Reconstruction of HDR Video by Transformed Tone Mapping

Abstract

To enable better quality decoding of the HDR image communicated as the corresponding LDR image with the luma mapping function applied, the applicant teaches an encoder for encoding a high dynamic range image (Im_HDR_PSEU). The high dynamic range image is first represented by a matrix of pixel colors (Y_SDR, Cb_SDR, Cr_SDR) of an image (Im_SDR) with a lower dynamic range than the high dynamic range image, and the low dynamic range image is compressed to be communicated as a compressed low dynamic range image (Im_C). Second, it includes a luma mapping function for calculating the high dynamic range pixel luma (Y_HDR) of the high dynamic range image by applying the luma mapping function to the pixel luma of the low dynamic range image, and is represented by the image metadata (SEI). The encoder is provided with an input 997 for receiving the luma mapping function (LMF) from the inverse tone mapping system 200, and the inverse tone mapping system 200 is configured to derive a luma mapping function for constructing the corresponding high dynamic range image (Im_HDR_PSEU) based on the analyzed characteristics of the master low dynamic range image (Im_LDR_mastr). In the encoder, the encoder is configured to convert the luma mapping function into an extended luma mapping function (LMF_HS) having a shape that maps the maximum normalization input to the maximum normalization output and is configured to determine a scaling value (SCAL), and is characterized by having a formatter 104 configured to output this extended luma mapping function (LMF_HS) and this scaling value (SCAL) as the metadata of the low dynamic range image (Im_SDR) output similarly.

Description

Technical Field

[0001] The present invention relates to a method and apparatus for reconstructing a high dynamic range (HDR) image from a received low dynamic range (LDR) image including information and a tone mapping function necessary for decoding a faithful reconstructed image of a master HDR image created and encoded on a transmission side at a reception side of a video communication system, and a corresponding encoding method.

Background Art

[0002] Until the first surveys around 2010 (and before the introduction of the first purchasable HDR decoding TVs in 2015), at least regarding video, all videos were made according to a common low dynamic range (LDR) encoding framework also known as standard dynamic range (SDR). This has several characteristics. First, there is only one video made, and this video was good for all displays. This system is a relative system, and white is the maximum (100%) signal encoded using the maximum luma code corresponding to the maximum non - linear RGB values R’ = G’ = B’ = 255 (255 in 8 - bit YCbCr encoding). There is nothing brighter than white, and everything was a typical reflective color (for example, paper either reflects all incident light at most, or absorbs some of the red wavelengths and reflects blue and green towards the eye, resulting in a cyan subtractive color, which by definition is somewhat darker than the white of the paper). Each display was technically constructed to display this whitest white (as a "drive requirement") as the brightest color, for example, at 80 nits (a simplified notation name for the SI quantity Cd / m^2) on a computer monitor and at 200 nits on an LCD display with a TL backlight. The observer's eye compensated for differences in brightness immediately, and thus, all observers saw almost the same image at home (despite differences in display) when not next to a store.

[0003] Rather than simply the colors printed or painted on paper, it was desired to improve the perceivable appearance of an image by creating pixels that emit a real light much brighter than the "white of the paper", also known as "diffuse white".

[0004] Some systems, such as BBC's HLG, achieve this by defining values in the encoded image that exceed white (where white is given a reference level of "1" or 100%), for example up to 10 times white, and can be defined such that pixels that are 10 times brighter are displayed.

[0005] Today, most systems are moving to a paradigm where video creators can define the absolute nit value of their images (i.e., a value that is neither twice nor ten times the undefined white level that is converted to the variable actual nit output on each endpoint display) based on the dynamic range capabilities of the selected target display. The target display is a virtual (intended) display on the video creation side, e.g., a 4000 nit (ML_C) target display for defining a specific video of 4000 nit, and the actual consumer endpoint display has a lower display maximum luminance (ML_D), e.g., 750 nit. In such a situation, the end display still needs to include hardware or software for luminance remapping, which is typically realized as a luma mapping that somehow matches the pixel luminance in the HDR input image (specifically its maximum luminance) that is too high to be faithfully displayed to a value on the dynamic range of the end display. The simplest mapping is just to clip all luminance above 750 nit to 750 nit, but in that case, for example, the beautiful structure of a 4000 nit sunset image with sunlit clouds in the range of 1000 - 2000 nit is clipped and lost, and is displayed as a white uniform 750 patch, so this mapping is the worst way to handle dynamic range mapping. In a better luma mapping, a subrange of 1000 - 2000 nit in the HDR input image is shifted, by an appropriately determined function, e.g., to 650 - 740 of the dynamic range of the end display (the function can be automatically determined within a receiving device such as a TV, STB, etc., or determined by the video creator as the most suitable for the video creator's artistic video or program and communicated with the video signal). Luma means encoding luminance, e.g., in 10 bits, using a function that assigns luma codes from 0 - 1023 to video luminance from, e.g., 0.001 - 4000 nit, by means of the so-called electro-optical transfer function (EOTF).

[0006] The simplest system is just to send the HDR, e.g., an image itself with a maximum brightness of 4000 nits (along with a properly defined EOTF). This is what is implemented in the HDR10 standard. More advanced systems such as HDR10+ also communicate a function for downmapping a 4000-nit image to a lower dynamic range such as 750 nits. These systems define a mapping function between two different maximum brightness versions of the same scene image and then use an algorithm that calculates a modified version of that function to calculate an end-point function regarding the maximum brightness of other displays, thus making this easier. For example, if the SDR image is newly interpreted as an absolute nit image rather than a relative image and there is an agreement to always define the SDR image such that the maximum pixel brightness is 100 nits, the video creator can communicate together by defining a function that specifies how to map the luminance from 0.001 (or 0), the first reference image grading, to 4000 nits, to the corresponding desired SDR 0 to 100 nits luminance (the secondary reference grading), which is called display tuning or adaptation. For both the 4000-nit ML_C input image (horizontal axis) and the 100-nit ML_C secondary grading / reference image, in a plot normalized to 1.0, if a function is defined to increase the darkest 20% of the colors, e.g., three times, i.e., when going from 4000 nits down to 100 nits, and there is a need to go down to 750 nits in a particular end-user TV, the required increase will be, for example, only twice. (This varies depending on which EOFT definition is used for luma. As mentioned above, luminance mapping is typically actually implemented as luma mapping in the color processing IC / pipeline. For example, using a psychovisual-uniform EOTF makes it possible to define the effect of the luminance change along the range to be more visually uniform, i.e., more relevant and visually impactful for humans.)

