JPS6351256B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a dissonance measuring device that automatically measures the dissonance that affects the psychological quality of general audio transmission systems, including a wide range of electroacoustic equipment, and a method for measuring the dissonance. This paper relates to a dissonance level psychological quality measurement method that objectively measures psychological quality by measuring the dissonance level used, and its aims are roughly as follows.

(1) Design of the transmission system so that both psychological and physical quality can be measured simultaneously to clarify the correspondence between the two, and to omit psychological tests by viewers;
Streamline monitoring and operations.

(2) Various benefits not previously available can be obtained by measurement using a program sound model signal as a test signal.

In general, to evaluate the quality of transmission systems, physical characteristics such as amplitude and phase frequency characteristics are usually measured, but psychological quality cannot be fully inferred from the measurement results. conducted separate sound quality evaluation experiments using music and speech as test signals, supplementing quality evaluations based on physical characteristics.

However, even in that case, psychological quality and physical properties are measured on separate test signals, so
There is no direct correspondence between the two, and therefore, complex statistical processing is required to establish the correspondence.

Furthermore, when music or speech is used as a test signal, in addition to the disadvantage that industrial parameters cannot be displayed, there is also the disadvantage that accurate evaluation is difficult due to differences in personal preference.

Furthermore, the specific implementation of psychological quality evaluation has conventionally required viewing experiments with a large number of viewers and long periods of time, and in some cases requires offline post-processing of data, including complex statistical processing. It was necessary.

These conventional shortcomings can be overcome by using a program sound model signal, which is a combination of sine wave signals that accurately represents the statistical characteristics of the program sound with timbre in the test signal, for both psychophysical and psychophysical quality measurements. However, the following problems still remained, especially in measuring the psychological quality of the transmission system itself.

(1) Quantitative measurement of both input and output sound quality of the transmission system under test (2) Expression and measurement of the psychological quality of the transmission system under test (3) Online real-time processing The purpose of the present invention is to measure the psychological quality of the transmission system itself. The objective is to solve the above-mentioned problems (1), (2), and (3) regarding measurement.

In the present invention, the above problem (1) is solved by limiting the sound quality to be measured to the dissonance felt due to nonlinear distortion and noise generated in the transmission system to be measured.

In other words, the method of objectively measuring the feeling of dissonance using a psychological scale called the degree of dissonance is described in, for example, "Consonance Theory," Toshiba Review, Vol. 25, No. 4, p. 481-486.
(1972), and it can be objectively measured using this method.

However, the above-mentioned "consonance theory" does not provide a sufficient measurement method for low-level distortion and noise near the minimum audible sound pressure level generated in the transmission system, and the present invention does not provide a sufficient measurement method for low-level distortion and noise near the minimum audible sound pressure level that occurs in the transmission system. This makes it possible to measure the degree of dissonance of the total sound pressure level.

Regarding issue (2) above, the psychological quality of the transmission system itself determined from the input sound quality and output sound quality of the transmission system under test is defined as the "psychological quality of the transmission system" as described later. , the problem is solved by providing a new measurement method.

Furthermore, the above problem (3) is solved by providing a method for processing spectral data, a further improved method for measuring the psychological quality of the transmission system, etc., as will be described later.

That is, the dissonance level psychological quality measurement method according to the present invention measures dissonance strengths corresponding to the input and output signals of the transmission system under test, and converts the difference between these dissonance strengths according to a power law. The method is characterized in that the psychological quality of the transmission system to be measured is expressed by degrees.

Further, the dissonance degree measuring device according to the present invention includes at least a dissonance strength calculation means, a dissonance strength-dissonance degree conversion means, and a dissonance degree display means. The present invention is characterized in that a means for truncating below the minimum audible level and a masking means are provided preceding the calculation means.

The present invention will be described in detail below with reference to the drawings.

First, to explain the conventional dissonance measurement method known from the above-mentioned "consonance theory" literature, a fast Fourier A transform (FFT) analysis device decomposes the sound to be measured into frequency spectra that can be considered as a set of discrete pure tones with the required frequency resolution, and then calculates each chord for each of n sets of two frequencies of the pure tone spectrum. The degree of dissonance of the chords in the set D _j (j=1, 2,...
..., n) are determined according to the physical-psychological experimental formula described in the above-mentioned "consonance theory" paper.

