Among people with a critical outlook on the world around them, a stereotype has taken root that high-resolution audio formats (with a sampling frequency of more than 44.1 Hz and quantization of more than 16 bits) in consumer audio equipment are just a marketing ploy that is created for those who do not know about the thresholds of perception. Like, there is improvement beyond the thresholds of human perception.
For a long time I was inclined towards a similar opinion, until I began to periodically note that I subjectively (for an inexplicable reason) like some of the highres records more. I can’t say that I’ll easily pass a blind test and determine with high accuracy where the highs are and where the mp3 is at 320 kbit / s. But listening to the records in AIFF format with a sampling frequency of 192 kHz and 32 bits, it seemed to me that I notice subtle improvements in the dynamic range and at low reproduction.
I do not really trust my ears. I began to doubt my own conclusions and decided to unearth something about the theoretical possibility of hearing these differences.
To his surprise, he discovered not marketing experts, but quite authoritative evaluations of AES (Audio Engineering Society) specialists. My search did not give an unequivocal answer, but I became less categorical about Hayrez.
Relying on personal experience, I have repeatedly written that Hi-res is needed only for people who do not have enough to listen, but you need to know that the sound quality is impeccable. After analyzing the findings of audio engineers and psychoacoustics specialists, I realized that there is a theoretical possibility to hear the difference.
Frequency and other “resolution” parameters
The main question that researchers have been facing over the past couple of decades is: does it make sense to increase the sampling rate of recordings and reproducing audio equipment? The question was posed in connection with the fact that the physiological threshold of hearing is limited to the range from 16 Hz to 21 kHz.
According to one of the pioneers of the domestic psychoacoustics, a professor at St. Petersburg University, a doctor of technical sciences, a member of the Coordination Council for Acoustics of the Russian Academy of Sciences, the chairman of the AES St. Petersburg department, Irina Aldoshina, no one has yet given an answer to the question.
The fact is that, despite the lack of evidence-based results of blind tests, where the subject unequivocally determined hayrez, there are a lot of theoretical prerequisites to assume that high “resolution” can affect the subjective perception.
Thus, V.Voychik, a professor at the University of Montreal McGill, wrote that one of the ways to increase the realism of a musical signal and create a “presence” effect, increase “transparency” is to increase the sampling rate above 44.1 kHz. Accordingly, the expansion of the upper frequency range beyond 20 kHz. (after reading this, I questioned his expert status, but in vain)
He also noted that in order to create a realistic signal, in addition to the sampling frequency, it is necessary to increase the resolution of reproducing and (!) Recording (!) Systems in the time, spatial and dynamic regions. This conclusion was made by a Canadian professor on the basis of the annual reports of AES, as well as based on knowledge of the complexity of the physiology of hearing and the neurophysiology of auditory perception.
ADC: Sampling Interval, Filters and Transparency
It is known that in order to transform an analog audio signal into a digital one, it is necessary to perform sampling, quantization and encoding. These processes occur during digital recording or when digitizing analog material. To do this, use the ADC. The process of such a transformation takes place in accordance with the Kotelnikov-Shannon-Nyquist theorem.
In accordance with the same theorem, accurate reconstruction of the original signal during the inverse transformation is possible only when the sampling frequency is higher than twice the maximum frequency in the spectrum of the original signal.
Thus, it would seem that the sampling frequency of 44.1 kHz (CD DA format) in theory should suffice for accurate transmission of the entire audible spectrum, i.e. fd> 2fb. But not everything is so simple, the problem lies in the sampling intervals.
AES experts write that at the time of sampling the signal should be constant and pass through a low-frequency so-called. anti-aliasing filter, which cuts off the signal at a frequency of fd / 2 to prevent the appearance of artifacts. Such filters are in all ADC. This filter causes the dispersion of the pulse characteristics of the original signal, which is due to the non-uniformity of the AF and AF characteristics and phase nonlinearity in the passband.
The result of this side effect of the filter is the temporary dispersion of the signal and the presence in each sample of information elements from the previous ones. Due to the dynamism and complexity of the musical signal, this dispersion may have some effect on the subjective perception. Although experts note that for this you need to be very attentive and experienced listener with outstanding "golden" ears.
When converting with a frequency of 44.1, a sampling interval of 22.7 μs is used. Wojczyk, Aldoshina and other AES experts pay attention to the problems of the temporal characteristics of the format. The dynamic capabilities of the pipe when performing forte allow you to reach peaks of 120-130 dB for 10 µs, dulcimers allow you to get an instant rise to 136 dB in 7 µs.
Accordingly, the sampling time intervals used for recording CDs are very far from the dynamic capabilities of the instruments. In modern format hayrez, these intervals are shorter (from 1 to 0.16 μs) and therefore, at the level of dynamics, they are able to more accurately and more realistically transmit sound.
Also, as an important characteristic of realism (“naturalness”) of sound when played back, AES experts consider so-called. "transparency". This is a subjective timbral sound feature inherent in the sound, which according to many experts closer to the natural. The researchers note that this feature also depends on the recording time parameters, i.e. from the sampling interval.
Ultrasound: need or not?
Scientists and enthusiasts in conducting recordings of “live” instruments using precision instruments, since the 1970s, have begun to note the presence of ultrasonic waves in the spectrum of these instruments.
As an example, a spectrum of a pipe can be cited, where components with a frequency of 40 kHz and a level of up to 60 dB are recorded, violin and viola have ultrasonic components up to 100 kHz, with a level of up to 85-90 dB.
Despite the fact that human hearing is not able to perceive waves with a frequency higher than 20 kHz (in rare cases 22 kHz - as a rule, in children), the presence of pronounced high-frequency components changes the temporal structure of the signal.
Fletcher, Kuznetsov and other authors who investigated the sounds of live instruments, noted that such a temporary effect can significantly affect the subjective perception of sound, despite the fact that the frequency itself is not perceived by ear.
Such information in theory answers the question about the expediency of increasing the sampling rate. At least when recording playback of live acoustic instruments.
Dry residue
Despite the fact that blind tests do not yet confirm significant differences in the perception of conventional and high-resolution formats, at the theoretical level they have a higher fidelity of reproduction. Moreover, sampling intervals, a reduction in the level of temporal dispersion, and the ability to reproduce the ultrasonic components of the signal suggest that the difference between the classical CD DA and high “resolution” formats may be noticeable and essential for subjective perception.
JeansOur
catalog contains
multiroom systems ,
media players and other equipment capable of playing music in high resolution formats.