing creates IM products with the speaker that are audible, too.

So ... there are 2 things: We respond to high frequency content, though it isn't throug the ear canals. Also, the distortion of anti-aliasing filters are measurable and audible when pushed through a conventional speaker since it created intermod products that are rather high.

So before you try to do a hatchet job on some science that disagrees with you - it would behoove you to understand what the "controversy was" what parts weren't controversial, and to educate yourself about things flanking this argument that might just shed additonal light that we can all use to get better sound.

I see 2 very strong arguments for higher rez digital and analog, both backed by scientific findings.

Michael Fremer's picture
That AP only recently came across that paper. So we published it.

If "science" claims measurements prove CD is transparent to the original source, then the science is wrong. Science is not infallible. It's then back to the drawing board.

Whatever the reasons, I cannot stand listening to CDs for more than a few minutes. I can listen to and be engaged in music from vinyl for hours on end. At the end of the listening session I am refreshed. After a few minutes of CD listening, I become irritated and unhappy and definitely not relaxed.

You can say I am "prejudiced"or "delusional" or whatever but I went into the "digital experience" back in 1979 or so with only the most enthusiastic mindset. I'd read how GREAT it was going to be and given all of the REASONS: wider dynamic range, no wow and flutter, flat response, no noise and hiss etc.


I bought the first rock record recorded digitally on the 3M machine good to 50K sampling rate (Ry Cooder's "Bob 'til You Drop"). I bought it the DAY IT WAS RELEASED. I brought it home and played it expecting it to be GREAT. I was a BIG Ry fan and had all of his previous, well-recorded albums.

It literally made me SICK. It sounded WRONG to my ears. I had to stop listening.

I spoke to my engineering friends who said "Wait until there's a true digital carrier, cutting lathes don't like the wide dynamics (etc.) of the digital recording."

So I waited. And then came CDs and IT SUCKED EVEN WORSE.

"But it's perfect" they said. You just don't like "perfect".

But it SUCKED.

EVERY CD reissue I heard sounded WORSE than the vinyl I knew and loved.

Then the CD lovers came out with a lot of excuses for why it originally SUCKED (they finally admitted what they originally denied but what I heard) was the wrong source, it was the primitive A/D converters it was the U-Matic tape storage mechanism etc.


So that is where I am now. I have a very good CD player (not that it matters according to the digital enthusiasts) and I do listen to a lot of CDs transferred to my Meridian Music Server. I have 3000+ ok?

But I find I can only listen to these files as background.... when I try to sit down and listen the way I look forward to doing with records, I last a few minutes and that's it. I get bored and/or agitated and I am moved to do something else.

I am not alone in having this experience. I think science should be applied to understanding this phenomenon instead of having CD advocates explaining it by saying I'm "prejudiced" or whatever.

tnargs's picture

That's a sad story, Michael, but I think there is zero chance that it is a story of anything but your personal, psychological response to something other than sound waves.

In fact, your story is almost a textbook example of 'tainted by bad first experience, ruined forever'. Very simple, basic science of psychology, and no need to argue that advances in science are needed to post-rationalise your history of CD experience as being in the sound waves. Why don't you accept that? As per your story, you built up a huge expectation, had a massive disappointment, and never recovered. Stop trying to ruin CD for the rest of us who enjoy both mediums, with your campaign of made-up and cherry-picked "science" trying to prove that what you don't enjoy has to be in the sound waves, and those CDs that some of us do enjoy to be inferior, and not 'noticing' its inferiority has to be our cloth ears. I know, that's how ego works in all of us, but for the sake of unbiased reportage, not to mention being convivial towards a fair segment of your audience who like both mediums, you need to let it go.

I think I have already acknowledged the loudness war and its corrosive effect on the quality of sound waves from a very large number of CDs, but I will repeat it here just to be sure that you realise I agree on that one. Bad mastering has been a curse on CD that it didn't deserve.

P.S. that Cooder album has been documented as incompetently produced (so your initiation was spoiled by genuinely bad sonics). The rock/pop sound is of course a 'sound effect', not anything pure, and the analog taping/mixing/squashing/mastering-to-tape was critical to the sound of rock/pop as a manufactured thing. Production crews were highly competent at it with analog recording gear, it was semi-automatic. They simply didn't know how to do that with the all-digital process in 1979, so mucked it up totally, not to mention the other step of getting it onto vinyl appropriately. It was a mess. Cooder himself has said that he hated the vinyl, much preferred the CD. But the CD's production in 1985 I cannot comment on, e.g. did they re-work it having figured out how to make the rock/pop sound on all-digital, or not?

All very sad for you, Michael, but no need to go on a mission to generalize and objectify what is really a personal, subjective story. I am not alone when I implore you to return to what we enjoy reading: your record reviews where you share with us your love of good music well made.

