JP4819507B2 - Apparatus sound quality evaluation method and apparatus sound quality evaluation apparatus - Google Patents

Description

  The present invention relates to a sound quality evaluation method for an apparatus and a sound quality evaluation apparatus for the apparatus, and more particularly to a method and apparatus for evaluating sound generated by an apparatus such as an image forming apparatus.

  As an apparatus, in an office or the like, various apparatuses equipped with an image forming apparatus such as a copying machine, a printer or a facsimile, and a multifunction machine of these are installed. In this type of image forming apparatus, many parts are mechanically connected. In addition, the image forming apparatus has a motor for driving these mechanisms and the like, and during the image forming operation, an operation sound of each part of the apparatus is generated, and this operation sound causes discomfort to the user. It can be a problem. In the past, these devices only had to be convenient. However, as the environment in the office has improved, there has been an increasing demand for solving noise problems for OA devices. Therefore, the noise reduction of the image forming apparatus has been promoted, and a considerable noise reduction has been achieved as compared with before.

  Currently, an image forming apparatus generally uses a sound power level or a sound pressure level (ISO 7779) as a method for evaluating noise. However, since these are values of acoustic energy generated from office equipment such as copying machines, printers, and multifunction machines, there is a case where the correlation with human subjective discomfort with noise is not so good.

  For example, when sounds with the same sound pressure level (equivalent noise level Leq: energy averaged over the entire measurement time) are compared and heard, there is a difference in discomfort due to the difference in sound frequency distribution and the presence or absence of impact sound. There may be. Moreover, even if the value of the sound pressure level is small, it may be uncomfortable if a high frequency component, a pure sound component, or the like is included.

  Therefore, in order to improve the office environment, it is necessary not only to evaluate and reduce the sound power level and sound pressure level of OA equipment, but also to simultaneously evaluate and improve sound quality.

  In order to evaluate and improve sound quality, it is necessary to quantitatively measure the sound quality for grasping the current situation and to measure how much it has been improved before and after the improvement. However, since sound quality is not a physical quantity, quantitative measurement cannot be performed. Therefore, it is difficult to set a target value.

  In the case of human sound quality evaluation, the expression is qualitative, such as “sound quality improved a little” or “substantially improved”. Furthermore, due to individual differences, it may be difficult to determine whether the evaluation varies from person to person or whether the obtained results are generally available. If the sound quality is not quantitatively expressed by physical characteristics, it is impossible to objectively evaluate whether the measure was effective or how effective.

  For this reason, it is necessary to perform subjective evaluation experiments and perform statistical processing on the results to quantify sound quality. For subjective evaluation, a method of asking the subject about the impression of the test sound using the SD (seminar differential) method has been often performed. Specifically, for a large number of stimuli (test sounds), stage evaluation of the adjectives that are paired is performed, factor analysis is performed on the results from principal component analysis, and quantitative explanation of the stimulus with multiple factors I do. The adjectives used for subjective evaluation are also prepared for content related to discomfort, and the relationship between the obtained multiple factors and discomfort is obtained by multiple regression analysis. Furthermore, factor analysis is also performed on the relationship between the factor and the acoustic physical quantity of sound, and finally, discomfort is predicted by physical quantity, such as physical quantity → factor → discomfort.

  By the way, there is a psychoacoustic parameter as a physical quantity for evaluating sound quality. Non-patent document 1 describes typical psychoacoustic parameters as follows (the units are shown in parentheses).

・ Loudness (sone): loudness ・ sharpness (acum): relative distribution of high frequency components ・ tonality (tu): articulation, relative distribution of pure tone components ・ roughness: coarse sound Feeling, Fractation Strength (vacil): Fluctuation intensity, Beatiness In addition, devices that can measure psychoacoustic parameters such as impulsiveness (iu): impact have emerged.

  As the value of any parameter increases, discomfort tends to increase. Among them, only loudness is standardized by ISO532B. However, the basic concept and definition of the other parameters are the same, but since the programs and calculation methods differ depending on the original research by the measuring instrument manufacturer, the measured values are usually slightly different depending on the manufacturer. In addition, like impulsiveness, there are original parameters developed by measuring instrument manufacturers.

  Noise generated from image forming devices such as copiers, printers, and multifunctional machines is composed of many timbres due to the complexity of the mechanism. For example, low-frequency heavy sounds, high-frequency high-pitched sounds, Generated sound is generated while changing in time from a plurality of sound sources such as a motor, paper, and solenoid.

  Although humans judge these sounds comprehensively to determine whether they are uncomfortable, it is considered that they are determined by weighting which part of the sound is particularly related to discomfort. That is, there are psychoacoustic parameters that have a large influence on discomfort due to the timbre of the machine and psychoacoustic parameters that have a small influence. These psychoacoustic parameters and sound pressure levels are combined to explain factors and predict discomfort.

  As described above, the following techniques have been proposed as techniques for evaluating the sound quality of image forming apparatuses and other sound sources. Patent Document 1 provides a sound quality evaluation apparatus and a sound quality evaluation method capable of performing a comprehensive sound quality evaluation in consideration of the influence on the subjective sense of a person, such as noise generated from office equipment. When the digital signal of the sound to be evaluated is input to the sound quality index calculation unit, the waveform preprocessing unit performs preprocessing necessary for calculation of various acoustic indexes, and then the acoustic index calculation unit In addition to calculating the loudness index in the loudness index calculation unit, the high frequency pure tone index and the broadband noise index are considered to be high frequency in consideration of the effect on the subjective annoyance and discomfort of noise. Based on the magnitude index, high frequency pure tone index, and broadband noise index obtained by the pure tone index calculation unit and broadband noise index calculation unit, the total sound quality index calculation unit calculates the total sound quality index, and the calculation result display unit In It has described a Shimesuru.

  Patent Document 2 describes psychological discomfort by improving and making it possible to rationally evaluate unpleasant sound sources for image forming apparatuses from low speed machines to medium to high speed machines with values that are easy to understand and understand. Image discomfort probability value calculated by a predetermined formula using sound pressure level value, loudness value of psychoacoustic parameter, sharpness value, tonality value, impulsiveness value, and image formation satisfying a predetermined condition. By providing the apparatus, a sound quality evaluation formula capable of calculating a probability of discomfort of sound emitted from an image forming apparatus operating at a low speed to a high speed is derived, and the allowable value of the speed and discomfort of the image forming apparatus is calculated. What approximates the relationship is described.

In Patent Document 3, the frequency analyzer analyzes the physical characteristics of the sample music reproduced by the player, as an automatic music scoring that combines objective evaluation by a machine and subjective evaluation by a person. The input unit inputs a person's “good”-“bad” evaluation of the sample song and the result of a questionnaire conducted by the person to obtain a reference element for evaluating the sample song. Then, the CPU obtains the reference element from the questionnaire result, and the CPU obtains the relationship between the physical characteristics of the song, the evaluation of the person, and the evaluation element, and stores these relations in the storage unit to store the music. The automatic scoring device uses at least one of the physical characteristics of the song sung by the singer, the listener's evaluation of the song, and the elements of the listener's evaluation, and the relationship stored in the storage unit And song Who have been described what scoring the songs sang.
JP 2001-336975 A JP 2004-219976 A JP 2000-99014 A The Japan Society of Mechanical Engineers "7th Design Engineering and System Division Lecture" Aiming for an innovative leap in design and systems for the 21st century! "" November 10th and 11th, 1997 "Sound / vibration and design, color and design (1)" Section 089B "Sound quality evaluation system using dummy head"

  By the way, there are the following two problems in the method of calculating the relationship between impression and factor, factor and discomfort, factor and physical quantity step by step and predicting the discomfort from the physical quantity by finally connecting them. .

(1) The error of the entire model cannot be estimated.
Since the discomfort is estimated in multiple steps, such as the relationship between impression and factor, factor and physical quantity, factor and discomfort, each error such as the relation between factor and physical quantity, the relation between factor and discomfort Although it is possible to calculate the physical model, factors, and discomfort, the overall model error cannot be estimated because the overall evaluation is not performed. Therefore, the conventional method cannot determine whether the accuracy of the entire model is good or bad.