[0007] A third class of more advanced HDR encoders raises to the next level by re-imaging these two reference grading images in a different way. If limited to the use of mostly reversible functions, for example, an LDR image that can be calculated on the transmitting side by downmapping the luminance or luma of a 4000 nit HDR image to an SDR image can actually be transmitted as a proxy for the actual master HDR image created by a video creator, such as a Hollywood studio for BD or OTT delivery or a broadcaster of a sports program. Then, the receiving device can apply the inverse function to reconstruct a faithful reconstruction of the master HDR image. A system that communicates the HDR image itself (as it was created) is called "Mode HDR", and a system that communicates an LDR image is called "Mode LDR coder".

[0008] FIG. 1 shows a typical example (summarizing principles for which the applicant has previously obtained patents, for example, in WO2015 / 180854), and FIG. 1 includes the decoding function itself and subsequent display adaptation as blocks, and these two techniques should not be confused.

[0009] Figure 1 schematically shows a video coding and communication as well as processing (display) system. On the creation side, an embodiment of a typical encoder (100) is shown. Those skilled in the art will first understand that a pixel-by-pixel processing pipeline for luminance processing (i.e., all pixels of the input HDR image Im_HDR (typically one of a plurality of master HDR images of a video created by a video creator, and details of creation such as camera capture and shading, or offline color grading are understandable to those skilled in the art and are omitted here as they do not deepen this explanation) are sequentially processed through the processing pipeline), and then that a video processing circuit operating on the entire image is shown. For example, a similar compressor such as MPEG's DCT compression or AV1 acts on blocks of pixels. Without wishing to limit the teaching, assume that the input image is a master HDR grading created by a human color grader. After the human color grader selects the maximum luminance of the master grading video, they determine where to position various video objects (with respect to average luminance) within the available range so that the image has the optimal impact on the consumer. For example, conventional averagely illuminated objects have luminance comparable to that which those objects would obtain in an LDR grading, while various types of HDR effect objects (such as clouds illuminated by a bright sun, the fireball of an explosion, a lamp shade, an object illuminated by a flashlight or sunlight) obtain various luminances in a higher sub-range of the available luminance range, for example exceeding the white level of the local scene. Some of these luminance levels change dynamically in various images, such as when a dark corridor is gradually illuminated by successively switching on a continuous row of ceiling TL lamps.

[0010] In FIG. 10, two different scene images from the video (dragon and suk, see FIG. 10A) are shown, and the creator selected that it is best to define this scene image in the luminance range of the master HDR maximum luminance (ML_C_MH) of 1500 nits (i.e., the master is, for example, secondary and not optimal, but it is necessary to perform (re)grading as similar as possible. However, the creator judged it to be the best because it brings a good HDR effect to, for example, the dragon's fiery breath. This is the optimal expression of this video). This can be imagined as first selecting the canvas shape for drawing and then starting the drawing with the optimal configuration. Therefore, in reality, in the case of the master HDR video (PRIM_GRAD), a human grader (or automaton) selects various average luminances in which the pixel luminances of different image objects are distributed around. For example, a diffuse reflection object is usually bright at a selected level LowH, for example, less than 220 nits (or if the corresponding secondary (re)grading (SEC_GRAD) ends at the secondary grading maximum luminance ML_C_secG, the corresponding secondary limit level (lowS) of a usually bright object is regarded as, for example, 120 nits (roughly speaking, this level corresponds to most of the pixel colors existing in the SDR image when reinterpreted in their respective luminance ranges). The dragon, for example, becomes black with an average luminance of 10 nits, while the tree is about 50 nits. To obtain the maximum effect with a large area, the grader gives a luminance distribution of about 800 nits to the flame. The specular reflection in the metal vase within the sunlight rays in the suk image reaches a pixel luminance of up to 1300 nits, for example. The hanging lamp should also have above-average brightness, but since the hanging lamp is not attention-grabbing, it is more conservative, for example, 350 nits (in contrast to the dragon's flame, which can be seen by the observer even if they are only 20 nits, but in that case, the flame does not seem very realistic or impressive).

[0011] As shown in the two luminance range projection representations of FIG. 10B, for example, when re-grading a secondary image with a narrower luminance dynamic range, the grader needs to find the corresponding luminance positions of all objects (any image). For example, if a very impressive flame is desired within a limited range and, by chance, there are not many objects in this video that are brighter and more prominent than the flame, the flame is projected to the 400 nit level by the object projection Fo_regrad. The tree is projected with equal luminance, and the dragon is made somewhat brighter, for example, considering that the secondary grading SEC_GRAD is expected to have less ability to produce dark black.

[0012] As can be seen in FIG. 10C, at least for practical consumer video communication (i.e., broadcasting, movies, etc.) or for purposes such as video conferencing, this re-grading can typically be summarized in a 2D plot by the shape of some luminance mapping function FL_regrad (it may also be shown normalized to a 1.0 plot). It is also shown on the left side of FIG. 10B that it may not start from the master HDR image created as the starting image by the video creator. In some cases, start from an SDR image called the SDR master (MST_SDR) and use some up-grading algorithm to obtain the master HDR image (i.e., PRIM_GRAD in FIG. 10B). For example, use a strong increasing function (MeffBoo) for the brightest object in the SDR image of the dragon. It should be noted that, in principle, when assigning luminance in the master HDR luminance range, the SDR image has no maximum luminance, but for example, it can be assumed that the SDR image reaches up to 100 nit.

[0013] Returning to FIG. 1, assume that we start with the HDR pixel luminance L_HDR sent through the selected HDR inverse EOTF in the luma conversion circuit 101 (note that some systems may already start working from luma). The corresponding HDR luma Y_HDR is obtained. It should be noted that the EOTF is a (typically fixed) function for obtaining luminance from the luma code that encodes luminance, and should not be confused with a re-grading function that typically has different optimal shapes for different scene images. For example, a perceptual quantizer EOTF is used. This input HDR luma is luma mapped by the luma mapper 102 to obtain the corresponding SDR luma Y_SDR. In this unit, colorimetry knowledge is applied. That is, there is an input connection UI to the appropriately determined shape of the luma mapping function (LMF). Broadly speaking, there are two possible classes. Offline systems employ a human color grader to determine the best LMF according to their artistic preferences through color grading software. The LMF is assumed to be defined as a LUT defined using the coordinates of several nodes (for example, in the case of the first node, (x1, y1)), but is not limited to this. For example, if there is dark content in the image and the human grader wants this content to be clearly visible when displayed on a low dynamic range display, specifically, on a 100 nit ML_C image for a 100 nit ML_D LDR display, the human grader sets the slope of the first line segment, that is, the position of the first node.