Michael Fremer's picture
It is not a sad story because there is a vinyl revival underway worldwide where people who hear as I do fought back and did not accept unpleasant sound masquerading as enjoyable sound. We did not accept a future in which we could not enjoy listening to music as we used to.

I am not trying to "ruin CD" for anyone but every time there's a story about the vinyl revival there are people there trying to "ruin vinyl" for others and trying to dissuade others from having a listen.

Nothing whatsoever sad for me....I'm having a great time....

isaacrivera's picture

It is the job of science to look for explanations to human experience and not the other way around. We do not have to make our experience match the current state of scientific knowledge. Science was invented to expand our awareness of our experience not to constrict it. If some human experience is repeatable and can't be explained by current science, then it has some work to do.

I recently had a similar argument with someone when I explained the sonic improvements I was experiencing when adding an SDS speed controller to my turntable. They told me it was impossible, because the science this and the science that. I replied I was not the gullible kind and the improvements I was noticing were not subtle. It is not my job to fit my experience in the current state of knowledge. It is up to engineers to go and try to figure out why I experience these things. A couple of weeks later, the person apologized and confessed, they had bought a speed controller (with money back guarantee) just to prove me, and others with like opinions, wrong, but when they hooked it up they could not believe the sonic improvements!!! They could not understand why, but it did work.

I am always shocked by those having strong opinions of things they have no experience with.

andrewmorgan's picture

And buy what you like, regardless if someone thinks that it sounds better recorded on paper wrapped round a brick, or if someone else thinks your new Pono is crap cause you can't get Bruno Mars in HD (I have 2 Astell & Kerns players not a Pono, but think I might buy one just because I cant get Bruno Mars).

I am not sure I can understand half that article, but I can read an MRI, but neither are going to tell my more than my rusty ungolden ears. I like my Vinyl, I like my hires audio for travelling, Sony MDR 7506 are not nice headphones, and Bruno Mars makes me gag on my porridge.


Jon's picture

I've lived a much less stressful life since I stopped participating in audio forums and arguing the merits of formats, sample rates and bit depths. I long ago discovered there is simply no point doing so since no one ever changes their mind and in any case, what other people say or think does not change what I hear, nor does it stop me buying vinyl or high resolution digital audio downloads and having a marked preference for those as compared to commercial CDs.

There is a very long running thread on a forum I belong to that not only argues against vinyl but also promotes the belief that all digital sounds identical because bits are bits (so in other words, the digital output of a $200 portable player matches that of a $10,000 CD transport). I could have died from stress trying to argue otherwise but it is pointless, bad for my health and gives me less time to actually enjoy my music.

In any event, I don't really understand why any serious audiophile (I don't mean the new generation "head fiers") would put any credence upon a test done through a cheap line switch, a $400 source and using headphones which (inherently) have such a lumpy and inadequate top end response that it is 100% certain nothing useable beyond 22 Khz ever got to the listener's skull to begin with. Just look at the response graphs of almost every headphone regardless of price out there (Headroom) and you won't see a remotely flat response curve, let alone a curve where the extreme HF is remotely preserved relative to the midrange or the lower top end. I'd get a far more accurate impression of the relative frequency content of a music sample using a basic LS3/5A monitor than I would a pair of headphones costing several thousand dollars, let alone the ones used in that test. And the LS3/5A was not exactly renowned for a sparkling and extended top end. Ignoring extreme top end products, most mainstream headphones are for practical purposes sonically "dead" beyond around 10 -12 KHz such is the roll-off compared to even a lower end dedicated audio speaker system. It's one reason why personally I have to spend a good amount of cash on a headphone just to get the same satisfaction as I do from a $1,000 pair of basic two way speakers.

So why bother to even argue. It's pointless and it just feeds the trolls. It's nice to see that my brain changes thanks to all this HF content and it basically vindicates what I have thought myself ever since listening to my very first Audio Fidelity LP in the late 60s (the Sidney Frey label where they put a big effort into marketing the ultrasonic frequencies on their disks).

So Michael, save yourself the stress and just don't bother replying to these trolls. You are doing far more by giving us useful information such as the study you linked to, let alone all your superb reviews.

bdiament's picture
azmoon's picture

...looks like a bunch of the audio no minds are now posting here.

Journeyman's picture

after all those "no minds" bring revenue to this site right?
Some users forget that tinny detail about pageviews, ads served by page and money! I'm new around here so I might take offense about your comment (not really ;-) I comment for fun most of the time, and actually really love innerfidelity from this same network). Best wishes.

gbougard's picture


please correct me if I'm wrong.
My understanding of DAC's is that these devices:

1. are fed with, say, a 44.1 signal
2. somewhat "recreate" what is lost (by what miracle, I don't understand, but anyway) in the signal
3. provide an enriched 96 or 192 or even 384 signal

if that is the case, then the signal emanating from such DAC devices should alleviate the problem raised by Michael, whom I'm tempted to believe in this discussion.

