JP2021064917A - Video signal conversion device and program - Google Patents

Abstract

To enable a signal level of human skin not to be lowered, a white level not to be changed, white blown highlights to be suppressed, and the tone of a dark area to be corrected according to a production intention when a tone conversion is performed between an HDR image and an SDR image.SOLUTION: An RGB/tristimulus value and chromaticity conversion unit 10 of a video signal conversion device 1 obtains an HDR luminance signal YHDR and chromaticity coordinates x, y from an HDR linear RGB signal EHDR. A compression processing unit 11 compresses the dynamic range of the HDR luminance signal YHDR to obtain an SDR luminance signal YSDR by using a tone mapping function defined by a gamma value γdark for dark area correction for correcting the tone of the dark area in addition to parameters that take into account a skin level of a person and other factors. A tristimulus value and chromaticity/RGB conversion unit 12 obtains an SDR linear RGB signal ESDR from the SDR luminance signal YSDR and chromaticity coordinates x, y.SELECTED DRAWING: Figure 1

Description

The present invention relates to a video signal conversion device and a program that perform gradation conversion between an HDR (High Dynamic Range) video and an SDR (Standard Dynamic Range) video.

Since the start of the new 4K8K satellite broadcasting in Japan, broadcasters have been studying efficient integrated production to simultaneously produce HLG (Hybrid Log Gamma) video and SDR video with different dynamic ranges. There is. The HLG image is an image obtained by one of a plurality of HDR methods.

The HDR video is produced in consideration of the HDR standard white level specified in the ITU-R report (see, for example, Non-Patent Document 1). As a conversion method from HDR video to SDR video in the integrated production of HDR video and SDR video, it is assumed that the standard white level of HDR video is white even in SDR video, and an HDR compatible camera is used. There is a conversion method (see, for example, Patent Document 1) that generates an SDR image from the linear signal of the scene light acquired in. There is also a conversion method in which only the HDR video is produced and the SDR video is generated from the linear signal of the display light at the final stage of the production system.

In the former conversion method in which conversion is performed by the camera system, the color reproducibility of the HDR image and the SDR image finally obtained will be different. This is because there is a difference in EOTF (Electro Optical Transfer Function) when converting a video signal (RGB signal) in HDR and SDR into display light.

On the other hand, in the latter conversion method in which the conversion is performed in the final stage, if the conversion is performed in consideration of only the reference white level of the HDR image, the brightness of the human skin becomes darker than that of the conventional SDR image, and the HDR There is a problem that the signal in the highlight region exceeding the standard white level of the image is reproduced as overexposure.

Comparing these two conversion methods, since the latter conversion method is efficient, it is necessary to solve the above-mentioned problems in the latter conversion method.

In order to solve the above-mentioned problem, the gain value is determined from the correspondence between the skin levels of the HDR video and the SDR video, and the signal in the highlight region is not reproduced as overexposure as much as possible while considering the reference white level of the HDR video signal. A conversion method was proposed and formulated as an ITU-R report (see, for example, Non-Patent Document 2 and Non-Patent Document 3). This conversion method is also described in Japanese Patent Application Nos. 2018-191560 and Japanese Patent Application No. 2019-045293, which were made by the same applicant of the present patent application and were not published at the time of the present patent application.

Here, the tone mapping function used for the compression process when converting the HDR image into the SDR image is often composed of an S-shaped curve similar to the characteristics of the photographic film (see, for example, Patent Document 2). This is because, in an image having a limited dynamic range such as an SDR image, the difference in brightness from the halftone is increased by reducing the gradation of the dark part, and the sense of detail or the sense of contrast is increased.

Japanese Unexamined Patent Publication No. 2016-197854 Japanese Unexamined Patent Publication No. 2018-05093

Report ITU-R BT.2408-0, “Operational practices in HDR television production” Report ITU-R BT.2446-0, “Methods for conversion of high dynamic range content to standard dynamic range content and vice-versa” K. Nomura, Y. Ikeda, Y. Kusakabe, Y. Nishida, “Method for HDR to SDR Conversion Considering HDR Reference White”, IDW’18, VHF1-4L, p.945-948, (2018)

In the video signal converter for converting HDR video to SDR video, the conversion methods of Non-Patent Document 2 and Non-Patent Document 3 described above do not lower the signal level of the human skin, do not change the white level, and highlight. It is intended to suppress overexposure.

However, with this conversion method, the gradation of the dark part cannot be adjusted. Therefore, for example, there is a problem that the brightness difference between the dark part and the intermediate part cannot be increased, and the detail feeling or the contrast feeling cannot be emphasized in the dark part.

It is assumed that the conversion method of Patent Document 2 described above is used to adjust the gradation of the dark portion, but the conversion method of Patent Document 2 is applied as it is to the conversion methods of Non-Patent Document 2 and Non-Patent Document 3. It is difficult to do.

This is because the conversion method of Patent Document 2 considers overexposure of highlights by using a function composed of S-shaped curves in order to convert an input luminance histogram into an output luminance histogram. It does not consider the signal level and white level of the skin. On the other hand, the conversion methods of Non-Patent Document 2 and Non-Patent Document 3 take into consideration the signal level, white level, and whiteout of highlights of a person's skin, and the purpose of conversion is basically different.

Therefore, the present invention has been made to solve the above problems, and an object of the present invention is that the signal level of a person's skin does not decrease, the white level does not change, and overexposure of highlights is suppressed. Further, it is an object of the present invention to provide a video signal conversion device and a program capable of correcting the gradation of a dark portion according to a production intention.

In order to solve the above problems, the video signal conversion device according to claim 1 is a video signal conversion device that converts HDR (High Dynamic Range) video into SDR (Standard Dynamic Range) video. The linear RGB signal of the HDR image is converted into a tristimulus value, the tristimulus value is converted into a chromaticity coordinate, the brightness signal of the three stimulus values is output as an HDR brightness signal, and the chromaticity coordinate is output. Using the output RGB / tristimulus value and chromaticity conversion unit and a predetermined tone mapping function, the dynamic range of the HDR brightness signal output by the RGB / tristimulus value and chromaticity conversion unit is compressed, and the dynamic range is compressed. A compression processing unit that obtains the brightness signal of the SDR image as an SDR brightness signal, the SDR brightness signal obtained by the compression processing unit, and the chromaticity coordinates output by the RGB / tristimulus value and the chromaticity conversion unit. It is provided with a tristimulus value and a chromaticity / RGB conversion unit that convert the tristimulus value into a linear RGB signal of the SDR image, and the compression processing unit is the value of the HDR brightness signal. The tone mapping function having the characteristic that the value of the SDR brightness signal increases with respect to the increase, and the characteristic of the skin level of the SDR image corresponding to the skin level of the HDR image, and the reference white level of the HDR image. The white level characteristic of the SDR image and the value of the SDR video signal corresponding to the linear RGB signal of the SDR image are the SDR video signal with respect to the upper limit value of the HDR video signal corresponding to the linear RGB signal of the HDR image. Using the tone mapping function, which has a characteristic of increasing up to the upper limit value of, and further, the characteristic of the dark portion differs depending on the gamma value for dark portion correction in which the gradation of the dark portion in a predetermined region of the SDR brightness signal can be adjusted. The dynamic range of the HDR brightness signal is compressed to obtain the SDR brightness signal.

Further, in the video signal conversion device according to the second aspect, in the video signal conversion device according to the first aspect, the compression processing unit is a region smaller than a predetermined turning point in the HDR luminance signal as the tone mapping function. Then, the dynamic range of the HDR luminance signal is compressed by using a 冪 function having the gamma value for dark portion correction as a 冪 exponent and using a logarithmic function in the region above the predetermined variation point in the HDR luminance signal. It is characterized in that it is obtained as the SDR luminance signal.