[0014] The second class is an embodiment where the current technology presented below is important and uses automata. These automata analyze an image (such as which luminance exists, where it exists, and to what extent it exists) and propose the shape of the best LMF function. A particularly interesting automaton, the so-called ITM ("Inverse Tone Mapping"), analyzes the input LDR image on the creation side instead of analyzing the master HDR image, and creates a pseudo-HDR image of this LDR image. The term "pseudo" here does not mean a low-quality HDR image, but rather, for example, by incorporating computer graphics objects, etc. into the RAW coding of the original camera capture of a digital video by a human color grader and grading it, it is calculated from what starts as an LDR image (in some cases, a high-quality LDR image that does not have too much clipping and has a digitization higher than the usual 8 bits, for example, 10 bits or 12 bits of luma, not too many DCT or other compression artifacts, etc.) instead of being directly generated as the original master HDR. Since most of the video is presented as LDR and may be created as SDR today or in the near future, this is very useful (at least, for example, some of the cameras in multi-camera production output SDR, and for example, a drone captures side video of a sports game, and this side video needs to be converted to the HDR format of the main program). By appropriately combining the capabilities of the mode LDR coding system with the ITM system, the inventors and their technical partners were able to define a system that can perform a double inversion. That is, the up-grading function of the pseudo-HDR image generated by analyzing the original LDR input is substantially the inverse function of the LMF used when coding the LDR communication proxy, and the LDR communication proxy is usually an SDR image broadcast or unicast to the receiver, and when applying the luminance mapping function communicated together in the metadata, it can be re-graded to something that faithfully approximates the master HDR grading by some of the receivers.In this way, it is possible to create a system that actually communicates not only the original (master) LDR image but also the information for creating a good HDR image for it (automatically if the customer creating the content desires, or in other versions with human input, for example, fine-tuning of the automatic settings). Therefore, in practice, (although by communicating the proxy SDR video image) the HDR master image of this video is communicated to the receiver.

[0015] The automaton uses all kinds of rules, such as identifying the locations where light sources exist in the image. However, the exact details are irrelevant to the description of the present invention, and it is simply that the automaton can generate some function LUP (the inverse function of LMF). Similarly, the function of the automaton can be input via the connection UI and applied in the lumamapper 102. Note that in the description of the simplest embodiment, there is only one (downscaling) lumamapper 102 that exists. This does not necessarily have to be a limitation. Since both the EOTF and the lumamapping usually map the normalized input domain [0,1] to the normalized output domain [0,1], there is one or more intermediate normalization mappings that (substantially) map 0 input to 0 output and 1 to 1. In such a case, the former intermediate lumamapping then functions as the basic mapping, and (secondly) the lumamapper 102 then functions as a correction mapping based on the first mapping.

[0016] The quantizer has a set of LDR image luminances Y_SDR corresponding to the HDR image luminance Y_HDR. For example, the darkest pixel in the scene is defined to be displayed with substantially the same luminance on the HDR display and the SDR display, but brighter HDR luminances are scaled within the upper limit of the SDR image as shown by the convex shape of the LMF function displayed within the lumamapper 102, thereby reducing the slope (or converging back towards the diagonal of the normalized axis system). Note that normalization is easily understood by those skilled in the art. It suffices to divide the luma code by power(2;number_of_bits). The normalized luminance can also be normalized, if necessary, by dividing any pixel luminance by the maximum value of its associated target display ML_C, e.g., 4000 nits.

[0017] Thus, an example of an indoor and an outdoor scene can be envisioned. In the real world, since outdoor luminance is typically 100 times brighter than indoor pixels, in an old-style LDR image, indoor objects are shown brightly and vividly colored, but everything outside the window is hard-clipped to uniform white (i.e., invisible). Here, when communicating HDR video with a reversible proxy image, the bright outdoor area visible through the window is made brighter (and possibly less saturated) in a controlled manner such that sufficient information remains available for reconstruction to HDR. This has advantages for both outputs, as systems that want to use the LDR image as is display an excellent depiction of the outdoor scene to the extent possible with the limited LDR dynamic range.

[0018] Therefore, in the sense that it is not necessary for the subsequent circuit to consider whether this LDR image is generated smartly or simply captured directly from the camera like an old LDR system, a set of Y_SDR pixel luminances form a "conventional LDR image" (along with their chrominance, details of which are not required in this explanation). Therefore, the video compressor 103 applies an algorithm such as MPEG HEVC or VVC compression. This algorithm is a set of data reduction techniques that use, among other things, the discrete cosine transform to convert, for example, an 8×8 pixel block into a limited set of spatial frequencies, and this technique does not require much information to represent the spatial frequencies. The amount of information required is adjusted by determining a quantization coefficient that determines how many DCT frequencies are retained and how accurately they are represented. The drawback is that the compressed LDR image (Im_C) is not as accurate as the input SDR image (Im_SDR), especially the occurrence of block artifacts. Depending on the choice of the broadcaster, the block artifacts can become severe, for example, some blocks in the air are represented only by their average luminance, thereby appearing as uniform squares. Since the compressor determines all its settings (including the quantization coefficient) in such a way that the quantization error is hardly visible or at least not a problem in the human system, this block artifact usually does not pose a problem.

[0019] The formatter 104 performs all the signal formatting required for the communication channel (the signal formatting is different, for example, when communicating via storage on a blu-ray disk and when, for example, in the case of DVB-T broadcast). Generally, all the variants have the characteristic that the compressed video image Im_C is grouped into the output image signal S_im together with a luma mapping function LMF that varies (or does not vary) for each image.

[0020] The formatter 151 performs the reverse of the formatting so that the compressed LDR image and the function LMF can be executed by subsequent circuits for reconstructing the HDR image or performing other useful dynamic range mapping processes. The decompressor 152 obtains a sequence of approximate LDR luma Ya_SDR, for example, based on VVC or VP9 decompression and sends it to the reverse HDR image reconstruction pipeline. In addition, the upscaling luma mapper 153 converts the SDR luma to the reconstructed HDR luma YR_HDR (this conversion uses the inverse luma mapping function ILMF which is (substantially) the inverse of the LMF). One of the explanatory diagrams shows two possible receiver (150) devices, where the receiver (150) device exists as a dual function within one physical device (the end user selects which parallel process to apply), or some devices have only one of the parallel processes (for example, some set-top boxes only perform the reconstruction of the master HDR image and store the image in a memory 155 such as a hard disk).