Then, wouldn't it make sense to explore the best DAC possible, ie the one that fills the gaps and missing info as well as possible and makes a signal more agreeable to our brains?

Please treat this question kindly and with sympathy: I am mostly an ignoramus. But I'm willing to learn, if the information is presented simply for laypersons like me. Sound for Dummies would be something I'd read!

Mark UK's picture


No. If any information is missing you cannot re-create it.
There are two issues, the sample rate and 'oversampling'.
The latter takes the incoming sample and multiplies it by a factor of 2, 4, or whatever. This makes it easier for the following parts to accurately regenerate the original signal. Most DACS do it, but not all. These are called 'NOS' DACs (for No Oversampling).

There are differences in opinion as to which is preferable, but it's just that, subjective opinion.

As for the sample rate, 44.1, 96, etc. Human hearing doesn't go above 20KHz, less in most people, so 44.1 is adequate. It's MOSTLY just a numbers game. BUT - the higher the sample rate the further the essential filter can be removed (within reason) from the range of human hearing the better. So 96 is maybe worthwhile as it allows a 48 KHz filter. Above that it is total nonsense.

Sampling accuracy.
Provided the sample rate is twice the range of human hearing (at at 44.1 it is) and the incoming signal has no freguencies exceeding half that (22.05 KHz) the signal is reproduced PERFECTLY. No information is missed. None at all. This is widely misundersttod. You do NOT 'miss' any 'peaks' because the process is NOT a step function as drawn in many stupid pictorial representations.

So 44.1 SHOULD be ok. If it makes you happy, 96 is fine and it MAT be better due too the filter. Any higher is rubbish. It's the same for DSD. Don't waste your money on anything higher than DSD 64, which relates roughly to 'regular' 96.

Michael Fremer's picture
Most CDs to my ears sound like SHIT. I don't like what they sound like. I didn't in 1984 yet everyone said they sounded FANTASTIC. BIG IMPROVEMENT. But they SUCKED. The people who said they sounded GREAT had their heads up their asses. They were listening to the supposed measured "PERFECTION". Why should I take any of those clowns seriously? They are still at it. Now they claim that 16 bit/44.1K is PERFECT. That would be you. This is the only technology where people claim an invention that skirts the limits of being BAD is PERFECT and anything that is clearly BETTER technologically is either "rubbish" or they grudgingly say "it can't hurt." I think trying to defend the perfection of CD---a 30 year old goat-roped invention--is insane.
Michael Fremer's picture
I have heard many DACs. I have heard many digital fillters. They mostly sound different from one another for a variety of reasons. It has taken DECADES to improve filter technology so they don't RING or produce PHASE SHIFT that's AUDIBLE. The CD apologists keep citing flat frequency response from 20Hz to 20kHz. There is information below 20Hz that's audible and lopped off by this "perfect" technology. There is information about 20kHz that this "perfect" technology lops off in a most unnatural way. In doing so the brick wall filters produce artifacts that intrude upon the range we can hear....what's more, the study we reproduced here suggests our "hearing" is a result of ear/brain interaction and perhaps just measuring frequency response doesn't provide a full measure of how we hear, especially in the time domain.
Mark UK's picture

I mostly agree.

But there is a whole lot of BS talked about 'digital', and it is encouraged by high-priced manufacturers.

Most of it succeeds because most people don't have a clue about how it all works. Particularly when they are talking about computers, streamers, and DACs, which of course is near enough all of it. And also about human hearing.

1) 'Mastering' is all. Buy a good 'record'.
2) Buy some good loudspeakers.
3) Buy an amplifier good enough to drive them properly.

Only then worry about all the BS on 'digital'.

Before anyone says "He's a troll from HA" I'm not. And how much do YOU know about computers? I've been doing them professionally for near on 40 years. I don't assume to know how to do at heart bypass or whatever you do for a living, you don't know much about what I do.

Michael Fremer's picture
DACs sound different from one another. Digital filters sound different from one another. The problems associated with DACs become known as people began learning how to measure them. What was measured was first heard. Issues include: jitter, ringing, pre-ringing, phase shift, etc. These issues are audible just as wow and flutter are audible. Which are more pernicious and annoying? That is a matter of opinion.
Mark UK's picture

Funny how a report is suddenly 'unearthed' that supports higher sample rates so the 'believers' say 'science' proves their point so 'science' is good.

But when a report appears that doesn't support their beliefs they say science doesn't know everything and the report is rubbish.

And as Bromo33333 says, perceptions are all over the place. So believe what you want. It won't kill you if you are wrong.

Me? I buy what I want for all sorts of reasons. And I am not 'price limited'.

