Before to start speculations about idea that high frequencies are or not are 'snake oil' level of things...
Unfortunately CD audio frequency standard (up to 20kHz) was created before 1980, on the basis of comparative tests of 'short listening sessions (every audio sample was auditioned no more than few minutes or less)', now it is recognized, that 'prolonged hearing' tests give other comparative results, especially in audio HF region.
It is generally accepted that 'humans ear' cannot hear/perceive air vibrations in the frequency range above 20 kHz, but new 'controlled lab level' tests it is proved that there is significant increase in the regional cerebral blood flow (rCBF) in the brain, when heared complex audio contain 20kHz to 22 kHz (or even higher) audio components... so increase of blood flow is surely influencing some processes in listeners brain... so it alters the 'mind' anyway. Another question is, is every listener somehow 'feeling' this difference in brains blood flow or is aware of 'changing subtle processes' in his mind... Other researche show, that power spectra of the alpha frequency range of the spontaneous electroencephalogram (alpha-EEG) recorded from the occipital region, when subjects were exposed to sound containing both an HFC and an LFC, compared with an otherwise identical sound from which the HFC was removed (i.e., LFC alone).
Psychological evaluation indicated that the subjects felt the sound containing an HFC to be more pleasant than the same sound lacking an HFC. These results suggest the existence of a previously unrecognized response to complex sound containing particular types of high frequencies above the audible range. We term this phenomenon the “hypersonic effect.”
details
http://jn.physiology.org/content/83/6/3548
http://www.icad.org/Proceedings/2002/YagiNishina2002.pdf
about HF harmonics
http://www.cco.caltech.edu/~boyk/spectra/spectra.htm
wider overview about music and human ear
http://www.silcom.com/~aludwig/EARS.htm
A couple of weeks ago I took some time to write up a criticism of this study on a website whose mention is verboten here on head-fi, so here's a copy:
In response to the oft-cited Oohashi high frequency study - http://jn.physiology.org/content/83/6/3548
First, Oohashi is an odd and brilliant man, an artist and a scientist. Here's an interview which gives you a gist of why he does what he does - http://www.sidianersatzvanes.com/supplemental/dr-tsutomu-oohashi/
To him the pursuit of what he calls the "hypersonic effect" is something of a lifelong endeavor. His hypothesis is deeply rooted in decades of experience in the music business and research in biology. Finding evidence to substantiate his ideas on human hearing is deeply important to him. When, in the late 90s, Japanese researchers attempted to substantiate claims of the superior audio quality of high sampling rates and high bandwidth recordings (96khz & DSD) in preparation for the arrival of new commercial formats (SACD, DVD-A, etc.), enthusiasm sometimes overrode practical sense. Improvements were quantified by soliciting listening reports in the form of questionnaires, as opposed to objective observation of perception.
That's not to say that such research was unfounded. One of the major revolutions in engineering was brought about by Floyd Toole's questionnaire-based research from the 1970s at NRC. Without those subjective reports we would not have a firm grasp of what metrics of speaker design actually matter to listeners.
But Oohashi, in true researcher fashion, did not buy the hypothetical explanations and positive results of those late 90s studies. Being a neuroscientist he wanted to apply techniques from medical diagnostics - namely brain imaging. It's important to remember that to the 70-something Oohashi of the late 90s these techniques were still very new, unavailable at the time of his doctoral research from decades back. Enter the hypersonic study.
The study was laboriously designed to include real live music with ultrasonic content at high sampling rates and with high bandwidth transducers. Recordings were made by lab mics and converted to digital @ 1.92mhz single bit delta-sigma modulation. Not being an idiot Oohashi was of course aware of the possibility of intermodulation distortion messing up his results.
In simple terms, intermodulation distortion is composed of non-harmonic products of interference between two or more frequencies. Say you measure 18 & 19khz tones, there you're looking for the level of the 1khz tone that is the difference of the frequencies. However in real life IMD is even more complex than that because we are dealing with complex non-linear systems. Even with the 18 & 19khz test tones you will get a pattern of tones on both immediate sides of the respective frequencies as well as all sorts of other hash. And this is simply on the electronic end of things! Transducers are even worse, the distortion products of musical signals may add up to be a degree of magnitude larger than that of single test tones. Yes, that speaker with .3% THD @ 1khz could still produce 3% or more total distortion with complex material, just from electromechanical limitations.
In order to minimize the possibility of IMD on the digital side of things Oohashi and crew split the signals into LFC (low-frequency content, under 22khz) and HFC (high frequency content) by using very steep high-order digital filters. Both signals were then rooted to separate amplifiers and separate transducers. The amplifiers were by Accuphase, some of the best that you can get. The LFC content was reproduced by a horn tweeter and two woofers and HFC content was reproduced by a diamond dome tweeter. Diamond domes have a high breakup frequency, typically 40+ khz, but up there when they ring they truly ring, 10-20+ db.
Anyway, the test protocol was simple and straightforward. There were four conditions: LFC content only, HFC content only, LFC+HFC content and silence for a baseline measurement. The first measurement tool borrowed from medical diagnostics was an alpha wave electroencephalograph, the thing where they stick a bunch of electrodes all around the skull and measure changes in electric potential. It is by necessity a very broad measurement, indicative of massive neuronal activity but indiscriminate in terms of specifics. High levels of alpha waves occur during meditation and when alpha waves intrude on sleep it is typically during REM cycles, vivid wakeful-like states. They indicate a kind of active idleness, when a person is frantically thinking you tend to have a decrease in alpha wave activity. This activity can also be willfully regulated by a person through neurofeedback training or meditation. What does it mean? Well...
The results indicated LFC and HFC signals individually offered only minute, insignificant changes in alpha activity, but that LFC+HFC signals showed a notable increase. This is the red flag: if ultrasonic components of musical signals can be interpreted to have an effect by boosting alpha potential then why do they do so only in the presence of regular 20-20,000hz signals?
The second measurement tool was positron emission tomography, a state-of-the-art three dimensional imaging system of cell metabolism. The results showed markedly decreased metabolic brain activity with HFC signals when compared with LFC and baseline signals and, again, markedly increased metabolic activity with LFC+HFC signals. Ultrasonic components this time are shown to not have much of an effect unless they are combined with LFC.
Oohashi concluded that only the combination of the two signals produced the hypersonic effect, moreover that the combination of the two signals was responsible for
novel brain activity outside of the areas normally associated with auditory processing.
Here's the rub, though: previous research has shown that people will in fact react differently to music once a certain level of distortion is introduced. Moreover, if these are experienced listeners primed for critical listening, they are likely to become actively annoyed at the presence of unnatural distortion. Oohashi's setup was inadequate to prevent the likelyhood of added distortion, whether through unstated digital filtering methods, or transducer limitations, or even the physical positioning of the speakers (high frequency wavelengths are so small by themselves and in comparison to a regular dome tweeter that moving your head even a centimeter can have an effect).
Now these concerns could've and would've been allayed if Oohashi's results were repeatable. Independent teams have failed to reproduce both the detection of ultrasonic frequencies by the ear and brain. They've also discounted the purported synergistic properties of adding ultrasonic content to regular recordings, since subjects simply couldn't tell the difference (Oohashi claimed that LFC+HFC reports indicated increased enjoyment).
The tough thing is that without being familiar with the details of Oohashi's setup or with prior brain imaging studies, it is difficult to draw any sort of definitive conclusion from his study. But being skeptical one can say that his study has flaws, that it has proven non-reproducible and that there is a sizable amount of contrary evidence.