[0021] When a display panel is connected to the receiver embodiment, for example, in the case of a 750 nit ML_D end-user display 190, the receiver has a display adaptation circuit 180 that calculates an output image of 750 nit instead of a reconstructed image of 4000 nit (this is shown by a dotted line to indicate that it is an optional component not related to the teachings of the present invention, but is often used in combination). Without going into detail about the many variations that can achieve display adaptation, there is usually a function determination circuit 157 that proposes an adapted luma mapping function F_ADAP (usually close to the diagonal) based on the shape of the inverse of the LMF. This function is loaded into the display adaptation luma mapper 156 and calculates a low-intensity HDR luma L_MDR that usually has a smaller dynamic range and ends at ML_D = 750 nit instead of ML_C = 4000 nit. Summary of the Invention Problems to be Solved by the Invention

[0022] Problems arise when the ITM automaton determines an LMF function that has soft clipping or, in the worse case, hard clipping for the brightest luminance. Soft clipping or hard clipping means that the slope of the steepest line segment (or tangent to the curve) is small, especially when measured with respect to typical settings of a compressor. For example, if the compressor creates blocks in the sky, which is the brightest object in the LDR image, the inverse LMF function of the decoder will have a large slope for the steepest line segment or part of the curve. This will increase the visibility of artifacts that should not be visible according to normal compression principles, leading to undesirable artifacts. It is possible to force a human grader or an automaton to use only functions with a sufficiently large slope for the brightest part of the curve, but this forcing is undesirable for some images from an artistic point of view or, in some cases, for automata that cannot be easily programmed within a set of rules.

[0023] Therefore, a general solution to the problem, regarded as a problem to be addressed by the inventors, is desired. This solution also has the advantage of not requiring the modification of existing algorithms for determining the shape of the desired luminescence mapping function (LMF), the brightness characteristics of its constituent objects, and its composition for any HDR scene image, for example, when obtained from ITM, using artificial intelligence.

[0024] EP3621307 defines a system that can encode a higher-quality master HDR image by calculating a proxy HDR image with a lower maximum encoded luminance (for example, the proxy video transmitted to the receiver may reach or potentially reach a pixel maximum luminance value of 800 nits, which is assumed to represent the pixel luminance at the same original location up to a maximum of 2000 nits). Further, based on some scale value of a luminance mapping function that determines how to downscale the received 800-nit proxy image to an even lower dynamic range, i.e., the maximum luminance image, a transformation is stretched upward to also define a regrading relationship between two HDR images, namely the creator's master HDR image and the proxy image for the corresponding presentation time. This transformation typically moves a curve (while maintaining its shape) closer to the diagonal of the luma plot, i.e., making the curve more linear or approximating the identity curve, over the entire input and output ranges (i.e., this transformation has nothing to do with clipping behavior. Note also that whether any clipping behavior is exhibited on some receiving-side display is independent of the clipping behavior on the creation or encoding side).

[0025] WO2014 / 128586 describes one possible approach to HDR video encoding, namely, creating a technically imperfect SDR proxy that is visually acceptable and adding a function for converting it to a secondary, visually appealing SDR video to the metadata communicated together.

Means for Solving the Problem

[0026] An encoder (100) for encoding an input high-dynamic range image (Im_HDR_PSEU) as encoded data (S_im), wherein the encoded data includes, firstly, a matrix of pixel colors (Y_SDR, Cb_SDR, Cr_SDR) of an image (Im_SDR) having a lower dynamic range than the high-dynamic range image, and secondly, image metadata (SEI) including a luma mapping function for calculating a high-dynamic range pixel luma (Y_HDR) of the high-dynamic range image by applying the luma mapping function to the pixel luma of the lower-dynamic range image. The encoder comprises an input (997) for receiving a luma mapping function (LMF) from a connectable or provided inverse tone mapping system (200), the inverse tone mapping system being configured to derive an upscaling function (LUP) for calculating a high-dynamic range image (Im_HDR_PSEU) from a master low-dynamic range image (Im_LDR_mastr) by applying an upscaling function to the luma of the master low-dynamic range image based on analyzed characteristics of the master low-dynamic range image, the inverse tone mapping system being configured to obtain the luma mapping function by inverting the upscaling function (LUP). In the encoder (100) for encoding a high-dynamic range image, The encoder comprises a mapping split unit (901) configured to compose a luma mapping function (LMF) with two data items, wherein the first data item defines an extended luma mapping function (LMF_HS) by mapping a clipping point (PtCli) at which the luma mapping function (LMF) first reaches the maximum output value (Vomax) (in the immediate vicinity of the maximum output value, or in many cases exactly the maximum output value) to an end point (Ptmax) having the maximum value of the input values and the maximum value of the output values corresponding to the horizontal scaling factor as coordinates, and by mapping all points of the luma mapping function having input coordinates lower than the input coordinates of the clipping point to extended input coordinates equal to the input coordinates multiplied by the horizontal scaling factor while maintaining the output coordinates of the respective curve points, the second data item is a scale factor preferably equal to the reciprocal of the horizontal scaling factor, and the encoder has a formatter (104) configured to output this extended luma mapping function (LMF_HS) and this scaling value (SCAL) as metadata of the similarly output low dynamic range image (Im_SDR). An encoder for encoding a high dynamic range image, and a corresponding decoder that uses the inverse operation of the first luma mapping and then uses scaling with the received parameters defining the operation When used, decoding artifacts are easily avoided.