Further, the video signal conversion device according to claim 3 is a video signal conversion device that converts an SDR (Standard Dynamic Range) video into an HDR (High Dynamic Range) video, and is a linear RGB of the SDR video. An RGB / tristimulus that converts a signal into a tristimulus value, converts the tristimulus value into a chromaticity coordinate, outputs a brightness signal of the three stimulus values as an SDR brightness signal, and outputs the chromaticity coordinate. Using the value and chromaticity conversion unit and a predetermined tone mapping function, the dynamic range of the SDR brightness signal output by the RGB / tristimulus value and chromaticity conversion unit is expanded, and the brightness signal of the HDR image is obtained. The expansion processing unit obtained as an HDR brightness signal, the HDR brightness signal obtained by the expansion processing unit, and the RGB / tristimulus value and the chromaticity coordinates output by the chromaticity conversion unit are converted into tristimulus values. The HDR is provided with a tristimulus value and a chromaticity / RGB conversion unit that convert the tristimulus value into a linear RGB signal of the HDR image, and the expansion processing unit responds to an increase in the value of the SDR brightness signal. The tone mapping function having the characteristic of increasing the value of the brightness signal, the characteristic of the skin level of the SDR image corresponding to the skin level of the HDR image, and the characteristic of the skin level of the SDR image corresponding to the reference white level of the HDR image. The white level characteristic and the value of the SDR video signal corresponding to the linear RGB signal of the SDR video increase to the upper limit of the SDR video signal with respect to the upper limit of the HDR video signal corresponding to the linear RGB signal of the HDR video. Using the tone mapping function, which has characteristics and further, the characteristics of the dark portion differ depending on the gamma value for dark portion correction which can adjust the gradation of the dark portion in a predetermined region of the SDR brightness signal, the SDR brightness signal The HDR brightness signal is obtained by expanding the dynamic range.

Further, in the video signal conversion device according to claim 4, in the video signal conversion device according to claim 3, the compression processing unit is a region smaller than a predetermined turning point in the SDR luminance signal as the tone mapping function. Then, the dynamic range of the SDR luminance signal is expanded by using the 冪 function having the inverse of the gamma value for dark portion correction as the 冪 exponent and using the exponential function in the region above the predetermined variation point in the SDR luminance signal. Then, it is obtained as the HDR luminance signal.

Further, the program of claim 5 is characterized in that the computer functions as the video signal conversion device according to claim 1 or 2.

Further, the program of claim 6 is characterized in that the computer functions as the video signal conversion device according to claim 3 or 4.

As described above, according to the present invention, the signal level of a person's skin does not decrease, the white level does not change, overexposure of highlights is suppressed, and the gradation of dark areas is further adjusted according to the production intention. Can be corrected.

It is a block diagram which shows the structure of the video signal conversion apparatus by embodiment of this invention. It is a flowchart which shows the processing of a video signal conversion apparatus. It is a block diagram which shows the structure of the RGB / tristimulus value and the chromaticity conversion part. It is a block diagram which shows the structure of a tristimulus value and a chromaticity / RGB conversion part. It is a figure explaining the correspondence relationship between the skin level of an HLG video signal and the skin level of an SDR video signal. It is a figure explaining the parameter of the tone mapping function used in the compression processing part. It is a figure explaining the tone mapping function used in the compression processing part. It is a figure which shows the characteristic of the display luminance when the HDR luminance signal Y _HDR is compressed into the SDR luminance signal Y _SDR. It is an enlarged view which shows the characteristic of the dark part in FIG. It is a figure which shows the characteristic of the display luminance at the time of compressing the HDR luminance signal Y _HDR into SDR luminance signal Y _SDR when the corresponding point of HDR reference white level Y _{HDR, Ref is changed.} It is a block diagram which shows the structure of the video signal conversion apparatus which performs the reverse conversion by embodiment of this invention. It is a flowchart which shows the process of the reverse conversion of a video signal conversion apparatus.

Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the drawings. The present invention is for correcting the gradation of dark areas in addition to the parameters that consider the skin level of a person, the reference white level, and the color reproduction of highlights when compressing the _{HDR luminance signal Y HDR} into the SDR luminance signal Y _SDR. It is characterized by using the tone mapping function defined by the parameter (gamma value for dark area correction γ _dark). The same applies to the case where the SDR luminance signal Y _SDR is extended to the HDR luminance signal Y _HDR.

As a result, _{when compressing the HDR luminance signal Y HDR} into the SDR luminance signal Y _SDR, it is possible to correct the gradation from the dark part to the halftone, and it is dynamic using a function having the characteristics of a commonly used S-shaped curve. The range can be compressed. Further, when the SDR luminance signal Y _SDR is extended to the HDR luminance signal Y _HDR , the gradation correction from the dark part to the halftone becomes possible, and the dynamic range is expanded by using a function having the opposite characteristic of the S-shaped curve. be able to. The opposite characteristic of the S-shaped curve will be described later.

Therefore, the signal level of the human skin does not decrease, the white level does not change, overexposure of highlights is suppressed, and the gradation of dark areas can be further corrected according to the production intention.

[Video signal converter]
First, a video signal conversion device that compresses the HDR luminance signal Y _HDR into the SDR luminance signal Y _{SDR will be described.} FIG. 1 is a block diagram showing a configuration of a video signal conversion device according to an embodiment of the present invention, and FIG. 2 is a flowchart showing processing of the video signal conversion device.

The video signal conversion device 1 is a device that converts HDR linear RGB signal E _HDR into SDR linear RGB signal E _SDR . The video signal conversion device 1 includes an RGB / tristimulus value and chromaticity conversion unit 10, a compression processing unit 11, and a tristimulus value and chromaticity / RGB conversion unit 12.

The RGB / tristimulus value and chromaticity conversion unit 10 _{inputs the HDR linear RGB signal E HDR} (step S201). Then, the RGB / tristimulus value and chromaticity conversion unit 10 converts the HDR linear RGB signal E _HDR into HDR tristimulus values X _HDR , Y _HDR , and Z _HDR (step S202).

The RGB / tristimulus value and chromaticity conversion unit 10 converts the HDR tristimulus values X _HDR , Y _HDR , and Z _HDR into chromaticity coordinates x and y (step S203). Then, RGB / tristimulus values and the chromaticity converter 10 outputs HDR tristimulus values X _HDR, Y _HDR, the HDR luminance signal Y _HDR of Z _HDR to the compression processing unit 11. Further, the RGB / tristimulus value and chromaticity conversion unit 10 outputs the chromaticity coordinates x and y to the tristimulus value and chromaticity / RGB conversion unit 12. Details of the RGB / tristimulus value and the chromaticity conversion unit 10 will be described later.

The compression processing unit 11 inputs the _{HDR luminance signal Y HDR} from the RGB / tristimulus value and the chromaticity conversion unit 10. Then, the compression processing unit 11 compresses the dynamic range of the _{HDR luminance signal Y HDR} by using a tone mapping function defined by a parameter such as _{a predetermined dark portion correction gamma value γ dark} , and obtains the _{SDR luminance signal Y SDR.} (Step S204). The compression processing unit 11 outputs the SDR luminance signal Y _SDR to the tristimulus value and the chromaticity / RGB conversion unit 12.

Here, parameters such as a predetermined dark portion correction gamma value γ _dark may be input as preset data. Further, some of the parameters such as the predetermined dark portion correction gamma value γ _dark may be calculated by a predetermined arithmetic process as described with reference to FIG. 6 to be described later. The details of the compression processing unit 11 will be described later.

The tristimulus value and the chromaticity / RGB conversion unit 12 _{input the SDR luminance signal Y SDR} from the compression processing unit 11, and input the chromaticity coordinates x and y from the RGB / tristimulus value and the chromaticity conversion unit 10. Then, the tristimulus value and chromaticity / RGB conversion unit 12 converts the SDR luminance signal Y _SDR and the chromaticity coordinates x and y into SDR tristimulus values X _SDR , Y _SDR and Z _SDR (step S205).

