Adjustment of display optimization behavior of HDR images

Abstract

To achieve a better and more adjustable way of HDR video display adaptation, the inventors of the applicant have contributed an image pixel brightness adaptation device (500), comprising: a connection (501) for receiving a brightness mapping function (F_ct; FL_50t1_1), the brightness mapping function specifying the offset of the brightness of a secondary image corresponding to the encoded high dynamic range image compared to the brightness at the same pixel position encoded in the encoded high dynamic range image, as well as a decoded high dynamic range image (Im_RHDR) and the brightness mapping function; a display adaptation unit (401), which is arranged to receive a value of a display maximum brightness (PB_D) that a specific display can display as the brightest pixel color, and an input brightness mapping function, and is arranged to apply an algorithm, which calculates a display adaptation brightness mapping function based on the input brightness mapping function and the display maximum brightness (PD_D), wherein the display adaptation brightness mapping function corresponds in shape to the input brightness mapping function but is closer to the input brightness mapping function. Near a 45 degree increasing diagonal, depending on the difference between the display maximum brightness (PB_D) and the value of the first maximum encodable brightness (PB_H) relative to the difference between the second maximum encodable brightness (PB_S) and the first maximum encodable brightness (PB_H); wherein the image pixel brightness adaptation device includes an alternative brightness mapping function determination unit (502), the alternative brightness mapping function determination unit is arranged to determine an alternative brightness mapping function (ALT_FL_50t1_1), and wherein the display adaptation unit (401) includes a combination unit (503), the combination unit is arranged to combine the brightness mapping function (F_ct; FL_50t1_1) and the alternative brightness mapping function (ALT_FL_50t1_1) into a combined brightness mapping function (CMB_FL_50t1_1), and wherein the display adaptation unit is arranged to apply its algorithm to the combined brightness mapping function as an input brightness mapping function.

Description

Technical Field

[0001] The present invention relates to a method and apparatus for adapting the high dynamic range image pixel luminance of an HDR video for a specific display on a display having a specific luminance dynamic range and in particular its maximum displayable luminance (PB_D). Background Art

[0002] A few years ago, new techniques for high dynamic range video coding were introduced, notably by the applicant (see for example WO2017157977).

[0003] The coding and processing of HDR video contrasts sharply with the traditional video coding according to which until a few years ago all videos were coded, nowadays called Standard Dynamic Range (SDR) video coding (aka Low Dynamic Range video coding; LDR): PAL in the analog era, and e.g. Rec.709MPEG2 in digital. In fact, starting from the need for brighter and possibly also darker image objects that can be coded (i.e. a greater range of pixel luminances in the starting, master HDR image created by the content creator), all the rules of video technology from which one experiences it have been revisited one by one and often reinvented.

[0004] With regard to encoding, the difference between HDR and SDR is not only a physical one (more different pixel brightnesses to be displayed on displays with greater dynamic range capabilities), but also a technical one in the development of different luma code allocation functions (OETFs), additional dynamically-per-image-changing metadata that specifies how to rescale the various image object pixel brightnesses to obtain an image with a secondary dynamic range different from the initial image dynamic range, etc.

[0005] The Rec.709 SDR luminance (luma) code definition, due to its approximately square root OETF function shape (luminance: Y = sqrt(Luminance L)), is capable of encoding (using 8 or 10 bits of luminance) only about 1000:1 of luminance dynamic range, which is encoded for the typical rendering capabilities of all displays at the time. Although no one in the SDR era was willing to specify or use a maximum luminance, that is, coding Peak brightness (brightness) PB_C, but the brightness of various LDR displays on the market in the 20th century was between about 0.1 the darkest display brightness (simply "black")) and 100 nits ("white"), the latter being the so-called show Peak luminous brightness (PB_D).

[0006] The first HDR codec to be introduced to the market, the “HDR10” codec, used e.g. to create the new black-band jewel box HDR Blu-ray, simply changed the OETF to a more logarithmically shaped perceptual quantizer (PQ) function standardized in SMPTE 2084, which allows for the definition of more luminances, i.e. between 1 / 10000 nit and 10000 nits, which is sufficient to meet all practical HDR image specifications for video products (e.g. movies, TV broadcasts, etc.).

[0007] HDR should not be overly simplistically confused with a large (larger) number of bits for the luma codeword (e.g. 16 bits instead of 8). This may be true for linear systems like the number of bits in an analog-to-digital converter, but since the code allocation function can have a very non-linear shape, it is now practical to define an HDR image with 10 bits of luma, which leads to advantages in reusability of deployed systems (e.g. an IC may have a certain bit depth, or a video cable, etc.).

[0008] After calculating the luminance, there is only one 10-bit pixel plane (or, more precisely, a 10-bit plane with the two chrominance planes Cb and Cr 3), which can classically be processed further down the line "as if" they were a mathematically SDR image, e.g. with MPEG-HEVC compression or the like.

[0009] Of course, the receiver should know that it is getting an HDR image instead of an SDR image, otherwise it will be rendered incorrectly. For example, if only linear mapping ( coding If the image maximum brightness PB_C is mapped to the SDR display peak brightness PB_D = 100 nits), then an image with PB_C = 1000 nits will appear 10 times too dark, which means that night scenes become unwatchable. It is possible to make brighter pixels by using another brightness map, but usually you should be very careful about best practices.

[0010] Due to the logarithmic nature of the PQOETF, HDR10 images are in principle viewable (if the luminance codes are interpreted as normal SDR codes, i.e., they are displayed after applying an approximate square power EOTF), but have an ugly degraded contrast, making them look especially washed out and with incorrect light levels (e.g. a criminal who is supposed to be hiding in the dark will suddenly look as if he is illuminated by a ceiling light).

[0011] The problem with just encoding and processing such a "pure HDR" video image (also called an HDR master grade, the term grade indicating how bright the various scene objects / pixels should be in an image representation with a PB_C of e.g. 1000 nits (or more) in order for the HDR scene to look its best in that representation) is that it will only render the image in a HDR master image with a PB_C of e.g. 1000 nits (or more) same Because PB_C of both the display and various content tends to vary ("cannot be controlled"), it soon became apparent that, for example, HDR Blu-ray discs do not always appear perfect, and in this case, night scenes may be unwatchable.

[0012] Furthermore, more advanced HDR video encoders encode two different dynamic range graded images, i.e. the graded HDR scene: an image of higher dynamic range, e.g. PB_C=5000 nits, and an image with lower dynamic range, typically an SDR image with PB_C=100 nits, since this image can then be immediately displayed on a conventional LDR display. The reader should understand, as will become clear below, that transmitting two differently graded images to a receiver does not mean actually transmitting two images, i.e. two sets of DCT transform matrices of YCbCr pixel colors: if one is transmitted together it allows all the mathematical information of the secondary image to be calculated from one of the pair (per video image presentation time instant) which is actually transmitted as, e.g., an MPEG-HEVC image.

[0013] Real-world scenes (although evenly lit scenes have only a smaller dynamic range due to the 100:1 object reflectivity ratio) can have quite high dynamic range. For example, a cave with a small opening, with sunlight outside, could be graded on a 10,000 nit PB_C reference representation, which contains an HDR grade of the scene suitable for home TV viewing, containing brightnesses well below 1 nit for the cave pixels, and up to 10,000 nits for at least some of the outdoor pixels. Such challenging HDR images are not so simple to translate to quite low dynamic range (e.g., at least 100x when going to SDR), especially if the content creator wishes to convey a still fairly similar HDR look, also in the SDR grade, but with Figure 1 In principle, this can be done.

[0014] For the convenience of the reader and to bring him quickly up to speed on some of the aspects involved, Figure 1 Several prototype illustrative examples are shown of the many possible HDR scenarios that a future HDR system (e.g. connected to a 1000 nit PB_D display) may need to be able to handle correctly, i.e. by rendering appropriate brightness for all objects / pixels in the image.

[0015] We showed typical problems involved in dynamic range mapping, such as going from a larger dynamic range of pixel brightness to a smaller one (where we can assume that at least the peak luminance capability will change). One can compare this problem to the task of packing a large number of items in smaller and smaller suitcases. In the largest suitcase, you can throw everything in there in a mess, and it will fit anyway. For a medium-sized suitcase, some optimizations may need to be decided, for example one may find that if clothes are packed tightly in smaller compartments, such as separate bags, they may take up less space and everything can be easily repacked. When using only the smallest suitcase, a new set of more severe actions may be needed. For example, if a book about the destination is packed, when the same information is better explained in another book, one may tear out some pages of the second book, again somehow reducing the amount of things to be packed (then it is "comparable things", but not as much as what is packed in the largest suitcase). If only clothes are packed, it is better not to adopt the best written material compression method described above, because it is better not to tear out parts of expensive clothes. Video coding technology is similar in that a stable, universally applicable, and fast-to-compute method had to be found to do “image packaging” on the fly, even though the needs of various sub-ecosystems might differ (e.g. high-artistic-quality Hollywood films, vs. crudely produced news material, perhaps even from lay contributors).

[0016] For example, ImSCN1 is a sunny outdoor image from a western movie (characterized by mostly bright areas or even only bright areas, which ideally should be rendered a bit brighter on an HDR display than on a 100 nit display to give more of a sunny look rather than an overcast and rainy day look, e.g. with an average brightness of say 500 nits), while ImSCN2 on the other hand is a nighttime image.

[0017] What makes one photo sunny and another dark? Not necessarily relatively Brightness, at least not in the SDR paradigm. What makes HDR image rendering different from the SDR era is that the dynamic range of SDR is so limited (about PB=100 nits, and minimum black (MB) level of about 0.1 to 1 nit), most of the time only the intrinsic reflectivity of the object can be displayed in SDR (which will be between 90% good white and 1% good black). This will be good for identifying objects (with a certain reflective brightness, and of course their chromaticity) under uniform technically controlled lighting, but not so good for conveying the beautiful variations in lighting itself that people can have in natural scenes and what effect that will have on the viewer.