[0027] When this operation processes luma, this solution can function in both a relative HDR system, i.e., a system that encodes a certain degree of excessive brightness as a multiple of the (LDR) white level diffused by the maximum luma code, and an absolute HDR coding system defined by the display that encodes the exact pixel luminance (by the corresponding luma code) up to a maximum coded luminance ML_C (e.g., ML_C = 5000 nits, or 1000 nits) as intended to be shown on some target display. The proxy SDR image is an image of such a nature that all HDR colors can be calculated with sufficient accuracy by using a luma mapping function. Of course, there are minor rounding errors, but when, for example, 3×10 bits are used for the image of the pixel colors of Y, Cb, and Cr, the system has been demonstrated to function properly except for issues addressed by current improvement techniques. Other communication standards or channels such as HDMI (registered trademark) have their own placeholders for communicating metadata instead of the SEI message, which is the standard MPEG mechanism for communicating specific metadata as needed. The encoder exists within specialized systems such as television production studios or within consumer devices such as uploading videos taken with a mobile phone to a social networking site. An inverse tone mapping (ITM) system performs the "inverse" (strictly speaking, not a mathematical inverse but an inverse as a functionality) of conventional tone mapping, which is defined to reduce high dynamic range images with a wide range of luminance to standard or low dynamic range luminance. Thus, the ITM system creates some corresponding HDR image (which is also a pseudo-HDR image since it was not originally created as an HDR image) from an input LDR image called the master LDR image (using a similar nomenclature as the master HDR image) in this document. Thus, for example, a system that uses other techniques such as a program of discovery rules or machine learning techniques to maintain the luminance (or their coding luma) of reflective objects as they are, but increase the luminance of emissive objects such as the sun or enhance the luminance of pixels of outdoor objects compared to indoor pixels can be called ITM.In the present invention, the ITM is limited to a variant form that defines the upscaling to an HDR image by using a luma mapping function, typically by using only the luma mapping function. The extended function maintains the shape, i.e., the offset above or below, for example, compared to the diagonal, up to at least some selected endpoints (losing the clipping part), but is a function extended in the size of a certain dimension. For example, the extended function may be stretched horizontally.

[0028] Advantageously, an encoder for encoding a high-dynamic-range image comprises a clipping detection circuit (904) configured to detect whether (Im_HDR_PSEU) has a portion clipped to the maximum output in the input range. According to the core principle, the original LMF function is divided into two luma processing operations. That is, in one operation, a normal function that is mapped to the output maximum only when the input maximum is reached is used, whereby, as a secondary characteristic, it advantageously becomes closer to the diagonal, improving coding and reversibility. It is possible to construct a variant form that processes some bad LMF curves by checking whether there is clipping, or performs some processing on all curves without checking, but continues the mapping in this way when, for example, encountering a function that has already mapped 1.0 to 1.0.

[0029] Advantageously, an encoder for encoding a high-dynamic-range image functions for an image defined on an absolute nit dynamic range where the maximum luminance is the end.

[0030] Advantageously, an encoder for encoding a high-dynamic-range image functions in a coding system in which a low-dynamic-range image (Im_SDR) is predefined as a low-dynamic-range image having a maximum luminance equal to 100 nits.

[0031] Advantageously, an encoder for encoding a high dynamic range image has its mapping division unit (901), and the mapping division unit (901) determines an extended luma mapping function (LMF_HS) by performing a horizontal stretching consisting of a linear scaling such that the input coordinate (XC) where clipping first occurs is mapped to the maximum normalized input coordinate.

[0032] A useful new technical principle can also be embodied as a method for encoding a high dynamic range image (Im_HDR_PSEU). The high dynamic range image is, firstly, represented by a matrix of pixel colors (Y_SDR, Cb_SDR, Cr_SDR) of an image (Im_SDR) with a lower dynamic range than the high dynamic range image, and the low dynamic range image is compressed to be communicated as a compressed low dynamic range image (Im_C), and secondly, the high dynamic range pixel luma (Y_HDR) of the high dynamic range image is calculated by applying a luma mapping function to the pixel luma of the low dynamic range image, and is represented by the image metadata (SEI) including the luma mapping function. In a method in which a luma mapping function (LMF) is received from an inverse tone mapping system (200), and the inverse tone mapping system (200) is configured to derive a luma mapping function for constructing a corresponding high dynamic range image (Im_HDR_PSEU) based on the analyzed characteristics of a master low dynamic range image (Im_LDR_mastr). The method includes the step of splitting the luma mapping function into two consecutive operations, wherein the second of the two consecutive operations includes converting the luma mapping function into an extended luma mapping function (LMF_HS) having a shape that maps a maximum normalized input to a maximum normalized output, the first operation includes linear scaling using a scaling value (SCAL), and the method is configured to output this extended luma mapping function (LMF_HS) and this scaling value (SCAL) as metadata of a low dynamic range image (Im_SDR) that is also output.

[0033] The method has the step of detecting whether (Im_HDR_PSEU) has a portion that is clipped to the maximum output among the input ranges.

[0034] The method functions for an image defined in an absolute nit dynamic range where the maximum luminance is the end. More specifically, the low dynamic range image (Im_SDR) is a low dynamic range image having a maximum luminance equal to 100 nit (also, the HDR image has an ML_C equal to that selected by a video creator, for example, a person who adjusts the settings of ITM, such as ML_C = 2000 nit or 10,000 nit, so as to result in a quality / impression HDR image). The method determines the extended luma mapping function (LMF_HS) by performing a horizontal stretch consisting of linear scaling such that the input coordinates (XC) at which clipping first occurs are mapped to the maximum normalized input coordinates.

[0035] In particular, those skilled in the art will understand that these technical elements can be embodied in various processing elements such as ASICs (application-specific integrated circuits, i.e., typically, causing a method to be executed by an IC (part of an IC) by an IC designer), FPGAs, programmed processors, etc., and can be present in various consumer or non-consumer devices, whether equipped with a display (e.g., a mobile phone that encodes consumer video) or a non-display device that can be externally connected to a display. They will also understand that the image and metadata can be communicated via various image communication technologies such as wireless broadcasts, cable-based communications, etc., and that the device can be used in various image communication and / or usage ecosystems such as television broadcasts, Internet-based on-demand, video surveillance systems, video-based communication systems, etc.

[0036] These and other aspects of the methods and apparatuses according to the present invention will become apparent from and will be elucidated with reference to the implementations and embodiments described below and the accompanying drawings, which function only as non-limiting specific illustrations exemplifying more general concepts. The dashed lines in the accompanying drawings are used to indicate that a component is optional, and a component without a dashed line is not necessarily essential. Dashed lines may also be used to indicate elements that are described as essential but are hidden inside an object, or intangible things such as, for example, the selection of an object / region.