The tristimulus value and chromaticity / RGB conversion unit 12 converts the SDR tristimulus values X _SDR , Y _SDR , and Z _SDR into SDR linear RGB signal E _SDR (step S206), and outputs _{SDR linear RGB signal E SDR (step S206).} Step S207). Details of the tristimulus value and the chromaticity / RGB conversion unit 12 will be described later.

(RGB / tristimulus value and chromaticity conversion unit 10)
Next, the RGB / tristimulus value and the chromaticity conversion unit 10 shown in FIG. 1 will be described in detail. As described above, the RGB / tristimulus value and chromaticity conversion unit 10 converts the HDR linear RGB signal E _HDR into HDR tristimulus values X _HDR , Y _HDR , Z _HDR , and HDR brightness signal Y _HDR and chromaticity coordinates x. , Y is calculated.

FIG. 3 is a block diagram showing the configuration of the RGB / tristimulus value and the chromaticity conversion unit 10. The RGB / tristimulus value and chromaticity conversion unit 10 includes an RGB / tristimulus value conversion unit 20 and a tristimulus value / chromaticity conversion unit 21.

The RGB / tristimulus value conversion unit 20 inputs the _{HDR linear RGB signal E HDR} = {R _HDR , G _HDR , B _HDR}. Then, the RGB / tristimulus value conversion unit 20 _{obtains the HDR} tristimulus values X HDR, Y _HDR , and Z _{HDR from the HDR linear RGB signal E HDR} = {R _HDR , G _HDR , B _{HDR} by the following equation.} ..

Here, the HDR linear RGB signal E _HDR corresponds to both HLG and PQ (Perceptual Quantization) in the HDR method, and the scene light in each method and the scene light are OOTF (Opto). Optical Transfer Function) Corresponds to both processed display light. The SDR linear RGB signal E _SDR , which will be described later, also supports both scene light and display light.

RGB / tristimulus value converter 20 outputs HDR tristimulus values X _HDR, Y _HDR, the HDR luminance signal Y _HDR of Z _HDR to the compression processing unit 11. Further, the RGB / tristimulus value conversion unit 20 outputs the HDR tristimulus values X _HDR , Y _HDR , and Z _HDR to the tristimulus value / chromaticity conversion unit 21.

The tristimulus value / chromaticity conversion unit 21 inputs _{HDR tristimulus values X HDR} , Y HDR, and Z _HDR from the RGB / tristimulus value conversion unit 20. Then, the tristimulus value / chromaticity conversion unit 21 obtains the chromaticity coordinates x and y from the _{HDR tristimulus values X HDR} , Y HDR, and Z _{HDR by the following equation.} The tristimulus value / chromaticity conversion unit 21 outputs the chromaticity coordinates x and y to the tristimulus value and the chromaticity / RGB conversion unit 12.

(Compression processing unit 11)
Next, the compression processing unit 11 shown in FIG. 1 will be described in detail. As described above, the compression processing unit 11 uses a predetermined tone mapping function _{to compress the dynamic range of the HDR luminance signal Y HDR} converted by the RGB / tristimulus value and the chromaticity converter 10, and the SDR luminance signal Y. _{Find SDR} .

This tone mapping function is composed of a combination of a linear function and a logarithmic function. This linear function includes a power function for correcting the detail feeling and the contrast feeling of the dark part gradation in addition to the function having the characteristic of a straight line. The linear function is used for the purpose of matching the appearance of the dark part and the halftone of the HDR image, and the logarithmic function is used for the purpose of storing the highlight gradation in the SDR image while maintaining it as much as possible.

The compression processing unit 11 compresses the dynamic range of _{the HDR luminance signal Y HDR by} the following equation to obtain the _{SDR luminance signal Y SDR.} The function shown in the upper part of the following equation (3) is a linear function (a power function whose exponent is a _{preset gamma value γ dark for dark area correction (when γ dark} = 1, a function having linear characteristics)). And the function shown in the lower row is a logarithmic function.

Here, the gain value k ₁ , the correction values k ₂ , k ₃ , k ₄ , the gamma value for dark area correction γ _dark , the SDR inflection point Y _{SDR, ip} and the HDR inflection point Y _{HDR, ip} are given by the above equation (3). ) Is a parameter of the tone mapping function.

The tone mapping function of the above equation (3) is an _{equation for obtaining the} SDR luminance signal Y _{SDR from the HDR luminance signal Y HDR} , and as shown in FIGS. 7 to 10 described later, with respect to an increase in the value of the _{HDR luminance signal Y HDR.} It has the characteristic of an S-shaped curve in which the value of the SDR luminance signal Y _{SDR increases.} More specifically, in this tone mapping function, as _{the value of the HDR luminance signal Y HDR} increases from 0 to the upper limit, the amount of change in the value of the _{SDR luminance signal Y SDR is small at first, then increases, and then} It has the characteristics of an S-shaped curve that becomes smaller.

In particular, the formula (3) upper Formula regions (Y _HDR <Y _{HDR, ip)} of the dark portion of the area which is, as the gradation of the HDR luminance signal Y _HDR remains, SDR from HDR luminance signal Y _HDR Conversion to the luminance signal Y _{SDR is performed.}

Specifically, this tone mapping function has the skin level characteristics of the SDR image corresponding to the skin level of the HDR image, and the white level characteristics of the SDR image corresponding to the reference white level of the HDR image. It has a characteristic that the value of the SDR video signal increases up to the upper limit of the SDR video signal corresponding to the upper limit of the HDR video signal. Here, the SDR video signal is _{a signal corresponding to the SDR linear RGB signal E SDR} , and the HDR video signal is a signal corresponding to the HDR linear RGB signal E _HDR . Further, this tone mapping function is preset to have a _{gamma value γ dark} for dark portion correction that can adjust the gradation of the dark portion in _{the region (Y HDR} <Y _{HDR, ip} ) in the upper equation of the equation (3). It is a function that is set so that the characteristics of the area differ depending on the parameters.

The gain value k ₁ is a parameter set from the correspondence between the HDR image and the skin level of the SDR image. The correction value k ₂ is a parameter calculated from the continuous condition of the first derivative values at _{the inflection points (Y HDR, ip} , Y _{SDR, ip} ) of the linear function and the logarithmic function. The correction value k ₃ is a parameter obtained from the constraint conditions of the HDR reference white level Y _{HDR and Ref.} The correction value k ₄ is a parameter calculated from the continuous condition of the values at _{the inflection points (Y HDR, ip} , Y _{SDR, ip} ) of the linear function and the logarithmic function.

The gamma value γ _dark for dark part correction is a parameter for correcting the gradation of the dark part, and takes a real number satisfying _{γ dark ≧ 1.} The SDR inflection point Y _{SDR, ip} is a parameter representing the linear signal value of the SDR luminance signal Y _{SDR corresponding to the preset inflection point.} The HDR inflection point Y _{HDR, ip} is a parameter representing a linear signal value of the HDR luminance signal Y _HDR corresponding to the SDR inflection point Y _{SDR, ip} in the SDR luminance signal Y _SDR.

Of these seven parameters, five parameters (gain value k ₁ , correction values k ₂ , k ₃ , k ₄ and HDR inflection point Y) _{excluding the gamma value γ dark} for dark area correction and SDR inflection point Y _{SDR, ip. HDR, ip} ) is unknown. However, these unknowns can be reduced to one parameter of the _{correction value k 3.} Hereinafter, the five parameters will be described in the order of gain value k ₁ , HDR inflection point Y _{HDR, ip} , correction value k ₂ , correction value k ₄ , and correction value k ₃ .

(Gain value k ₁ )
As described above, the gain value k ₁ is a parameter set from the correspondence between the HDR image and the skin level of the SDR image.

