When consumers seek to purchase a product, they compare many different factors before making the final decision, including design, cost, performance, and brand, to name a few. Among numerous such factors, there is one unique factor that clearly sets audio equipment apart from every other product: its "sound".

Perhaps it is because this "sound" is intangible, but each person has his own approach to sound. Each individual feels differently about it, and each individual has a unique attitude towards it. Frequently, audiophiles cannot even agree on the terminology used to describe sound.

For instance, two audiophiles listening to the same product may disagree on the amount of bass that it has - one might think that it is 'thin', while the other might argue that it is 'excessive' or 'rich'. You might say that a DAC is 'clear and accurate', while to me the same unit sounds 'harsh and clinical'. One man's 'separation' is another's 'disjointedness'. Such disagreements are ubiquitous in every area pertaining to evalua-tion, from conception of space to resolution.

In addition, some reviewers speak of the ‘ambience’ or ‘soundstage’ only accurately presented by a well-installed pair of loudspeakers in their reviews of earphones and headphones. (Personally, I wonder whether it would be more appropriate to label this quality ‘openness’.)

While there are numerous causes for this subjectivity, the primary culprit is likely that people evaluate other products based on what is familiar to them; the problem with this, of course, is that everyone has a different opinion of the same product since no two people have identical setups.

Furthermore, our brains tend to mess with our perception of sound as well: when we listen to more expensive or better-designed products, we tend to perceive their sounds to be relatively superior to their cheaper and less attractive counterparts, even when the sound is, in fact, identical. Thus, a blind test is necessary for an accurate evaluation of the sound and sound alone, isolated from all other variables. But as those who have tried it should know, it is painstakingly annoying and rather difficult to set up with rigor. To most people, it just isn't worth it to go through the process to analyze the sound of a single product.

After all, we listen to music to enjoy, not to test.

At Golden Ears, we try to straighten out these problems by measuring all areas of the sound of a product from many perspectives using various methods.

The empirical measurements of sound are analogous to a blueprint of a building: just as a solid understanding of the architecture is aided by varied perspectives - from the top, left, and front - our comprehension of sound may be augmented by data regarding frequency response, impulse response, cumulative spectral decay and so forth. Once we understand what each measurement means, we immediately have an idea of what to expect from a particular product, and an objective standard of comparison to boot.

The oft-brought up argument against this paradigm goes that ‘the data don’t matter; it’s what you hear’. When I first founded Golden Ears, I too doubted how much relevance the measurements had to the perceived sound - but in reviewing about 200 products to date (August 2011), there hasn’t been a single exception to the rule that a unit that measures great also sounds great.

This is not to suggest that data are infallible - they have their own blindspots that need to be taken with a grain of salt. For instance, the THD(Total Harmonic Distortion) is, strictly speaking in the hi-fi sense, a corruption of the original signal and thus a blemish preferably eliminated, but the numbers do not always match what we hear (tube amps). This is because the measurement of the THD does not take the Masking Effect (where smaller sounds are ‘masked’ by their larger counterparts) into consideration - this is one of the many places where further research could lead to improved methodology. (c.f. Earl Geddes & Lidia Lee’s work on Sound Quality Perception)

‘Does ‘listening’ to music belong to art? Or does it belong to science?’

While each of us might have slightly different answers, it’s pretty safe to assume that most would say that listening to music is closer to art. Now, if we were asked a slightly different question,

‘Does ‘creating’ music belong to art? Or does it belong to science?’

It would also be likely that most of us, again, would agree that creation of music belongs to art. But if we were to tweak the question again to read,

‘Does ‘reproducing’ recorded music belong to art? Or does it belong to science?’

We would suddenly have quite a bit of discord, unlike as in the previous questions - the right answer to this question is, in fact, the key to constructive discussions of audio products.

Though everyone is entitled to their own opinions, the Golden Ears’ answer to this riddle is that ‘reproducing’ recorded music belongs to the domain of science.

Perhaps it is because the acts of ‘listening’, which belongs to art, and ‘reproducing‘, which belongs to science, are simultaneous that we are tempted into thinking that reproduction of music is art as well - but it is important to note that while the listener is engaging in art, equipments that are reproducing prerecorded music should not be doing the same.

If an equipment introduces new signals to a completed track (Figure 1), the listener is presented with something entirely different from the intent of the original creator. Thus, the equipment ought to play back precisely what is on the record, and an equipment that meets such a criteria is said to be ‘faithful’ to the original recording, from which the term High-Fidelity, and the abbreviation ‘Hi-Fi’ derives.

In art, modifying what exists or creating something new is ‘creative’ and recommended; but in science, this creativity is called ‘distortion’, and is frowned upon. And it is when we approach audio with the misguided mindset that ‘reproduction’ is art that we get a product like ‘24k gold SATA interconnect’ or ‘magic audiophile stones’ that we see when we google ‘audiophiles are retarded’.

Again, reproduction belongs to the domain of science.

Because reproducing music is science that strives to preserve the original recording, the manufacturers ought to develop their products scientifically, and a review site such as Golden Ears also has a duty to evaluate products under this same paradigm of ‘faithfulness to the original recording’.
For the manufacturer, the scientific approach not only leads them towards developing a flat response expected of a Hi-Fi device, but also allows them to introduce purposeful distortions according to preference, and for the reviewer, it leads to credibility as a uni-form standard ensures fairness and by extension consensus.

This, however, does not preclude or prescribe a personal preference in listening, which belongs to art. For instance, I have a preference for flat response from loudspeakers (distinguished from a flat response from earphones), but Basshead Bob might like something warmer and heavier on the lower end, while Delicate Daisy might prefer some brassy sparkle in the treble. And such preferences ought to be respected.

To find out where your own preferences are, you could look up your current setup in the GE database (mainly the frequency response), and if you think that it lacks bass, then another product with a higher line at the bass end of the graph might suit your preference, and vice versa. Or, you could simply look at the graph of a product that you liked - this represents your preference.

But it is important to remember that personal preferences can change over time, and that nobody has such a thing as ‘one-size-fits-all’ preference that happens to be every-one’s cup of tea. Thus, from the Golden Ears’ perspective, we can only recommend products that are towards flat response that most people will find closer to their personal preference and acceptable if not fantastic.