[0018] If the display allows, and therefore the image encoding and processing technology should do so, one would actually see the sun shining through the trees when walking in a forest, i.e., instead of a more yellowish impression of some patches as on an SDR display, one would expect to see bright and colorful sunlit clothing as one walks from shadow to sunlight. And fires and explosions should also have optimal visual effects, at least as far as PB_D allows.

[0019] In SDR, it is possible to make night images somewhat darker in the brightness histogram, but not too dark, otherwise it will just render too dark and uglify the image (and probably make it at least partially unwatchable). Furthermore, in a 100 nit TV or 100 nit encoding, there is no room for anything too bright. Therefore, one must display objects independently of their illumination, and cannot simultaneously faithfully display all of the sometimes highly contrasting illuminations of the scene that may occur. In practice, this means that a highly bright sunny scene must be rendered using roughly the same display brightness (0-100 nits) as an overcast and rainy day scene. You have to figure out what is going on from other cues, such as the viewer's expectation that a cactus may be illuminated by sunlight and therefore bright. Even night scenes cannot be rendered too dark, otherwise the viewer will not be able to distinguish the darkest parts of the image very well, so again these night brightnesses are rendered in the range of about 1 to 100 nits. A traditional solution is to color the night scene blue, so that the viewer understands that he is not looking at a daytime scene. These are actually severe limitations of SDR imaging (other examples are e.g. clipping everything outside a window to a single white, so there's nothing to see there either), but the fact that audiences and industrial users are somehow "used to it" doesn't mean there isn't room for improvement.

[0020] Now, of course, in real life human vision also adapts to the amount of light available, but not as much (most people in real life do realize it's getting darker, or they're in a darker or very bright environment). Also, technical TV systems don't adapt the same way the human eye and brain do, and adapting to home display content isn't the same as adapting to a pristine scene in the desert.

[0021] So one would like to render images with all the spectacular local and temporal lighting effects that can be artistically designed to get more realistic rendered images, at least if one has an HDR monitor available. Exactly what is the appropriate luminance for a lightsaber in a dark room, we will leave it to the color grader creating the master grade to decide (when we say color grader, we mean the equivalent concept for each ecosystem, so not necessarily humans spending a lot of time defining pixel luminance for master HDR and SDR images, but also automated grading for live broadcasts, etc.), and this application will focus on the technical elements needed to create and process such images.

[0022] Figure 1 The left axis is the brightness of objects one would expect to see in a 5000 nit PB master HDR grade for a 5000 nit PB_D display (i.e. the grader makes the image assuming that a typical high quality HDR TV at home will have 5000 nit PB_D, and he might actually be sitting in such a representation of a home viewing room, and grades on such a 5000 nit PB_D grading reference display). If one wants to convey not just an illusion, but the real feeling of a cowboy in a bright sunlit environment, then these pixel brightnesses must be specified and rendered bright enough (but not too bright to be annoying, which is a typical pitfall of HDR image creation and processing), e.g. around 500 nits (compared to the previous scene, e.g. an interior scene). For night scenes, one would mainly want the brightness to be darker, but the main character on the motorcycle should be easily identifiable, i.e. not too dark (e.g. around 5 nits), while there can be quite high brightness pixels, such as street lamps, e.g. around 3000 nits on a 5000 nit display, or around peak light brightness on any HDR display (e.g. 1000 nits).

[0023] The third example, ImSCN3, shows what can now also be achieved on HDR displays: very bright and very dark pixels can be rendered simultaneously. It shows a dark cave with a small opening through which sunlight can be seen outside. For this scene, one might want to make objects like trees that are lit by sunlight slightly less bright than for scenes where one wants to render the impression of a bright sunny landscape, e.g. around 400 nits, which should be more in harmony with the essentially dark features of the cave interior (since one also does not want scattering in the human eye to visually deteriorate objects in a dark cave). The color grader might want to optimally coordinate the brightness of all objects (already in the PB_HDR=5000 nit master HDR image) so that nothing looks out of place dark or light and the contrast is good, e.g. a person standing in the dark in this cave might be encoded in the master HDR graded image at around 0.05 nits (assuming that the HDR rendering is capable of rendering not only bright highlights but also dark areas).

[0024] But now the question is, if one has these master HDR object pixel luminances, where in the dynamic range should they be, say the dynamic range ends at 1500 nits (all luminances below 1500 nits can also be faithfully represented in this smaller dynamic range, but what about pixel luminances above 1500 nits? In typical behavior they will all be clipped to the same 1500 nits PB_C value, which is far from ideal).

[0025] Thus, as in our suitcase packing analogy, in principle the content creator has a task to define a large number of re-graded images starting from a 5000 nit HDR master image, e.g. a 3000 nit PB_C image, 2000, 1500, 1000, 750, 500, 300 and 100 nit images. Such a large task is of course not commercially viable. Therefore, the applicant has invented a technique that allows the content creator to grade only his master HDR image and the corresponding SDR image, and all intermediate images can be determined based thereon by an automatic technique called a display optimization system (see WO2017108906).

[0026] Just to illustrate some technical video encoding possibilities, in order to clarify some components of the inventive concepts described in detail below, which are important for a good understanding, we describe an exemplary HDR video encoding system designed by the applicant for HDR images, and in particular HDR video encoding (whereby the reader should understand that the principles of the invention are also applicable to other systems than the exemplary system chosen for simplicity of explanation).

[0027] This video encoding system can not only handle the communication (encoding) of a single standardized HDR video (e.g., a 10-bit perceptual quantizer used as a luminance code defined for EOTF encoding) for a typical single type of display in the field (e.g., the image is defined as PB_C=1000 nits, assuming that each end viewer has a 1000 nit PB_D display), but it can simultaneously transmit and process videos that have optimal appearance / grading for a variety of other display types that may be in the field with various other peak luminances, especially SDR images for 100 nit PB_D SDR displays.

[0028] That is, although in such an HDR video communication system only one type of graded image is actually communicated as the transmitted pixelated image, which has various variants, we will illustrate in this example one of communicating an SDR image to a receiver via any video communication system (but a variant of communicating an HDR image may also be used), because one or more brightness remappings, i.e., regrading functions, defining the HDR image pixel colors, in particular the brightness from those SDR image pixel brightnesses, can also be added to the metadata, and people transmit HDR images at the same time and also look for scenes (there is actually no need to transmit an HDR image, such as dual image communication, or at least a second layer of pixelated HDR image data).

[0029] In addition, a set of appropriate reversible color transformation functions F_ct is defined on the encoding side, for example by a human color grader, such as Figure 2 shown.

[0030] These functions define how to obtain a plausible looking SDR image (Im_LDR) corresponding to this HDR master image MAST_HDR from the master HDR pixel luminances (or equivalently their luminance codes), while ensuring that the original master HDR (MAST_HDR) image can be reconstructed with sufficient accuracy as a reconstructed HDR image (Im_RHDR) at any receiving end by using the inverse function IF_ct. The IF_ct function can be determined from the transmitted forward HDR to SDR mapping F_ct function, or the system can even transmit the IF_ct function directly, for example via the MPEG mechanism of SEI messages, or any other suitable metadata communication mechanism.

[0031] The color converter 202 typically applies a F_ct brightness mapping of the brightness of the pixels of the master HDR image (MAST_HDR), which we assume are normalized so that the maximum brightness is 1.0 (note that both HDR and SDR color domains can then be covered, which means that the DR brightness transformation corresponds to an upward or downward shift of colors in the normalized color domain of the encodable colors). In order to understand the concept of the present invention in a simple way, for the sake of simplicity, it can be assumed that the F_ct HDR to SDR brightness mapping is a 1 / 4 power brightness mapping function (L_out_SDR = power (L_in_HDR; 1 / 4)) for deriving the normalized SDR output brightness of the pixels of the 100 nit PB_C SDR output image Im_LDR (i.e. Figure 1 right brightness).

[0032] Since the receiver must be able to reconstruct the master HDR image from the corresponding SDR image received, or at least an approximate reconstruction, but with some compression-related artifacts, in addition to the actual pixelated image, the color mapping function must then enter the video encoder 203. Without limitation, we can assume that the video is compressed by this video encoder 203 using an MPEG HEVC video compressor, producing a coded (SDR) output image Im_COD, and that the function is stored in metadata, for example via a SEI mechanism or similar techniques.

[0033] Thus, following the actions of the content creation device 221, the rest of the communication chain, from the perspective of the image communication technology, pretends that it gets a "normal SDR" image as input. So, for example, the transport formatter 204 can apply all the necessary transformations to format the data to pass through some transmission medium 205 (e.g. channel coding for storage on a BD disc, or frequency coding for cable transmission, etc.).

[0034] The image data is then transmitted via some transmission medium 205, such as satellite or cable or the Internet, for example according to ATSC 3.0, or DVB, or any video signal communication principle, to one or more receiving ends, which can be consumer video devices such as televisions, or set-top boxes, or professional systems such as cinema receiving units, etc.

[0035] At any consumer or professional end, a receiver deformatter 206, which may be incorporated in various physical devices such as a set-top box, TV or computer, removes the channel coding (if any) by applying deformatting and channel decoding. A video decoder 207 then applies, for example, HEVC decoding, generates a decoded SDR image Im_RLDR, and unpacks the color transformation function metadata F_ct. A color converter 208 is then arranged to convert the SDR image to any non-SDR dynamic range image (i.e. PB_C is above 100 nits, and typically at least 4 times higher).