Michael Fremer's picture
1) Science does not know everything. Anyone claiming that science knows "everything" knows nothing about science. And science is sometimes wrong. 2) This report was not "unearthed" any more than other reports that you agree with are "unearthed".
Mark UK's picture

I don't really have an opinion. I am trying to see George Orwells 'doublethink' five fingers but I can't. My next purchase will be a better DAC, probably a dCS as I won't pay high money for a box with a ten dollar high street DAC chip in it. The DAC so-called 'designers' are merely implementing someone elses ideas as best they can. dCS do it all from scratch.

I am just amused that when a "high res file is worthwhile" report comes out the value of 'science' is applauded.

When a "high res file is a waste of time" report comes out 'science' is derided.

Michael Fremer's picture
Their ring dad technology is a step above... I reviewed the big stack and it was the best digital I've heard along with MSB's equally unique tech. But I still prefer listening to records.
Mark UK's picture

What's all this stuff from 'amateurs' that we can 'perceive' (by various mysterious unstated methods) frequencies that we can't hear with our ears?

Anyone got a reference? (And not one from self-appointed HiFi 'experts',)

Michael Fremer's picture
Or are you an "amateur" reader?
Mark UK's picture


Mark UK's picture

I purchased one of the first CD players available.
I was glad to get rid of the background noise but it was strangely 'uninvolving' compared to vinyl.

Mastering? If several 'record' prodcere came to my house, I played a couple of their files, CD, or whatever and they all said "exactly how I meant it to sound" I never need spend anymore on the equipment, do I?

Michael Fremer's picture
That might actually happen.
Mark UK's picture

Having several 'record' producers listen. As audio 'quality' is about accuracy and transparency to the available source. It is not what we might think is 'nice'

Speaker? yes, Amplifers etc too.

The more you go up the 'quality' scale the more things should sound the same. Cheap and nasty stuff has excuses.

Michael Fremer's picture
Let's see Mark: I've had Eddie Kramer visit to play Hendrix recordings, I've had Greg Calbi and Steve Berkowitz come by too. (Greg is veteran STERLING SOUND mastering engineer, Steve Berkowitz oversees Dylan and Miles Davis catalog reissues and most recently did same for Beatles mono box). And you?
BillK's picture

Most everyone introduced to CD in 1983/1984 was listening not to what was there - how the music sounded - but rather to what was not - clicks, pops, horrific 1980s vinly pressing artifacts, tape hiss and the like.

If that's what you spent all your time listening to from that point on, you never realized what you lost until the first time you heard vinyl on a good setup anew.

I remember even at the first few Rocky Mountain Audio Fests, turntables were few and far between, but for those who bothered to truly listen

Finally, even as a vinyl fan, I have to relate a quick story of what happened to me this past year:

I've long been a fan of Mannheim Steamroller's first Christmas album, having heard it for the first time in college, off LP, and just loved it. Wonderful music, I listened to it over and over.

But over the past few decades I've always felt it was just "eh" - some of that was due to it being overplayed on radio but I thought it just didn't touch me anymore.

This past Christmas I pulled out that LP and put it on the turntable and fell in love with it all over again. It was beautiful and heartfelt.

Then as an experiment I played the album as I had been for the past few decades - from CD - and I lost all interest. Everything that made me stop in my tracks and listen when playing from vinyl was just gone.

Not scientific by any means, but certainly a huge revelation to me.

Ortofan's picture

..."if the engineer used good microphones and minimum processing even standard CDs can sound very much like music. What sounds horrible are bad microphones, bad processing, too much compression, too much manipulation of the data."
No criticism of the resolution, in and of itself, being a particular limitation.

BillK's picture

This being the same Peter McGrath who almost always uses vinyl or high bit rate digital for demos he does of Wilson Audio products while rarely if ever using Red Book CDs?

Ortofan's picture by Mr. Mcgrath at a local hi-fi store, all the music played was said to be rips of redbook CDs.

Michael Fremer's picture
I've never attended a Peter McGrath demo where he didn't use high resolution files of recordings that he's made himself—he's a highly accomplished recording engineer with an impressive catalog of analog and digital recordings issued by a few labels including Harmonia Mundi.
Ortofan's picture

...of the various audio files/clips being used for the demo and Mr. McGrath replied that they were redbook CD resolution. There was a general sense of surprise or disappointment that they weren't hi-res. None of the music was to my particular taste, so it didn't matter that much to me. While they may have been chosen to illustrate the capabilities of the speakers, the demo pieces didn't exactly seem to be audience favorites, either.

Following the demo, I asked Mr. McGrath if he had any of his Audiofon piano recordings (such as those by Earl Wild or Valentina Lisitsa) with him, but he did not. Maybe this wasn't his usual collection of sample recordings?