According to the "consonance theory," the psychological quantity _D
It can be converted into a physical quantity I _j (j=1, 2, . . . , n). That is, I _j = (D _j /K) ^1/ 〓 (j = 1, 2, ..., n) (k, β are constants) The dissonance strength I _j of each of these sets has additive property. Therefore, if we add up the dissonance intensity I _j for the entire frequency spectrum of the sound to be measured, we get _o 〓 ^j=1 I _j = (D/k) ^1/ 〓 Therefore, D=k( _o 〓 ^j=1 I _j ) The following relationship is obtained, and the degree of dissonance D of the sound to be measured can be calculated according to the flowchart shown in FIG. 1 showing the above-mentioned calculation process.

However, according to the above-mentioned "consonance theory" literature, the sound pressure level of the target sound to which the above-mentioned conventional dissonance measurement method can be applied is 40±
It is limited to the medium level range of about 20dB (SPL),
It cannot be applied over the entire range of normal sound pressure levels.

The dissonance measuring device according to the present invention having the configuration shown in FIG. 2 is similar to the conventional dissonance measuring device having the configuration corresponding to the flowchart shown in FIG. The frequency spectrum obtained by introducing the frequency spectrum into the frequency spectrum converter 2 is sequentially supplied to a two-frequency dissonance strength calculation device 6 and a summing device 7, and the dissonance strength calculated for the entire frequency spectrum is converted into a dissonance degree by a converter 8 and displayed. Device 9
A two-frequency dissonance strength calculation device 6 in a dissonance measurement measurement device configured to display a value of 0 to 30dB (SPL)
The conventional "consonance theory" is extended and applied to the low sound pressure level range, and the degree of dissonance D _j can be calculated and determined using the above-mentioned calculation formula.

In the first place, with regard to nonlinear distortion and noise generated in the transmission system under test, as the characteristics of the transmission system have improved, distortion and noise in the extremely low sound pressure level region near the minimum audible sound pressure level have an impact on the quality of the transmission system. As a result, it has become important to accurately measure the degree of dissonance caused by distortion and noise in the low sound pressure level range.

Therefore, in the present invention, the scope of application of the above-mentioned calculation formula shown in the paper on "consonance theory" is extended to the low sound pressure level range up to the minimum audible sound pressure level. The frequency dissonance intensity calculation device 6 is configured, but if the scope of application is expanded to measure the degree of dissonance based on distortion and noise in the low sound pressure level range of 0 to 30 dB (SPL), the following two methods will be realized. The dissonance level will also be measured for sound components that cannot actually be heard due to auditory phenomena, and the results of the dissonance level measurement will no longer match the results of the rating experiment.

One of the auditory phenomena mentioned above is an auditory phenomenon expressed by the minimum audible level frequency characteristic, in which pure tones below a certain sound pressure level cannot be heard, and the other is an auditory phenomenon expressed by the minimum audible level frequency characteristic, in which pure tones below a certain sound pressure level cannot be heard. This is an auditory phenomenon expressed by masking frequency characteristics, in which pure tones at a low sound pressure level are masked by pure tones at a high sound pressure level, making them inaudible.

Therefore, the minimum audible level or below truncation device 4 refers to the conventionally known minimum audible level frequency characteristics to determine the frequency spectrum components of the frequency spectrum of the sound to be measured that are below the minimum audible level and cannot be heard in reality. , is not input to the two-frequency dissonance strength calculation device 6.

In addition, in the masking device 5, by referring to the conventionally known masking level frequency characteristics related to the relative sound pressure level and absolute sound pressure level of two pure tones, the masking device 5 calculates the difference between the frequency spectra that can be treated as a set of pure tones of the sound to be measured. Frequency spectrum components that are masked and cannot be heard in reality are determined and are not input to the two-frequency dissonance intensity calculation device 6. It goes without saying that the above-mentioned masking phenomenon includes not only two pure tones that are input at the same time, but also so-called forward masking and backward masking that occur between two input pure tones that are adjacent in time.