(2) The model is not properly interpreted.
That is, it is assumed that the model interpretation of FIGS. 15 and 17 is performed although the model interpretation of FIGS. 16 and 18 is performed. In the conventional models of FIGS. 16 and 18, discomfort cannot be controlled by physical quantities or impressions. This is because the direction of the arrow is not suitable for Y because of physical quantity or impression. (However, it is correct to infer Y from the factor.) Factor analysis is a method for finding the underlying true value lurking from the impressions of various adjectives. In other words, the factor is the cause of the causal relationship, and the result is an impression (adjective). This will find potential true values not only in the case of multiple adjective impressions on a stimulus, but also on factor analysis of multiple physical quantities on the stimulus.

Here, it is proved that FIG. 17 and FIG. 18 are different.
Now, the acoustic physical quantity data of 36 test sounds as shown in Table 1 is prepared. This is a sample sound actually used in the SD method experiment described later.

Using this data, try the factor analysis from the principal component analysis. Using statistical analysis software JMP (registered trademark), principal component analysis and factor analysis were mechanically performed with three principal components and three factors.
The results are shown in Table 2.
An image of the factor analysis result is shown in FIG.

  Here, we explain what happens when the factor analysis arrows are reversed. From FIG. 19, the true value of sound pressure level, loudness, and tonality is factor 1. Therefore, here, for example, the loudness and the first factor will be described as an example.

In relation to loudness and factor 1,
x⇒y, y⇒x
Compare the regression coefficients at.
In other words, the regression coefficient is compared from the scatter diagram of the relationship between the factor and loudness. FIGS. 21 and 22 are shown with the x and y axes reversed.
From FIG.
Factor 1 = 2.0311 × Loudness + 7.8969 (a)
From FIG.
Loudness = 0.38889 × factor−3.0713 (b)
Is obtained.
Equation (a) is converted to a loudness equation.
Loudness = 0.443234405 × factor-3.887991729 (c)
When the equation (a) is converted into (c) and compared with the equation (b), it can be seen that the analysis is different from the case where the factor is the objective variable and the loudness is the explanatory variable.

  Therefore, in the regression analysis, it is understood that the roles of the objective variable and the explanatory variable are reversed and the roles should not be reversed on the obtained formula. That is, if FIG. 18 is used as FIG. 17, an incorrect result is obtained.

  In order to solve the problems (1) and (2) above, a model is analyzed in one step, and a model in which the arrows are also directed in the directions of FIGS. 15 and 17 is provided. For that purpose, covariance structure analysis (Structural Equation Modeling: SEM) is used.

  By using the multi-index model of covariance structure analysis, the relationship that has been analyzed and analyzed in several steps can be quantified and obtained in one analysis, and the hypothesis logic as a whole is evaluated and statistically analyzed. It is possible to verify.

  That is, in the multi-index model of covariance structure analysis, a model in which a plurality of factor analysis and multiple regression analysis are interlaced can be collectively analyzed. If the factor analysis result is further subjected to regression analysis, errors accumulate and the accuracy of the entire analysis decreases, and errors in the entire model cannot be clarified.

  On the other hand, in the covariance structure analysis, the entire model can be analyzed at once, the estimation accuracy can be improved, and the degree of fit between the data and the model can be evaluated.

The present invention creates a discomfort model based on the effect in the factor space when moving a plurality of acoustic physical quantities, and predicts the discomfort accurately with respect to the problem of unpleasant sound as described above. The purpose is to clarify the discomfort from physical quantities.
However,

  In the first aspect of the invention, a plurality of types of sounds emitted when the device is operated are evaluated by the SD (Semantic Differential) method, the analysis result is subjected to factor analysis, and the acoustic physical quantities, factors, and discomfort of the sound are further analyzed. Covariance structure analysis is performed, and from the results, the discomfort model and sound quality discomfort are expressed as discomfort = a × impact factor + b × metallic factor (a and b are constants, 0 <a <1, 0 <b It is a sound quality evaluation method for a device characterized by evaluating discomfort from the acoustic physical quantity of the sound generated by the device obtained in <1).

  According to a second aspect of the present invention, in the sound quality evaluation method for an apparatus according to the first aspect, a and b in the formula are a = 0.51 and b = 0.32.

  According to a third aspect of the present invention, in the sound quality evaluation method of the apparatus of the second aspect, the impact factor and the metal factor of the formula are expressed as impact factor = 0.88 × standardized loudness + (− 0.16) × standardized sound. Pressure level + 0.28 × standardized impulsiveness metallic factor = 0.42 × standardized loudness + (− 0.32) × standardized sound pressure level + 0.21 × standardized sharpness + 0.61 × standardized tonality + 0.2 × standardized in Pulsiveness standardization: (individual data-data average value) / standard deviation.

  The invention of claim 4 is characterized in that an impression of discomfort is predicted from the distribution of the impact sensation factor of the sound generated by the device and the metallic factor from the discomfort model according to any one of claims 1 to 3. It is the sound quality evaluation method of the apparatus to perform.

  According to a fifth aspect of the present invention, in the sound quality evaluation method for a device according to any one of the first to fourth aspects, the device is an image forming apparatus.

  The invention according to claim 6 is a means for evaluating a plurality of types of sounds emitted when the apparatus is operated by an SD (semantic differential) method, a means for performing factor analysis on the evaluation results, and an acoustic physical quantity and factor of sound. The covariance structure analysis for the discomfort is obtained, and the discomfort model and the sound quality discomfort are expressed by the discomfort = a × impact factor + b × metallic factor (a and b are constants, 0 < a sound quality evaluation device for a device, characterized in that it evaluates discomfort from the acoustic physical quantity of the sound generated by the device, and includes means for obtaining a <1, 0 <b <1).

  The invention of claim 7 is the sound quality evaluation apparatus of the apparatus of claim 6, wherein a and b in the above equation are set to a = 0.51 and b = 0.32.

  According to an eighth aspect of the present invention, in the sound quality evaluation apparatus for the apparatus of the seventh aspect, the impact factor and the metal factor of the formula are expressed as impact factor = 0.88 × standardized loudness + (− 0.16) × standardized sound pressure. Level + 0.28 × standardized impulsiveness metallic factor = 0.42 × standardized loudness + (− 0.32) × standardized sound pressure level + 0.21 × standardized sharpness + 0.61 × standardized tonality + 0.2 × standardized impulsive Nest standardization: (individual data-data average value) / standard deviation.

  The invention of claim 9 is characterized by comprising means for predicting the impression of discomfort from the distribution of the impact factor and the metallic factor of the sound generated by the apparatus based on the discomfort model of claims 6 to 8. Is a sound quality evaluation device for the device.

  According to a tenth aspect of the present invention, in the sound quality evaluation apparatus according to any one of the sixth to ninth aspects, the apparatus is an image forming apparatus.

  According to the present invention, in the covariance structure analysis, the entire model can be analyzed at once, the estimation accuracy is increased, and furthermore, the degree of matching between the data and the model can be evaluated. Therefore, it is possible to create a discomfort model based on the effect in the factor space when moving multiple acoustic physical quantities and to accurately predict discomfort for the problem of unpleasant sound generated by the device. Become.

  Embodiments as the best mode for carrying out the present invention will be described below with reference to the drawings.

  Hereinafter, a sound quality evaluation method for an image forming apparatus as a sound quality evaluation method for an apparatus according to the present invention will be described with reference to the drawings.

A. Configuration of Image Forming Apparatus There are various forms of image forming apparatuses with different image forming speeds, structures, and apparatus sizes, and typical ones are shown below.
A-1. Image Forming Apparatus Having Multiple Modes FIG. 1 is an explanatory diagram showing an example of the overall configuration of an image forming apparatus (apparatus having multiple modes) according to an embodiment of the present invention. The present invention is a method for evaluating noise generated by such a general image forming apparatus. Prior to the description of the evaluation method, the configuration of the general image forming apparatus will be described.

