JP2014077970A - Harmonic distortion rate measuring method and apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a harmonic distortion rate measuring method applied for an analog acoustic signal and a digital sound signal, which has no difference between hearing and perception sound quality and measured harmonic distortion rate.SOLUTION: Various harmonic distortion rates are measured by executing subtraction processing of harmonic components inducing a hearth effect (Time Discrimination) from a measured signal. Various Vintage instruments, vacuum tube instruments, coaxial loudspeakers and the like which obtain the evaluation of "sound quality is good" in spite of high distortion rate by various conventional harmonic distortion rate measuring methods are adjusted between superiority in specification and hearing and perception sound quality. In addition, the harmonic components are added to measured data to thereby weight ultra-low distortion rate measurement of a resistor, a capacitor and the like which is difficult to be measured in a normal environment and to perform measurement as normal level (voltage amplitude) distortion rate data subjected to scientific acoustic feeling correction. Thus, a measuring device and an acoustic signal measuring method are inexpensively provided, which can match the conventional distortion rate measurement theory and the sensitivity of "golden ears".

Description

The present invention relates to various harmonic distortion measurement methods and devices that match human audibility applied to analog sound signals and digital sound signals, and in particular, the determination criteria for 'good sound' and 'bad sound' in automatic measurement. It is systematized from various acoustic theories.

'Good sound' and 'bad sound' are greatly related to individual tastes, and it has been said that theoretical systematization has been difficult. However, in a listening experiment of sounds that are always heard in everyday life, such as complex sounds such as natural sounds and musical sounds, or human voices, it is a form that mainly focuses on harmonic distortion, but it is a good sound A trend has been confirmed in the criteria for 'and' bad sound '. Therefore, it is considered possible to scientifically formulate 'good sound' and 'bad sound' with a focus on harmonic distortion.

This systematic formulation of scientific criteria was first initiated by Japanese researcher Seiya Nikaido in the 1970s (see Non-Patent Document 1). In this way, the criteria for 'good sound' and 'bad sound' specify the sound source to some extent, and if it is limited to a form that focuses on the harmonic distortion rate, a specific tendency is confirmed, and scientific research has a history of over 40 years Exists. However, the judgment of “good sound” and “bad sound” of actual sound signals and various sound devices is left to the 100% skilled Golden Year.

With the advent of high-spec audio equipment and audio media in recent years, various types of audio equipment and audio media that appear to exceed human perception limits in terms of measurement values have appeared. In particular, a 1-bit audio equipment group represented by DSD feels that there is less difference between a “good sound” and a “bad sound” between price differences compared to an old audio medium such as a CD. However, it is hard to say that the sound quality for everyone is higher than that of the excellent specs when the listening test is performed in consideration of the preference of lovers or considering “good sound” and “bad sound”.

This trend has been confirmed more prominently in loudspeakers. The US-based ALTEC coaxial speaker 604, which has been sold for more than 60 years, has measured distortion and frequency characteristics compared to the HIFI bottom line product group. Unfortunately, it is extremely difficult to know the high sound quality from the measurement data. However, since its basic design has spread all over the world and is still followed by the latest loudspeakers, it can be understood that good characteristics and good sound quality do not necessarily match.

In addition, loud speakers called 'special machines' supported only by some enthusiasts have a relatively large proportion of coaxial speakers and full-range speakers that generate a large amount of divided vibration distortion, and are multiway inline types with excellent measurement data. -It has been confirmed that loudspeakers are not necessarily high quality and equal.

In recent years, measurement of ultra-low distortion has become possible, and harmonic distortion measurement of electronic components such as resistors and capacitors has become common. However, there is no clear correlation between the measurement data of the 'brand' parts that are considered to have good harmonic distortion and sound quality.

Non-Patent Document 2 is a paper that mathematically discusses how the human auditory system induces auditory masking with reference to pure medical physical phenomena.

The commentary in the paper is a discussion and no conclusion has been reached, but as is clear from mathematical considerations, if the auditory masking elements described in the paper can be offset by pure physical phenomena, the listening sound is triggered when auditory masking is triggered. It can be understood that the discrimination performance can be improved.