[0036] For example, a 5000 nit original master image Im_RHDR can be reconstructed by applying the inverse color transform IF_ct of the color transform F_ct used on the encoding side to generate Im_LDR from MAST_HDR. However, a display adaptation unit 209 may also be included, which transforms the SDR image Im_RLDR to a different dynamic range, for example, Im3000 nits is optimally graded in the case where the display 210 is a 3000 nit PB display, or a 1500 nit PB or 1000 nit PB image, etc. We have assumed, without limitation, that the video decoder and the color converter are located in a single video redefinition device 220. It will be appreciated by a reader in the art that one may similarly design a topology that transmits, for example, an HDR image with PB_C=10000 nits, and for a corresponding TV or display, the color converter causes the output HDR image to have, for example, PB_C=2500 nits, and the various units may be operated via a network connection, for example, on different servers, and so on.

[0037] The existing technical components (according to the innovative components of the present teachings and / or the existing technical components that they can be connected, cooperate, integrate, etc.) can be embodied or implemented as various technical systems typical in image or video technology, that is, for example, in various hardware devices. For example, the video re-determination device 220 can have any technical video supply output 231, such as an HDMI cable that can be connected to a TV display, etc. (also such as a storage device, etc.; or even a network cable for transmitting the output image Im_RHDR, corresponding to Im3000 nits, to another potential remote device or system, etc.). Depending on the physical variant chosen, there may be an image or video output signal formatter 230 that converts the image into a single image suitable for any technical situation (e.g., although we clarify below that the core chrominance calculations may, for example, produce as output a linear R, G, B representation of the pixel color, depending on the technical configuration, the final image signal (I_out) sent to, for example, the display 210 may, for example, be HLG formatted, uncompressed, or MPEG or AV1 compressed, etc., and the signal formatter 230 may include units such as a typical integrated circuit responsible for handling all such signal derivation processes (whether fixed as a single option, or configurable).

[0038] Figure 3 Only one example has been illustrated in which it is advantageous to recognize what is typically the case for the color converter 202 (and accordingly, the color converter 208) of the present invention, and that this particular color computation core can elegantly implement all variants as one configurable core, but the elements can also be designed in various different topologies as long as the innovative contribution of the present innovation to the technology is considered. The color is input in a color format that is very classic for video color representation: YCbCr (by the way, in various embodiments this space can be defined according to various nonlinear R, G, B color triples, such as gamma-2.0 definition, PQ definition, etc.).

[0039] The chroma multiplier determiner 301 determines an appropriate multiplicative scaling factor based on the luminance (Y) value of any consecutive image pixel being processed, which is used by the multiplier 302 to multiply the scaling factor s(Y) with the two input color chromaticity coordinates Cb and Cr, i.e. the output red chromaticity Cro=s(Y)*Cr of the input color, and the output blue chromaticity Cbo=s(Y)*Cb, using the same s(Y) factor to keep the hue of the output color the same as the hue of the input color (while appropriately affecting the saturation of the pixel color). The chroma multiplier determiner 301 can be arranged to read from metadata, which is typically communicated in conjunction with the transmitted SDR (or other) image MET (F_CLUT), which contains, for example, a LUT of S factors for each possible Y value that the SDR image may have. The output chrominance values ​​Cro and Cbo are input to a matrix calculator 303 together with the luminance Y, which uses standard 3x3 matrix coefficients (according to known chromaticity methods, depending on the selected color primaries, such as Rec.709 or Rec.2020, etc.) to calculate the normalized RGB representation therefrom, i.e. the normalized red input component RnI, the normalized green input component GnI and the normalized blue input component BnI. These will be converted to the (normalized) required output RGB values ​​RnO, GnO, BnO, i.e. in this case the (normalized) HDR reconstruction components.

[0040] Given any input luminance Y that the pixel being processed happens to have, this, e.g., brightening of the pixel triplet is achieved by multiplying the three input components by the same luminance multiplier value g(Y). We have shown in the prior art above that any given normalized luminance can be converted to a luminance mapping function shape (e.g., a parabola starting at (0,0) and ending at (1,1)) into a corresponding set of g multipliers for all possible normalized luminances Yn. This action is performed by a luminance multiplier determiner 304, which reads as input metadata MET(F_PLUT) transmitted by the content creator, which encodes the shape of the brightness mapping curve, or equivalently the relative brightness mapping curve, or whichever is selected.

[0041] After the multiplier 305, the pixel brightnesses are correctly moved to their HDR_reconstruction relative positions, although still normalized to the 1.0 color space. Finally, the output color calculator 306 can do all the necessary calculations to format the color according to the needs of the technical output component, such as a display connected through, for example, an HDMI cable, a wireless communication channel, etc. It can determine the output color format RGB_DF as, for example, a PQ-RGB format, but it may also apply, for example, various optimizations when it knows that the connected display is of a certain physical type (it may not do or do something different if the output is, for example, a hard disk record for storage for later viewing, etc.), but these details are not relevant to illustrate the present invention.

[0042] Figure 4 It is clarified that for a content creator who may have optimized the contrast ratio for a particular HDR scene image (for example, an image in which the left portion is relatively dark, and therefore requires relative brightening in the standardized SDR representation of the darker area, so that a person lurking in the shadows is semi-visible in both the HDR image display and the SDR image display; while a person in the fog has sufficient contrast in the brighter parts of the scene and its successive images, this may result in a hyperbola that optimizes the contrast of the two areas, such as Figure 4 FL_50t1_2 curve shown) for any choice of re-grading function shape (and case; encoding example), showing how optimization (aka display adaptation) works.

[0043] Check Figure 4 A, assuming that the current image (or a sequence of consecutive images) is such that (e.g. a received HDR image) is optimally regraded into a corresponding SDR image via a particular normalized brightness mapping function FL_50t_1 (note that a skilled reader in the video art can understand how to formulate teachings in normalized brightness and normalized luminance simply by changing the axes and corresponding curve shapes).

[0044] Without wanting to restrict ourselves unnecessarily, we will continue to assume that the examples explained so far use axes that are all transformed to perceptually uniform brightness (i.e., equally spaced steps on the horizontal and vertical axes correspond to approximately perceptually equal differences in brightness).

[0045] According to the applicant, the following formula can be used to convert brightness (or any number of color coordinates, such as the linear contribution of the red primary, if desired) to this perceived uniform luminance v:

[0046] v(Ln; Lmax)=log[1+(RHO(Lmax)-1)*power(Ln; 1 / 2.4)] / log[RHO(Lmax)]

[0047] RHO(Lmax)=1+32*power(L / 10000;1 / (2.4))[Formula 1]

[0048] In these formulas, L is the (normal, absolute, nits == Cd / m^2) brightness of the pixel; Ln is the maximum brightness normalized to 1.0, i.e., by having Lmax equal to the peak luminance of the encoded image PB_C, e.g. 5000 nits, and then dividing: Ln=L / PB_C.

[0049] So if we need to map, for example, 5000 nits of content to SDR brightness, Figure 4 Shown on the vertical axis in B, will be the output luminance v_SDR, which can be converted to brightness by using the inverse of Equation 1. We show this so that due to the approximately logarithmic nature of the visually uniform luminance representation of brightness, the tens before PB_C_SDR = 100 nits are approximately equidistant on the vertical axis, i.e., v = 0.6 corresponds approximately to 10 nits, and so on.

[0050] Similarly, if we have a 5000 nit PB_C defined HDR input, the normalized luminance on the horizontal axis will correspond to approximately equidistant 0; 1; 10; 100; 1000; the end point 5000 being a bit closer to where the 10k position falls off.

[0051] Corresponding to this axis system, the content creator can then define the shape of the brightness remapping function for the current image as needed (for example, a dark street with some street lights requires some brightening of the darkest parts of the houses or bushes on the street to keep them sufficiently visible in the darker PB_C image corresponding to the main HDR image, for rendering them on the darker PB_D display, which can be seen by the greater than 45 degree slope of the dark end of the first exemplary brightness mapping (equivalent to brightness mapping) curve FL_50t1_1), as well as a display adaptation strategy that automatically derives the shape of the curve required to map 5000 nits brightness to, for example, 650 nits brightness (in the case where a 650 nit PB_D connected display needs to provide a suitably optimized / re-graded version of the received main HDR image, or reconstruct it in the case where an SDR representative image is transmitted and received), as we will now explain.

[0052] While there can be multiple display adaptation variants, all of which will work with the current innovative technical additions of this patent application, in order to keep a complex discussion simple, we assume that the specific display adaptation mechanism used is the one jointly standardized by applicant Royal Philips (Koninklijke Philips) and Technicolor in ETSI TS103 433-2V1.1.1 (2018-01), and for present purposes, we briefly re-summarize it and be more general.

[0053] The idea is that a content creator, such as a human color grader, determines the mapping curve shape FL_50t_1 or at least one of the master HDR and SDR graded images (the technical reader will appreciate that if the grader only makes two images, the applicant can also use techniques to deduce how the brightness of said SDR image relates to one of the HDR images through an automatically derived FL_50t_1 function, but again these details are not relevant enough for the current discussion, so we assume that the grader, for example, draws the shape of the FL_50t1_1 curve using a color processing user interface tool and checks whether this indeed gives an SDR image that looks correct, or he changes the function again until it has a shape that produces the desired SDR image output), without much time available.

[0054] So he doesn't want to make lots of different regrading curves (e.g., having made a master HDR image of optimal quality at 5000 nits, he doesn't want to make a function of how to optimally regrade to 2000 nits, and another function of how to regrade to 1000 nits, because since these functions operate on the same master HDR image, with the same image object, with vastly comparable luminance regrading needs, usually the two functions will look relatively similar: a 5000 to 2000 mapping curve will perform a "slightly weaker" normalized luminance shift than a 5000 to 1000 mapping, etc.).

[0055] So, while theoretically a lot can be said about the various re-grading functions and images, from a technical practical point of view one could argue that for most, if not all, HDR images a sufficiently good quality medium dynamic range (MDR) image can be obtained for a PB_D display between 100 nits and PB_C_master_HDR, which in this case is 5000 nits, if an automatic display adaptation algorithm is used which automatically calculates an intermediate function for mapping v_HDR to v_MDR (such as the first exemplary display adaptation function F_DA50t6_1 used to create a 600 nit PB_C MDR image), which reduces the workload for the grader since he knows that only one single HDR to SDR mapping function, such as the first SDR mapping function FL_50t1_1, needs to be produced (of course, for each different prototype image case, as a cave will require a very different luminance mapping function shape than, for example, a barber shop with a blue light panel).