  The image forming apparatus shown in FIG. 1 is a digital color printer that employs an electrophotographic system, and includes an optical unit 1, a photoreceptor unit 3, a developing unit 4, a transfer unit 5, a fixing unit 46, and a paper feeding unit. 110.

  At the time of image formation, an image formation target sheet (including printing paper, OHP sheet, etc., but hereinafter referred to as paper) accommodated in a paper feeding unit 110 disposed at the bottom of the image forming apparatus is shown from the lower right side of the figure. It is transported along a predetermined transport path that rises diagonally to the left. The sheet thus transported is fed out from the sheet feeding unit 110 and is conveyed along the oblique conveyance path from the lower right side to the upper left side of the drawing to the upper side of the sheet feeding unit 110. During this time, the paper is similarly passed between the four photosensitive units 3, the developing unit 4 and the transfer unit 5 arranged side by side along the conveyance path, and a predetermined image is transferred. The sheet on which the image has been transferred is conveyed to a fixing unit 46 disposed on the upper left side of the photosensitive unit 3, the developing unit 4, and the transfer unit 5, and the transferred image is fixed by the fixing unit 46.

  As shown in FIG. 2, the optical unit 1 is a unit that extends along a sheet conveyance path that is an oblique direction such as the lower right to the upper left direction of FIG. 2, and includes a housing 11 disposed along that direction. Have. Laser diodes (LD) 17 (Bk: black), 18 (C: cyan), 19 (M: magenta), and 20 (Y: yellow) for each of four colors are attached to the upper portion of the housing 11.

  Further, the housing 11 includes a polygon mirror motor 2 for main operation line operation, two-layer fθ lenses 21 and 22 for dot position correction, and long WTL lenses 23, 24, 25 for surface tilt correction. 26, a cylinder lens or the like for correcting a laser beam diameter (not shown) is attached.

  The polygon mirror motor 2 is integrally formed with two upper and lower polygon mirrors 27, and the polygon mirror 27 is irradiated with laser light emitted from the LDs 17, 18, 19, and 20.

  The LDs 17, 18, 19, and 20 corresponding to the respective colors emit light in accordance with the conveyance timing of the paper, and the light (indicated by a thick line in the figure) is a cylinder lens, a polygon mirror 27, two-layer fθ lenses 21 and 22, a long length. The photosensitive drum 28 of each color is irradiated via the WTL lenses 23, 24, 25, and 26.

  Note that it is preferable to adopt a two-beam type LD unit corresponding to black. In other words, by adopting a two-beam LD, two beams can be written simultaneously during monochrome image formation, and rapid writing can be performed while suppressing the number of rotations of the polygon mirror motor 2. By reducing the rotation speed of the polygon mirror motor 2 in this way, it is possible to obtain an effect of suppressing noise and an effect of extending the life of the motor. For example, when printing in the color mode, the rotation speed of the polygon mirror 27 is 29528 rpm (revolutions per minute) and the printing speed is 28 ppm (pages per minute). Despite being small, the printing speed is 38 ppm.

  Returning to FIG. 1, the structure of the photosensitive unit 3, the developing unit 4, and the transfer unit 5 in the image forming apparatus will be described. As shown in the figure, this image forming apparatus employs a four-drum tandem image forming method, and this method improves the printing speed in the full-color printing mode and the monochrome printing mode. . Further, as described above, the photoconductor unit 3, the developing unit 4 and the transfer unit 5 are arranged obliquely, thereby reducing the installation space, thereby reducing the size of the entire apparatus.

  The photosensitive unit 3 and the developing unit 4 are independent units for each color. That is, the photosensitive unit 3 and the developing unit 4 for magenta (M), the photosensitive unit 3 and the developing unit 4 for cyan (C), the photosensitive unit 3 and the developing unit 4 for yellow (Y), black (Bk ) Photosensitive unit 3 and developing unit 4, which are arranged in the above order from the lower right side to the upper left side of FIG. The M, C, and Y photoconductor units 3 except for Bk have exactly the same structure, so that any new unit (M, C, Y) may be used. Good.

  The transfer unit 5 is a unit extending along the oblique direction below the photosensitive unit 3 and the developing unit 4 arranged in the oblique direction in the above-described order, and is disposed along the oblique direction. Yes. The transfer unit 5 includes a plurality of rollers and an endless transfer belt 29 wound around the rollers. When the roller is rotated by a motor (not shown), the transfer belt 29 is rotated counterclockwise in the figure, and the sheet fed from the paper feeding unit 110 is placed on the transfer belt 29 and is moved from the lower right side to the upper left side of the figure. To the side. Further, a P sensor 6 is disposed on the downstream side in the transport direction of the transfer unit 5 (upper left side in the figure), and the P sensor 6 detects the density of the P sensor pattern formed on the transfer belt 29. The detection result is used for control.

  Here, sectional views of the photosensitive unit 3 and the developing unit 4 corresponding to the colors shown in FIG. 3 are shown. As shown in the figure, the photosensitive unit 3 has a photosensitive drum 28 (for example, φ30). The photosensitive drum 28 has a hollow cylindrical shape, and is rotated clockwise in the drawing by a driving mechanism described later.

  A charging roller 36 (for example, φ11) is disposed above the photosensitive drum 28. The surface of the charging roller 36 is disposed at a position separated from the surface of the photosensitive drum 28 by about 0.05 mm. The charging roller 36 is rotated in the opposite direction to the photosensitive drum 28, that is, counterclockwise in the drawing, and applies a uniform charge on the surface of the photosensitive drum 28.

  A cleaning brush 37 is disposed above the charging roller 36. A cleaning brush 39 and a counter blade 38 are disposed on the upper left side of the photosensitive drum 28 to clean the photosensitive drum 28.

  A waste toner collection coil 40 is disposed on the left side of the cleaning brush 39, and the waste toner collected by the waste toner collection coil 40 is conveyed to the waste toner bolt 16 shown in FIG. ing.

  Next, the developing unit 4 adopts a dry two-component magnetic brush developing system, and includes a developing roller 30, a developing doctor 31, a conveying screw left 32, a conveying screw right 33, a toner concentration sensor 34, A cartridge 35.

  Next, the drive mechanism of the photoreceptor unit 3 will be described with reference to FIG. The photoconductor unit 3 is provided for each color, and there are four units. There are three photoconductor units 3 for M, C, and Y (for color), and a photoconductor unit 3 for Bk. Are driven by separate drive mechanisms. That is, the color photoconductor unit 3 is driven by the color drum drive motor 41 as a drive source and the gears 43 and 44 and the joint 45 that transmit this drive force.

  On the other hand, the black photoconductor unit 3 is driven by another gear 44 and a joint 45 which use another black drum drive motor 42 as a drive source and transmit this drive force. Accordingly, during color mode printing, only the color drum drive motor 41 operates and the black drum drive motor 42 stops. On the other hand, during monochrome mode printing, only the black drum drive motor 42 operates and the color drum drive motor 41 stops. The color drum drive motor 41 and the black drum drive motor 42 are stepping motors.

  As shown in FIGS. 5 and 6, the fixing unit 46 adopts a belt fixing method, and the belt has a smaller heat capacity than the fixing roller. There are advantages over the method used, such as shortening the warm-up time and lowering the roller set temperature during standby.

  The fixing unit 46 heats and pressurizes the sheet on which the image has been transferred, and fixes the toner image on the sheet. The fixing unit 46 includes a fixing belt 13 and an oil application unit 47. In the oil application unit 47, the gel oozes out from the oil and is supplied from the application felt 48 to the application roller 49. A small amount of silicone oil is applied to the fixing belt 13 while the application roller 49 rotates.

