Fact-checked by Grok 2 weeks ago

RGB color model

The RGB color model is an model that represents colors in , , and electronic displays by specifying the intensities of three primary components: , , and (RGB). This model operates on the principle of additive color mixing, where combining varying amounts of these primaries produces a wide of colors—mixing all three at full intensity yields white , while equal mixtures of two create secondary colors like , , and . Originating from the trichromatic theory of human , which posits that the eye perceives color through three types of cells sensitive to , , and wavelengths, the RGB model became foundational for (CRT) displays in the mid-20th century and has since been adapted for liquid-crystal displays (LCDs), LEDs, and digital cameras. Key specifications of the RGB model include defined coordinates for the primaries, a reference (typically CIE D65 illuminant simulating daylight), and —a nonlinear (often approximated as 2.2 for ) that adjusts signal values to better match human nonlinear perception of brightness, improving efficiency in digital encoding. For (HDTV), the BT.709 standard establishes precise RGB primaries (red at x=0.64, y=0.33; green at x=0.30, y=0.60; blue at x=0.15, y=0.06) with a D65 , enabling consistent color reproduction across broadcast and production workflows. The variant, proposed by and in 1996 and formalized as IEC 61966-2-1 in 1999, serves as the default for the , browsers, and consumer devices, using 8 bits per for 16.7 million possible colors and a viewing of 80 cd/m² under 64 ambient light. Other variants, such as Adobe RGB, expand the color for professional and by shifting primaries to cover more of the . Despite its device-dependency—colors can vary across hardware—the RGB model's simplicity and alignment with light emission make it indispensable for real-time rendering in and video.

Core Concepts

Additive Color Mixing

The RGB color model operates on the principle of mixing, where colors are produced by the superposition of , , and intensities, enabling the creation of a wide of perceptible colors. In this system, from these three primaries is combined such that increasing the intensity of each component brightens the resulting color, with equal maximum intensities of all three primaries yielding white . This approach leverages the of addition, allowing any color within the model's to be approximated by adjusting the relative intensities of the primaries. The theoretical foundation for additive color mixing in RGB is provided by Grassmann's laws, formulated in , which describe the empirical rules governing how mixtures of colored lights are perceived. These laws include (scaling intensities scales the perceived color), additivity (the mixture of two colors added to a third equals the sum of their separate mixtures with the third), and the invariance of matches under certain conditions, ensuring that color mixtures behave as vector additions in a . For instance, pure is represented by intensities R=1, G=0, B=0, while varying these values—such as R=1, G=1, B=0 for —spans the through linear combinations, approximating the full range of human color perception enabled by the visual system's . In contrast to additive mixing, subtractive color models like (cyan, , ) used in pigments and absorb wavelengths, starting from and yielding darker colors as components are added, which is why additive mixing is particularly suited for light-emitting devices such as displays where is directly projected and combined. The perceived color in additive mixing can be conceptually expressed as the linear superposition of the primary spectra weighted by their intensities:
\mathbf{C} = r R + g G + b B
where \mathbf{C} is the resulting , r, g, and b are the spectral distributions of the , , and primaries, and R, G, B are their respective intensity scalars (normalized between 0 and 1). This equation underscores the model's reliance on additive principles without deriving from .