[0056] The display adaptation unit 401 (typically an integrated circuit that performs the technical mathematics explained in the ESTI standard; or an equivalent system) will then calculate the required first exemplary display adaptation function F_DA50t6_1 for mapping the HDR brightness to the appropriate 600 nits PB_C MDR brightness based on the input brightness mapping function and the PB_D value (600 nits in the example) (i.e., accurately respecting the specific brightness re-grading requirements of the current image, as transmitted by the content creator as the first SDR mapping function FL_50t1_1co transmitted together with the encoded HDR image in the metadata as received by the demodulator 206). Similarly, if for another image the second SDR mapping function FL_50t1_2 has a different shape as shown (typically with the same PB_D value), then the correctly optimized second exemplary display adaptation function F_DA50t6_2 has the same general shape (indicating specific re-grading needs for different areas or objects in the image, such as a dark corner, an area on a table under a strong light, etc.), but will result in a display that is a little "weaker / in-between" by an appropriate amount to take into account the differences between the PB_C_master_HDR and PB_C_MDR=PB_D_available displays.

[0057] Figure 4How this automatic display adaptation generally works in an elegant way is shown in B. For any point on the diagonal, corresponding to some input luminance, for example for the first input luminance VnH1 or the second input luminance VnH2, a metric direction can be followed, for example orthogonal to the diagonal, until it falls on an input function, in the example the second SDR mapping function FL_50t1_2. This point corresponds to this (tilted 45 degrees to the left relative to vertical upward) metric to 100 nits position, because it corresponds to the function required to re-grade the master HDR input luminance to a corresponding 100 nits output luminance. The PB_C_master_HDR position, in this case 5000 nits, corresponds to the diagonal (starting) position (as the skilled reader will understand, because the re-grading of Ln_5000 to Ln_5000 corresponds to an identity transformation). If a metric is now defined for the intermediate positions, for example a logarithmic metric, so that starting from the 100 nit position on the input luminance mapping curve there are larger steps and towards the diagonals smaller steps, said metric can locate all positions of the desired PB_C (e.g. 600 nits), thus producing all points of the curve for a non-dashed display, and thus this is the desired shape of the second exemplary display adaptation function F_DA50t6_2. The only thing that must then be done is to send this ("new" / optimized) curve shape to the luminance multiplier determiner 304, which can then be used Figure 3 The color calculation engine shown in Figure 1 is used to calculate all pixel brightness of the 600 nit MDR image starting from the master_HDR image pixel colors as input.

[0058] The system works well and has been demonstrated satisfactorily many times on many types of HDR content. However, one problem with it is that it is rather static, due to its automatically fixed algorithmic nature: in addition to the variant chosen for clarification, one could use another variant, for example one that uses another metric, i.e. the distance between the input luminance mapping function position along the diagonal and the 100…PB_C_master_HDR position, a different positioning, or another orientation other than 45 degrees of the metric axis along the diagonal distribution, or more complex changes to the display adaptation algorithm, but given any such chosen algorithm, the algorithm can usually be embedded in the integrated circuit of the receiving device that performs the display adaptation, since reconfiguration on the fly may be considered quite cumbersome, the display adaptation function result will always remain "fixed" unchanged. This may at least be inconvenient for some customers, as they may still find some images, for example, too dark, or too low in contrast, etc., at least for some of the possible display peak luminance range, such as PB_D < 350 nits. It is important to understand that one may want to keep the grader's luminance mapping function unchanged as a constraint of the technical framework. One could argue that if brighter images are wanted, content creators should first create a steeper function FL_50t1_1, assuming there will never be any other issues, such as the possibility that the display on a brighter display (e.g. PB_C>2500) might again be slightly or more than slightly too bright, or other issues with the display being near the top of the gamut, etc.

[0059] But a content creator might argue that these functions are "his gold". For the total set of all MDR images, the general specification on which all secondary processing is based might be seen as too important to modify, or in any case essentially "just right" (i.e., maybe someone might find a display a little too dim under specific conditions, but that doesn't mean that the reference display of a 100 nit image in particular is incorrect, let alone that the 100 nit image itself, i.e., as an image that defines content, is incorrect, nor does it mean that the content creator optimized that image in a way that is not correct). Therefore, from a practical point of view, we would like to have a simple way to adjust or improve the generation of MDR images on the receiving end by changing the display adaptation, while keeping all the incoming image information (i.e., including the SDR mapping function received in the SEI metadata for each consecutive video image) unmodified. Summary of the invention

[0060] A practical method is implemented through an image pixel brightness adaptation device (500) to solve the static nature of display adaptation in the prior art and provide some further customizability at the receiving end, including:

[0061] a connection (501) to a contained or connectable video decoder (207), the video decoder being arranged to receive a coded high dynamic range image (Im_COD), the coded high dynamic range image being coded according to a first maximum codable brightness (PB_H), and the video decoder being arranged to receive metadata specifying at least one brightness mapping function (F_ct; FL_50t1_1), the at least one brightness mapping function specifying an offset of brightness of a secondary image corresponding to the coded high dynamic range image compared to the brightness of the same pixel position as the pixel position coded in the coded high dynamic range image, the secondary image having a second maximum codable brightness (PB_S), the second maximum codable brightness being less than or greater than the first maximum codable brightness (PB_H), and the video decoder being arranged to output a decoded high dynamic range image (Im_RHDR) and the brightness mapping function;

[0062] a display adaptation unit (401) arranged to receive a value of a display maximum brightness (PB_D) that a specific display can display as a brightest pixel color and an input brightness mapping function, and the display adaptation unit is arranged to apply an algorithm, the algorithm calculating at least one display adaptation brightness mapping function based on the input brightness mapping function and the display maximum brightness (PB_D), characterized in that the at least one display adaptation brightness mapping function corresponds in shape to the input brightness mapping function but is closer to a 45 degree increasing diagonal on a perceptually uniform axis of a graph of the input brightness mapping function, depending on the difference between the values ​​of the display maximum brightness (PB_D) and the first maximum encodable brightness (PB_H) relative to the difference between the second maximum encodable brightness (PB_S) and the first maximum encodable brightness (PB_H);

[0063] The device is characterized in that the image pixel brightness adaptation device comprises an alternative brightness mapping function determination unit (502), which is arranged to determine an alternative brightness mapping function (ALT_FL_50t1_1), and

[0064] wherein the display adaptation unit (401) comprises a combining unit (503) arranged to combine the at least one brightness mapping function (F_ct; FL_50t1_1) and the alternative brightness mapping function (ALT_FL_50t1_1) into a combined brightness mapping function (CMB_FL_50t1_1), and wherein the display adaptation unit is arranged to apply its algorithm as an input brightness mapping function to the combined brightness mapping function;

[0065] The image pixel brightness adaptation device comprises a brightness mapping unit (510) arranged to receive pixel brightnesses of a decoded high dynamic range image (Im_RHDR) and to apply an adapted combined brightness mapping function (ADJ_F_DA50t6_1) to these pixel brightnesses to obtain an output brightness of an output image (Im_DA);

[0066] The image pixel brightness adaptation device comprises an output image or video communication cable or wireless channel to which a display can be connected, and an output signal formatter (230) arranged to send the output image (Im3000 nits).

[0067] Firstly, with regard to the algorithmic or hardware details of the display adaptation unit, the experienced reader is informed that there may be a number of alternative ways to calculate the general technical essence of the HDR to PB_D optimized medium dynamic range display adaptation, namely based on determining a function closer or further away from the HDR to SDR re-grading brightness mapping function, depending on where the value of the display maximum brightness (PB_D) is located between the first maximum encodable brightness (PB_H) and the second maximum encodable brightness (PB_S), for which the display adaptation unit (401) must calculate the best adapted image and its pixel brightness or luminance, and thus similarly the proximity of the resulting display adaptation brightness mapping function to the diagonal. Furthermore, there may be some variations in how exactly to define (approximate) perceived uniform brightness: although such a function will essentially be approximately logarithmic, as shown in the example, the parameters of the logarithmic function may be varied and the algorithm will still work the same and produce an image that looks good. Therefore, the skilled person understands that these details do not constitute the essence of our current innovative contribution and may be varied while providing the same type of recognizable device. It is also understood by those skilled in the art that the device may be a stand-alone image color optimization device, i.e. it will be operably connected to a separate video decoder (which will obtain and decode to provide all data as required, i.e. image pixel color data, which we may assume, for example, in linear RGB representation, or for simplicity of illustration in YCbCr, which is calculated using a well-known matrix based on R, G, B non-linear component values, which may be PQ encoded, for example; and metadata, which comprises at least one brightness mapping function used by the final brightness mapping calculation algorithm of the display adaptation unit), or the device may be an all-inclusive overall system (which may, for example, be entirely embodied as a television display, etc.). The second (graded reference) image is the image at the other end of the desired dynamic range to be covered, so for e.g. a 4000 nit PB_C master HDR image this is typically a 100 nit PB_C SDR image, no matter which of the two is actually transmitted (although it is also possible to calculate the display-adapted final MDR image starting from an SDR image, for simplicity we will assume in the description that it is calculated from the HDR image, i.e. to a lower dynamic range (or at least maximum brightness, assuming the minimum black of the input and output images are the same), unless there is an extrapolation to a higher dynamic range that is more impressive than the image originally created). Typically it makes sense to have the dynamic range of the two reference gradings of the content creator (i.e. typically their PB_C values) differ by at least a multiplication factor of 4, otherwise there is not much point in making at least technically high quality display optimizations, although in principle the same principle can be applied even if the differences between the various image gradings are smaller. The display maximum brightness can again be input to the display adaptation unit (401) in various ways, depending on the technical implementation variant.For example, the set-top box may poll which of the various displays is connected, which may then communicate its value of PB_D back to the STB (or the user may input a value which he considers to be a good working value for his TV via the STB user interface, etc.) before commencing color optimization and outputting an image or video to the connected TV display. On the other hand, if the device is the TV itself, then it's unique PB_D value (e.g. relating to the backlight behind an LCD display panel, or a value which the TV manufacturer considers safe to use without overheating an OLED panel, etc.) may have been stored in an unmodifiable memory within the display adaptation unit (401), i.e. in contrast to our general purely schematic illustration, in which case the unit 401 receives the PB_D value from itself. Again, this mutable aspect is not really important for identifying when a device is of a type as described in the present invention and herein.