  By thus applying oil to the fixing belt 13, the fixing belt 13 and the paper are easily peeled off. The application operation by the oil application unit 47 is performed every time one sheet is conveyed, and oil application is performed each time one sheet is conveyed by a mechanism having a solenoid or a spring (not shown). The unit 47 is driven and brought into contact with the fixing belt 13. On the other hand, when one sheet of paper passes, the oil application unit 47 is separated from the fixing belt 13 by the above mechanism.

  Further, as shown in FIG. 5, a cleaning roller 50 is provided on the upstream side of the fixing belt 13 in the sheet conveyance direction, and the cleaning roller 50 adsorbs dirt on the fixing belt 13, thereby performing belt cleaning. Made.

  The above is the configuration of the fixing unit 46, and the sheet that has passed through the fixing unit 46 is conveyed to the main body tray 510 shown in FIG.

  Next, the configuration of the paper feeding unit 110 will be described with reference to FIG. The paper feeding unit 110 has three trays, a first tray 9, a second tray 10, and a manual paper feeding tray 8. Each of these trays employs an FRR paper feeding system as a system for feeding out the paper stored in the tray. The feed mechanism using the FRR paper feed system is reversely rotated with respect to the paper feed roller that is driven to rotate in the paper feed direction in order to separate the paper fed from the paper bundle stacked in the paper feed tray one by one. It has a configuration in which a roller is in contact.

  Under this configuration, the reverse roller is applied with a weak torque in the opposite direction to the paper feed roller via the torque limiter, so that the reverse roller is in contact with the paper feed roller, or one sheet of paper is both In the state where it enters between the rollers, it rotates with the paper feeding roller, while in the state where it is separated from the paper feeding roller, or in the state where two or more sheets enter between the rollers, it rotates in the reverse direction. For this reason, when the multi-feed paper enters, the paper in contact with the reversing roller is returned to the downstream side in the paper feeding direction, and multi-feed is prevented.

The sheets stored in the first tray 9 are separated by the first sheet feeding unit 51 and sent out from the first tray 9. Then, the fed sheet is transported by the relay roller 53 and reaches the transport roller 55. Here, the sheet is conveyed toward the upper left registration roller 7 while being turned by the conveyance roller 55.
The conveyed sheet hits the resist roller 7 that is stopped, and thereby the skew of the sheet is corrected. Then, timing adjustment with the image forming process by the photoconductor unit 3 or the like is performed, a registration clutch (not shown) is connected at a predetermined timing, the registration roller 7 is driven, and the sheet is conveyed toward the transfer unit. Thereafter, the sheet is conveyed by the transfer belt 29 as described above, and is subjected to processing such as predetermined image transfer.

  The paper stored in the second tray 10 is transported toward the transport roller 55 by the second paper feed unit 52 and the relay roller 54, and thereafter the same as the paper stored in the first tray 9. It is. The paper set in the manual paper feed tray 8 is transported toward the registration roller 7 by the paper feed unit 56, and the subsequent processing is the same as the paper transport from the first tray 9.

Next, a configuration for driving the first paper feed unit 51 and the second paper feed unit that send out paper from the first tray 9 and the second tray 10 as described above will be described.
As shown in FIG. 8, both these units are driven by a single stepping motor 83, and driving force is transmitted to each unit via a first paper feed clutch 57 and a second paper feed clutch 58. Is called. That is, when the paper is sent out from the first tray 9, only the first paper feed clutch 57 is connected, and when the paper is sent out from the second tray 10, only the second paper supply clutch 58 is connected.

  As described above, this image forming apparatus has the color photoconductor unit 3 and the like, and can perform color printing as well as monochrome printing. More specifically, as shown in Table 1, the printer has four printing modes such as “monochrome mode”, “color mode I”, “color mode II”, and “OHP / thick paper mode”, which are operated by the user. When a mode is selected by operating a unit or the like, a control unit (operation control unit) of the image forming apparatus (not shown) controls each unit according to the selection and operates each unit in the operation mode. In this image forming apparatus, three types of image forming speeds (182.5 mm / s = 38 ppm (pages per minute), 125.0 mm / s = 28 ppm, 62.5 mm / s = depending on the mode selected by the user by the control unit). 14 ppm).

  That is, control is performed so that the rotational speeds of the motors such as the stepping motor 83, the black drum driving motor 42, and the color drum driving motor 41 are changed according to the selected operation mode. In this specification, “ppm” is the number of output sheets per minute of A4 landscape paper.

  Here, in the high-resolution “Color Mode II” and “OHP / Thick Paper Mode”, the printing speed (image forming speed) is 14 ppm, whereas in the “Monochrome Mode”, the printing speed is 38 ppm, which is nearly three times higher. There is a speed difference. In order to realize such a large speed difference with a single motor, this image forming apparatus employs a stepping motor as a drive source for a paper transport mechanism system and the like.

A-3. Configuration of Large (Console Type) Image Forming Apparatus FIG. 13 is an explanatory diagram showing a configuration example of an image forming apparatus (console type) according to an embodiment of the present invention. That is, the overall height is designed to be used by installing on the floor surface, and the whole is the upper part (ADF (automatic document feeder) 510, scanner 520, writing unit 530, image forming engine 540) 500, lower part (bank). A console type digital MFP (multi-function printer) composed of a paper feed unit 570) is shown. In other words, there may be provided a normal copy function, a printer function according to an instruction from a personal computer, and a facsimile function. This type of image forming apparatus is generally a high speed machine.

  The upper part 500 includes an optical unit in which optical elements (scanner 520 and writing unit 530) are housed in a casing, an image forming engine 540 positioned below the optical unit, and an ADF 510 disposed on the upper part of the casing.

  In FIG. 13, reference numeral 501 denotes a photosensitive drum as an image carrier on which an electrostatic latent image is formed, reference numeral 502 denotes a charging charger, reference numeral 503 denotes a developing unit, reference numeral 504 denotes a transfer / separation charger, reference numeral 505 denotes a cleaning unit, Reference numeral 506 is a fixing unit, reference numeral 507 is a registration roller, reference numeral 511 is a document table, reference numeral 512 is a contact glass, reference numeral 513 is an exposure lamp, reference numeral 514 is a first mirror, reference numeral 515 is a second mirror, and reference numeral 516 is a third mirror. , 517 is an imaging lens, 518 is a CCD, 519 is a mirror, and 590 is a caster with a lock function.

  That is, the scanner 520 includes a contact glass 512 on which a document is placed and a scanning optical system. The scanning optical system includes a first carriage on which an exposure lamp 513 and a first mirror 514 are mounted, a second carriage that holds the second mirror 515 and the third mirror 516, an imaging lens 517, and a CCD 518. Yes. A first carriage that is driven by a stepping motor and moves at a constant speed when reading an original is provided, and a second carriage that is driven at a speed half that of the first carriage.

  An original (not shown) on the contact glass 512 is optically scanned by the first carriage and the second carriage, and the reflected light obtained there is an exposure lamp 513, a first mirror 514, a second mirror 515, a third mirror. An image is formed on the CCD 519 via the mirror 516 and the imaging lens 517, and photoelectrically converted.

  The writing unit 530 includes a laser output unit, an fθ lens, a mirror (all not shown), and the like. The laser output unit is provided with a laser diode and a polygon mirror which are laser light sources.

  The image signal output from the image processing unit is converted into a laser beam having an intensity corresponding to the image signal by the writing unit 530, and is shaped into a light beam having a fixed shape by a collimating lens, an aperture, and a cylinder lens, and then converted into a polygon mirror. Irradiated and output. The laser beam output from the writing unit 530 is applied to the photosensitive drum 501 through the mirror 519. The laser beam that has passed through the fθ lens is applied to a beam sensor (not shown) that generates a main scanning synchronization detection signal arranged outside the image area.

  The ADF 510 conveys originals set on the original table 511 one by one to the contact glass 512, and discharges them after reading. That is, the document is set on the document table 511 and the width direction is aligned by the side guide. Documents on the document table 511 are fed one by one from the bottom document by a feed roller, and are fed onto the contact glass 512 by the transport belt 553. After the reading on the contact glass 512 is finished, the original is discharged onto the discharge tray by the transport belt and the discharge roller.