In other words, the detection characteristics of the cochlear hair cells function as a 'capture window' for discriminating the fundamental tone in the state where the fundamental tone and the delayed tone are fused. Since perceptible air vibrations are only periodic waveforms, it can be concluded that all perceptible acoustic stimuli can be decomposed into symmetrical waveforms.

Therefore, the detection characteristic of cochlear hair cells is to extract a half period of the listening sound, so that even if the fundamental sound is the same, a constant fundamental sound is obtained from a good periodic waveform in which the period of the symmetric waveform varies depending on the complex situation of various signals I understand that it works to make it easier to extract.

Also, as explained in the paper, in this situation, auditory masking occurs due to the influence of the half-period mask of the listening sound due to the cochlear detection characteristics and the delay of the auditory neuron, but this is rather Considering the fact that perceptual conditions are unique and various animals with developed hearing have similar auditory systems, combining the delay characteristics of auditory neurons makes it easier for auditory to distinguish fundamental sounds under more complex delayed sound perception. It seems that the delay characteristic appears in the auditory neuron. (If the delay characteristic is unnecessary in the auditory neuron, the compensation system should be developed).

As described above, Non-Patent Document 2 has the same content as the explanation of the medical aspect of the Haas effect, but it is difficult to obtain the medical source of the Haas effect and talking about the Haas effect source itself is a part of the taboo. In addition, many of the various similar papers are extremely difficult to understand, and because of the strong nuances of human experimentation, I tried non-patent literature 2 as a reference and tried to explain it abstractly.

Patent Documents 1 and 2 are inventions that cancel out the auditory masking shown in Non-Patent Document 2 by the Haas effect induced by controlling harmonic components electronically. Although known as CMR (unmodulation masking) in a unique perception situation, artificial CMR (unmodulation masking) forced induction techniques have not been developed so far.

At first glance, it seems that CMR (comodulation masking cancellation) in this unique perception situation does not make sense, but in fields such as 'music production', the volume of each part is adjusted to match the inflection of the music to the composer's intention. Performers and staff operate special effects such as sound quality, reverberation, etc. However, even under such circumstances, most producers and composers can hear all the sounds on the score with inflection. I hope.

Although this situation is a major contradiction, Patent Document 1 and Patent Document 2 are the only inventions (as of 2012) that can scientifically improve the contradiction.

Non-Patent Document 3 is a scientific explanation of Patent Document 1 and Patent Document 2. In the past, it was thought that artificial CMR (co-modulation masking removal) was impossible as described above, but it was not possible to accurately interpret linguistic interconversion between CMR (co-modulation masking removal) and actually perceived acoustic phenomena. This is because it has not been carried out, and Non-Patent Document 3 is the first document showing it in science. The paper suggests that the improvement of discrimination by the Haas effect is equivalent to the cancellation of co-modulation masking, taking the perceptual improvement effect by flutter echo as an example, and presents an electronic countermeasure method.

As for the original Hearth effect, an experiment was conducted on air vibration (amount of electric power), and the perception improvement of 10 dB at the maximum was confirmed. Patent Documents 1 and 2 are inventions that induce the Haas effect purely (voltage amount), and the perceptual improvement when this principle is used for conventional harmonic distortion measurement is 20 dB (10 times). . This means that 10 times weighting is possible for measurement data of extremely low distortion such as harmonic distortion measurement of electronic parts.

Therefore, by adding and subtracting the harmonic distortion obtained by the Hearth effect when measuring the distortion factor, it is said why the distortion inherent in the electronic component and the mechanical distortion such as microphonic noise that is confirmed when the electronic component vibrates affect the sound quality. Despite the problems and time-axis distortions such as time jitter in digital audio are very small, the question of why it affects 'good sound' is more detailed from the correlation between hearing and harmonic distortion. It becomes possible to digitize.
Japanese Patent No. 4427672 JP2008-219844 Journal of the Acoustical Society of Japan 28 (9), 485-495, 1972-09-01 Seiya Nikaido Consideration on detection limit and measurement method of non-linear strain Kazuichi Kita Faculty of Engineering, Kyoto University Neural network model for tone recognition based on time-series firing of auditory nerve AES Tokyo Convention 2009 Yasuo Sano Methods for Sound Improving Arts and Discrimination Performances

In the present invention, the harmonic group obtained by the Haas effect shown in Japanese Patent No. 4427672 is subjected to addition / subtraction processing at the time of analysis of an acoustic device or an acoustic signal, thereby performing a combination of various measurement data and a hearing evaluation by a golden ear, We propose a new method for realizing audibility with electronic devices.