Therefore, the below-minimum audible level truncation device 4 and the masking device 5, which are added to the dissonance measuring device having a conventional configuration according to the present invention, each have, for example, each data of the conventionally well-known minimum audible level frequency characteristic or masking frequency characteristic. is accessed and read out from the lookup table memory stored in advance using required data such as the frequency of the input frequency spectrum component and the relative sound pressure level, and the gate circuit is controlled according to the result of level comparison with the input frequency spectrum component one by one. It can be relatively easily configured by This makes it possible to accurately measure the contribution to increase in dissonance intensity in the low sound pressure level range.

Therefore, below the minimum audible level truncation device 4
As shown in FIG. 2, the masking device 5 is placed in front of the input side of the two-frequency dissonance intensity calculation device 6, but the relative sound pressure of two pure tones is In order to reduce as much as possible the amount of data processing by the masking device 5, which requires complex data processing taking into consideration the level, absolute sound pressure level, etc., it is preferable to place the below-minimum audible level cutting device 4 in front of the masking device 5, as shown in the figure. .

Next, in order to solve the conventional problems (1) to (3) regarding psychological quality measurement mentioned above, we calculated the psychological quality of the transmission system itself, which is not mentioned at all in the conventional "cooperation theory", by measuring the dissonance level. To explain the method of measuring psychological quality according to _the present invention, which measures the _{psychological} quality using The increment of dissonance strength (I _p −I _i ) is
This is caused by linear and nonlinear distortion and noise in the transmission system. In the present invention, the relative psychoacoustic deterioration of the output sound quality with respect to the input volume caused by passing through the transmission system where such nonlinear distortion and noise occurs is determined by the dissonance degrees D _p and D _i of the input and output of the transmission system, respectively. does not match the difference between
Dissonance degree obtained by converting the increment of dissonance strength (I _p −I _i ) according to the power law Q=k(I _p −I _i )〓 It was found that k and β are constants. Ta. Therefore, in the present invention, this degree of dissonance Q is defined as the psychological quality of the transmission system.

This transmission system psychological quality Q can be easily calculated as described above with reference to FIGS. 1 and 2, and moreover, good correspondence can be obtained with the psychological quality based on conventional evaluation experiments.

However, in order to construct a transmission-based psychological quality measuring instrument with excellent practical performance, it is also necessary to consider the form of introduction of the signal to be measured from the transmission system to the measuring instrument.

In other words, from the perspective of using psychological quality measuring instruments, the third
There are two possible ways to introduce the signal under test, as shown in Figures a and b.

Figure a shows a signal introduction form in which the input/output signals of the transmission system can be switched as they are and introduced as input to a psychological quality measuring instrument, and the configuration is suitable for this signal introduction form. is used as a psychological quality measuring instrument. In addition, Figure b shows a signal introduction form when only the output of the transmission system can be introduced as an input to the psychological quality measuring instrument because the input and output ends of the transmission system are far apart. A psychological quality measuring instrument is one with a configuration that suits the form.

One of the other objects of the present invention mentioned above is that it exhibits the same performance in the latter case as in the former case, and can be used in common in both cases. This is achieved by realizing a psychological quality measuring instrument that also provides the following advantages.

That is, in the signal introduction form shown in FIG. 3a, after measuring the degree of dissonance on the input side of the transmission system,
Since the degree of dissonance on the output side is subsequently measured, the degree of dissonance of the output of the transmission system corresponding to the input of the transmission system delayed by the time required for measurement on the input side is measured, which is useful for measuring the quality of the transmission system. Differences occur in the signals to be measured, and measurement errors based on these differences greatly affect the measurement performance of psychological quality measuring instruments, especially when dynamic test signals are used. On the contrary, Fig. 3b
In the signal introduction mode shown in , since the degree of dissonance is measured only on the output side of the transmission system, there is no problem of measurement time difference, and the above-mentioned measurement error does not occur in psychological quality measurement.

In addition, in psychological quality measuring instruments, it is possible to avoid redundantly performing dissonance calculations for input signal components included in transmission system output signal components on both the input and output sides, and it is possible to improve processing speed. You also get benefits.

In addition, since the psychological quality measuring instrument does not directly handle the transmission system input signal, it has the advantage of being able to measure the transmission system psychological quality even when the transmission system input and output terminals are far apart. .