  The recording sheets stacked on the first tray 571, the second tray 572, the third tray 573, and the fourth tray 574 of the bank sheet feeding unit 570 are the first sheet feeder 575, the second sheet feeder 576, and the second sheet feeder 576, respectively. The paper is fed by a third paper feeding device 577 and a fourth paper feeding device 578 and further conveyed by a bank vertical conveyance unit 579 and a main body vertical conveyance unit 580. When the leading edge of the recording paper is detected by a registration sensor (not shown), the recording paper is conveyed for a predetermined time, and then temporarily stops at the nip portion of the registration roller 507 and enters a standby state.

The waiting recording paper is sent to the photosensitive drum 501 side in accordance with the timing of the image valid signal, and is brought into close contact with the photosensitive drum 501 by the transfer on of the transfer / separation charger 504, and the image is transferred. Further, the recording paper is separated from the photosensitive drum 501 by separation on from the photosensitive drum 501. The recording paper onto which the toner image has been transferred is conveyed by a conveying device, fixed by the heat and pressure action of the fixing unit 506 including a fixing roller and a pressure roller, and discharged to the outside by the paper discharge roller 581. . The image forming speed of the image forming apparatus according to the above embodiment is, for example, about 362 mm / s.

  As described above, in the image formation on the photosensitive drum 501, an electrostatic latent image is formed by irradiating the charge charged on the photosensitive drum 501 by the charging charger 502 with the laser beam, and the developing unit 503 forms the photosensitive drum 501. Form an image on top.

  When duplex printing is performed using the duplex unit 585, the recording sheet after fixing is guided to the duplex conveyance path 586 by the switching claw 528, passes through the feed roller 532 and the separation roller 533, and is accumulated on the duplex tray. The recording paper accumulated in the tray is brought into contact with the feed roller when the tray is raised, and is fed to the main body vertical conveyance unit 580 when the feed roller rotates, and is re-fed to the registration roller 507 and then back to the back surface. Is printed.

  When performing reverse paper discharge, the switching claw 567 guides the recording paper in the direction of the reverse special tray 564, and when the rear end of the recording paper passes the reverse detection sensor 568, the transport roller 569 is reversely rotated, and the paper discharge tray direction. And discharge the paper to a preset tray.

B. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention evaluates noise generated by the image forming apparatus as described above, and based on the evaluation, evaluates the sound quality of the image forming apparatus and identifies the unpleasant sound source for reducing the unpleasant feeling given to the person by the noise. The flow of deriving the sound quality evaluation formula is as follows.

(1) Collection of operation sound of image forming apparatus (2) Analysis of operation sound (3) Creation of test sound from collected operation sound (4) Acoustic physical quantity (psychoacoustic parameters, sound pressure level) of test sound Measurement (5) Adjective selection for SD method experiment (6) Collecting subjects, SD method experiment by presenting test sound (7) Principal component analysis of SD method experiment result (8) Factor analysis of SD method experiment result (9 ) Factor estimation by synthesis of variables (10) Covariance structure analysis (11) Comparison of predicted and measured values (12) Discomfort mapping

Details of the above process will be described below.
(1) Collection of operation sound of image forming apparatus The operation sound of the image forming apparatus was recorded using binaural (binaural) recording using a dummy head HMS (Head Measurement System) III manufactured by Head Acoustics. This is because by performing binaural recording and reproducing with dedicated headphones in this way, it can be reproduced as if a human actually heard the sound generated by the machine.

  Further, the image forming apparatus to be subject to subjective evaluation uses an image forming apparatus having a plurality of modes (image forming operations are performed at three types of image forming speeds) and an image forming apparatus having a single high image forming speed. In the two models, operation sounds at four image forming speeds were collected and measured. The image forming speed of the image forming apparatus having a plurality of modes is 14 ppm, 28 ppm, 38 ppm, and the image forming speed of the high speed image forming apparatus is 65 ppm. That is, presentation sounds from the low speed range to the high speed range are prepared. The image forming apparatus having a plurality of modes is a color image forming apparatus, and an apparatus having a specification of a color mode at 14 ppm and 28 ppm and a monochrome mode at 38 ppm was used.

The measurement conditions are as follows (see FIG. 9).
・ Recording environment ・ ・ ・ Semi-anechoic room ・ Dummy head 203 ear position (sound collecting position) 204 ・ ・ ・ 1.2m height
Horizontal distance 1 m (1 ± 0.03) from the end face of the device 201 to be measured, the width direction is the center position of the device and the recording direction.・ Recording mode ... FF (Free field: for anechoic room)
・ HP filter 22Hz

The reason why the height of the dummy head is 1.2 m is that, as a recent method of using the image forming apparatus, it is considered that there are many cases in which an instruction for image formation is issued from a personal computer or the like while the user is seated. Is. Of course, a dummy head may be installed at a height of 1.5 m 2 in consideration of a state where a human is standing.

  By the way, the sound emitted by the image forming apparatus is usually different for each direction. This is because the positions of various motors, the sheet conveyance path, and the position of the paper discharge port are not located at the center of the apparatus but are distributed. Therefore, the sound produced by a certain sound source (such as a motor) can be heard well on the right side, but the sound collected for each direction is different, such as not being heard well on the left side. The sample sound used for the experiment described below may be collected in any direction, but when performing the sound quality evaluation experiment, it is better to unify the sample sound in any one direction. Therefore, in this experiment, it is considered that the user has the most chances to hear on the front side, but the rear side of the image forming apparatus is usually installed along the wall, so there is little opportunity to hear the sound on the rear side. Therefore, it was decided to use what was collected on the front side as the test sound.

(2) Analysis of operation sound Next, the operation sound of the image forming apparatus collected as described above was analyzed.
First, when an image forming apparatus having a plurality of modes was analyzed for noise when operating at a color of 28 ppm, that is, at a printing speed of 28 ppm, an analysis result as shown in FIG. 10 was obtained. The upper side of FIG. 10 represents the collected sound on the time axis, and the lower side of FIG. 10 represents the collected sound on the frequency axis. From this result, seven main sound sources were extracted. First, the fixing oil application impact sound of the fixing unit 46 was extracted on the time axis. On the frequency axis, color development drive system sound, paper feed stepping motor sound, charging sound, drum drive stepping motor sound, polygon mirror motor sound, and paper sliding sound were extracted.

  Next, when the noise was analyzed when the color was 14 ppm, that is, when the printing speed was 14 ppm, an analysis result as shown in FIG. 11 was obtained. The upper side of FIG. 11 represents the collected sound on the time axis, and the lower side of FIG. 11 represents the collected sound on the frequency axis. From this result, as the main sound source, the fixing oil application impact sound is extracted on the time axis, and the feeding stepping motor sound, charging sound, drum drive motor sound, polygon mirror motor sound, paper sliding sound are extracted on the frequency axis. Extracted.

  Next, when the noise at the time of monochrome 38 ppm, that is, when operating at a printing speed of 38 ppm, was analyzed, an analysis result as shown in FIG. 12 was obtained. The upper side of FIG. 12 represents the collected sound on the time axis, and the lower side of FIG. 12 represents the collected sound on the frequency axis. From this result, as a main sound source, fixing oil application impact sound was extracted on the time axis, and development drive system sound, charging sound, drum driving stepping motor sound, and paper sliding sound were extracted on the frequency axis.

  Next, when the noise when the high-speed image forming apparatus (65 ppm) was operated was analyzed, an analysis result as shown in FIG. 14 was obtained. The upper side of FIG. 14 represents the collected sound on the time axis, and the lower side of FIG. 14 represents the collected sound on the frequency axis. From these results, paper impact sounds and metal (clutch / solenoid) impact sounds are extracted on the time axis as main sound sources, and paper impact sounds, bank motor sounds, development motor sounds, main motor sounds, paper on the frequency axis. The sliding sound was extracted.