FIG. 2 shows a basic configuration of a general distortion factor meter, that is, a distortion factor meter input buffer / calibration signal output means (7), a band elimination filter and a filter fc tuning means (8), and a calibration signal / band elimination signal switching. The harmonic distortion content of the measurement object is measured by the measurement result display means (10) such as a circuit, detection, amplification means (9), various meters, video signal modulation circuit and the like.

In such a conventional measurement method, a vacuum tube amplifier, a coaxial loudspeaker, or the like is measured with a high distortion. In addition, chorus sounds and guitar sounds with the best adjustment of the oral exciter and the addition of harmonics (oral exciters cannot be expected to have any effect on all sound sources) are audible and have no perceived distortion. Although it is perceived, a high distortion rate is similarly measured in the conventional measurement method.

FIG. 3 shows the basic configuration of the distortion meter according to the present invention. Distortion meter input buffer / calibration signal output means (7), band elimination filter and filter fc tuning means (8), calibration signal / band elimination signal switching circuit and detection, amplification means (9), ESSENCE harmonic component addition / subtraction The harmonic distortion content to be measured is measured by the measurement result display means (10) such as means (11), various meters, video signal modulation circuit and the like.

The difference from FIG. 2 is the ESSENCE harmonic component addition / subtraction means (11). ESSENCE is an acronym for English Supersonic Sonic with Noise Corresponding Ear Parameter in Japanese Patent No. 4427672, and is used as an abbreviation for the same expression used in the 2009 AES Tokyo Convention.

FIG. 1 shows an actual circuit derived from the embodiment in Japanese Patent No. 4427672. In addition to various linear amplification systems such as a buffer / phase inversion 1 circuit (1), a phase inversion 2 circuit (2), and an ESSENCE harmonic addition / subtraction circuit (6), spectrum control and harmonic generation 1 circuit (3 ), Spectrum control and harmonic generation 1 circuit (3), asymmetric differential operation of input signal / harmonic generation 2 and audible high-frequency filtering circuit (4), ESSENCE harmonic addition / subtraction polarity adjustment volume (5) Weighting is applied to harmonic components that induce CMR (comodulation masking cancellation), which is important for '. The trimmer in the spectrum control and harmonic generation 1 circuit (3) sets the phase shift start threshold of the harmonic component necessary for inducing CMR (comodulation masking cancellation), and automatically adapts when using the DSP. However, the actual circuit is set so that the measured value shows the minimum value according to various measurement conditions.

According to the present invention, it is possible to measure a harmonic distortion rate that matches human perception. As described above, along with the old-style auditory correction values related to the amplitude and frequency characteristics represented by Robinson's loudness curve, scientific weighting to the measured harmonic distortion has become possible. Furthermore, the establishment of this technology means that the importance of 'uniform power spectrum radiation in the sound field' that some speaker designers have long advocated can be realized by the radiation of harmonic distortion that induces the Haas effect. It was possible to present simultaneously from various aspects such as physical, psychological, medical, etc., and it was proved by presenting a working circuit that electronic engineering could be approached to the same conclusion.

The DSP is optimal for delicate adaptive operation, but the measurement target is harmonic distortion and you want to make the bandwidth as wide as possible, or you can specify the measurement method to the user like a semi-automatic measurement device. If parameter operation is possible for each, an analog circuit composed of operational amplifiers is optimal for both cost and maintainability.

FIG. 3 shows an embodiment of a distortion meter using the present invention. It should be noted that differences in various standards such as CCIR and IHF and differences in measurement methods such as total harmonic distortion and noise distortion do not affect the essence of the present invention, and are therefore omitted.