Generally, gain processing is performed to make the HDR luminance signal 200 [cd / m ² ] correspond between the SDR luminance signals 88 to 100 [cd / m ^2]. As a result, it is possible to realize conversion from the HLG video to the SDR video in consideration of the reference white level. As an example of preventing the gradation of the highlight region of the HLG from being suddenly lost when the reference white level is taken into consideration, a gain that makes the reference white level 75% of the HLG video signal correspond to the white level 96% of the SDR video signal. Processing is used.

FIG. 5 is a diagram illustrating a correspondence relationship between the skin level of the HLG video signal and the skin level of the SDR video signal. In this figure, the results of analyzing the correspondence between the skin levels of the same person in the HLG video and SDR video, which are one of the HDR videos of the same content produced individually, are shown as plots (circles). See). This HLG video program is content produced based on the reference white level of 75% (0.75 in FIG. 5) of the HLG video signal.

The horizontal axis shows the level of the HLG video signal, and the vertical axis shows the level of the SDR video signal. The solid line shows the characteristics of the prior art, and the broken line shows the characteristics of the example considering the skin level. The point α2 will be described with reference to FIG. 8 described later, and the point β2 will be described with reference to FIG. 10 described later.

The HLG video signal is a signal corresponding to _{the HDR linear RGB signal E HDR} input by the RGB / tristimulus value and the chromaticity conversion unit 10. The SDR video signal is a signal corresponding to the tristimulus value and the SDR linear RGB signal E _SDR output by the chromaticity / RGB conversion unit 12.

From the analysis results shown in FIG. 5, _{the average value of the skin level Vh skin} of the HLG video signal is about 50% of the HLG video signal, whereas _{the average value of the skin level Vs skin} of the SDR video signal is the SDR video signal. It can be seen that it is about 67% of.

The characteristic of the prior art is that the gain processing for associating the HLG luminance signal 200 [cd / m ² ] with the SDR luminance signal 88 [cd / m ² ] is applied, which is about _{the skin level Vh skin of the HLG video signal.} The skin level of the SDR video signal corresponding to 50% is about 55%. Therefore, the skin level of the SDR image in the prior art is lower than the skin level of the SDR image in the example considering the skin level.

That is, in order to consider the skin level, it is necessary to perform the gain processing for matching the skin levels as shown by the broken line in FIG. Therefore, for the gain value k ₁ , one of the following three equations is used according to the assumed linear signal.

The above equation (4) shows _{the gain value k 1} when the normalized signal of the display light is used as a reference. The above equation (5) shows _{a gain value k 1} ^{when the brightness value [cd / m 2} ] of the display light is used as a reference. The above equation (6) shows _{the gain value k 1} when the normalized signal of the scene light is used as a reference. _{L W} of the above formula (5) is the peak brightness of the display.

Here, from the analysis results shown in FIG. 5, as an example of the gain processing in consideration of the skin level, the skin level Vh _sink 50% of HLG is made to correspond to the skin level Vs _{skin 67% of SDR.} The gain value k _{1 at} that time is 7.5439 in the above formula (4), and 0.7544 in the above formula (5) when the display peak brightness of the HLG video signal is 1000 [cd / m ^2]. In the above formula (6), it becomes 5.4272. In terms of device mounting, this gain value k ₁ may be a parameter arbitrarily set by the user.

(HDR inflection point Y _{HDR, ip} )
As described above, the HDR inflection point Y _{HDR, ip} is a parameter representing the linear signal value of the HDR luminance signal Y _HDR corresponding to the SDR inflection point Y _{SDR, ip} in the SDR luminance signal Y _SDR.

The HDR variation point Y _{HDR, ip has the SDR variation point Y SDR, ip} corresponding to the knee point when converting the HDR luminance signal Y _HDR to the SDR luminance signal Y _SDR , and the skin level by the following equation. It is calculated from the reflected gain value k _1.

Here, the SDR inflection point Y _{SDR, ip} is a value obtained by converting the inflection point corresponding to the knee point in the SDR video signal into the SDR luminance signal Y _{SDR of the scene light or the display light.} When this SDR inflection point Y _{SDR, ip} is set to a high value, the signal above the reference white level of the HDR video signal is compressed from the inflection point to 109% which is the upper limit value of the HDR video signal. , The banding effect with reduced gradation is likely to occur in the highlight area. Therefore, it is desirable to set the inflection point to a point slightly higher than the skin level of the SDR image shown in FIG. Therefore, the range of possible values of the SDR inflection point Y _{SDR, ip} is set from 80% or more of the SDR video signal to the signal level corresponding to the corresponding point of the reference white level described later, and can be arbitrarily set by the user. You may.

(Correction value k ₂ )
As described above, the correction value k ₂ is a parameter calculated from the continuous condition of the first derivative values at _{the inflection points (Y HDR, ip} , Y _{SDR, ip} ) of the linear function and the logarithmic function.

Since the first derivative values at the inflection points (Y _{HDR, ip} , Y _{SDR, ip} ) match in the upper linear function and the lower logarithmic function of the above equation (3), the following equation is derived.

The equation (8) is subjected to an unknown correction value k _{3 , a gain value k 1} set by the user, an SDR inflection point Y _{SDR, ip} , and an HDR inflection point Y calculated by the equation (7). _{When organized using HDR and ip} , the following equation is derived.

(Correction value k ₄ )
As described above, the correction value k ₄ is a parameter calculated from the continuous condition of the values at _{the inflection points (Y HDR, ip} , Y _{SDR, ip} ) of the linear function and the logarithmic function.

In the tone mapping curve which is the characteristic of the lower equation of the equation (3), the gain value k ₁ set by the user, the SDR inflection point Y _{SDR, ip} , and the HDR inflection calculated by the equation (7) are calculated. The following equation is derived by arranging using the points Y _{HDR, ip and the correction value k 2} calculated by the equation (9) above.

For the four parameters shown so far, the gain value k ₁ , the HDR inflection point Y _{HDR, ip} , and the correction values k ₂ , k ₄ , the gain value k ₁ and the SDR inflection point Y _{SDR, ip} are set in advance by the user. It is uniquely determined by calculating the _{correction value k 3} which is set and unknown.

(Correction value k ₃ )
As described above, the correction value k ₃ is a parameter obtained from the constraint conditions of the HDR reference white level Y _{HDR and Ref.}

The correction value k ₃ is calculated by determining the correspondence between the HDR reference white level Y _{HDR, Ref} and the SDR white level Y _{SDR, WP.} Specifically, the correction value k ₃ is a knee point HDR change point Y _{HDR, ip} and an SDR _{change point Y SDR, ip} change point (Y _{HDR, ip} , Y _{SDR, ip} ) and HDR. The amount of change in the SDR luminance signal Y _SDR between the reference white level Y _{HDR, Ref} and the reference white level points (Y _{HDR, Ref} , Y _{SDR, WP} ) of the SDR white level Y _{SDR, WP} , and the above equation (3). It is calculated by the following formula that matches the amount of change in _{the SDR luminance signal Y SDR} of the tone mapping curve in the lower formula.

The HDR reference white level Y _{HDR, Ref} is a linear signal value of the scene light or display light of 75% to 109% of the HLG video signal in addition to the linear signal value of the scene light or display light of the HLG video signal 75%. Of these, the parameters may be arbitrarily set by the user.

_{The correction value k 3} calculated by the above formula (11) is uniquely obtained. _{As a result, all five parameters (gain value k 1} , correction values k ₂ , k ₃ , k ₄ and HDR inflection point Y _{HDR, ip} ) out of the seven parameters in the above equation (3) are set. Can be done.

(Parameter of tone mapping function)
The tone mapping function is as shown in the above equation (3), and its parameters are gain value k ₁ , correction values k ₂ , k ₃ , k ₄ , gamma value for dark area correction γ _dark , SDR inflection point Y _{SDR, ip.} And HDR inflection point Y _{HDR, ip} , total 7

These seven parameters may be preset by the user in the video signal conversion device 1 shown in FIG. 1 in consideration of the above equations (4) to (11).