Therefore, in a psychological quality measuring instrument that can obtain such numerous advantages, it is possible to calculate the dissonance intensity increment (I _p −I _i ) necessary for measuring the psychological quality of the transmission system without directly handling the transmission system input signal. How to perform the measurement will be explained in order below.

First, to discuss the differences in required performance of measuring instruments due to differences in the physical properties of signals handled by sound transmission systems, generally speaking, sound transmission systems do not directly input or output sound waves, such as microphones or speakers. There are transmission systems, such as amplifiers and tape recorders, that input and output electromagnetic signals. Therefore, when both the input and output are sound waves, the psychological quality of the transmission system itself can be determined by measuring the dissonance intensity increment (I _p −I _i ) using the absolute values of the input and output sound pressures as they are. . In this case, it is necessary to calibrate the sensitivity of the spectrum analyzer 2 with respect to sound pressure, but the sensitivity calibration can be performed using the electrical signal output of a standard microphone in response to a known sound pressure input.

On the other hand, in order to determine the psychological quality of the transmission system itself, which does not handle sound waves, it is necessary to provide an appropriate standard sound pressure corresponding to electromagnetic signals.For example, a standard of around 80 dB (SPL), which is close to the average sound pressure of an orchestra, is required. You can choose the sound pressure.

In the present invention, the effective value of the transmission system input signal and the effective value of the output signal are obtained from the sum of the power spectrum components, and the spectral level is corrected so that the effective value corresponds to the reference sound pressure level. shall be carried out.

Such signal processing can be realized by inserting an effective value calculation circuit at point P in the configuration of the apparatus of the present invention shown in FIG. In this case, the input signal component and the distortion component included in the transmission system output signal are separated by threshold value judgment, and the circuit that calculates the ratio between the sum of these input signal components and the sum of the distortion component is shown in Fig. 2. It is also possible to arrange a branch at point P of , and to calculate and display the physical strain rate at the same time.

Next, in general spectrum analyzers that use fast Fourier transform (FFT), the spectrum of the input signal is usually analyzed through a time window such as a Hanning window, so even if the input signal is a sine wave, the analysis result will be , a spectrum spread occurs around the input signal frequency. Furthermore, even when no time window is provided, it is rare that a single frequency spectrum is obtained as an analysis result.

Therefore, in the present invention, when a program sound model signal is used as a test signal as a condition for using the above-mentioned psychological quality measuring device, each sine wave is spectral analysis means for obtaining purely line spectral amplitudes only at wave frequencies; and spectral processing for processing the spectrum into a form suitable for input to the two-frequency dissonance intensity calculation device 6 in the configuration shown in FIG. As shown in FIG. 4, it is necessary to newly add a processing means as a spectrum processing device 21 at point P in the configuration shown in FIG. 22 and peak level sorting device 23 represent these means, respectively.

Therefore, in order to convert the sinusoidal sum spectrum analysis result into a line spectrum, fast Fourier transform (FFT) is required.
In the analysis, we focus on the signal levels at discrete frequencies at three adjacent points, and if the signal level at the center is the peak, we perform spectrum processing in which the spectral amplitudes before and after it are reduced to 0, by stepping through the frequencies of the three points of interest one step at a time. It is conceivable to do this repeatedly while progressing. However, in such spectrum processing, there is a possibility that peaks are created in addition to the peak spectrum, and in order to eliminate such undesired peaks, it is necessary to focus on frequency points other than the three frequencies mentioned above. It is necessary to determine the magnitude relationship between levels, which complicates signal processing and requires a lot of processing time.

In the present invention, in order to eliminate the above-mentioned drawbacks and increase the signal processing speed, the following steps are taken.

(1) Detect only the frequency address of the peak spectrum.

(2) Do not set the amplitude of the frequencies around the peak to 0.

(3) Peak level data is not handled at the peak spectrum detector stage.

In the present invention, this aspect of signal processing is performed by the peak spectrum detector 2 in the configuration shown in FIG.
2, and furthermore, in order to efficiently obtain the minimum data necessary for signal processing after the two-frequency dissonance strength calculation device 6 at point P in the configuration shown in the second figure, the following peak level sorting device is used. 23, information on the order of frequency addresses arranged in descending order of peak level is obtained.