The method of fully automating FIG. 3 is that the band elimination filter is automatically tuned with an analog PLL or the like and the two trimmers in FIG. This is possible.

The following is the configuration and explanation of each part. The distortion meter input buffer / calibration signal output means (7) directly connects the dedicated signal source output to the calibration signal / band-removal signal switching circuit and detection / amplification means (9) in addition to the input buffer required as a measuring instrument. By doing so, it is a CAL (calibration) function that constitutes the measurement result display means (10) such as various meters and video signal modulation circuits.

The band elimination filter and the filter fc tuning means (8) are selective band elimination means for manually or automatically tuning to a bandwidth equal to the input spectrum and removing only the input signal. The ESSENCE harmonic component adding / subtracting means (11) is tuned to manually or automatically adjust the two internal trims so that the value displayed by the measurement result display means (10) such as various meters, video signal modulation circuits, etc. shows the lowest value. It is a means for adding and subtracting harmonic components that induces CMR (comodulation masking cancellation), which has been conventionally considered as “good distortion” but difficult to grasp in measurement, and is therefore a factor of high distortion. Various meter, video signal modulation circuit, etc. measurement result display means (10) converts a CAL (calibration) signal or signal to be measured into various meter drive signals or video signal modulation signals and displays predetermined numerical values on predetermined display means To do.

Applying this technology to active noise canceling technology attenuates only unpleasant sounds that affect the human spirit, or changes the way the noise is heard (for example, sports cars) It is possible to efficiently realize a human-friendly sound environment, such as sound leakage from portable music devices, wind noise from high-speed railways, and engine sound from various airplanes. It is done.

It is an example of an ESSENCE harmonic addition / subtraction circuit. It is a block diagram of a configuration example of a general harmonic distortion meter. It is a block diagram of a configuration example of a harmonic distortion meter according to the present invention.

DESCRIPTION OF SYMBOLS 1 Buffer / phase inversion 1 circuit 2 Phase inversion 2 circuit 3 Spectral control and harmonic generation 1 circuit 4 Differential operation, harmonic generation 2 and audible limit high frequency filter circuit 5 ESSENCE harmonic balance adjustment volume 6 ESSENCE harmonic addition / subtraction circuit 7 Distortion meter input buffer circuit, calibration signal output circuit block diagram notation 8 Band elimination filter and filter fc tuning circuit block diagram notation 9 Calibration signal / band elimination signal switching circuit and detection, amplification circuit block diagram notation 10 Various meters, video Signal modulation circuit display means block diagram notation 11 ESSENCE harmonic component addition / subtraction means block diagram notation

Claims (5)

In each process of recording, transmitting, amplifying, and reproducing an acoustic signal, an odd-order harmonic group and an even-order harmonic obtained by controlling a delay time sufficient to induce a Haas effect (Time Discrimination) as an amplitude function of the acoustic signal. By adding or subtracting two types of harmonic groups such as wave group, sine phase harmonic group and cosine phase harmonic group independently or in a mixed state, the acoustic noise level can be reduced without increasing the audible noise level. A measurement technique that generates harmonic components that improve discrimination performance and measures total harmonic distortion by adding and subtracting spectral information extracted from these harmonic components. A measuring instrument to which claim 1 is applied. In each process of recording, transmitting, amplifying, and reproducing an acoustic signal, an odd-order harmonic group and an even-order harmonic obtained by controlling a delay time sufficient to induce a Haas effect (Time Discrimination) as an amplitude function of the acoustic signal. By adding or subtracting two types of harmonic groups such as wave group, sine phase harmonic group and cosine phase harmonic group independently or in a mixed state, the acoustic noise level can be reduced without increasing the audible noise level. A measurement technique that measures the noise distortion rate by generating harmonic components that improve discrimination performance and adding / subtracting spectral information extracted from the harmonic components. A measuring instrument to which claim 2 is applied. Software and media including binary code that implements claims 1, 2, 3, and 4 by digital signal processing (DSP).

Images