It has been a very busy week at work and obviously the world this week has been going through all kinds of turmoil on account of Coronavirus and the very real effects it's having on travel, production, the economy, and of course health care!
Over the last week, I spent some time on Audiophile Style chatting with the folks there in the thread on "Why are objective assessments important...". As you can see, many topics were covered including subjectivity, objectivity, what is a "perfect" reproduction chain, whether objective results correlate with subjective experience, all the way to a discussion of neuroscience, the mind, etc...
Instead of repeating too much from the thread, let's hit on a few key points and elaborate. The problem with forums I find is that often thoughts become lost in threads, requiring that at some point, a review be done, otherwise topics just end up being recycled ad nauseum (remember back in the day when people kept asking "does FLAC sound the same as WAV?"). Furthermore, it's hard to elaborate on ideas in the midst of discussions compared to a self-contained blog article.
I. The Big PictureNo matter what we do, we should never lose sight of the "Big Picture". This principle applies to everything IMO. Knowledge is made much more useful when we maintain the broad perspective and can appreciate how it all fits together. Only then can we have a sense of relative importance whenever we talk about specifics. I don't know about you guys, but it's rather frustrating in life dealing with obsessive-compulsive folks who spend extreme amounts of time on minutiae of insignificance!
The diagram above is I hope a straightforward representation of how I see the audiophile hobby which as I've expressed before, is a combination of the "Music Lover" component with the "Hardware Audiophile" element.
While as audiophiles, we often do not speak much of this, we must never forget that unless one records one's own music (!), the PRODUCTION part of music has an intrinsic relevance for us. "Pro" audio people for the most part are NOT like audiophiles ;-).
The PRODUCTION side is complex. It incorporates music labels, producers, financial interests, audio engineers, and of course the artists themselves. The music can be "made" totally virtually on computers, at home, in a state-of-the-art studio like Abbey Road, and anywhere in between. The music can come from live feeds, well controlled intimate "natural" recordings, to complex studio productions, to completely synthetic data constructed in a DAW like Avid Pro Tools. Just think of the varieties of microphones, placement techniques, instruments used, room acoustics of performances, and of course mind-boggling variety of synthetic plug-ins, sampling manipulations and digital tools available!
We as consumers generally do not have much to say about the music production. This is why I have the red barrier between the "Pro Audio" and "Consumer Audio" sides. As I've expressed over the years, it sure would be nice if consumers and the press provided more feedback to the PRODUCTION side about aiming for better sounding content like demanding higher dynamic range. One way is to have different masterings of music released on the various formats (for example, a "Standard Resolution" master in CD-quality, and an "Advanced Resolution" for hi-res).
There are some audio/mastering engineers who do speak out and cater to the audiophile segment (people like Steve Hoffman, Barry Diament, and Mark Waldrep), but for the most part, audio engineers don't typically accept the values or apparent superstitions promoted by audiophile companies. Likewise, I know a few musicians, and when I tell them one of my hobbies is as an "audiophile". For the most part, they don't seem that fascinated by speakers, what amps, which DACs or spin vinyl for that matter. The love for music is obviously there, but their focus is not on the hardware.
To the far right of that diagram is PERCEPTION. Whereas the PRODUCTION side is complex, consisting of numerous people including artists, engineers and such, PERCEPTION is ultimately solely determined by the "One" - each of us. We determine our own preferences. We decide what is worthwhile "art", which song sounds "good", what artist "speaks" to us. This is the domain of subjectivity. Even if one day, some kind of brain scan is found to correlate nicely to "prove" that a person is "enjoying" the sound... So what? It was never up to anyone else how the "subject" experienced the "qualia" of an artistic impression, or the joy elicited.
Between these two poles of PRODUCTION where the common audiophile / music lover has essentially no power, and that of PERCEPTION where the "subject" is all that matters, lies the products we use for REPRODUCTION. As audiophiles, this is our battleground. This is where the Industry battles for our mindshare and money as well!
Do we go for the low-end budget "consumer" gear at Walmart, or are we going to climb the rungs of quality to the "mid fi" tier, maybe crossover to see what the "Pro" world has to offer, or go all the way to "hi-fi", perhaps soar to the stratospheric level of the "hi-end" luxury products?!
In straddling between the two poles, REPRODUCTION acts as the conduit for the music that is produced and the subjective world of perception. The question of course, is just what kind of conduit do we want? And this leads us to the next topic...