Next, for a normal transmission system without linear distortion, from the additive nature of the dissonance intensity mentioned above, the output side dissonance intensity I _p can be decomposed as I _p = I _i + I _d . Therefore, the dissonance intensity I _d of the nonlinear distortion can be determined as I _p −I _i =I _d . The dissonance intensity I _d of this nonlinear distortion is further decomposed into the dissonance intensity caused by the combination of the input sine wave signal component and the distortion component, and the dissonance intensity caused by the combination of the distortion components. I can do it. Therefore, these dissonance strengths can be found as the sum of the dissonance strengths of the combined sounds of two sine waves, so
By extracting and calculating only the combination of spectral data necessary for calculating the dissonance intensity, duplication of dissonance intensity meter calculations can be avoided and signal processing time can be significantly shortened.

Therefore, the information necessary to extract the spectral data combination is the input signal component included in the output signal of the transmission system, and the input signal component can be identified by (a) a method using a known input frequency; b) A method of determining based on the number of input signal components where the input signal component>distortion component, and (c) a method of determining the input signal>>distortion component using a threshold that can separate the two.

When the above-mentioned conditions (b) or (c) are satisfied, if the spectra are sorted in order of level magnitude by the peak level sorting device 23 in the configuration shown in FIG. The first few line spectral amplitudes that are lined up are necessarily input signal components, and extraction of combination data of those input signal components and distortion components is based on the input signal components and other line spectral amplitudes. This can be done by mechanically combining these in sequence. In addition, if conditions (b) or (c) do not hold, if the input signal components are identified using the method (a) and then the input signal components and distortion components are sorted individually, the above-mentioned method can be applied. Similarly, all the minimum required combination data can be mechanically specified.

In actual measurement of distortion component dissonance strength,
Even when the amplitude-frequency characteristics of the transmission system are substantially flat, that is, even when there is some linear distortion, large measurement errors will not occur by applying the various measurement methods described above. In addition, if the magnitude of nonlinear distortion is 40 dB or less of the input signal component, the sum of dissonance strengths due to the combination of distortion components will be smaller than the sum of dissonance strengths due to the combination of input signal components and distortion components. Since it is often much smaller, omitting it will not cause a large measurement error.

Finally, FIGS. 5 and 6 show examples of the configuration of the dissonance measuring device used in the psychological quality measuring method according to the present invention, which is constructed by integrating the details described above, as the psychological quality measuring device and the dissonance degree measuring device according to the present invention, respectively. Each will be shown and explained.

In the configuration example shown in FIG. 5 corresponding to the above-mentioned psychological quality measuring device, the input signal and output signal of the transmission system to be measured are switched by a switching device 31 and supplied to the spectrum analysis processing device 32 having the configuration shown in FIG. A two-frequency dissonance intensity calculation device 6 with the configuration shown in FIG.
The dissonance intensities I _i and _{I p} on the input side and output side of the transmission system, which _are calculated based on
The dissonance intensity increment due to the transmission system itself is calculated by finding the difference between I _p and I _i , and the dissonance intensity increment is transferred to the converter 3.
4 to obtain the degree of dissonance due to the transmission system itself, and display it on the display device 9.

On the other hand, the 6th instrument corresponding to the psychological quality measuring instrument mentioned above
In the illustrated configuration example, the output terminal 4 of the transmission system under test
1 is supplied to the spectrum analysis processing device 32 having _the configuration shown in FIG _. The combination of spectral data necessary for calculating the consonance strength is extracted, and the spectral data obtained as a result of the extraction is used to calculate the dissonance strength increment I _p −I _i due to the transmission system itself. Display device 9 displays the degree of dissonance due to the transmission system itself obtained in the same manner as
to be displayed.

As is clear from the above description, according to the present invention, the following remarkable effects can be obtained.

(1) The psychological quality of the transmission system can be automatically measured at high speed.

(2) By extending the conventional consonance theory and introducing a masking device and a device that cuts below the minimum audible level, it is possible to measure the psychological quality of distortion and noise in the low sound pressure level range, which is about the detection limit level. .