(3) Creation of test sound from collected operation sound Next, as described above, the sound collected at the position on the front side of the device was collected using “ArtemiS” which is sound quality analysis software manufactured by Head Acoustics. The sound was processed. As a sound processing method performed in this experiment, a sound during one cycle of the printing operation was cut out from the collected original sound. Then, among the sounds in one cycle period, a filter process was performed on the frequency source or the time axis on the part related to the main sound source extracted as described above, and a process of attenuating or enhancing these parts was performed. That is, three levels of sound (emphasis / original sound / attenuation) were created for the sound of the sound source extracted in one mode. When presenting the prepared test sound to the subject, it is difficult to make a judgment because it is too short if one cycle of the sound is heard, so it is presented as a sound of several cycles.

  As described above, the sounds collected on the front, rear, left and right sides of the image forming apparatus are different from each other, but on the front side created in this experiment rather than the range of psychoacoustic parameter values obtained from such four-direction sounds. It has been confirmed that the range of psychoacoustic parameter values obtained from the three test sounds obtained by emphasizing and attenuating the collected sounds is wider. In other words, a subjective evaluation experiment using three sounds obtained by emphasizing and attenuating the sound collected on the front side as in this experiment covers the characteristics of the sound obtained from the sound collected in four directions. it can.

  Based on the sound collected on the front side as described above, if three levels of sound (emphasis, original sound, attenuation) are created from the sound emitted by the main sound source extracted for each image forming apparatus, the sound generated during each image formation Nine sounds were created based on the L9 orthogonal table with different levels of sound sources extracted for.

  First, an image forming apparatus having a plurality of modes will be described. Table 4, Table 5, Table 6, and Table 7 are descriptions of each mode of the image forming apparatus having a plurality of modes. Here, Table 4 shows three-level sounds created for the main sound sources (seven) extracted from sounds collected when the image forming apparatus having a plurality of modes operates at a color of 28 ppm, that is, at a printing speed of 28 ppm. The result of having created nine test sounds by allocating based on the L9 direct table is shown. In the table below, the test sounds are distinguished from each other by adding a number with a circle. By assigning to the orthogonal table in this way, there is no correlation between the level changes of the sound source, and therefore analysis can be performed while ignoring changes of other sound sources.

  In the above table (the same applies to the following tables), “−1” is a sound created by attenuation until the sound is almost inaudible, “0” is a sound at the level of the original sound, and “1” is the original sound. The sound is emphasized until the difference in level is clearly seen. For example, the test sound “color 28 ppm9” in Table 2 is marked with “0” for all sound sources, indicating that all remain as the original sound.

  Next, Table 5 shows the three-level sounds created for the main sound sources (six) extracted from the sounds collected when operating at 14 ppm color, that is, at a printing speed of 14 ppm, based on the L9 orthogonal table. Results are shown. However, since the charging sound, the drum driving stepping motor sound, and the polygon mirror motor sound are sounds of the same tonality component, each test sound was set to the same level. The paper feed stepping motor sound is also a tonality component, but since this is an intermittently generated sound, it was decided to swing the level separately from the motor sound.

  Table 6 shows the result of allocating three levels of sounds created for the main sound sources (five) extracted from the sound collected when operating at 38 ppm monochrome, that is, at a printing speed of 38 ppm, based on the L9 orthogonal table. Indicates. However, since the charging sound and the drum driving stepping motor sound have the same tonality component, each test sound has the same level.

  Similarly, similar operations were performed on image forming apparatuses at other speeds. Table 7 shows the result of allocating three levels of sounds created for the main sound sources (six) extracted from the sound collected when operating at 65 ppm monochrome based on the L9 orthogonal table.

  The above is the details of the process of creating a test sound from the collected operating sound and creating an experiment.

(4) Measurement of acoustic physical quantities (psychoacoustic parameters, sound pressure level) of the test sound Next, the test sound created as described above was psychologically analyzed using the sound quality analysis software “ArtemiS” manufactured by Head Acoustics. Acoustic parameters were determined. In this sound quality analysis software, various settings can be selected when obtaining psychoacoustic parameters. In this experiment, default settings were adopted.

  For example, for loudness, “FFT / ISO0532”, “Filter / ISO0532”, and “FFT / HEAD” can be selected as “Calculation method”, but the default “FFT / ISO0532” is adopted, and “Spectrum Size” is the default. No. “4096”. Regarding the sharpness, the default “FFT / ISO532” was adopted for the “Calculation method”, and the default “Aures” of “Aures” and “von Bismark” was adopted for the “Sharpness method”. “Spectrum Size” was performed with the default “4096”. Other psychoacoustic parameters correlate with loudness and automatically change depending on the loudness setting.

  Using the sound quality analysis software set as described above, the psychoacoustic parameter values of the test sound created in the process of (3) above were obtained. The results are shown in Table 8. In addition, the result of Table 8 is a result of rounding off the PPM value to the first decimal place and the other values to the second decimal place.

(5) Adjective selection for SD method experiment The hypothesis of the discomfort model of the operation sound of the image forming apparatus was made, and the assumed factors and adjectives expressing the factors were picked up. Three factors were assumed as sound quality evaluation factors of the operation sound of the image forming apparatus. The right notation of each factor is a psychoacoustic parameter used as a hint.
(I) Factor 1 (impact factor): loudness, impulsiveness (ii) Factor 2 (metallic factor): Tonality, (sharpness)
(Iii) Factor 3 (rubbing factor): Sharpness Furthermore, a hypothesis that there is a factor “total discomfort” as a comprehensive evaluation from the above three factors was established. The concept is shown in FIG.

  While listening to the printer sound used for the actual subjective evaluation, we selected adjective candidates that are considered to represent each factor. The relationship between adjective candidates and each factor is as shown in FIG. In FIG. 20, the unpleasant impression factor is “true value”, which is a potential variable that cannot be directly observed. Ask the subject to answer the discomfort impression with an adjective. This adjective is an “observable variable” that can be directly observed, but “error” accompanies the observation. An external effect that cannot be explained only by factors (true values) that are common components of adjectives (observable variables) is “error”, and errors cannot be observed directly.

The adjectives were finally narrowed down to the following 16 words.
・ Intense ・ Metallic ・ Large ・ Stuffed ・ Kicking ・ Rubbing ・ Rubbing ・ Irritation ・ Impact ・ Early ・ Powerful ・ Sharp ・ Sharing ・ Noisy ・ Swelling ・ Unpleasant Irritation and unpleasant were assumed to be adjectives that evaluate overall discomfort.

(6) SD method experiment by collecting subjects and presenting the test sound The SD (semiferential differential) method is generally a bipolar method using adjective pairs, but there are many adjectives that are difficult to form pairs, so this SD method Then, the unipolar method was adopted.
In the monopolar method, a test sound is presented to a subject, and the adjectives of 16 words are evaluated on a five-point scale based on the impression they hear. For example, the adjective “Intense” was rated as close to 5 as the impression I heard was intense, and if it was not so intense, it was evaluated as close to 1.

41 subjects gathered. The score data obtained from the experiment was monitored using a scatter diagram matrix, residual plot, etc., and outliers were excluded. As a result, it was decided to use 40 subjects excluding one person's data. The total number of score data used in the analysis is as follows.
Number of score data: 40 subjects x 36 sounds x 16 adjectives = 23040

(7) Principal component analysis of SD method experiment results Principal component analysis can be performed using various commercially available spreadsheet software and statistical analysis software. For example, statistical analysis software “JMP (registered trademark of SAS Institute Inc)” or “SPSS (registered trademark of SPSS Inc)” can be used. In the present invention, JMP was used.

  We analyzed adjectives other than the general discomfort index, “Irritation” and “Uncomfortable”. In order to remove individual errors, analysis was performed using average data of subjects.