II. IntentAs you can see below each box in the diagram, I've listed possible intentions that we may need to consider. These intentions, or values will interact across the domains.
On the PRODUCTION side, there is of course the intended sound of the artist working with the producer, recording engineer, mixing engineer, and mastering engineer before the music shows up on your CD, vinyl, or download. Along the way, I guess "corporate" has a hand in making decisions on how much money is available to support a project, considerations on how well the product will sell, etc. Online, you'll see many articles talking about creating an "optimal mix", or mastering targeted for small/mobile speakers. In fact, as an audiophile, check out Gearslutz at some point and read some of the headline articles, see what the "pro" audio guys talk about on their discussion forums.
As audiophiles, the intent when we PERCEIVE music is hopefully that it "sounds good" - satisfying the illusion of "space", the sound of a "real" instrument, the dynamics of a "true" performance, and the emotions that come with this experienced sense of "success" in creating that illusion. Since perception always is filtered by our biological mechanisms - ear and mind - human hearing is "psychoacoustic", the field of interdisciplinary study that has allowed us to find such elegant solutions for lossy compression for example by understanding the physiological, psychological, and cognitive limitations of auditory perception. Remember that when people ask questions like "When do measurements correlate with subjective impressions?", the answer to that question will come through studies of psychoacoustics correlating with whatever objective measure is being performed. Realize that there will likely be a range of answers or at least we must appreciate that any one person's perception must be graded along a population curve with varying norms likely affected by age, gender, intellectual ability, musical ability/experience, perhaps even down to genetically determined traits.
Remember that to enjoy good, "euphonic" sound, one does not have to be an "audiophile" or what I believe is more appropriately labeled "hardware audiophile". "Music lovers" already perceive beauty without wringing hands over whether USB cables make a difference, if their speaker cables are good enough, or whether one might want a tube amp! In fact, one could argue that loving music, enjoying the sound from whatever device one has is a blessing, and the neurotically obsessive "hardware audiophilia", the curse.
So then, what about that middle box that we audiophiles wring our hands over, the elusive "perfect" REPRODUCTION gear?
To me, the answer is straight forward because my intent has always been to achieve "high fidelity" and "transparency". By taking this position, it automatically determines for me how I adjudicate the adequacy of many devices. Testing DAC quality is very straight forward. Whether computer streamers, playback software, digital cables make a difference is elementary (as you know, "bits are bits"). Understanding this intent and with objective testing, it's quite obvious that digital-to-analogue conversion these days is simply one of life's "solved problems". Furthermore, there is obviously nothing much to cables. Don't worry about bit-perfect streaming devices. It's better to focus our attention on what might be the "least perfect" of system components like vinyl playback (if one is into that although I consider LP's ultimately qualitatively limited), amplifiers perhaps (these too perform very well these days), definitely speakers and room acoustics.
Other than aiming for transparency, one could aim for a more "colored" sound intentionally. Nothing wrong with intentionally using EQ for example, or liking the sound of a tube amp with known high levels of distortion so long as one is not strictly adhering to the maintenance of high-fidelity or transparency.
Finally, let's also remember that we must consider the concept of "value". All consumers must consider this; surely even multi-millionaires and billionaires are wise enough to make sure that money is well spent! It would be foolish if one did not consider this intentionally in every purchase we make.
I believe that if we can express our intent for the kind of reproduction we want, then there will be much less conflict between audiophiles.
III. In the Big Picture, not "everything matters"When we look at a painting, do we need to analyze every brush stroke to make sure it's perfectly "correct"? Must we be distracted by a slightly misplaced dab of color in that painting? Do we need to "pixel peep" a digital photograph in order to be satisfied? Of course not. So too it is with audio reproduction and the choice of components we pick. While the ideal for me is "transparency" for my main system, I'm happy to spend time and disposable income to get things "right", what I've come to learn is that there's no need to sweat the small stuff. There is certainly a point where I can say the system is "more than good enough" for what I can hear, in the service of allowing me to enjoy music in the evenings!
Various audiophile articles, especially Industry sponsored ones like this one about Nordost, claim that "everything matters". I'm sure it matters to them to impress upon audiophiles that garnishing a system with cables "north of $350,000" would be good for the company's bottom line; but I seriously hope audiophiles don't feel that this is even close to being true!
The PRODUCTION does not need to be perfect to be enjoyed. Likewise, our PERCEPTION does not need to be (and is not!) perfect. In fact, not everything matters to a subjective "music lover" - there will be many artists one will have no interest in, likewise there are probably musical genres we might never even explore in our lifetimes.