[0068] The alternative brightness mapping function determination unit (502) may use a number of (simple) fixed strategies to determine a good alternative brightness mapping function (ALT_FL_50t1_1). For example, for many applicants of the technology, it may be sufficient to do something as simple as brightening the dark, e.g. by Figure 7 The Para Shadowgain ("Para" is the name of a specific brightness mapping curve with a linear slope at the dark end 0,0 and the bright end 1,1, with a smooth parabola segment connecting those linear parts in the middle) alternatives shown, or simple contrast modification, etc., but now, depending on the PB_D, i.e., rather than "on the content" (i.e., a pair of extreme HDR and SDR reference gradings, and the brightness mapping function linking them), can be controlled according to which part of the span of the PB_D-enabled display needs more or less correction. There is an advantage to doing the adjustment before the display adaptation is fixed instead of after (e.g., it is also possible to display the Para received from the content creator in the metadata of the communication signal and the alternative para proposed by the receiving end, i.e., the TV itself, and then correct the display adaptation's para using the display adaptation alternative), because doing this before still has the function of passing through the standard display adaptation behavior, such as the logarithmic position of the metrics of the various PB_Ds, which vary more around SDR requiring a lot of adjustment, and have an "almost perfect HDR" behavior for high PB_D displays.

[0069] When there can be multiple functions, a skilled reader should not have much difficulty understanding (generally, it should be understood that the claims are written with the least detailed embodiments, i.e., only one image is processed, only one original, content creator-derived brightness mapping function, and only one alternative brightness mapping function, and only one final, optimal display-adapted brightness mapping function is derived, ready to be loaded or applied to the color processor to obtain an optimal MDR image corresponding to the HDR master image in appearance (i.e., the relative brightness positions and general color impression of various image objects) for a specific unique display of the PB_D that needs to provide a display-adapted MDR image.

[0070] Multiple different luminance mapping functions may be created for successive images to be presented one after the other in time, and the display adaptation unit would then make multiple successive display adapted luminance mapping functions corresponding to each of the input luminance mapping functions, respectively, as they are created by the content creator. Multiple output display adapted luminance mapping functions may also be made for a single input luminance mapping function if the display adaptation unit needs to serve two output video streams, e.g. one to a high quality 1500 nit PB_DHDR display and another simultaneously to a portable display on which another member of the family is watching the show in the kitchen.

[0071] What may be a little complicated to understand (at least from a patent application perspective, which is not necessary to understand the present technical contribution, but is useful for completeness), is that a luminance mapping function may consist of many partial luminance component mapping functions (defined as applied in succession). For example, a content grader may apply Para to perform a coarse grading, roughly balancing dark and bright areas. That is, when mapping a scene with two completely different lighting areas, such as indoors, which are typically 100 times darker than outdoors in nature, and may be, for example, 10 times darker in the HDR reference grading depending on how the content creator maps the real world to his master HDR image, Para can be used to brighten the darks relative to the brights, thereby compressing the contrast of the bright areas to some extent, which is a very good simple way to make a lower dynamic range version of a higher dynamic range image (and is usually very good on a lot of content). However, there may be a commercial sign outside, embodied as white text applied (e.g. sandblasted) on a glass plate. When the contrast is reduced by an overly simplified method, such as the upper slope of Para must be very small to make room for all the darker object / pixel brightness within the small brightness range of the SDR second reference grading, this text may become difficult to read. According to the applicant's principle, the content creator / encoder can solve this problem by applying a customizable curve (CC) after Para, which brings more contrast again exactly around the luminance position of the white text and the luminance position of the slightly white colors reflected around / behind / on the glass, making the text very readable again (also note the difference between calculation precision and encoding word length, etc., but such details do not need to be explained here).

[0072] For the new teachings at present, one can understand that the continuity of many brightness mapping curves is itself a brightness mapping curve, so it is possible to pretend that there is only one "complete" curve (in fact, if the reader wants to keep things simple for the purpose of understanding this patent application, he may assume that the grader used only uses one Para, without CC). However, not only can display adaptation be performed on the complete brightness mapping function, but specific mathematical operations can be performed on the partial brightness mapping curve itself, for which we recommend that interested readers refer to ETSI TS 103 433-2 V1.1.1 (2018-01) paragraph 7.3 "metadata recomputation".

[0073] Regardless of the details, the current innovation improvements are about having a very elegant adjustment mechanism that, due to the limitations of the technical framework, tends to be implemented relatively statically (as explained).

[0074] In a more advanced embodiment, the function determination unit 502 can analyze various details of the input image (overall brightness distribution characteristics, segmented into various regions and analyzed geometric structure and contrast metrics, such as integral derivatives, texture characteristics, etc.) and derive adjustments therefrom, expressing them as alternative brightness mapping function shapes.

[0075] The combination unit 503 can also apply the combination in multiple variants, which the skilled reader can understand based on our examples, but usually this simple linear weighted combination is good enough in practice (the skilled reader likes simpler variants that require less transistors and power, but of course alternative embodiments can also be derived), and the more interesting part is to control how this combination will depend on the PB_D situation. Combination does not refer to the broadest possible concept, such as only part of one curve is exchanged by part of another curve, but all or most of the brightness will get a new curve output, which depends on the output formulated in the first curve and the output formulated in the second curve or prescription.

[0076] A good advantage arises when the image pixel brightness adaptation device (500) has a combination unit that determines, based on the value of the display maximum brightness (PB_D), that the combined brightness mapping function is more similar in shape to the alternative mapping function or to at least one brightness mapping function, respectively. Then, for example, more hybrid alternative behaviors can be controlled to correct specific relative brightness relocations for smaller PB_D values, and the reader will understand that we may want to do this in a variety of ways, and complex formulas may be needed to specify what to do in which case (for example, the alternative brightness function can be corrected for two brightness relocation aspects, the first aspect being controlled to a first degree, such as only within a certain sub-range of the total PB_D range to be processed and not higher than the PB_D_limPB_D value, while the second aspect, typically corresponding to another brightness sub-range, is controlled in a different way, i.e., has a different combination behavior that depends on PB_D, etc.). For example, the combination unit can check the PB_D value, and then no matter which combination mapping function behavior it applies, it can check whether the PB_D value is within a certain percentage deviation from the maximum PB_D (such as the PB_HDR of the input image), and then may slightly perturb the original function based on the shape of the alternative function, but for example keep it within a certain range around the first function. For PB_D values ​​below PB_D1 it may start to deviate severely, while below PB_D2 it even mostly follows the shape of the alternative function. Other algorithms for implementing this PB_D dependent function combination behavior are possible. This allows for better control over the difficult task of displaying high dynamic range images when only the lower dynamic range (actually displaying peak luminance) is to be displayed.

[0077] A practically simple but well-working variant of the image pixel brightness adaptation device 500 has a combined brightness mapping function determined by a combination unit by linear weighting of each brightness value, defined as: CMB_FL_50t1_1(Vn) = (1-A)*FL_50t1_1(Vn)+A*ALT_FL_50t1_1(Vn), where Vn is a perceptually uniform brightness representation of the pixel brightness, which can be applied by applying a logarithmic function to the brightness, and A is a weight value between 0 and 1, which is derived based on the value of the display maximum brightness (PB_D) by applying a function, the function sets A to zero below the low display maximum brightness (PLOW) and sets A to 1 above the high display maximum brightness (PHIG), and if the display peak luminance is between the low display maximum brightness (PLOW) and the high display maximum brightness (PHIG) according to a preset weighting function shape, then A is set to be equal to a value between 0 and 1. The values ​​of PLOW and PHIG can then be fixed, or optimized in an intelligent way, depending on what produces good results, either generally as an average for all images, or for specific types of images (e.g., detail classification based on the brightness histogram, such as very bright small dark objects, i.e. a small number of dark pixel brightness, contrasting very dark images, etc.).

[0078] The reader understands that other formulas for the desired adjustments may be formulated as alternative brightness mapping functions and weight definitions.

[0079] In this case, the function determination unit 502 (in conjunction with the operations to be performed by the combining unit) typically takes care that the substitute function also mainly has the shape required for a proper HDR to MDR regrading. Typically, the method is applied anyway with some minor adjustments. For (near) HDR MDR images, the work of the display adaptation itself will guarantee the correct behavior (close to diagonal) regardless of which substitute or final function is used, also because of the logarithmic nature of its metric.

[0080] A practical and simple embodiment of the image pixel brightness adaptation device may use the following as a preset weighting function shape: when defined on the input axis measured in said perceptually uniform brightness representation, the preset weighting function shape is a linearly increasing shape between zero and one. Likewise, one may fix one of several (substantially similar working) perceptual brightness representations, and the device will operate in the same manner, the selection depending on variables not relevant to the present patent application (so without wanting to be limiting, the reader may assume that it is to represent brightness as a perceptually uniform brightness that can be calculated by the example formula 1).

