[Someone in a Facebook group asked me more or less the question below.]

TK, how do you feel about ‘high resolution’ digital audio? Can you hear the difference between 16/44.1 and 24/96??

I have not seen any formal, properly designed experimental/testing data which demonstrates the ability of adult humans to hear sounds in free air (as opposed to bone conduction, which has been explored as part of hearing assistance R&D) significantly higher than the commonly accepted nominal bound of human hearing (20 Hz – 20 kHz) and have never seen credible scientific evidence presented of outliers who could hear above 22 kHz.

(There is a study by a Japanese researcher, Oohashi, that produced experimental data suggesting a respondent in their testing, as i recall it, did show such outlying capability, but that study’s methodology was criticized as compromised in peer review. As far as I know, that finding has never been replicated.)

There is certainly no evidential reason I know of to think there is any widespread ability to hear above the scientifically accepted human threshold.

So, what do we make of people who claim they can hear performance differences at high sample rates on a given converter– of whom there are quite many?

The first and most ‘obvious’ factor in the common view of perceptual scientists is the old experimental bugaboo, cognitive bias. Cognitive bias can manifest in a number of forms, most prominently, as confirmation bias (the so-called placebo effect).

Of course, no one likes to be told they are imagining perceptions, particularly perceptions they believe set them apart from the common ability — but confirmation bias of this nature is extremely well documented in the scientific literature and has been studied extensively.

It is the fundamental reason that perceptual testing is virtually always done in double bind conditions to prevent extraneous cues that can influence the subject to believe one way or the other.

Fundamentally, we humans did not evolve to be rational creatures who innately analyze the world through logic. (Logical analysis and reasoning are essentially learned capabilities.)

On the contrary, humans, like other animals, evolved in such a way as to facilitate their survival in a primordial wilderness, making snap decisions based on heuristic shortcuts in order to be able to respond to immediate threats in their environment.

BUT WAIT, THERE’S MORE…

There is a substantial body of unscientific (but quite sincere) assertions by a number of fairly experienced practitioners of audio recording who have claimed to be able to tell the difference between sample rates in their recording and playback systems.

This is not necessarily suggestive that these people are imagining their ability to differentiate!

The most commonly asserted, scientifically credible reason is that different ADC/DAC devices can be demonstrated to perform with varying sonic characteristics at different sample rates — characteristics that CAN manifest in the nominal audible range — in large part because of differences in anti-alias filter design and application at different sample rates. (Judging from anecdotal evidence, this was probably a bigger factor in early converter designs. However the rise of multi-bit over-sampling in modern converter design has generally greatly lessened perceivable differences between well-designed converters.)

So, it is, indeed, entirely possible that a recordist or audiophile might be able to perceive performance characteristic differences in a given converter operating at different sample rates. This, however, can be explained without having to resort to overthrowing scientifically accepted nominal frequency bounds of human hearing.