So too in the objective world. Not every measurement matters. As I've referred to many times, jitter is measurable but not necessary worth paying attention to from a sound quality perspective. This doesn't mean jitter measurements are useless though because they could be a reflection of the quality of engineering which is worth knowing about when considering a high-priced component.
It is a sign of the hardware audiophile's neuroticism that he should even consider purchasing high-priced, objectively unsubstantiated, lacking-in-common-sense cables "north of $350,000". That this would somehow provide a valuable contribution to his system considering the significantly more substantial limitations across each of these Big Picture parts of audio and music.
IV. High-Fidelity from the vantage point of modern audio electronics...Finally, let's remind ourselves that we are living in a time when modern audio electronic devices are already highly refined. Pay no mind to some audiophiles who might thumb their nose at $100 DACs, OFC 12AWG zip-cord cables, or "cheap" Raspberry Pi music streamers.
While the question mentioned above: "When do measurements correlate with subjective impressions?" is still worth asking and be explored through the psychoacoustic lens, let's not be surprised if the answer seems difficult to find these days!
We are living in a time when other than some of the cheapest components, much of what we have already perform at very low distortion levels! When inexpensive DACs are routinely way less than 0.01% THD+N, an excellent amplifier like the Hypex NC252MP is capable of -90dB THD+N @ 1W into 4Ω for less than $500, would it be at all surprising to consider that these might be below the threshold for human hearing already when subjected to controlled, blinded testing? Is it any surprise that even iPhones and old CD players are indistinguishable from an Oppo UDP-205 with a flagship ESS Technology ES9038Pro Sabre DAC (as tested last year)?
Remember, on the PRODUCTION end, we routinely have music these days that are highly dynamically compressed with typically high noise floors. The other day, I was listening to Canadian Andy Shauf's new album Neon Skyline. Here's the title track's noise level during the fade out at the end of this DR10 track:
For a modern rock/pop song, DR10 is quite decent, and as you can see, the noise level is down around -110dB (64k-point FFT) - better than most modern productions if you have a look at the data in an audio editor. In fact, this is better than many so-called "high-res" recordings in 24-bits.
Compare this to the collection of DAC noise level measurements below. Even an inexpensive <US$90 Topping D10 can rather easily provide lower noise level than what the music demands:
As for human PERCEPTION, realize that with blind tests, like MP3 vs. FLAC lossless, results are not suggesting that most people can significantly differentiate between high bitrate 320kbps MP3 from the original lossless. Objective techniques can easily show us the difference between lossless and lossy such as using a null test. This is the power of psychoacoustic modeling of human hearing; even when we drastically reduce the amount of data, perceptually, the sound can remain practically "transparent". Compared to MP3, how much "change" do you think USB cables can impart? Do you think objectively "good" DACs at similar volume levels will differ to the point of removing masked frequencies that MP3 encoding performs? Oh, yeah, what about power cables? ;-)
So... If on both sides we can show that there are limitations to music PRODUCTION, and we can show that human PERCEPTION likewise has limitations, just how much fidelity does the REPRODUCTION components have to be to consider it as "good enough" especially since we as consumers actually have the power to measure these things and listen to them in a controlled fashion?
The data suggest that devices like DACs have already reached a level where measurements like noise and harmonic distortion are low enough that there is actually no point asking "do they correlate"! I believe human hearing is actually inadequate as a gauge for the true resolution. If we were living at a time where devices could only achieve 1% THD+N at best, and routinely 5+% THD+N, I bet nobody would be having a debate as to whether the measurements and sound quality correlated and audiophiles would very much celebrate the arrival of a new source component able to reach the next lower level of distortion like 0.5%THD+N!
Those days of gradual, meaningfully improving resolution for most devices are long gone. We would have to look back at the advances of the 1940's to early 1980's. Notice that those years also happen to be the era when audiophiles actually also appreciated the importance of objective measurements and "high-fidelity" audio was still the cool technology to have at home. I bet you it has been a long time since JBL ran an ad in Playboy (here's one from 1976):
We can see cases in audiophile writings online like this one by Peter Qvortrup (of Audio Note) and Rafe Arnott about "High Fidelity, the Decline of the Decades" where they seem to be grasping at similar thoughts as this post although obviously with different perspectives and conclusions. While I agree from a PRODUCTION perspective that many of the recordings from the 50's and 60's (like the works of Rudy Van Gelder) sound great, this is a reflection of the skills of the engineer who crafted the art of capturing and producing material with more natural techniques which REPRODUCE well using 2-channel audio. There is nothing stopping modern artists and producers from doing the same and achieving even better results. It's a happy coincidence that Qvortrup/Arnott recognized replay equipment and amplification to have improved over the years (I consider this a coincidence because they claim amplifiers got better because of SET triode reintroduction - that's IMO ludicrous).