[0081] The following specific embodiment may already provide sufficient display adaptation quality adjustment for many customers and / or market situations: an image pixel brightness adaptation device (500) as described above, wherein the at least one brightness mapping function (F_ct; FL_50t1_1) is at least partially defined by a brightness mapping function, the brightness mapping function comprising: a first linear segment for the darkest sub-range of the total input brightness range, the linear aspect being satisfied in the perceptually unified brightness representation; a second linear segment for the brightest sub-range of the total input brightness range, and a non-decreasing segment of a non-linear shape for an intermediate sub-range between the darkest sub-range and the brightest sub-range, the two ends of which are connected to the linear segment, wherein the alternative brightness mapping function (ALT_FL_50t1_1) comprises at least a first alternative linear segment for the darkest sub-range, whose slope is different from the slope of the first linear segment for the darkest sub-range of the at least one brightness mapping function. Therefore, when there is a finely graded CC second part brightness mapping curve, only the coarse Para needs to be corrected, noting that this is a fine grading of the brightness of the same object, i.e. the fact of shifting to a new brightness sub-range compared to the originally specified sub-range, showing that the adapted Para performs a different coarse balance, i.e. a relative repositioning of the said sub-ranges.

[0082] The ending brightness of the darkest (and lightest) linear segments may or may not be the same for the original, creator, and replacement Paras.

[0083] Note that PB_H is the first maximum brightness of the main HDR image (even if communicated as a corresponding SDR image) and is typically transmitted, and the SDR maximum brightness can be fixed in advance and therefore known, typically equal to 100 nits, but can also be changed and transmitted, and the present embodiment works similarly.

[0084] Various technical implementations can also be reflected as follows:

[0085] A pixel brightness adaptation method, comprising:

[0086] Receiving a coded high dynamic range image (Im_COD) coded according to a first maximum encodable brightness (PB_H) and receiving metadata specifying at least one brightness mapping function (F_ct; FL_50t1_1) specifying an offset of brightness of a secondary image corresponding to the coded high dynamic range image compared to the brightness of the same pixel position encoded in the coded high dynamic range image, the secondary image having a second maximum encodable brightness (PB_S), the second maximum encodable brightness preferably being at least 4 times smaller or greater than the encodable brightness (PB_H);

[0087] Decoding a coded high dynamic range image (Im_COD) into a decoded high dynamic range image (Im_RHDR);

[0088] receiving in a display adaptation step a value of a display maximum brightness (PB_D) that a particular display is capable of displaying as the brightest pixel color and the brightness mapping function, and applying an algorithm for calculating at least one display adaptation brightness mapping function based on the brightness mapping function and the display maximum brightness (PB_D), wherein the at least one display adaptation brightness mapping function corresponds in shape to the input brightness mapping function but is closer to a 45 degree increasing diagonal of a graph of the input brightness mapping function on a perceptually uniform axis, the proximity to the diagonal being dependent on a difference between the display maximum brightness (PB_D) and the first maximum encodable brightness (PB_H) relative to a difference between the second maximum encodable brightness (PB_S) and the first maximum encodable brightness (PB_H);

[0089] Characterized in that the method comprises determining an alternative brightness mapping function (ALT_FL_50t1_1), and

[0090] wherein the display adaptation step comprises combining the at least one brightness mapping function (F_ct; FL_50t1_1) and the alternative brightness mapping function (ALT_FL_50t1_1) into a combined brightness mapping function (CMB_FL_50t1_1), and wherein the display adaptation step is arranged to apply its algorithm as an input brightness mapping function to the combined brightness mapping function, producing an adapted combined brightness mapping function (ADJ_F_DA50t6_1);

[0091] Receiving pixel intensities of the decoded high dynamic range image (Im_RHDR) and applying an adapted combined brightness mapping function (ADJ_F_DA50t6_1) to these pixel intensities to obtain an output brightness of an output image (Im_DA);

[0092] Output an output image (Im3000 nits), which is produced by applying the adapted combined brightness mapping function to the pixel brightness of the decoded high dynamic range image (Im_RHDR) on an image or video communication cable or wireless channel, and the display can be connected to the image or video communication cable or wireless channel.

[0093] A general type of image pixel brightness adaptation method, wherein, additionally, depending on the value of the display maximum brightness (PB_D), the combined brightness mapping function is determined to be more similar in shape to the alternative mapping function or respectively at least one brightness mapping function.

[0094] A method for adapting image pixel brightness as described above, wherein the combined brightness mapping function is determined by linear weighting of each brightness value, defined as: CMB_FL_50t1_1(Vn)=(1-A)*FL_50t1_1(Vn)+A*ALT_FL_50t1_1(Vn), wherein Vn is a perceptually uniform brightness representation of pixel brightness, which can be applied by applying a logarithmic function to the brightness, A is a weight value between 0 and 1, and the weight value can be derived based on the value of the display maximum brightness (PB_D) by applying a function, wherein the function sets A to zero when it is lower than the low display maximum brightness (PLOW), sets A to one when it is higher than the high display maximum brightness (PHIG), and if the display peak brightness is between the low display maximum brightness (PLOW) and the high display maximum brightness (PHIG) according to a preset weighting function shape, then A is set to a value between 0 and 1.

[0095] A method for adapting image pixel brightness, wherein when defined on an input axis measured in terms of perceptually uniform brightness representation, the preset weighting function shape is a linearly increasing shape between zero and one.

[0096] A method for adapting the brightness of image pixels, wherein the at least one brightness mapping function (F_ct; FL_50t1_1) is at least partially defined by a brightness mapping function, the brightness mapping function comprising: a first linear segment for the darkest sub-range of the total input brightness range, the linear aspect being satisfied in a perceptually uniform brightness representation; a second linear segment for the brightest sub-range of the total input brightness range; and a non-linearly shaped non-decreasing segment for an intermediate sub-range between the darkest sub-range and the brightest sub-range, both ends of which are connected to the inner ends of the linear segments, and wherein the alternative brightness mapping function (ALT_FL_50t1_1) comprises at least a first alternative linear segment for the darkest sub-range, the slope of which is different from the slope of the first linear segment for the darkest sub-range of the at least one brightness mapping function.

[0097] The maximum encodable brightness refers to the physical brightness corresponding to the maximum encodable pixel color, i.e., the highest brightness encoding (e.g., 1023 in 10 bits), i.e., the actual brightness of the whitest white as it should ideally be displayed on any display and can be displayed on the corresponding virtual display associated with the image. That is, the virtual display may indicate that the brightest encodable brightness of the image is, for example, 1200 nits, and ideally, if the actual receiving end (e.g., consumer) display is 1200 nits maximum displayable white or more, such a display should present the highest brightness encoded achromatic pixel with a display brightness of 1200 nits. The commonly communicated function can then have a function shape optimized by the content creator, indicating how such image brightness (in fact, brightness in general) should be displayed on a display with a smaller maximum brightness capability (e.g., a 600 nit display), such as by mapping the maximum image brightness to the highest possible displayable brightness, and using a larger sub-range below this for darker colors rather than brighter colors, etc. It is understood that the image creator may combine the relative brightness positions of various image objects differently, depending on the maximum encodable brightness PB_C, e.g. for a 4000 nits PB_C, a lamp may be made brighter compared to a non-luminous image object pixel brightness compared to a 900 nits PB_C. One may also think of it as the pixel brightness of an object, e.g. a lamp, being shifted to a new relative position compared to, e.g., a chair object pixel brightness in a room, for various PB_C image encodings, on an axis normalized to 1.0 brightness (or even luminance, by defining an EOTF of brightness corresponding to various brightness codes). BRIEF DESCRIPTION OF THE DRAWINGS

[0098] These and other aspects of the method and apparatus according to the invention will become apparent and elucidated by reference to the embodiments and examples described below and with reference to the accompanying drawings, which are intended only to illustrate more general non-limiting specific illustrative concepts, wherein dashed lines are used to indicate that components are optional, and non-dashed line components are not necessarily essential. Dashed lines may also be used to indicate elements that are interpreted as essential but are hidden inside an object, or for intangible things, such as the selection of objects / areas (and how they are displayed on a display).

[0099] In the attached picture:

[0100] Figure 1schematically illustrating several typical brightness transformations that occur when optimally mapping a high dynamic range image to a corresponding optimally color graded and similar looking (given the difference of the first and second brightness dynamic ranges DR_1, respectively DR_2, as similar as desired and feasible) lower dynamic range image (e.g. a standard dynamic range image of 100 nits peak luminance), which in the reversible case (e.g. SL-HDR1 encoding, which conveys an SDR variant of the HDR image, which still needs to be reconstructed as an HDR image by the receiver by applying the inverse of the brightness mapping function that creates SDR brightness from HDR brightness to the received SDR image brightness) will also correspond to the mapping of the received SDR image of the actual encoded HDR scene to the reconstructed HDR image of the scene;

[0101] Figure 2 schematically showing a satellite view example of a technique for encoding a high dynamic range image, i.e. an image capable of having a brightness of typically at least 600 nits (i.e. at least 6 times the PB_C of an SDR image) or higher (typically 1000 nits or higher, e.g. 2000 nits maximum occurring / encodable pixel brightness PB_C or 10000 nits PB_C), recently developed by the applicant, which in practice makes it possible to communicate HDR image(s) as an SDR image plus metadata encoding a color transformation function, said color transformation function comprising at least a suitably determined brightness transformation of the pixel colors (typically embodied technically as a brightness mapping function, in a perceptually uniform brightness domain), which a decoder will use to convert the received SDR image(s) into HDR image(s) which are faithful reconstructions of the original master HDR image(s) created at the image creation end;

[0102] Figure 3 A particularly useful dynamic range alteration color processing core is shown (without limitation as to the applicability of presently useful improvements), which the present applicant has standardized according to his various versions of HDR video encoding / decoding methods; this computational circuit can be used elegantly not only for HDR image decoding, but also for display adaptation to obtain an optimal medium dynamic range image for a particular display with maximum displayable brightness PB_C that happens to be present at the location of any particular audience, for example by using as input a received or decoded version of the content creator's master HDR image, using YCbCr encoded pixel colors;