Brief Description of the Drawings

[0037]

Figure 1

Figure 2

Figure 2

Figure 3

Figure 3

Figure 4

Figure 4

Figure 5

Figure 5

Figure 6

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Best Mode for Carrying Out the Invention

[0038] In FIG. 2, the encoding side is shown (as already explained in FIG. 1), and here the ITM (Inverse Tone Mapping) system (200) is also shown together. As explained, the ITM system (200) starts from the master LDR image (Im_LDR_mastr) instead of from the master HDR image (i.e., an image originally created as HDR, for example, at 4000 nits), and ITM can create a good-looking HDR image (Im_HDR_PSEU) from that master LDR image. Since the core system does not actually need to calculate Im_HDR_PSEU and only needs to calculate the calculated best upscaling function LUP, the arrows of the ITM system (200) are shown as dotted lines. However, it should be noted that some variants of ITM may actually calculate the HDR output image and calculate some of their characteristics (e.g., the histogram of the region, the relationship between the brightness of different regions, texture measurements, the profile of the illumination, etc.). The input device that provides the master LDR image is also shown as a dashed line because there may be several systems in which the present invention should be incorporated into the input device. A typical application example is to acquire real-time images from a camera (or formally one or more cameras, and the images of the cameras are mixed, for example, in a mixed master LDR video), and their luminance characteristics (e.g., average luminance) are roughly set by the shader 205. Another example is the case where a completed LDR video is acquired from memory, for example, an old video. The luma mapping function derivation unit 201 derives the optimal upscaling function (LUP) of the incoming LDR image. This function is inverted by the inverter 202, and the inverter 202 provides the inverse function of the upscaling function LUP as the output function LMF_out, and this inverse function functions as the luma mapping function LMF of the LDR-based encoder in the mode. However, (again in the case of the dotted lines 101 and 102), note that no actual downscaling from HDR to LDR is required.The reason is that the ITM system already processes this, and the master LDR image Im_LDR_mastr functions as a proxy image that is communicated instead of the native HDR image. The inverse function of the LUP function is communicated together as metadata (e.g., SEI message) in order to reconstruct a close version of the (pseudo) HDR image IM_HDR_PSEU as the HDR image.

[0039] As schematically shown in FIG. 3, the inventors have found that some ITM deformation forms give an LMF curve with hard clipping to some input images. That is, when the normalized input of the luma mapper 102 is equal to xc (for example, xc = 0.85 in the perceptually uniform luma which is approximately logarithmic), it is mapped to the maximum output of the LDR proxy image (that is, normalized and represented as 1.0, which corresponds to the 8-bit code value Y = R’ = G’ = B’ = 255). In such a function, for example, all inputs higher than 0.85 are similarly written into the LDR image matrix with Y = 255, and the communicated function LMF has a horizontal slope exceeding XC which is, for example, 0.85. As shown in FIG. 4, a typical mode LDR decoder will have related problems when reconstructing the HDR output image. Theoretically, there is an infinite slope at LDR luma Y = 255 or normalized 1.0, which means that in principle, for such an ILMF input function received in the metadata of the image to be processed, the decoder cannot function correctly. In practice, heuristic relaxation measures are taken in the decoder. For example, the decoder uses a quadratic curve ILMF2 derived from the ILMF (or actually, in many cases, the LMF which is the actual function communicated in the metadata). Such an ILMF2 curve typically has the same values (x, y) as the ILMF for all its curve trajectory points except the maximum value. An embodiment is shown in which the highest sub-range of the curve is divided into two parts, one with a small slope and the other with a steep (but not infinite) slope. However, this still leads to significant visual artifacts, especially when the quality of the MPEG compressor of the LDR proxy video is set low, that is, many compression artifacts occur, and these compression artifacts increase significantly especially at the brightest end of the luma range and become more visually annoying.

[0040] As explained with reference to FIG. 5, the inventors understood that when the LMF function is rotated up to the diagonal (beyond the angle ROTH), as a result, the clipping point where the horizontal division starts is on the diagonal and has equal x and y coordinates (a rotated luma mapping function LMF_R is obtained). Since the curve is now closer to the identity transformation, this already clearly gives a good invertible function. However, the problem is that its output is not exactly the output required for this image or this particular HDR scene. For example, instead of the pseudo-HDR luma value XC becoming the whitest white on an LDR display (as it normally is), a higher value XE will only become as bright as YE (e.g., 240 in 8 bits). Only a brighter value, i.e., 1.0 (which may not even exist in the pseudo-HDR image depending on the situation), becomes the whitest LDR color. Thus, the resulting image is too dark. One of the advantages or even purposes of the mode LDR coding / decoding system is to enable consumers still viewing on old-fashioned LDR displays to obtain a very appropriate simulated LDR image corresponding to the ultimately communicated HDR image. So, at this point, since the consumer will see an image that is too dark, this purpose is blocked. It is possible to check how the clipping point (the first time the curve reaches 1.0, or, in a scenario where, for example, soft clipping is performed, the point where it reaches a value very close to 1.0 when some relaxation is allowed, such as deleting some of the brightest values) can be projected onto the diagonal by rotation (ROTH) to obtain the diagonal point PtDi and ultimately the end point Ptmax (shown by the extended arrow starting from the point, which can equally be represented by an arrow starting from (0,0)).

[0041] Therefore, as shown in FIG. 6, it was considered to perform further stretching (DSTR) along the diagonal and brighten the resulting output LDR image again (therefore approaching the originally determined optimal LMF function, but with a clipping part at the upper limit). By actually further studying this system, it was found that the rotation ROTH and diagonal stretching DSTR substantially form an equilateral triangle, so that the Y value of the curve node is substantially reacquired by the operation of stretching after rotation, and the rotated and stretched luma mapping function LMF_RS is obtained. Only the X points are at different positions, which is not yet perfect. Other characteristics of this curve, except for the horizontal clipping part, (although stretched) still maintain the original shape, and as explained, this is at least not desirable for decoding the HDR image from the communicated LDR proxy image with good quality, which is the ultimate goal (instead of building a high-quality and innovative HDR codec only to be able to receive a good-quality LDR image, it is more important to build it so that a nearly perfect HDR image can be viewed on future highest-quality HDR displays).

[0042] However, this substantial equivalence of the y values led to a further insight of removing the second dimension and performing everything in one dimension. That is, the horizontal stretching HSTR of the curve can be easily performed.

[0043] This brings the characteristic that 1D mapping can be used as shown in FIGS. 7 and 8, and further, an improved coding technique (for corresponding mirror decoding) can be designed from the 1D mapping.

[0044] As shown in FIG. 7, the entire re-grading operation in the plot between the normalized input Norm_in (this normalized input has HDR characteristics. That is, these luma can be obtained, for example, from a pseudo-HDR image of 2000 nits corresponding to the master LDR image by ITM and encoded) and the normalized output (this normalized output is also the normalized 8-bit or 10-bit LDR luma) is shown.

[0045] If one wants to map, for example, XC = 0.8 to 1.0 instead of mapping 1.0 to 1.0, the characteristic of this LMF_HS mapping curve (corresponding to LMF_RS in the rotation and stretching embodiment, but in the preferred embodiment, it only results from the horizontal stretching that maps the XC point of the LMF curve obtained from ITM to 1.0 on the horizontal axis of the input range), one should define the point of 0.8 as the new "1.0" in some way.