In addition, the gain value k _{1 of these seven parameters, the gamma value γ dark} for dark area correction, and the SDR inflection point Y _{SDR, ip} are preset by the user, and the correction values k ₂ , k ₃ , k ₄ and HDR are set in advance. The inflection point Y _{HDR, ip} may be calculated by the compression processing unit 11 by the above equations (7), (9), (10), and (11). In this case, the reference white level points (Y _{HDR, Ref} , Y _{SDR, WP} ) of the HDR reference white level Y _{HDR, Ref} and the SDR white level Y _{SDR, WP} are also preset by the user.

Specifically, the compression processing unit 11 _{calculates the correction value k 3} by the above formula (11), calculates the HDR inflection point Y _{HDR, ip} by the above formula (7), and the above formula (9). The correction value k ₂ is calculated by the above equation (10), and the correction value k ₄ is calculated by the above equation (10). As a result, seven parameters are set, so that the tone mapping function of the above equation (3) can be constructed.

FIG. 6 is a diagram for explaining the parameters of the tone mapping function used in the compression processing unit 11. In this figure, the compression processing unit 11 has a preset gain value k ₁ , an SDR inflection point Y _{SDR, ip} , a gamma value for dark area correction γ _dark , an HDR reference white level Y _{HDR, Ref,} and an SDR white level Y _{SDR. , WP} is input, and the block diagram when the HDR inflection point Y _{HDR, ip} and the correction values k ₂ , k ₃ , and k _{4 are calculated is shown.}

The compression processing unit 11 includes HDR inflection points Y _{HDR, ip} and calculation units 24, 25, 26, 27 for calculating correction values k ₂ , k ₃ , and k _4.

The calculation unit 24 inputs the preset gain value k ₁ and the SDR inflection point Y _{SDR, ip} , performs the calculation of the above equation (7), and obtains the HDR inflection point Y _{HDR, ip} . Then, the arithmetic unit 24 _{outputs the HDR inflection point Y HDR, ip} to the arithmetic units 25, 26, 27.

The calculation unit 27 inputs _{preset gain values k 1} , SDR inflection point Y _{SDR, ip} , gamma value for dark area correction γ _dark , HDR reference white level Y _{HDR, Ref} and SDR white level Y _{SDR, WP.} At the same time, the HDR inflection point Y _{HDR, ip} is input from the calculation unit 24. Then, the calculator 27 calculates the correction value k ₃ performs operation of the equation (11), and outputs the correction value k ₃ to the arithmetic unit 25.

The calculation unit 25 _{inputs a preset gain value k 1} , an SDR inflection point Y _{SDR, ip,} and a gamma value γ _dark for dark portion correction, and inputs an HDR inflection point Y _{HDR, ip from the calculation unit 24.} , Further, the correction value k ₃ is input from the calculation unit 27. Then, the arithmetic unit 25 calculates the correction value k ₂ performs operation of the equation (9), and outputs the correction value k ₂ to the arithmetic unit 26.

The calculation unit 26 _{inputs a preset gain value k 1} , an SDR inflection point Y _{SDR, ip,} and a gamma value γ _dark for dark portion correction, and inputs an HDR inflection point Y _{HDR, ip from the calculation unit 24.} .. Further, the calculation unit 26 _{inputs the correction value k 2} from the calculation unit 25, and further inputs the correction value k ₃ from the calculation unit 27. Then, the arithmetic unit 26 calculates the correction value k ₄ performs operation of the equation (10), and outputs the correction value k _4.

As a result, in the compression processing unit 11, the preset gain value k ₁ , SDR inflection point Y _{SDR, ip} , gamma value for dark area correction γ _dark , HDR reference white level Y _{HDR, Ref} and SDR white level Y _{SDR. , WP} are used to calculate the HDR inflection point Y _{HDR, ip} and the correction values k ₂ , k ₃ , k ₄ . Then, the gain value k ₁ , the correction value k ₂ , k ₃ , k ₄ , the gamma value for dark area correction γ _dark , the SDR inflection point Y _{SDR, ip} and the HDR inflection point Y _{HDR, ip} are used for a total of seven parameters. The tone mapping function of the above equation (3) is configured, and the compression processing is performed by the compression processing unit 11.

FIG. 7 is a diagram illustrating a tone mapping function used by the compression processing unit 11. This tone mapping function has a tone mapping curve composed of seven parameters, and has the characteristic of an S-shaped curve whose value increases as a whole.

The inflection points (Y _{HDR, ip} , Y _{SDR, ip} ) of this tone mapping function are, for example, SDR inflection point Y _{SDR, ip} = 58.5357 and HDR inflection point Y _{HDR, ip} = 77.5924. The tone mapping function has the characteristics of the linear function in the upper part of the above equation (3) in the region of _{HDR brightness signal Y HDR} <Y _{HDR, ip = 77.5924, and HDR brightness signal Y HDR} ≧ Y _{HDR, ip} =. In the region of 77.5924, it has the characteristics of the logarithmic function in the lower part of the above equation (3).

More specifically, as the value of the HDR luminance signal Y _HDR increases from 0 to the upper limit, the value of the SDR luminance signal Y _SDR changes _{in the region of HDR luminance signal Y HDR} <Y _{HDR, ip = 77.5924.} The amount is small at first and then large. Further, in the region of HDR luminance signal Y _HDR ≧ Y _{HDR, ip} = 77.5924, the amount of change in the value of the _{SDR luminance signal Y SDR is large at first and then decreases.}

By using the tone mapping function having such characteristics, the linear function is used in the dark region, which is the region _{of the HDR luminance signal Y HDR} <Y _{HDR, ip = 77.5924, and the gradation does not deteriorate.} .. Therefore, it is possible to emphasize the feeling of detail or the feeling of contrast in the dark part.

As shown in FIGS. 8 and 9 described later, the characteristics of the linear function in the dark region can be changed according _{to the preset gamma value γ dark for dark region correction. Therefore, by changing the gamma value γ dark} for dark portion correction, the gradation expression of the dark portion can be adjusted according to the user's production intention.

(Example of parameters)
An example of the parameters of the tone mapping function will be described. For example, the gamma value for dark area correction γ _dark = 1, the SDR _{variation point Y SDR, ip} is set to a value corresponding to 80% of the SDR video signal, and the correspondence between the skin levels is 50% for the HLG video signal and 67% for the SDR video signal. As a constraint condition for the reference white level _{, it is assumed that the HDR reference white level Y HDR and Ref} correspond to 96% of the SDR video signal. In this case, HDR inflection point Y _{HDR, ip} = 77.5924, gain value k ₁ = 0.7544, correction value k ₂ = 17.1209, correction value k ₃ = 0.7075, correction value k ₄ = 79. It becomes 5822.

When the gamma value for dark area correction is γ _dark = 1.1, HDR inflection point Y _{HDR, ip} = 77.5924, gain value k ₁ = 0.7544, correction value k ₂ = 15.8855, correction value. K ₃ = 0.7533 and the correction value k ₄ = 80.7673.

When the gamma value γ _dark for dark area correction is adjusted in this way, the HDR inflection point Y _{HDR, ip} and the gain value k ₁ are not changed, but only the correction values k ₂ , k ₃ , and k ₄ are changed. ..

(Example of tone mapping curve)
Next, an example of the tone mapping curve will be described. FIG. 8 is a diagram showing the characteristics of the display luminance when the _{HDR luminance signal Y HDR} is compressed into the SDR luminance signal Y _SDR. Specifically, when the correction value k ₃ is derived, the characteristics when the constraint condition that the HDR reference white level Y _{HDR, Ref} = 203 becomes the SDR white level Y _{SDR, WP} = 90 (point α1) is given. The point α1 which is the reference white level point (Y _{HDR, Ref} , Y _{SDR, WP} ) in FIG. 8 corresponds to, for example, 75% of the HLG video signal and 96% of the SDR video signal (point α2 in FIG. 5). FIG. 9 is an enlarged view showing the characteristics of the dark portion in FIG.