[0103] Figure 4schematically illustrating the basic technical aspects of a display adaptation process, which is typically chosen to be a simple, fixed, automatic process, based on any possible input luminance mapping function shape (two examples FL_50t1_1 and FL_50t1_2 are shown), which function specifies how to re-grade luminances normalized to 1.0 from a first representative grading of an HDR scene to a second representative grading (typically a master HDR image, which has some maximum encodable brightness, PB_C = e.g. 5000 nits, and a 100 nit PB_C SDR image), deriving as output a corresponding (essence of the function shape preserved) final display adaptation luminance map for determining an MDR image from the HDR image pixel luminances, the MDR image having the correct relative luminances for optimally driving an MDR display with 100 <= PB_D <= PB_C, so that the MDR image looks reasonably similar to the HDR image when displayed on a high quality HDR reference display with PB_D_HDR_reference = PB_C, with respect to differences in display capabilities;

[0104] Figure 5 It is shown how such a display adaptation process or unit can be further improved to be adjustable according to the insights of the inventors, by providing it with not only the luminance mapping function of the original content creator, but also an alternative luminance mapping function of an appropriate shape (ALT_FL_50t1; for each case where re-grading of the image is required), so that the display adaptation unit can achieve a less static re-grading behavior for various possible actual PB_D values, and how, for example, on certain displays, e.g. PB_D < 550 nits, a certain image may still look a bit too dark, while on other displays, e.g. PB_D > 1500 nits, it may still be a bit too bright;

[0105] Figure 6 shows an elegant and simple way of specifying how the system should handle various PB_D values ​​that may appear at the receiving end, i.e. how the original and alternative brightness mapping functions can be combined into a single combined function before applying a fixed display adaptation algorithm (to whatever variants happen to exist, and therefore fixed in e.g. a TV set) to the combining function;

[0106] Figure 7further illustrating the general display adaptation adjustment approach by focusing on a simple example (but which may already address important most of the perceived or potentially occurring imperfect behaviors of static display adaptation), which proposes an alternative luminance curve, one with (at least) a different slope (called Shadowgain), for a particular parameterized regrading curve, called Para, e.g. Shadowgain_ALT = (1.k)*Shadowgain_content_creator, where k is e.g. 1, 2, 3, 4, and Shadowgain_content_creator is the Shadowgain of Para, which is selected as the best choice for SDR grading of the current master HDR image and is communicated in the receiver as metadata associated with the transmitted image or video, Shadowgain_ALT is the adjustment value proposed by the receiver, e.g. based on some measured luminance property of the current color-changing image; and

[0107] Figure 8 Shows Figure 6 Some examples of linear weight A determination functions based on PB_D, where some values ​​are in nits. DETAILED DESCRIPTION

[0108] Figure 5 It is generally shown what integrated circuits or similar units the improved display adaptation unit of the image pixel brightness adaptation device 500 of the present patent application will include.

[0109] As described above, the novel image pixel brightness adaptation apparatus 500 will include an alternative brightness mapping function determination unit 502. Depending on which device it resides in (e.g., a set-top box that prepares images for a specific TV, or in the TV itself, etc.), this unit can formulate the alternative brightness grading function shape ALT_FL_50t1_1 in a variety of ways, such as ranging from a function that has no or little dependence on the brightness adaptation function of the SEI communication of the image content or the content creator, or the function can largely follow the shape of the brightness adaptation function of the SEI communication, with only one aspect being different (e.g., slightly different), such as being a little higher in the v_input, v_output map, or having some shape perturbations in a small sub-range of v_input, etc.

[0110] This may be a fixed number designed by the technology provider or device manufacturer, or may be a variable number that may be determined on the fly, e.g., for each image, etc.

[0111] The alternative luminance mapping function (ALT_FL_50t1_1) and the original luminance mapping function (ALT_FL_50t1_1) determined by the content creator as a re-grading need indicating function and received (typically extracted by a decoder video 207, said video decoder being included in the device 500, or at least connected to it during operation, so that the function can be received via input 501), are combined in any way (in a combining unit 503) to produce a combined grading function CMB_FL_50t1_1, which implements a small part of the two functions according to the display adaptation needs, i.e. typically a specific value of PB_D for the connected display. Note that input 501 can double as an input for the decoded image Im_RHDR, or there can be a separate input for this image from an included or connectable decoder.

[0112] In the elucidation example we see that the original function is a somewhat crude shape (e.g. pure Para etc) which mainly implements a relative brightening of the darkest luminances (for lower PB_C MDR images, or more precisely for SDR grading). The alternative function implements some contrast stretching in the perceptually uniform input sub-range MR, e.g. because there is a key object that is not easily re-graded at the darkest, i.e. lowest PB_D displays. We see this behavior leaking into the combined brightness mapping function (ADJ_F_DA50t6_1) of the display adaptation, since the standard display adaptation algorithm is static and a known source, and can therefore be adjusted as needed.

[0113] Finally, the brightness mapping unit 510 uses an adapted combined brightness mapping function ADJ_F_DA50t6_1 (as input), which uses the specification to map the brightness of the reconstructed / decoded HDR image Im_RHDR to the best corresponding, display-adapted output brightness of the display-adapted image Im_DA, which can be sent to any image or video output depending on the specific technical implementation of a specific image or video processing device or system.

[0114] Figure 6 An advantageous way of weighting the two brightness functions for each brightness is generally shown. For example, for a particular PB_D value equal to PB_D1, the final weight factor A adjusted is A=0.6

[0115] For all perceptually uniform brightnesses (Vn) that may appear in the input image and are normalized to 1.0, the combining function can be calculated as:

[0116] CMB_FL_50t1_1(Vn; PB_D1)=0.6*FL_50t1_1(Vn)+0.4*ALT_FL_50t1_1(Vn)

[0117] This combined function can then be input into the standard algorithm for display adaptation as if, but it explicitly is not, it was the luminance mapping function of the original content creator (ie as if we were performing standard display adaptation).

[0118] Finally, this function can be used in the color processing core, for example with Figure 3 The color processing core is clarified so as to obtain the output brightness of the MDR image based on the input brightness of the HDR image (or equivalently the calculation can be implemented starting from the SDR image, in the case of a combined brightness mapping function with a different shape).

[0119] Figure 8 Some practical examples are shown with some numerical brightness values ​​superimposed. Typically the PB_D values ​​will be judged so the scale will end at the PB_C value for the content (e.g. a movie), in this case 4000 nits. We show that the weight factor determination curves may also vary (e.g. a first weight determination curve WP1 and a second weight determination curve WP2), e.g. depending on the type of image content being processed, as some content may be well graded on a larger subset of displays PB_D because it is less important, while other content may start to show some problems later, etc. While we illustrated the examples with a logarithmic definition of a perceptually uniform brightness representation, the definition of the adjustment, in particular the specification of the weight determination functions, may also work in other non-linear representations.

[0120] The horizontal axis may also be described as the relative value of PB_D (ie, PB_D / PB_H) in a logarithmic system with PB_H defined as the maximum value (ie, for example, by Equation 1, or a similar perceptually uniform formula).

[0121] By function shape or the shape of a function, we mean the trajectory of output points for various input points, such as a parabolic convex shape, which can be controlled by shape control parameters, such as the values ​​a, b and c for the formula y_out = a*x^2+b*x+c.

[0122] Figure 7 It is shown how the light brightness behavior of the receiver display adaptation behavior can be elegantly and simply influenced by using another Para-based adjustment.

[0123] The original Para of the content creator received as metadata has a dark segment SD ending at brightness Lsd, a bright linear segment SB starting from Lsb, and a middle (parabolic) segment SM controlling the gradation of the middle range brightness.

[0124] The alternative grading Para is generated by an alternative brightness mapping function determination unit 502, which has, for example, a slope SL that is 1.3 times steeper than the creator's Para slope SL (aka Shadowgain) for the alternative dark segment SDA (or typically 1.x, even 2.x, etc.). The rest of the function may not be a pure Para (e.g., a Para that follows a custom CC curve shape), but the alternative brightness mapping function determination unit 502 may also determine, for example, a highlight gain for the alternative brightest segment SBA, such as a highlight gain that largely corresponds to the slope of the SB region, while leaving some extra room for the darkest output brightness, coordinated with the determination of the alternative middle region regrading behavior of the segment SMA. Other Para function shape control parameters, such as the highlight gain may be equal for the alternative brightness mapping Para and the original one or different, etc.

[0125] The algorithm components disclosed herein may be implemented (in whole or in part) in practice as hardware (eg, part of an application specific IC) or as software running on a special digital signal processor or a general purpose processor or the like.

[0126] A person skilled in the art should be able to understand from our introduction which components may be optional improvements and may be implemented in combination with other components, and how the (optional) steps of the method correspond to the individual devices of the device, and vice versa. The term "device" in this application is used in its broadest sense, i.e. a group of devices that allow a specific goal to be achieved, and thus may be, for example, (a small circuit part of) an IC, or a dedicated device (e.g. a device with a display), or a part of a network system, etc. "Arrangement" is also intended to be used in the broadest sense, and thus may include in particular a single device, a part of a device, a (partial) collection of cooperating devices, etc.

[0127] The term computer program product is to be understood as including any physical realization of a set of commands for a general or special purpose processor, after a series of loading steps (possibly including intermediate conversion steps, such as translation into an intermediate language, and the final processor language) to input the commands into the processor and to perform any characteristic functions of the invention. In particular, a computer program product can be realized as data on a carrier, such as a disk or tape, data present in a memory, data transmitted via a network connection (wired or wireless), or program code on paper. In addition to the program code, the characteristic data required for the program can also be embodied in the computer program product.

[0128] Some steps required to operate the method may already exist in the functionality of the processor rather than being described in the computer program product, such as data input and output steps.