Look at the pictures and the words selected. It's obvious that Qvortrup has an issue with the fact that audio gear has achieved excellent quality at lower cost which speaks to his bias towards luxury and IMO poor value by targeting objective evidence, the rise of digital audio and advances like negative feedback as his unsubstantiated complaints. Most puzzling, I'm really not sure how Qvortrup/Arnott can justify the claim that "loudspeaker quality" has dropped down to 4/10 by the 1990's and peaked at 9/10 back in the 1930's and 40's!!!
While I disagree with Qvortrup that modern speakers should be ranked so poorly, it is still true that speakers and rooms are where we must look to squeeze out the highest levels of performance today.
V. Let's conclude...As you can see, we've hit a number of topics here. That is to be expected when we look at the "Big Picture". My hope is that this might be a reasonable way to see audio/music reproduction from a "bird's eye view". From that vantage point, I think we can look at the Big Picture "parts" - PRODUCTION, REPRODUCTION, and PERCEPTION - and imagine how the pieces interact, and more so, are actually interdependent when it comes to the final human experience. Furthermore, I hope by doing this, we can consider attribution. If we can understand that some issues we face are results of the PRODUCTION system, then we shouldn't try to fix them with our REPRODUCTION devices. In fact, knowing this means we stop trying to make every recording sound good because high-fidelity reproduction of a poor recording is supposed to sound poor. Likewise, there are all kinds of factors that will change PERCEPTION. Moods change, stress levels vary, there may be time-of-day effects. Noise exposure changes hearing acuity. Previous listening experiences affect how well and what we listen for. Perception also has its limits as we age. This perceptual variability is why we must always consider the validity of purely subjective testimony. Developing insight into our own limits and psychology is important, I do not believe everyone is capable of good insight.
Having said the above, I believe that if an audiophile claims that he can hear something scientifically questionable (eg. USB cable differences), then the only reasonable suggestion is to invite the person to try some kind of controlled blind listening to ascertain that the claim is true. This is the only way to move forward in a sober fashion with a reliable foundation if we are to honestly make declarations. Otherwise everyone will end up confused as to what to believe.
I have heard audiophiles claim that "objectivists are no fun". Well, I think learning and gaining knowledge is fun ;-). However, at the same time, reading a science fiction or fantasy story is fun as well! But is that what we're aiming at here with audiophile products? Again, only you can answer this for yourself.
I'm certainly not going to be dogmatic and insist that people cannot have "fun" discussions, sharing ideas about what products to get, what to try... In this world, people spend all kinds of money on ideas and claims that IMO are without merit (or frankly untrue). I think so long as one can afford it, everybody's aware that claims are based on opinions rather than substantiated fact, and apply a bit of common sense, it's "mostly harmless".
In life, there are certainly times when we should "regress in the service of the ego" in order to enjoy the pleasures that life may have in store. Go ahead, have fun buying, trying, and listening, but realize that there are times when one needs to get back to the world of adult thought and appreciate that unless connections can be made with controlled listening, the level of evidence of a significant effect is likely low.
Please don't ever get stuck just reading sci-fi and fantasy!
Notice that Munich is (was) scheduled for May 14-17. That's actually quite awhile off. Of course as the spread of infections continues over the months ahead, we'll have much more information about the risk factors around morbidity and mortality by mid May. The fact that Asian COVID-19 numbers are not looking that bad and they're ahead of North America and Europe by weeks, is encouraging news. The real question is whether health care systems are able to manage the most severe cases needing acute interventions. I suspect we'll look back on this as an important "stress test" of systems including fueling debate between countries with publicly-run "universal" vs. private health care models. Also, questions around regional and economic disparities likely will arise.
Stay healthy, friends...
March is here. I'm certainly looking forward to Spring Break coming up with the family. Hope you're enjoying the music!
Again, remember to try out the THD blind listening test. I believe it is only with doing more controlled listening tests that we can approach subjects like "How do objective measurements correlate with human perception?" with any kind of confidence.