FIG. 10 is a diagram showing the characteristics of the display luminance when the HDR luminance signal Y _HDR is compressed into the SDR luminance signal Y _SDR when the corresponding points of the HDR reference white levels Y _{HDR and Ref are changed.} Specifically, the characteristics when the constraint condition that the HDR reference white level Y _{HDR, Ref} = 1000 becomes the SDR white level Y _{SDR, WP} = 123 (point β1) is given when the correction value k _{3 is derived.} The point β1 which is the reference white level point (Y _{HDR, Ref} , Y _{SDR, WP} ) in FIG. 10 corresponds to, for example, 100% of the HLG video signal and 109% of the SDR video signal (point β2 in FIG. 5). These characteristics show the tone mapping curve when the _{gamma value γ dark} for dark area correction is adjusted in the range of 1 ≦ γ _{dark ≦ 1.5.}

As shown in FIGS. 8 to 10, _{by changing the gamma value γ dark} for dark area correction, the tone mapping curve is changed, and the characteristics of the tone mapping function corresponding to the value of the _{gamma value γ dark for dark area correction are obtained.} Can be done. In particular, in the dark region shown in FIG. 9, the lower linear function of the above equation (3) is used as the tone mapping function.

When the conventional S-curve function is used, the change in the SDR luminance signal Y _{SDR with respect to the change in the HDR luminance signal Y HDR} is small in the dark region.

On the other hand, when the tone mapping function of the embodiment of the present invention is used, the change of the SDR luminance signal Y _{SDR with respect to the change of the HDR luminance signal Y HDR} is larger than that of the conventional S-shaped curve in the dark region. In the dark region shown in FIGS. 8 to 10, _{the characteristic of 1 <γ dark} ≤ 1.5 is downward (HDR luminance signal Y) with respect to the characteristic of the straight line when the _{gamma value for dark portion correction γ dark = 1.} It is a curve that sinks in the (axial direction of _HDR). Therefore, when the HDR luminance signal Y _HDR is converted to the SDR luminance signal Y _SDR , the gradation of the HDR luminance signal Y _{HDR remains.}

In the tone mapping function of the embodiment of the present invention, the degree of change of the SDR luminance signal Y _{SDR with respect to the change of the HDR luminance signal Y HDR} can be set according to the value of the _{gamma value γ dark} for dark portion correction.

Therefore, it is possible to emphasize the sense of detail or the sense of contrast in the dark part, and _{by changing the gamma value γ dark} for correcting the dark part, the gradation expression of the dark part can be adjusted according to the production intention of the user.

(Tristimulus value and chromaticity / RGB conversion unit 12)
Next, the tristimulus values and the chromaticity / RGB conversion unit 12 shown in FIG. 1 will be described in detail. As described above, the tristimulus value and chromaticity / RGB conversion unit 12 has the SDR luminance signal Y _SDR obtained by the compression processing unit 11 and the chromaticity coordinates x, converted by the RGB / tristimulus value and chromaticity conversion unit 10. convert the y SDR tristimulus values X _SDR, Y _SDR, the Z _SDR, obtains the SDR linear RGB signal E _SDR.

FIG. 4 is a block diagram showing the configuration of the tristimulus value and the chromaticity / RGB conversion unit 12. The tristimulus value and chromaticity / RGB conversion unit 12 includes a chromaticity / tristimulus value conversion unit 22 and a tristimulus value / RGB conversion unit 23.

The chromaticity / tristimulus value conversion unit 22 _{inputs the SDR luminance signal Y SDR} from the compression processing unit 11, and inputs the chromaticity coordinates x and y from the RGB / tristimulus value and the chromaticity conversion unit 10. Then, the chromaticity / tristimulus value conversion unit 22 uses the following equation to obtain two of the SDR tristimulus values X _SDR , Y _SDR , and Z _{SDR from the SDR luminance signal Y SDR} and the chromaticity coordinates x and y. Obtain the stimulus values X _SDR and Z _SDR.

The chromaticity / tristimulus value conversion unit 22 outputs the SDR tristimulus values X _SDR , Y _SDR , and Z _SDR to the tristimulus value / RGB conversion unit 23.

The tristimulus value / RGB conversion unit 23 inputs _{SDR tristimulus values X SDR} , Y SDR, and Z _SDR from the chromaticity / tristimulus value conversion unit 22. Then, the tristimulus value / RGB conversion unit 23 obtains the SDR linear RGB signal E _SDR = {R _SDR , G _SDR , B _SDR } _{from the SDR tristimulus values X SDR} , Y _SDR , and Z _{SDR by the following equation.} .. The tristimulus value / RGB conversion unit 23 outputs the SDR linear RGB signal E _SDR = {R _SDR , G _SDR , B _SDR }.

Here, when there is a signal ( _RSDR , G _SDR , B _{SDR ) in which the SDR linear RGB signal E SDR} <0, the tristimulus value / RGB conversion unit 23 sets the signal to 0 and outputs the signal.

As described above, according to the video signal conversion device 1 of the embodiment of the present invention, the RGB / tristimulus value and chromaticity conversion unit 10 converts the HDR linear RGB signal E _HDR into the HDR tristimulus values X _HDR , Y _HDR , It _{is converted to Z HDR} , and the HDR tristimulus values X _HDR , Y _HDR , and Z _HDR are converted to chromaticity coordinates x and y.

The compression processing unit 11 provides a tone mapping function defined by a _{gamma value γ dark} for dark area correction for correcting the gradation of dark areas, in addition to parameters for considering the skin level of a person, a reference white level, and color reproduction of highlights. The HDR luminance signal Y _HDR dynamic range is compressed and the SDR luminance signal Y _SDR is obtained. This tone mapping function is composed of a combination of a linear function and a logarithmic function.

The tristimulus value and chromaticity / RGB conversion unit 12 converts the SDR brightness signal Y _SDR and the chromaticity coordinates x, y into SDR tristimulus values X _SDR , Y _SDR , Z _SDR , and SDR tristimulus values X _SDR , Y. _{Converts SDR} and Z _SDR to SDR linear RGB signal E _SDR .

As a result, _{the shape of the tone mapping curve can be changed by changing the gamma value γ dark} for dark portion correction, and gradation correction from the dark portion to the halftone can be performed. Then, the dynamic range of the _{HDR luminance signal Y HDR} can be compressed by using a function having the characteristics of a generally used S-shaped curve. Therefore, the signal level of the human skin does not decrease, the white level does not change, overexposure of highlights is suppressed, and the gradation of dark areas can be further corrected according to the production intention.

Further, in the integrated program production of the HLG video and the SDR video, the HLG video program is produced, and by passing through the video signal conversion device 1, the skin level is as if the program was produced with the SDR video, and the reference white level is It is possible to generate an SDR image that is considered and has suppressed highlight overexposure.

Further, for example, it is possible to convert a PQ signal in consideration of the reference white level to an SDR image, it is possible to correct the color reproduction of highlights according to the production intention, and the gradation of dark areas can be adjusted according to the production intention. It becomes possible to correct. Then, it can be introduced into the video signal conversion device 1 in an integrated production system that is expected to become widespread in the future, and further, the video signal conversion device 1 can be mounted on an external tuner or a receiver.

[Video signal converter that performs inverse conversion]
Next, a device that performs inverse conversion to the processing of the video signal conversion device 1 shown in FIG. 1, that is, a video signal conversion device that extends the _{SDR luminance signal Y SDR} to the HDR luminance signal Y _{HDR will be described.}

FIG. 11 is a block diagram showing a configuration of a video signal conversion device that performs reverse conversion according to an embodiment of the present invention, and FIG. 12 is a flowchart showing a process of reverse conversion of the video signal conversion device.