Choice of RGB Primaries

The RGB color model is grounded in the trichromatic theory of human , which posits that color arises from three types of cone photoreceptors in the , each sensitive to different ranges of . These include long-wavelength-sensitive (L) cones peaking around 564–580 nm (perceived as ), medium-wavelength-sensitive (M) cones peaking around 534–545 nm (perceived as ), and short-wavelength-sensitive (S) cones peaking around 420–440 nm (perceived as ). This physiological basis directly informs the selection of , , and as primaries, as they align with the peak sensitivities of these cones, enabling efficient representation of the through additive mixing. The choice of RGB primaries is further guided by physical principles aimed at optimizing color reproduction. In the , primaries are selected to maximize the —the range of reproducible colors—while balancing luminous efficiency, where green contributes the most to perceived brightness due to the higher sensitivity of the human visual system to mid-wavelength light (corresponding to the Y tristimulus value in CIE ). This selection ensures broad coverage of perceivable colors without excessive energy loss, as the primaries form a triangle in the chromaticity diagram that encompasses a significant portion of the spectral locus. Historically, the CIE standardized RGB primaries in 1931 based on color-matching experiments, defining monochromatic wavelengths at approximately 700 nm (red), 546.1 nm (), and 435.8 nm (blue) to establish the CIE RGB color space. These evolved into modern standards like , proposed in 1996 by and , which specifies primaries with chromaticities of red at x=0.6400, y=0.3300; green at x=0.3000, y=0.6000; and blue at x=0.1500, y=0.0600 in the CIE 1931 diagram, tailored for typical consumer displays and web use. A key limitation of RGB primaries is metamerism, where distinct spectral distributions can produce identical color matches under one illuminant (e.g., daylight) but appear different under another (e.g., incandescent ), due to the incomplete spectral sampling by only three primaries. Primary selection also involves conceptual optimization to achieve perceptual uniformity, such as minimizing color differences measured along ellipses—ellipsoidal regions in chromaticity space representing just-noticeable differences—to ensure even spacing of colors in human perception.

Historical Development

Early Color Theory and Experiments

The foundations of the RGB color model trace back to early 19th-century physiological theories of vision. In 1801, Thomas Young proposed the trichromatic hypothesis, suggesting that human color perception arises from three distinct types of retinal receptors sensitive to different wavelength bands, providing the theoretical basis for using three primary colors to represent the full spectrum of visible hues. This idea built on earlier observations of but shifted focus to the eye's internal mechanisms rather than purely physical properties of light. Hermann von Helmholtz refined Young's hypothesis in the 1850s, elaborating it into a more detailed physiological model by classifying the three cone types as sensitive to , , and light, respectively, and emphasizing their role in mixing to produce all perceivable colors. Helmholtz's work integrated experimental data on and spectral responses, establishing the trichromatic framework as a cornerstone for subsequent RGB-based theories. In 1853, formalized the mathematical underpinnings of color mixing in his paper "Zur Theorie der Farbenmischung," proposing that colors could be represented as vectors in a three-dimensional linear space where any color is a of three primaries, adhering to laws of additivity, , and superposition. This provided a rigorous for RGB representations, enabling quantitative predictions of color mixtures without relying solely on perceptual descriptions. James Clerk Maxwell advanced these ideas through experimental demonstrations of synthesis in the 1850s and 1860s. In his 1855 paper "Experiments on Colour," Maxwell described methods to mix colored lights to match spectral hues, confirming that , , and primaries could approximate a wide range of colors via superposition. Building on this, Maxwell's 1860 paper "On the Theory of Compound Colours" detailed color-matching experiments using a divided disk and lanterns, further validating the trichromatic approach. The culmination came in , when Maxwell projected the first synthetic full-color image by superimposing , , and filtered projections of black-and-white photographs of a tartan ribbon, demonstrating practical at the Royal Institution. Later in the 1880s, Arthur König and Conrad Dieterici conducted key measurements of spectral sensitivities in normal and color-deficient observers, estimating the response curves of the three types and confirming their peaks in the , , and regions of the . Their 1886 work, "Die Grundempfindungen in normalen und anormalen Farbsystemen," used flicker photometry on dichromats to isolate individual cone fundamentals, providing empirical support for the physiological basis of RGB primaries. Despite these advances, early RGB theories faced limitations in representing the full of human-perceivable colors, as real primaries like those chosen by could not span the entire chromaticity space without negative coefficients. This issue persisted into the , leading the (CIE) in 1931 to define the XYZ color space with imaginary primaries that avoid negative values and encompass all visible colors, highlighting the incompleteness of spectral RGB models for absolute color specification.