Using statistical analysis software “JMP”, the relationship between 14 adjectives was examined using a scatter diagram and a correlation coefficient matrix. FIG. 23 shows a correlation diagram matrix, and Table 9 shows a correlation coefficient matrix.
Strong, large, powerful, and noisy have a strong positive correlation with each other, and when intense, there is a positive correlation with sharp, noisy, shocking, sharp, noisy, and swell.
There is a strong positive correlation between metallic, kinky, and high, and these adjectives have a positive correlation with sharpness.
There is a weak negative correlation with undulations.
Other relationships are close to uncorrelated.
Also, the plot does not seem to have a large outlier.

  Here, principal component analysis was performed from the correlation matrix. Table 10 shows the results. (Displays up to the fifth principal component)

  From Table 10, the main component is interpreted. The eigenvalue of the first principal component is about 7.7, which is the largest. The contribution rate is about 55%. It is considered to be a component that expresses impact and sound volume because it contains eigenvectors that are intense, jerky, shocking, powerful, sharp, rubbing, and noisy little by little.

  The eigenvalue of the second principal component is about 4.2. The contribution rate is about 30%. It is considered to be a metal component because there is a little bit of eigenvector, which is metallic, kinky, and filled, and there is a negative eigenvector.

  The eigenvalue of the third principal component is about 1.1. The contribution rate is about 8%. From the size of the eigenvector, it is considered to be a component of a rubbing feeling. Since the eigenvalue below the fourth principal component is 1 or less and the contribution rate is about 4% or less, it is treated as a residual.

(8) Factor analysis of the SD method experimental results When interpreting the principal component analysis, it is difficult to attach a smart catchphrase to the conflicting concept. The main component is rotated using a method called varimax rotation. The new component obtained by the rotation is called a factor, and the individual ability-variate classification is derived to simplify the structure. In other words, it is possible to eliminate the concept of conflict by rotating the principal component, and the interpretation of the factors becomes easy. Such a method is called factor analysis.
Here, factor analysis is performed based on principal component analysis. Factor analysis was performed by setting up to 3 main components to be rotated.
First, as a result of factor analysis, Table 11 shows a list of factor loadings (varimax rotation).

  When the cumulative contribution rate up to the third component in Table 9 of the principal component and the cumulative contribution rate up to the third factor in Table 11 are compared, they agree at 92.31%. This means that the amount of information does not change with rotation.

  The cumulative contribution ratio of the third factor is 90% or more, and it is considered that the sound quality of the printer can be explained by three factors from the result of the factor analysis. Therefore, the number of factors is assumed to be 3.

  In FIG. 23, an adjective having a low factor loading and an adjective having a high factor loading but straddling a plurality of components are determined to be adjectives that are inappropriate for observation of the current impression and are excluded. Therefore, the adjectives “mom”, “sharp”, “noisy”, and “swelled” in the shaded part of Table 11 were excluded.

(9) Factor estimation by composition of variables Since each adjective contains an error, in order to analyze the structure between factors as a true value, it is necessary to estimate the factor from the adjective. Therefore, impression factors were estimated by defining the adjectives of each group obtained by partial correlation analysis by synthetic variability using factor analysis, and each factor was defined.

  FIG. 24 is an impact sensation factor, FIG. 25 is a metallic factor, FIG. 26 is a rubbing sensation factor, and FIG.

(10) Covariance structure analysis The estimated concept was verified. Covariance structure analysis using SPSS software Amos (structural equation modeling software) was performed. As a result of trial and error, the presence of the rubbing factor is expected to reduce the accuracy of the entire model, so the model was corrected as a two-factor model. In addition, the adjectives “gacha-gacha”, “impact”, and “kinkin” are excluded because they are not suitable for the whole model, and the accuracy of the reconstructed model is verified.

  FIG. 28 shows a reconstructed model. “E” in FIG. 28 is an error. The path coefficient is indicated by a standard partial regression coefficient, and the larger the absolute value, the greater the relation. Some pass coefficients have a negative sign, but this is considered to occur as an adjustment factor action because some psychoacoustic parameters are highly correlated with each other. For example, since the loudness and the sound pressure level are highly correlated with each other, it can be considered that the sound pressure level works as an adjustment factor for the influence of the two parameters on the impact factor.

  GFI and AGFI were used as a standard for the fitness of the entire model. The accuracy of fitting of the entire model obtained is 0.942 in AGFI, which is a highly accurate model. Therefore, the causal model of printer sound was estimated as a two-factor model. “Impact factor” and “metallic factor” were extracted as the main impression factors.

GFI (Goodness of Fit Index; fitness index)
AGFI (Adjusted Goodness of Fit Index; modified fitness index)
The closer the value is to 1, the better the fit to the data. The AGFI can be regarded as a value adjusted for the degree of freedom, and “GFI ≧ AGFI”, which is a stricter index than the GFI.

(11) Comparison between predicted value and actual value Here, based on the model of FIG. 28, calculation of total discomfort is performed.
The acoustic physical quantity is not a measured value itself but a standardized value.
Standardization is an operation for viewing data with different average values and standard deviations on the same ring. Specifically, (individual data-average value) / standard deviation is calculated.
From the model,
Total discomfort = 0.51 x impact factor + 0.32 x metallic factor (1)
Impact factor = 0.88 × standardized loudness + (− 0.16) × standardized sound pressure level + 0.28 × standardized impulsiveness metallic factor = 0.42 × standardized loudness + (− 0.32) × standardized sound Pressure level + 0.21 x standardized sharpness + 0.61 x standardized tonality + 0.2 x standardized impulsiveness

The discomfort calculated by the model is compared with the discomfort of the result of the SD method experiment by the subject. Here, a comparison is made for “uncomfortable”.
"Uncomfortable" is more than a model,
0.9 x total discomfort.
Since this is a standardized value, it is necessary to restore the original value in order to compare with the actual measurement value of the SD method.
“Uncomfortable” predicted by this model is the inverse transformation of standardization.
Therefore, the average value of “unpleasant” + standardized “unpleasant” × standard deviation is obtained.

Tables 12, 13, and 14 summarize these calculation results in a table.
Table 12 shows the results of standardizing the acoustic physical quantities, Table 13 shows the results of calculating each factor and discomfort using the standardized results, and Table 14 summarizes the average values and standard deviations.

FIG. 29 is a scatter diagram of the predicted values based on the model and the actually measured values based on the SD method.
The slope of the scatter diagram is 1, and the contribution rate R2 = 0.73. As a result of the SD method, the accuracy is higher. Therefore, the discomfort of the sound can be calculated from the acoustic physical quantity by the equation (1).

(12) Discomfort mapping Furthermore, according to the model obtained this time, if the acoustic physical quantity of the printer sound is known, it is possible to map an unpleasant impression using two factors (impact factor, metallic factor). It is.

  FIG. 30 is an unpleasant impression map. The arrow is the direction of discomfort. According to the model, the coefficient for the total discomfort factor is 0.32 for the metallic factor and 0.51 for the impact factor, so that the impact factor contributes to the discomfort at a ratio of about 3: 5.

  Depending on which region of the graph is plotted, it can be determined what factors are uncomfortable or not uncomfortable. Since the origin of the graph is an average value, the area shown in the figure, which is drawn so that the line perpendicular to the unpleasant arrow line intersects the origin, is an area that is less unpleasant than the average. It is desirable to bring the operating sound of the equipment to that area. Designing to bring it to that area can alleviate discomfort.

  In the above example, the image forming apparatus is described as an example of the apparatus. However, the apparatus may be an apparatus such as another OA device.