The video signal conversion device 2 is a device that converts an SDR linear RGB signal E _SDR into an HDR linear RGB signal E _HDR . The video signal conversion device 2 includes an RGB / tristimulus value and chromaticity conversion unit 30, an extended processing unit 31, and a tristimulus value and chromaticity / RGB conversion unit 32.

The RGB / tristimulus value and chromaticity conversion unit 30 _{inputs the SDR linear RGB signal E SDR} (step S1201). Then, the RGB / tristimulus value and chromaticity conversion unit 30 converts the SDR linear RGB signal E _SDR into SDR tristimulus values X _SDR , Y _SDR , and Z _SDR (step S1202).

The RGB / tristimulus value and chromaticity conversion unit 30 converts the SDR tristimulus values X _SDR , Y _SDR , and Z _SDR into chromaticity coordinates x and y (step S1203). Then, RGB / tristimulus values and the chromaticity converter 30 outputs SDR tristimulus values X _SDR, Y _SDR, the SDR luminance signal Y _SDR of Z _SDR to the expansion processing unit 31. Further, the RGB / tristimulus value and chromaticity conversion unit 30 outputs the chromaticity coordinates x and y to the tristimulus value and chromaticity / RGB conversion unit 32. That is, the RGB / tristimulus value and chromaticity conversion unit 30 performs the reverse processing of the tristimulus value and the chromaticity / RGB conversion unit 12 shown in FIG.

The internal configuration of the RGB / tristimulus value and chromaticity conversion unit 30 is such that the HDR linear RGB signal E _{HDR is changed} to the SDR linear RGB signal E _SDR in the configuration of the RGB / tristimulus value and chromaticity conversion unit 10 shown in FIG. Because it is the same as replacing the HDR luminance signal Y _HDR with the SDR luminance signal Y _SDR and replacing the HDR tristimulus values X _HDR , Y _HDR , Z _HDR with the SDR tristimulus values X _SDR , Y _SDR , Z _SDR. , The description is omitted.

The expansion processing unit 31 inputs the _{SDR luminance signal Y SDR} from the RGB / tristimulus value and chromaticity conversion unit 30. Then, the expansion processing unit 31 expands the dynamic range of the _{SDR luminance signal Y SDR} by using a tone mapping function defined by a parameter such as _{a predetermined dark portion correction gamma value γ dark} , and obtains the _{HDR luminance signal Y HDR.} (Step S1204). The expansion processing unit 31 outputs the HDR luminance signal Y _HDR to the tristimulus value and the chromaticity / RGB conversion unit 32. That is, the expansion processing unit 31 performs the reverse processing of the compression processing unit 11 shown in FIG.

Here, similarly to the compression processing unit 11 shown in FIG. 1, _{parameters such as a predetermined dark portion correction gamma value γ dark} may be preset by the user. Further, the gain value k ₁ or the like may be preset by the user, and the correction value k ₂ or the like may be calculated by calculation.

This tone mapping function is an inverse function of the tone mapping function used by the compression processing unit 11 shown in FIG. 1, and is composed of a combination of a linear function and an exponential function. This linear function includes a power function for correcting the detail feeling and the contrast feeling of the dark part gradation in addition to the function having the characteristic of a straight line.

The expansion processing unit 31 expands the dynamic range of _{the SDR luminance signal Y SDR} and obtains the _{HDR luminance signal Y HDR by the following equation.} The function shown in the upper part of the following equation (14) is a linear function ( _{a power function whose reciprocal of the preset gamma value γ dark} for dark part correction is a power exponential), and the function shown in the lower part is an exponential function. ..

Here, the gain value k ₁ , the correction value k ₂ , k ₃ , k ₄ , the gamma value for dark area correction γ _dark , the SDR inflection point Y _{SDR, ip} and the HDR inflection point Y _{HDR, ip} are given by the above equation (14). It is a parameter of the tone mapping function shown in), and is as described in the above equation (3).

The tone mapping function of the above equation (14) is an _{equation for obtaining the} HDR luminance signal Y _{HDR from the SDR luminance signal Y SDR} , and has a characteristic in which the horizontal axis and the vertical axis of FIGS. 7 to 10 are exchanged, that is, the SDR luminance signal. It has the opposite characteristic of an S-shaped curve in which the value of the _{HDR luminance signal Y HDR} increases with respect to the increase in the value of _{Y SDR.} More specifically, in this tone mapping function, as _{the value of the SDR luminance signal Y SDR} increases from 0 to the upper limit, the amount of change in the value of the _{HDR luminance signal Y HDR increases first, then decreases, and then} It has the opposite characteristics of an S-shaped curve that increases.

Specifically, this tone mapping function has the skin level characteristics of the SDR image corresponding to the skin level of the HDR image, and the white level characteristics of the SDR image corresponding to the reference white level of the HDR image. It has a characteristic that the value of the SDR video signal increases up to the upper limit of the SDR video signal corresponding to the upper limit of the HDR video signal. Here, the SDR video signal is _{a signal corresponding to the SDR linear RGB signal E SDR} , and the HDR video signal is a signal corresponding to the HDR linear RGB signal E _HDR . Further, this tone mapping function is preset to have a _{gamma value γ dark} for dark portion correction that can adjust the gradation of the dark portion in _{the region (Y SDR} <Y _{SDR, ip} ) of the upper equation of the equation (14). It is a function that is set so that the characteristics of the area differ depending on the parameters.

The tristimulus value and chromaticity / RGB conversion unit 32 _{inputs the HDR luminance signal Y HDR} from the expansion processing unit 31, and inputs the chromaticity coordinates x and y from the RGB / tristimulus value and chromaticity conversion unit 30. Then, the tristimulus value and chromaticity / RGB conversion unit 32 converts the HDR luminance signal Y _HDR and the chromaticity coordinates x and y into HDR tristimulus values X _HDR , Y _HDR and Z _HDR (step S1205).

The tristimulus value and chromaticity / RGB conversion unit 32 converts the HDR tristimulus values X _HDR , Y _HDR , and Z _HDR into HDR linear RGB signal E _HDR (step S1206), and outputs the _{HDR linear RGB signal E HDR (step S1206).} Step S1207). That is, the tristimulus value and chromaticity / chromaticity conversion unit 32 performs the reverse processing of the RGB / tristimulus value and chromaticity conversion unit 10 shown in FIG.

The internal configuration of the tristimulus value and chromaticity / RGB conversion unit 32 replaces the _{SDR luminance signal Y SDR} with the HDR luminance signal Y _HDR in the configuration of the tristimulus value and chromaticity / RGB conversion unit 12 shown in FIG. Because it is the same as replacing the SDR tristimulus values X _SDR , Y _SDR , and Z _SDR with the HDR tristimulus values X _HDR , Y _HDR , and Z _{HDR, and replacing the SDR linear RGB signal E SDR} with the HDR linear RGB signal E _HDR. , The description is omitted.

As described above, according to the video signal conversion device 2 of the embodiment of the present invention, the RGB / tristimulus value and chromaticity conversion unit 30 converts the SDR linear RGB signal E _SDR into SDR tristimulus values X _SDR , Y _SDR , It _{is converted to Z SDR} , and the SDR tristimulus values X _SDR , Y _SDR , and Z _SDR are converted to chromaticity coordinates x and y.

The expansion processing unit 31 expands the dynamic range of the _{SDR luminance signal Y SDR} by using a tone mapping function defined by a parameter such as a predetermined dark portion correction gamma value γ _dark , and obtains the _{HDR luminance signal Y HDR.} This tone mapping function is composed of a combination of a linear function and an exponential function.

The tristimulus value and chromaticity / RGB conversion unit 32 converts the HDR luminance signal Y _HDR and the chromaticity coordinates x, y into HDR tristimulus values X _HDR , Y _HDR , Z _HDR , and HDR tristimulus values X _HDR , Y. _{Converts HDR} and Z _HDR to HDR linear RGB signal E _HDR .