(3) Psychological quality and physical quality can be determined simultaneously for the same test signal, and the direct correspondence between the two can be known.

(4) By using a known program sound model signal as a test signal, the following effects can be obtained.

(i) It is possible to measure the psychological quality of the transmission system itself by observing only the output data of the transmission system, and therefore it is possible to realize remote location measurement and speed up measurement by eliminating redundant processing.

(ii) A psychological method that takes a probabilistically weighted average of the dynamic quality of the transmission system by setting a single level, without resetting the level of the test signal as a parameter as in conventional distortion measurement methods. You can expect quality.

(5) It is also possible to measure the psychological quality of a transmission system that does not directly input or output sound waves relative to the reference sound pressure.

(6) The following effects can be obtained with respect to each part of the dissonance measuring device.

(i) By inserting a peak spectrum detection device and a level sorting device as spectrum processing devices after the spectrum analyzer, the supply of unnecessary data to the subsequent dissonance intensity measurement device is eliminated and the data processing speed is increased. It can be realized. Further, by using these devices in combination, preparation for the next stage processing can be completed only by the order information extraction processing regarding the signal level of the frequency address, and high-speed line spectralization can be easily realized.

(ii) the introduction of sub-audible level truncation and masking devices to better approximate the auditory characteristics and enable the realization of the various effects mentioned above, and to eliminate the supply of unnecessary data to subsequent processing units; This also enables faster measurement.

(7) By defining the psychological quality of the transmission system itself based on the difference in dissonance strength between the input and output of the transmission system, even if noise is added to the test signal, the quality can be measured by largely canceling out the effect of noise. can be performed stably, and similar effects can be obtained even when the spectral configuration of the test signal is changed.

(8) If a spectrum analyzer having a spectrum averaging device and a subtracting device for the internal noise spectrum of the spectrum analyzer is used, a dissonance measuring device with excellent stability and accuracy can be realized.

[Brief explanation of drawings]

Fig. 1 is a flowchart showing a conventional dissonance calculation mode, Fig. 2 is a block diagram showing an example of the configuration of the dissonance measuring device of the present invention, and Fig. 3 a and b are dissonance measurement of the present invention. FIG. 4 is a block diagram showing an example of the configuration of the spectrum processing device according to the present invention, and FIG. 5 is a block diagram showing an example of the configuration of the psychological quality measuring device according to the present invention. 6 are block diagrams similarly showing other configuration examples. 1...signal source, 1'...program sound model signal source,
2...Spectrum analyzer, 4...Truncation device below minimum audible level, 5...Masking device, 6...
Two-frequency dissonance strength calculation device, 7... Addition device, 8
... Dissonance intensity dissonance degree converter, 9 ... Dissonance degree display device, 11 ... Transmission system to be measured, 12 ... Changeover switch, 13, 13' ... Psychological quality measuring device, 2
DESCRIPTION OF SYMBOLS 1...Spectrum processing device, 22...Peak spectrum detector, 23...Peak level sorting device, 31...Switching device, 32...Spectrum analysis processing device, 33...I _p -I _i calculation device, 34... ... Dissonance intensity dissonance degree converter, 35 ... Psychological quality display device, 41 ... Transmission system output to be measured, 42 ... I _p -
I _i data combination device for calculation.

Claims (1)

[Scope of Claims] 1 Dissonance strengths corresponding to input and output signals of the transmission system under test are measured respectively, and the difference between these dissonance strengths is converted according to a power law to determine the degree of dissonance. A dissonance level psychological quality measurement method characterized by expressing the psychological quality of. 2. Rearranging the spectral data of the output signal of the transmission system under test so that it is arranged in ascending order of level, and determining the difference in dissonance intensity by considering the first part of the spectral data as the input signal of the transmission system under test. A dissonance degree psychological quality measuring method according to claim 1, characterized in that: 3. In a dissonance degree measuring device comprising at least a dissonance intensity calculation means, a dissonance intensity-dissonance degree conversion means, and a dissonance degree display means, prior to the dissonance intensity calculation means, the dissonance level is truncated below the minimum audible level. A dissonance measuring device characterized by comprising a means and a masking means.