Adoption in Photography

The adoption of RGB principles in photography marked a pivotal shift from monochrome to color imaging, building on foundational additive color theory. In 1907, the brothers introduced the Autochrome process, the world's first commercially viable color film, which employed an additive mosaic screen of grains dyed red, green, and blue-violet—approximating RGB primaries—to filter light onto a panchromatic . This innovation, comprising about 4 million grains per square inch, allowed the capture and viewing of full-color transparencies by recombining filtered light, though it required longer exposures than film. By the 1930s, subtractive processes incorporating RGB separations gained prominence, exemplified by Eastman Kodak's 1935 launch of , the first successful multilayer for amateurs. Developed by Leopold Mannes and Leopold Godowsky Jr., it used three panchromatic emulsion layers sensitized to , , and wavelengths via color couplers, producing , , and dyes during controlled development to form positive transparencies. This RGB-based separation evolved from earlier additive experiments, enabling vibrant slides for 16mm cine and 35mm without the need for multiple exposures. In the 1940s, three-color separation techniques became standard in commercial printing, where RGB-filtered negatives were used to create subtractive overlays in processes like Kodak's Dye Transfer, facilitating high-volume color reproduction for magazines and advertisements. The transition to in the 1970s introduced sensor-based RGB capture, with Bryce E. Bayer's 1976 for the array revolutionizing image sensors. This mosaic pattern overlays red, green, and blue microlenses on a grid of photosites in and devices—twice as many green for sensitivity—capturing single-color data per , which algorithms then interpolate to yield complete RGB values. By the , this technology standardized in consumer digital cameras, such as early models from and , outputting RGB-encoded images that bypassed film processing and enabled instant for the masses. Early additive RGB films faced technical hurdles, including color fringing from misalignment between the filter mosaic and emulsion layers, which caused edge artifacts in Autochrome plates due to imperfect registration during manufacturing or viewing.

Implementation in Television

The implementation of the RGB color model in television began with early mechanical experiments, notably John Logie Baird's 1928 demonstration of a color television system using a Nipkow disk divided into three sections with red, green, and blue filters to sequentially capture and display color images additively. The transition to electronic television culminated in the 1953 NTSC standard approved by the FCC, which utilized shadow-mask cathode-ray tubes (CRTs) featuring arranged in triads on the screen interior. These CRTs incorporated three guns—one each for , , and —to generate and modulate the respective primary signals, with the shadow mask ensuring that each beam precisely excites only its corresponding phosphor dots, thereby producing the intended color at each location. For broadcast transmission within the limited 6 MHz channel , the correlated RGB signals were transformed into the color space, where the (Y) component, derived as a weighted sum of RGB (Y = 0.299R + 0.587G + 0.114B), occupied the full bandwidth for compatibility, while the (I and Q) components were modulated onto a 3.58 MHz subcarrier with reduced bandwidths (1.5 MHz for I and 0.5 MHz for Q) to exploit human vision's lower acuity for color details. Additionally, was introduced during this era to counteract the CRT's nonlinear power-law response (approximately γ ≈ 2.5), applying a pre-distortion (V_out = V_in^{1/γ}) in the camera chain to achieve linear light output matching scene reflectance. In the , European systems like PAL (introduced in 1967) and (1967) retained RGB primaries closely aligned with specifications for compatibility in international production, but diverged in encoding: PAL alternated the phase of the subcarrier (4.43 MHz) between lines to mitigate hue errors, while sequentially transmitted frequency-modulated blue-luminance and red-luminance differences. Studio equipment for these formats employed full-bandwidth RGB component signals—equivalent to sampling in modern digital parlance—enabling uncompressed color handling during production, effects, and editing before conversion to the broadcast-encoded form. The advent of digital high-definition television marked a key evolution, with Recommendation BT.709 (adopted in 1990) establishing precise parameters, including primaries (x_r=0.64, y_r=0.33; x_g=0.30, y_g=0.60; x_b=0.15, y_b=0.06) and D65 , optimized for progressive or interlaced displays in HDTV production and exchange. This standard facilitated the shift from analog RGB modulation to digital sampling while preserving the additive mixing principles for accurate color reproduction on CRT-based HDTV sets.