[0129] It should be noted that the above embodiments are illustrative rather than limiting of the present invention. In the case where a technician can easily implement the mapping of the presented examples to other areas of the claims, for the sake of brevity, we have not mentioned all of these options in depth. In addition to the combination of elements of the present invention as combined in the claims, other combinations of elements are also possible. Any combination of elements can be implemented in a single dedicated element.

[0130] Any reference signs between brackets in a claim are not intended to limit the claim. The word "comprising" does not exclude the presence of elements or aspects not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements.

Claims

1. An image pixel brightness adaptation device (500), comprising: a connection (501) to a contained or connectable video decoder (207), the video decoder being arranged to receive a coded high dynamic range image Im_COD, the coded high dynamic range image being coded to have a first maximum encodable brightness PB_H, and the video decoder being arranged to receive metadata specifying a brightness mapping function FL_50t1_1, the brightness mapping function specifying an offset of the brightness of a secondary image from the value of the brightness of a corresponding pixel at the same pixel position in the coded high dynamic range image, the secondary image having a second maximum encodable brightness PB_S, and the video decoder being arranged to output a decoded high dynamic range image Im_RHDR and the brightness mapping function to the image pixel brightness adaptation means; The image pixel brightness adaptation device (500) further comprises: a display adaptation unit (401) arranged to receive a value of a display maximum brightness PB_D that a specific display can display as the brightest pixel color and an input brightness mapping function, the display adaptation unit being arranged to apply an algorithm, the algorithm calculating a display adaptation brightness mapping function based on the input brightness mapping function and the display maximum brightness PB_D, wherein the display adaptation brightness mapping function corresponds in shape to the input brightness mapping function but is closer to a 45 degree increasing diagonal line on a perceptually uniform axis of a graph of the input brightness mapping function, the proximity to the diagonal line depending on a difference between the value of the display maximum brightness PB_D and the first maximum encodable brightness PB_H relative to a difference between the second maximum encodable brightness PB_S and the first maximum encodable brightness PB_H; The image pixel brightness adaptation device comprises an alternative brightness mapping function determination unit (502), which is arranged to determine an alternative brightness mapping function ALT_FL_50t1_1, and wherein the display adaptation unit (401) comprises a combination unit (503) arranged to combine the brightness mapping function F_ct;FL_50t1_1 and the alternative brightness mapping function ALT_FL_50t1_1 into a combined brightness mapping function CMB_FL_50t1_1, and wherein the display adaptation unit is arranged to apply its algorithm as an input brightness mapping function to the combined brightness mapping function and to produce as an output an adapted combined brightness mapping function ADJ_F_DA50t6_1; The image pixel brightness adaptation device comprises a brightness mapping unit (510) arranged to receive pixel brightnesses of the decoded high dynamic range image Im_RHDR and to apply the adapted combined brightness mapping function ADJ_F_DA50t6_1 to these pixel brightnesses to obtain output brightnesses of the output image Im_DA; The image pixel brightness adaptation device includes an output image or output video communication cable or wireless channel, a display can be connected to the output image or output video communication cable or wireless channel, and an output signal formatter (230) arranged to send the output image Im_DA through the output image or output video communication cable or wireless channel.

2. The image pixel brightness adaptation device (500) according to claim 1, wherein: The combination unit determines whether the combined brightness mapping function is more similar in shape to the alternative brightness mapping function or, on the contrary, more similar to the brightness mapping function depending on the value of the display maximum brightness PB_D. For lower values ​​of the display maximum brightness, the combined brightness mapping function is more similar in shape to the alternative brightness mapping function.

3. The image pixel brightness adaptation device (500) according to claim 2, wherein: The combined brightness mapping function is determined by linear weighting of each brightness value and is defined as: CMB_FL_50t1_1(Vn)=(1-A)*FL_50t1_1(Vn)+A*ALT_FL_50t1_1(Vn), wherein Vn is a perceptually uniform brightness representation of pixel brightness that can be obtained by applying a logarithmic function to the brightness, and A is a weight value between zero and one, the weight value being derived based on the value of the display maximum brightness PB_D by applying a function that sets A to zero when it is below a low display maximum brightness PLOW and sets A to one when it is above a high display maximum brightness PHIG, and sets A to a value between zero and one if the display peak luminance is between the low display maximum brightness PLOW and the high display maximum brightness PHIG according to a preset weighted distribution shape.

4. The image pixel brightness adaptation device (500) according to claim 3, wherein: When defined on an input axis measured in the perceptually uniform brightness representation, the predetermined weighted distribution shape is a linearly increasing shape between zero and one.

5. The image pixel brightness adaptation device (500) according to one of the preceding claims, wherein: The alternative brightness mapping function determination unit (502) is arranged to determine an alternative brightness mapping function ALT_FL_50t1_1 having a shape similar to that of the brightness mapping function FL_50Tt1_1 except for a shape perturbation within a sub-range of input values.

6. The image pixel brightness adaptation device (500) according to one of claims 1 to 4, wherein: The alternative brightness mapping function determination unit (502) is arranged to determine the shape perturbation differently for each image.

7. The image pixel brightness adaptation device (500) according to one of claims 1 to 4, wherein: The brightness mapping function FL_50t1_1 is defined at least in part by means of a brightness mapping function comprising: a first linear segment for the darkest sub-range of the total input brightness range, the linear aspect being satisfied in a perceptually uniform brightness representation; a second linear segment for the brightest sub-range of the total input brightness range; and a non-decreasing segment of a non-linear shape for an intermediate sub-range between the darkest sub-range and the brightest sub-range, the two ends of which are connected to the linear segment; And wherein the alternative brightness mapping function ALT_FL_50t1_1 comprises at least a first alternative linear segment for the darkest sub-range, whose slope is different from the slope of the first linear segment for the darkest sub-range of the brightness mapping function.

8. A method for adapting image pixel brightness, comprising: receiving a coded high dynamic range image Im_COD, the coded high dynamic range image being coded according to a first maximum encodable brightness PB_H, and receiving metadata specifying a brightness mapping function FL_50t1_1 specifying an offset of a brightness of a secondary image corresponding to the coded high dynamic range image compared to the brightness of the same pixel position encoded in the coded high dynamic range image, the secondary image having a second maximum encodable brightness PB_S; Decoding the coded high dynamic range image Im_COD into a decoded high dynamic range image Im_RHDR; receiving in a display adaptation step a value of a display maximum brightness PB_D that a specific display can display as the brightest pixel color and an input brightness mapping function, and applying an algorithm for calculating a display adaptation brightness mapping function based on the input brightness mapping function and the display maximum brightness PB_D, wherein the display adaptation brightness mapping function corresponds in shape to the input brightness mapping function but is closer to a 45 degree increasing diagonal of a graph of the input brightness mapping function on a perceptually uniform axis, the proximity to the diagonal depending on a difference between the value of the display maximum brightness PB_D relative to a difference between the second maximum encodable brightness PB_S and the first maximum encodable brightness PB_H and the first maximum encodable brightness PB_H; wherein the method comprises determining an alternative brightness mapping function ALT_FL_50t1_1, and wherein the display adaptation step comprises combining the brightness mapping function F_ct;FL_50t1_1 and the alternative brightness mapping function ALT_FL_50t1_1 into a combined brightness mapping function CMB_FL_50t1_1, and wherein the display adaptation step is arranged to apply its algorithm as an input brightness mapping function to the combined brightness mapping function, generating an adapted combined brightness mapping function ADJ_F_DA50t6_1; Receiving pixel intensities of the decoded high dynamic range image Im_RHDR and applying the adapted combined brightness mapping function ADJ_F_DA50t6_1 to these pixel intensities to obtain output intensities of the output image Im_DA; An output image Im3000 nits is output over an image or video communication cable or wireless channel, wherein the output image is generated by applying the adapted combined brightness mapping function to the pixel brightness of the decoded high dynamic range image Im_RHDR, and a display can be connected to the image or video communication cable or wireless channel.

9. The image pixel brightness adaptation method according to claim 8, wherein: Depending on the value of the display maximum brightness PB_D, the combined brightness mapping function is determined to be more similar in shape to the alternative brightness mapping function, or conversely, to be more similar in shape to the brightness mapping function, wherein for lower values ​​of the display maximum brightness, the combined brightness mapping function is more similar in shape to the alternative brightness mapping function.

10. The image pixel brightness adaptation method according to claim 9, wherein: The combined brightness mapping function is determined by linear weighting of each brightness value and is defined as: CMB_FL_50t1_1(Vn)=(1-A)*FL_50t1_1(Vn)+A*ALT_FL_50t1_1(Vn), wherein Vn is a perceptually uniform brightness representation of pixel brightness that can be obtained by applying a logarithmic function to the brightness, and A is a weight value between zero and one, which is derived based on the value of the display maximum brightness PB_D by applying a function, wherein the function sets A to zero when it is below the low display maximum brightness PLOW and sets A to one when it is above the high display maximum brightness PHIG, and if the display peak luminance is between the low display maximum brightness PLOW and the high display maximum brightness PHIG according to a preset weighted distribution shape, then A is set to a value between zero and one.

11. The image pixel brightness adaptation method according to claim 10, wherein: When defined on an input axis measured in the perceptually uniform brightness representation, the predetermined weighted distribution shape is a linearly increasing shape between zero and one.

12. The image pixel brightness adaptation method according to any one of claims 8 to 11, wherein: The brightness mapping function FL_50t1_1 is at least partially defined with the aid of a brightness mapping function comprising: a first linear segment for the darkest sub-range of the total input brightness range, wherein the linear aspect is satisfied in a perceptually uniform brightness representation; a second linear segment for the brightest sub-range of the total input brightness range; and a non-decreasing segment of a non-linear shape for an intermediate sub-range between the darkest sub-range and the brightest sub-range, both ends of which are connected to the inner ends of the linear segments, and wherein the alternative brightness mapping function ALT_FL_50t1_1 comprises at least a first alternative linear segment for the darkest sub-range, whose slope is different from the slope of the first linear segment for the darkest sub-range of the brightness mapping function.

Images