It is sectional drawing which shows the image forming apparatus used as the sound quality evaluation object which concerns on embodiment. FIG. 2 is a cross-sectional view illustrating a configuration of an optical unit of the image forming apparatus illustrated in FIG. 1. FIG. 2 is a cross-sectional view illustrating a configuration of a developing unit of the image forming apparatus illustrated in FIG. 1. FIG. 2 is a perspective view illustrating a driving mechanism of a photosensitive unit of the image forming apparatus illustrated in FIG. 1. FIG. 2 is a perspective view showing a fixing unit of the image forming apparatus shown in FIG. 1. FIG. 2 is a cross-sectional view illustrating a configuration of a fixing unit of the image forming apparatus illustrated in FIG. 1. FIG. 2 is a cross-sectional view illustrating a configuration of a sheet feeding unit of the image forming apparatus illustrated in FIG. 1. FIG. 2 is a perspective view illustrating a configuration of a sheet feeding unit of the image forming apparatus illustrated in FIG. 1. It is a figure which shows the conditions of the sound quality measurement of an image forming apparatus. 3 is a graph illustrating an example of a sound quality measurement result of the image forming apparatus illustrated in FIG. 1. 6 is a graph showing another example of the sound quality measurement result of the image forming apparatus shown in FIG. 1. 6 is a graph showing another example of the sound quality measurement result of the image forming apparatus shown in FIG. 1. It is sectional drawing which shows the other image forming apparatus used as the sound quality evaluation object which concerns on embodiment. It is a graph which shows the example of the sound quality measurement result of the image forming apparatus shown in FIG. It is a figure which shows the relationship between the impression of the model which estimates discomfort from a physical quantity, and a factor. It is a figure which shows the relationship between the impression of the model which estimates discomfort from a physical quantity, and a factor. It is a figure which shows the relationship between the impression of the model which estimates discomfort from a physical quantity, and a factor. It is a figure which shows the relationship between the impression of the model which estimates discomfort from a physical quantity, and a factor. It is a figure which shows the factor analysis result at the time of performing a factor analysis from a principal component analysis. It is a figure which shows the hypothesis that there exists the factor "total discomfort" as comprehensive evaluation from three factors. It is a graph for comparing a regression coefficient from the scatter diagram of the relationship between a factor and loudness. FIG. 22 is a graph for comparing regression coefficients from a scatter diagram showing the relationship between factors and loudness, in which the coordinates of factors and loudness in FIG. 21 are interchanged. Correlation diagram matrix showing relationships for 14 adjectives for principal component analysis It is a figure which shows the adjective of the shock sensation factor estimated using the factor analysis. It is a figure which shows the adjective of the metallic factor estimated using the factor analysis. It is a figure which shows the adjective of the rubbing sensation factor estimated using the factor analysis. It is a figure which shows the adjective of the explanatory drawing of the general discomfort estimated using the factor analysis. It is a figure which shows the model reconstructed as a 2 factor model. It is a scatter diagram of the predicted value by a model, and the measured value by SD method. It shows the relationship between the predicted value of the unpleasant impression and the measured value.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Optical unit 2 ... Polygon mirror motor 3 ... Photoconductor unit 4 ... Development unit 5 ... Transfer unit 6 ... P sensor 7 ... Registration roller 8 ... Paper feed Tray 9 ... first tray 10 ... second tray 11 ... housing 13 ... fixing belt 16 ... waste toner bolt 17 ... LD units 21, 22 ... two-layer fθ lens 23 , 24, 25, 26 ... lens 27 ... polygon mirror 28 ... photosensitive drum 29 ... transfer belt 30 ... developing roller 31 ... developing doctor 32 ... transport screw left 33 ... ..Conveying screw right 34 ... toner density sensor 35 ... cartridge 36 ... charging roller 37 ... cleaning brush 38 ... counter blade 39 ... cleaning brush 4 ... Waste toner collection coil 41 ... Color drum drive motor 42 ... Black drum drive motors 43, 44 ... Gear 44 ... Gear 45 ... Joint 46 ... Fixing unit 47 ... Oil application unit 48 ... application felt 49 ... application roller 50 ... cleaning roller 51 ... first paper supply unit 52 ... second paper supply unit 53 ... relay roller 54 ... relay Roller 55 ... Conveying roller 56 ... Paper feed unit 57 ... First paper feed clutch 58 ... Second paper feed clutch 83 ... Stepping motor 110 ... Paper feed unit 201 ... Covered Measuring device 202 ... operation unit 203 ... dummy head 204 ... position (sound collecting position)
500... Upper part 501... Photosensitive drum 502... Charger 503 .. Development unit 504... Separation charger 506. .... Document platen 512 ... Contact glass 513 ... Exposure lamp 514 ... First mirror 515 ... Second mirror 516 ... Third mirror 517 ... Imaging lens 518 ... CCD
519 ... Mirror 520 ... Scanner 528 ... Switching claw 530 ... Unit 532 ... Feed roller 533 ... Separation roller 540 ... Image forming engine 553 ... Conveying belt 567 ... Replacement claw 568... Reverse detection sensor 569... Conveyance roller 570... Bank feeding unit 571... First tray 572... Second tray 573. Tray 575... First paper feeding device 576... Second paper feeding device 577... Third paper feeding device 578... Fourth paper feeding device 579. Vertical transport unit 581... Discharge roller 585... Duplex unit 586.

Claims (10)

For the multiple types of sounds that are emitted when the device is activated, an evaluation is made by the SD (semantic differential) method,
Perform factor analysis on the evaluation results,
In addition, covariance structure analysis is performed on acoustic physical quantities, factors, and discomfort of sound.
From the results, the discomfort model and the discomfort of the sound quality are expressed. Discomfort = a × impact factor + b × metallic factor (a and b are constants, 0 <a <1, 0 <b <1)
In
A method for evaluating sound quality of a device, characterized by evaluating discomfort from an acoustic physical quantity of sound generated by the device. In the above formula, a and b are a = 0.51 and b = 0.32.
The sound quality evaluation method for an apparatus according to claim 1, wherein: The impact factor and the metal factor of the above formula are as follows: Impact factor = 0.88 × standardized loudness + (− 0.16) × standardized sound pressure level + 0.28 × standardized impulsiveness Metallic factor = 0.42 × Standardized loudness + (− 0.32) × standardized sound pressure level + 0.21 × standardized sharpness + 0.61 × standardized tonality + 0.2 × standardized impulsiveness Standardization: (individual data-data average value) / standard deviation The method for evaluating the sound quality of the apparatus according to claim 2. From the discomfort model according to any one of claims 1 to 3,
A method for evaluating sound quality of a device, wherein an impression of discomfort is predicted from a distribution of impact factors and metallic factors of sound generated by the device.   5. The sound quality evaluation method for an apparatus according to claim 1, wherein the apparatus is an image forming apparatus. Means for evaluating a plurality of types of sounds emitted when the apparatus is operated by an SD (seminar differential) method;
Means for performing factor analysis on the evaluation results;
In addition, a means to obtain a covariance structure analysis of acoustic physical quantities, factors, and discomfort of sound,
From the results, the discomfort model and the discomfort of the sound quality are expressed. Discomfort = a × impact factor + b × metallic factor (a and b are constants, 0 <a <1, 0 <b <1)
With the means to find in
An apparatus for evaluating sound quality of an apparatus, wherein discomfort is evaluated from an acoustic physical quantity of sound generated by the apparatus. In the above formula, a and b are a = 0.51 and b = 0.32.
The sound quality evaluation apparatus for an apparatus according to claim 6. The impact factor and the metal factor of the above formula are expressed as follows: Impact factor = 0.88 × standardized loudness + (− 0.16) × standardized sound pressure level + 0.28 × standardized impulsiveness Metallic factor = 0.42 × standardized Loudness + (− 0.32) × standardized sound pressure level + 0.21 × standardized sharpness + 0.61 × standardized tonality + 0.2 × standardized impulsiveness Standardization: (individual data-data average value) / standard deviation The sound quality evaluation apparatus for an apparatus according to claim 7. From the discomfort model according to any one of claims 6 to 8,
An apparatus for evaluating sound quality of a device, comprising means for predicting an impression of discomfort from a distribution of impact factors and metallic factors of sound generated by the device. 10. The sound quality evaluation apparatus according to claim 6, wherein the apparatus is an image forming apparatus.