Expansion to Computing

The expansion of the RGB color model into personal computing began in the late 1970s and early 1980s, as microcomputers transitioned from monochrome displays to basic color capabilities. Early systems like the (1977) supported limited RGB-based color through outputs, but the IBM Personal Computer's introduction of the (CGA) in 1981 marked a pivotal shift for the emerging PC market. CGA introduced a 4-color mode at 320×200 resolution (black, cyan, magenta, and white), using 2 bits per pixel with RGBI signaling to approximate additive colors, enabling simple graphics and text in color for business and gaming applications. This was followed by the (EGA) in 1984, which expanded to 16 simultaneous colors from a palette of 64 (2 bits per RGB channel), supporting resolutions up to 640x350 and improving visual fidelity for . By the mid-1980s, demand for richer visuals drove further advancements, culminating in IBM's (VGA) standard in 1987 with the PS/2 line. VGA introduced a 256-color palette derived from an 18-bit RGB space (6 bits per channel, yielding 262,144 possible colors) at 640x480 resolution, allowing more vibrant and detailed imagery through modes like Mode 13h. In the early 1990s, (SVGA) extensions from vendors like enabled (24-bit RGB, or 8 bits per channel, supporting 16.7 million colors) at higher resolutions, such as 800x600 or 1024x768, facilitated by chips like the S3 928 (1991) with up to 4MB VRAM for direct color modes without palettes. These developments standardized RGB as the foundational model for PC graphics hardware, bridging the gap from television-inspired analog signals to digital bitmap rendering. Key software milestones accelerated RGB's integration into computing workflows. Apple's Macintosh II, released in 1987, was the first Macintosh to support color displays via the AppleColor High-Resolution RGB Monitor, using 24-bit RGB for up to 16.7 million colors and enabling early desktop applications with full-color graphics. Microsoft Windows 95 (1995) further popularized high-fidelity RGB by natively supporting 24-bit color depths, allowing seamless rendering of 16.7 million colors in graphical user interfaces and applications. The release of 1.0 in 1992 by , managed by the , provided a cross-platform for 3D RGB rendering pipelines, standardizing vertex processing and operations for real-time graphics. Microsoft's 1.0 (1995) complemented this by offering Windows-specific APIs for RGB-based 2D and 3D acceleration, including for bitmap surfaces and for scene composition. The evolution of graphics processing units (GPUs) in the late 1990s amplified RGB's role in . NVIDIA's (1999), the first GPU, integrated transform and lighting engines to handle complex RGB pixel shading at high speeds, evolving from fixed-function pipelines to programmable shaders for dynamic color blending and texturing. This progression established RGB as the default for bitmap graphics in operating systems and software, profoundly impacting by enabling affordable color layout tools like (1985 onward), which leveraged RGB monitors for editing before CMYK conversion for print. The shift democratized visual content creation, transforming publishing from specialized to accessible workflows.

RGB in Devices

Display Technologies

The RGB color model is fundamental to the operation of various display technologies, where it enables the reproduction of a wide range of colors through the controlled emission or modulation of , green, and . In (CRT) displays, three separate electron beams, each modulated by the respective RGB signal, strike a phosphor-coated screen to produce light; the phosphors, such as those in the P22 standard, emit , green, and upon , with a ensuring precise alignment to prevent color fringing. This additive mixing allows CRTs to approximate the by varying beam intensities, achieving color gamuts close to the standard in consumer applications. Liquid crystal display (LCD) and light-emitting diode (LED)-backlit panels implement RGB through a of subpixels, each filtered to transmit red, green, or blue wavelengths from a white source. In these systems, (TFT) arrays control the voltage applied to liquid crystals, modulating light transmission per subpixel to form full-color pixels; post-2010 advancements incorporate quantum dots as color converters to enhance gamut coverage, extending beyond traditional limits toward DCI-P3. LED backlights, often using white LEDs with RGB phosphors, provide higher efficiency and brightness compared to earlier CCFL sources. Organic light-emitting diode () displays utilize self-emissive RGB pixels, where organic materials in each subpixel emit light directly when an is applied, eliminating the need for a and enabling perfect blacks through selective pixel deactivation. This structure offers superior contrast ratios and viewing angles, with white RGB (WRGB) variants—employing an additional white subpixel—improving power efficiency for brighter outputs without sacrificing color accuracy. OLEDs typically achieve wide color gamuts, covering up to 95% of (as of 2024), due to the precise emission spectra of organic emitters. To account for the nonlinear response of these display devices, where light output is not linearly proportional to input voltage, gamma encoding is applied in the RGB signal pipeline; a common gamma value of 2.2 compensates for this by encoding the signal such that the decoded output follows the device's power-law response curve. The decoding relationship is given by: V_{\text{out}} = V_{\text{in}}^{\frac{1}{\gamma}} where V_{\text{in}} is the encoded voltage (0 to 1), V_{\text{out}} is the linear light intensity, and \gamma is the display's gamma factor. This perceptual linearization ensures efficient use of and matches human vision's logarithmic sensitivity. In modern (HDR) displays, the RGB model is extended to support greater luminance ranges and bit depths, with the standard (published in 2012) defining wider primaries and a 10-bit or higher encoding to enable peak brightness exceeding 1000 nits while preserving color fidelity in both SDR and HDR content. These advancements, integrated into and quantum-dot-enhanced LCDs, allow RGB-based systems to render over a billion colors with enhanced detail in shadows and highlights.