[0046] This is executed by a pre-scaling operator in the new encoder, and this operator addresses inappropriate ITM LMF functions by correcting inappropriate ITM LMF functions as needed (that is, when there is a hard clipping part in the curve). Norm_in_NEW = Norm_in * (1 / XC) [Equation 1] A linear scaling operation defined by

[0047] This can be regarded as two units of luma mapping when the decoder behaves correspondingly (mirror symmetrically) as shown in FIG. 8.

[0048] If the decoder recognizes the scale factor SCAL of the encoder extension (or any value that can calculate SCAL, such as the horizontal coordinate XC of the clipping start of the LMF function), the decoder can use the block after mapping. Therefore, the decoder first maps the 1.0 input to the 1.0 output and applies the inverted horizontal stretch luma mapping function ILMF_HS having the shape of the inverted LMF_RS. Then, the second block performs compression scaling using the value SCAL communicated in the metadata (or any value related to SCAL that enables the calculation of SCAL on the receiving side), and as a result, the value 1 is finally remapped to XC again. This XC value is, for example, the HDR luma defined in a perceptual quantizer EOTF (a visually homogenized luma system, that is, a system designed so that the equal difference steps of luma give substantially equal visual effects to humans over the entire luma scale) that encodes a planned display luminance of 2500 nit. This value indicates how an HDR image is displayed, for example, when the user has a display of 4000 nit or 5000 nit that can show such an image by directly displaying the intention of the pixel luminance in its exact nit representation. In embodiments with display adaptation, the system takes this into account and calculates, for example, the best possible approximation of a 4000 nit image with the dynamic range capability of a 750 nit display. In fact, what is required as the output of the decoder is not a very high HDR output, but a value at or near the DES in FIG. 4. If there is an intermediate luma mapping, it should be noted that in this strategy, without correcting the infinite decoding slope behavior, the same result as the original clipping mapping function of the ITM cannot be obtained and there will be a slight difference. Research has shown that this difference is not a practical problem of concern in practice. If it is still considered that further improvement is needed, an innovative encoder can calculate a slightly different SCAL value, and as a result, even when there is an intermediate mapper, the LMF_HS curve will overlap more closely with the original LMF curve when used with the scaling process.In practice, when the concept of splitting is formulated, the details of the complete shape of the LMF_HS function are communicated to the decoder as metadata, so the encoder embodiment can also adjust the shape of the LMF_HS function somewhat as needed.

[0049] A preferred way to define the SCAL value is such that the SCAL value can be directly used on the decoding side to scale down based on some luma definition (e.g., PQ luma) (usually, upgrading the received SDR proxy image and then scaling it down to be lower than the theoretical maximum luminance). Thus, for example, when the value 0.89 is mapped to 1.0 on the encoder side, the SCAL value can be defined as the value to which 1.0 should be remapped on the decoding side, i.e., 0.89 (the x - coordinate of the first clipping point PtCli, which is usually a normalized input value and is normalized to have a maximum value of 1.0). It should be noted that the clipping detector can function, for example, by verifying that all higher input values are mapped to the maximum normalized output value or y - coordinate with the value 1.0. Note that since the function and the input SDR image define a pseudo - HDR image, i.e., it is the coding of a pseudo - HDR image, there is no need to check anything regarding the pseudo - HDR image that can be generated by ITM. The extended HSTR corresponds to multiplying the input - value coordinate by a multiplier value equal to 1 / SCAL for each point on the curve that has an output value (y - coordinate) fixed to a value Y between 0 and 1. Thus, for example, the x - coordinate equal to 0.5, also called the input value of the original LMF curve, is moved to the input value of the LMF_HS curve which is 0.5 * 1.1236 = 0.56 (by multiplication with the reciprocal 1 / SCAL, also called the inverse - scale factor), and that point has the same y - coordinate as the original point on the LMF curve.

[0050] Thus, as shown in FIG. 9, when the decoder uses these two consecutive luma mapping stages, the new ITM - improved encoder can handle this.

[0051] FIG. 9 shows an embodiment of an improved encoder according to current innovative insights. Also in this case, most of the technical units are substantially as described above.

[0052] The innovative mapping division unit 901 inputs, into the metadata associated with an image (or a set of similarly mapped images), instead of a mere luma mapping function, a luma mapping function adjusted using one of the embodiments described with reference to FIGS. 3 to 8, for example, a horizontally stretched luma mapping function LMF_HS (or a rotated and stretched luma mapping function resulting from its embodiment). The mapping division unit 901 also outputs the corresponding scale factor SCAL into the metadata, and as a result, the overall of these two operations substantially maps the result of mapping using only the original LMF obtained from the ITM, but the behavior in the decoder is improved. In addition, the mapping division unit 901 includes a function reforming circuit 902 configured to determine an extended version of the LMF such that the maximum normalization input (1.0) is mapped to the maximum normalization output. The mapping division unit 901 also includes a scaling factor determination unit 903 configured to determine and output the scale factor SCAL. Cb_SDR and Cr_SDR are the blue and red chroma components of the pixel color of the LDR proxy image Im_SDR in typical color encoding, respectively (a person skilled in the art recognizes how to reformulate the pixel color, for example, to R'G'B' encoding, using known colorimetric techniques). The image signal output S_im for communication to the receiver includes the compressed LDR image Im_C and the metadata (SEI) including LMF_HS and SCAL. Further, the function reforming circuit 902 is typically embodied to analyze whether the LMF function requires reforming. For example, the function reforming circuit 902 includes a clipping detection circuit 904 configured to detect whether the LMF function of the ITM is clipped at the maximum value, that is, typically, whether it is a strictly increasing function mainly, and whether an input value less than 1.0 is already mapped to the output 1.0. If function reforming is not required, the function reforming unit can be embodied to perform an identity transformation. Various units are physically combined.

[0053] The scale factor relates to the start of clipping (up to the maximum output value) and should not be confused with other scale factors in HDR technology. For example, it is a value that adaptively controls how much the curve should move towards or away from the diagonal in a normalized luma plot, which corresponds to the identity transformation. (For example, a maximum brightness image of 1000 nits is optimal for a 1000 nit maximum-capability display, and thus there is no need to apply a mapping to these image lumas to obtain the drive image of the display).

[0054] The algorithm components disclosed in this document are actually implemented either (wholly or partially) as hardware (e.g., part of an application-specific IC) or as software executed on a special digital signal processor or a general-purpose processor, etc.