As a result, _{the shape of the tone mapping curve can be changed by changing the gamma value γ dark} for dark portion correction, and gradation correction from the dark portion to the halftone can be performed. Then, the dynamic range of the _{SDR luminance signal Y SDR} can be extended by using a function having the opposite characteristic of the generally used S-shaped curve. Therefore, the signal level of the human skin does not decrease, the white level does not change, overexposure of highlights is suppressed, and the gradation of dark areas can be further corrected according to the production intention.

Although the present invention has been described above with reference to embodiments, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the technical idea.

As the hardware configuration of the video signal conversion devices 1 and 2 according to the embodiment of the present invention, a normal computer can be used. The video signal conversion devices 1 and 2 are composed of a computer provided with a volatile storage medium such as a CPU and RAM, a non-volatile storage medium such as a ROM, and an interface.

Each function of the RGB / tristimulus value and chromaticity conversion unit 10, the compression processing unit 11, and the tristimulus value and chromaticity / RGB conversion unit 12 provided in the video signal conversion device 1 is a program describing these functions. Is realized by causing the CPU to execute the above.

Further, each function of the RGB / tristimulus value and chromaticity conversion unit 30, the extended processing unit 31, and the tristimulus value and chromaticity / RGB conversion unit 32 provided in the video signal conversion device 2 also describes these functions. Each of these is realized by having the CPU execute the programs.

These programs are stored in the storage medium, read by the CPU, and executed. In addition, these programs can be stored and distributed in storage media such as magnetic disks (floppy (registered trademark) disks, hard disks, etc.), optical disks (CD-ROM, DVD, etc.), semiconductor memories, etc., and can be distributed via a network. You can also send and receive.

The video signal conversion device and program of the present invention are useful as a technique for converting a dynamic range in, for example, integrated production in which HDR video and SDR video are simultaneously produced. Further, the video signal conversion device and the program of the present invention can be mounted on the receiver.

1 Video signal conversion device 2 Video signal conversion device 10 RGB / tristimulus value and chromaticity conversion unit 11 compression processing unit 12 tristimulus value and chromaticity / RGB conversion unit 20 RGB / tristimulus value conversion unit 21 tristimulus value / color Degree conversion unit 22 chromaticity / tristimulus value conversion unit 23 tristimulus value / RGB conversion unit 24 arithmetic unit 25 arithmetic unit 26 arithmetic unit 27 arithmetic unit 30 RGB / tristimulus value and chromaticity conversion unit 31 extended processing unit 32 tristimulus Value and chromaticity / RGB converter E _HDR HDR linear RGB signal E _SDR SDR linear RGB signal X _HDR , Y _HDR , Z _HDR HDR tristimulus value X _SDR , Y _SDR , Z _SDR SDR tristimulus value x, y chromaticity coordinates Y _HDR HDR brightness signal Y _SDR SDR brightness signal k ₁ gain value k ₂ correction value k ₃ correction value k ₄ correction value γ _dark gamma value for dark area correction Y _{SDR, ip} SDR change point Y _{HDR, ip} HDR change point Y _{HDR, Ref} HDR standard white level Y _{SDR, WP} SDR white level Vh _skin HLG video signal skin level Vs _skin SDR video signal skin level

Claims (6)

In a video signal converter that converts HDR (High Dynamic Range) video to SDR (Standard Dynamic Range) video.
The linear RGB signal of the HDR image is converted into tristimulus values, the tristimulus values are converted into chromaticity coordinates, the luminance signal of the tristimulus values is output as an HDR luminance signal, and the chromaticity coordinates are output. Output RGB / tristimulus value and chromaticity conversion unit,
A compression processing unit that compresses the dynamic range of the HDR luminance signal output by the RGB / tristimulus value and chromaticity converter using a predetermined tone mapping function, and obtains the luminance signal of the SDR image as an SDR luminance signal. When,
The SDR luminance signal obtained by the compression processing unit, the RGB / tristimulus value, and the chromaticity coordinates output by the chromaticity conversion unit are converted into tristimulus values, and the tristimulus values are converted into the tristimulus value of the SDR image. It is equipped with a tristimulus value that converts to a linear RGB signal and a chromaticity / RGB conversion unit.
The compression processing unit
The tone mapping function having the characteristic that the value of the SDR luminance signal increases with respect to the increase of the value of the HDR luminance signal, and the characteristic of the skin level of the SDR image corresponding to the skin level of the HDR image, the said. The white level characteristic of the SDR image corresponding to the reference white level of the HDR image and the value of the SDR video signal corresponding to the linear RGB signal of the SDR image are the values of the HDR video signal corresponding to the linear RGB signal of the HDR image. It has a characteristic of increasing up to the upper limit of the SDR video signal with respect to the upper limit, and further, the characteristic of the dark portion differs depending on the gamma value for dark portion correction that can adjust the gradation of the dark portion in a predetermined region of the SDR luminance signal. A video signal conversion device characterized in that the dynamic range of the HDR luminance signal is compressed to obtain the SDR luminance signal by using the tone mapping function. In the video signal conversion device according to claim 1,
The compression processing unit
As the tone mapping function, in a region smaller than the predetermined turning point in the HDR luminance signal, a 冪 function having the gamma value for dark portion correction as a 冪 exponent is used, and is equal to or higher than the predetermined turning point in the HDR luminance signal. A video signal conversion device characterized in that the dynamic range of the HDR luminance signal is compressed and obtained as the SDR luminance signal by using a logarithmic function in the region of. In a video signal converter that converts SDR (Standard Dynamic Range) video to HDR (High Dynamic Range) video.
The linear RGB signal of the SDR image is converted into tristimulus values, the tristimulus values are converted into chromaticity coordinates, the luminance signal of the tristimulus values is output as an SDR luminance signal, and the chromaticity coordinates are output. Output RGB / tristimulus value and chromaticity conversion unit,
An expansion processing unit that expands the dynamic range of the SDR luminance signal output by the RGB / tristimulus value and chromaticity conversion unit using a predetermined tone mapping function, and obtains the luminance signal of the HDR image as an HDR luminance signal. When,
The HDR luminance signal obtained by the expansion processing unit, the RGB / tristimulus value, and the chromaticity coordinates output by the chromaticity conversion unit are converted into tristimulus values, and the tristimulus values are converted into the HDR video. It is equipped with a tristimulus value that converts to a linear RGB signal and a chromaticity / RGB conversion unit.
The expansion processing unit
The tone mapping function having a characteristic that the value of the HDR luminance signal increases with respect to the increase of the value of the SDR luminance signal, and the characteristic of the skin level of the SDR image corresponding to the skin level of the HDR image, the said. The white level characteristic of the SDR image corresponding to the reference white level of the HDR image and the value of the SDR video signal corresponding to the linear RGB signal of the SDR image are the values of the HDR video signal corresponding to the linear RGB signal of the HDR image. It has a characteristic of increasing up to the upper limit value of the SDR video signal with respect to the upper limit value, and further, the characteristic of the dark portion differs depending on the gamma value for dark portion correction that can adjust the gradation of the dark portion in a predetermined region of the SDR luminance signal. A video signal conversion device characterized in that the dynamic range of the SDR luminance signal is expanded to obtain the HDR luminance signal by using the tone mapping function. In the video signal conversion device according to claim 3,
The compression processing unit
As the tone mapping function, in a region smaller than the predetermined variation point in the SDR luminance signal, a power function having the inverse of the gamma value for dark portion correction as a power index is used, and the predetermined variation in the SDR luminance signal is used. A video signal conversion device characterized in that, in a region above a point, an exponential function is used to extend the dynamic range of the SDR luminance signal to obtain it as the HDR luminance signal. A program for causing a computer to function as the video signal conversion device according to claim 1 or 2. A program for causing a computer to function as the video signal conversion device according to claim 3 or 4.

Images