[0055] Those skilled in the art should be able to understand from this disclosure which components are optional improvements and can be implemented in combination with other components, and how the (optional) steps of the method correspond to the respective means of the device and vice versa. The term "device" in this application is used in the broadest sense, i.e., as a group of means that enables the realization of a specific purpose, and thus can be, for example, an IC (a small circuit part thereof), a dedicated device (such as a device with a display), or a part of a networked system. "Arrangement" is also intended to be used in the broadest sense, and thus "arrangement" includes, inter alia, a single device, a part of a device, an aggregate of cooperating devices (a part thereof), etc.

[0056] The term "computer program product" should be understood to encompass any physical realization of a set of commands that enable a general-purpose or special-purpose processor to input commands to the processor after a series of load steps, including intermediate conversion steps such as translation into an intermediate language and a final processor language, to execute any of the 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 existing in memory, data moving via a network connection (wired or wireless), or program code on paper. Apart from the program code, characteristic data necessary for the program is also embodied as a computer program product.

[0057] Some of the steps required for the operation of the method, such as data input and output steps, already exist in the function of the processor instead of being described in the computer program product.

[0058] It should be noted that the above embodiments are illustrative rather than limiting of the present invention. In order to be concise, not all of these options have been described in detail when those skilled in the art can easily realize the mapping of the examples presented in other areas of the claims. Apart from the combinations of the elements of the present invention combined in the claims, other combinations of elements are also possible. Any combination of elements can be realized in a single dedicated element.

[0059] The reference signs between parentheses in the claims are not intended to limit the claims in any way. The word "comprising" does not exclude the existence of elements or aspects not listed in the claims. A singular element does not exclude the existence of a plurality of such elements.

Claims

1. 1. An encoder for encoding an input high dynamic range image (Im_HDR_PSEU) as encoded data (S_im), the encoded data comprising: first, a matrix of pixel colors (Y_SDR, Cb_SDR, Cr_SDR) of an image (Im_SDR) having a lower dynamic range than the high dynamic range image; and second, metadata (SEI) of the image comprising a luma mapping function for calculating high dynamic range pixel lumas (Y_HDR) of the high dynamic range image by applying the luma mapping function to pixel lumas of the image with a lower dynamic range, 1. An encoder for encoding a high dynamic range image, the encoder comprising an input for receiving a luma mapping function (LMF) from an inverse tone mapping system connectable or provided thereto, the inverse tone mapping system deriving the upgrading function (LUP) for calculating the high dynamic range image (Im_HDR_PSEU) from a master low dynamic range image (Im_LDR_master) by applying an upgrading function to the luma of the master low dynamic range image (Im_LDR_master) based on analyzed characteristics of the master low dynamic range image (Im_LDR_master), and the inverse tone mapping system inverting the upgrading function (LUP) to obtain the luma mapping function, the encoder comprises a mapping division unit for defining the luma mapping function (LMF) as two data items, the first data item being a stretched luma mapping function (LMF_HS) defined by mapping a clipping point (PtCli) at which the luma mapping function (LMF) first reaches a maximum output value (Vomax) to an end point (Ptmax) having a maximum value of input values ​​and the maximum output value as coordinates, the end point corresponding to a horizontal scaling factor, and by mapping all points of the luma mapping function having input coordinates lower than the input coordinates of the clipping point to stretched input coordinates equal to the respective input coordinates multiplied by the horizontal scaling factor, while maintaining the output coordinates of the respective curve points; 10. An encoder for encoding a high dynamic range image, wherein the second data item is a scale factor preferably equal to the inverse of the horizontal scaling factor, and wherein the encoder has a formatter that outputs the expanded luma mapping function (LMF_HS) and scaling value (SCAL) as metadata of the low dynamic range image (Im_SDR) that is also output.

2. 2. The encoder for encoding high dynamic range images of claim 1, comprising a clipping detection circuit that detects whether a luma mapping curve (LMF) provided to the input by the inverse tone mapping system has a portion of the input range clipped to a maximum output.

3. 3. An encoder for encoding high dynamic range images according to claim 1 or 2, wherein the image is defined on an absolute nit dynamic range, the end of which is the maximum luminance.

4. 4. The encoder for encoding a high dynamic range image according to claim 3, wherein said low dynamic range image (Im_SDR) is a low dynamic range image having a maximum luminance equal to 100 nits.

5. encoding an input high dynamic range image (Im_HDR_PSEU) as encoded data (S_im), the encoded data including, first, a matrix of pixel colors (Y_SDR, Cb_SDR, Cr_SDR) of an image (Im_SDR) having a lower dynamic range than the high dynamic range image, and, second, metadata (SEI) of the image including a luma mapping function for calculating high dynamic range pixel lumas (Y_HDR) of the high dynamic range image by applying the luma mapping function to pixel lumas of the image with a lower dynamic range; receiving a luma mapping function (LMF) from a connectable or provided inverse tone mapping system, wherein the inverse tone mapping system derives the upgrading function (LUP) for calculating the high dynamic range image (Im_HDR_PSEU) from the master low dynamic range image (Im_LDR_master) by applying an upgrading function to the luma of the master low dynamic range image (Im_LDR_master) based on the analyzed characteristics of the master low dynamic range image (Im_LDR_master), and the inverse tone mapping system inverts the upgrading function (LUP) to obtain the luma mapping function (LMF); In a method comprising: the method includes a step of dividing the luma mapping function (LMF) into two data items, the first data item being a stretched luma mapping function (LMF_HS) defined by mapping a clipping point (PtCli) at which the luma mapping function (LMF) first reaches a maximum output value (Vomax) to an end point (Ptmax) having as coordinates a maximum value of input values ​​and the maximum output value, the end point corresponding to a horizontal scaling factor, and by mapping all points of the luma mapping function having input coordinates lower than the input coordinates of the clipping point to stretched input coordinates equal to the respective input coordinates multiplied by the horizontal scaling factor, while maintaining the output coordinates of the respective curve points; a second data item is a scale factor preferably equal to the reciprocal of said horizontal scaling factor; outputting the decompressed luma mapping function (LMF_HS) and scaling value (SCAL) as metadata of the low dynamic range image (Im_SDR), which is also output; A method comprising:

6. The method of claim 5 , comprising detecting whether the luma mapping function has a portion of the input range clipped to a maximum output.

7. 7. The method of claim 5 or 6, wherein the image is defined over an absolute nit dynamic range, the end of which is the maximum luminance.

8. The method of claim 7, wherein the low dynamic range image (Im_SDR) is a low dynamic range image having a maximum luminance equal to 100 nits.

Images