Image Capture Systems

In digital cameras, the RGB color model is implemented through single-sensor designs that capture light via a color array (CFA) overlaid on the . The most prevalent CFA is the pattern, which arranges red, green, and blue s in an RGGB mosaic, where green s occupy half the s to align with human visual sensitivity. This setup allows each photosite to record intensity for only one color channel, producing a mosaiced image that requires subsequent processing to reconstruct full RGB values per . To obtain complete RGB data, algorithms missing color values from neighboring pixels, employing techniques such as edge-directed interpolation to minimize artifacts like . For instance, estimates values based on adjacent samples, while more advanced methods, like gradient-corrected , adapt to local structures for higher fidelity. These processes ensure that the final RGB approximates the scene's mixing as captured by the . Scanners employ linear RGB CCD arrays to acquire color-separated signals, with mechanisms differing between flatbed and drum types. Flatbed scanners use a trilinear CCD array—comprising three parallel rows of sensors, each dedicated to , , or —mounted on a movable that scans beneath a glass platen, capturing reflected in a single pass for efficient RGB separation. Drum scanners, in contrast, rotate the original around a light source while a fixed head with tubes (one per RGB channel) reads transmitted or reflected , enabling higher resolution and reduced motion artifacts through precise color isolation via fiber optics. Post-capture processing in image acquisition systems includes white balance adjustment, which normalizes gains to compensate for varying illuminants, ensuring neutral reproduction of whites. This involves scaling the raw signals—such as multiplying red by a factor of 1.32 and by 1.2 under daylight—based on responses to a reference , thereby aligning the captured RGB values to a like D65. Algorithms automate this by estimating illuminant and applying per-channel corrections. Key advancements in RGB capture include the rise of sensors in the 1990s, which integrated analog-to-digital conversion on-chip, reducing costs and power consumption compared to traditional CCDs while supporting Bayer-ed RGB acquisition in consumer devices. Additionally, formats preserve the sensor's linear RGB data—captured as intensities per filter without —retaining 12-bit or higher precision for flexible post-processing. Challenges in RGB image capture arise from noise in low-light conditions, where photon shot noise and read noise disproportionately affect channels, leading to imbalances such as elevated green variance due to its higher sensitivity, which can distort color fidelity. Spectral mismatch between sensor filters and ideal RGB primaries further complicates accurate reproduction, as real-world filters exhibit overlapping responses (e.g., root mean square errors of 0.02–0.03 in sensitivity estimation), causing metamerism under non-standard illuminants.

Digital Representation

Numeric Encoding Schemes

In digital imaging and computing, RGB colors are encoded using discrete numeric values to represent intensities for each channel, with the bit depth determining the precision and range of colors. The standard color depth for most consumer applications is 8 bits per channel (bpc), yielding a total of 24 bits per pixel and supporting 16,777,216 possible colors (2^8 × 3 channels). This encoding maps intensities to values from 0 (minimum) to 255 (maximum) per channel, providing sufficient gradation for typical displays while balancing efficiency. Higher depths, such as 10 bpc for a 30-bit total, are employed in broadcast and professional video workflows to minimize visible banding in smooth gradients, offering 1,073,741,824 colors and improved for transmission standards like those in systems. Key standards define specific encoding parameters, including gamut, transfer functions, and bit depths, to ensure consistent color reproduction across devices. The standard, proposed by and in 1996 and formalized by the (IEC) in 1999 as IEC 61966-2-1, serves as the default for web and consumer electronics, using 8-bit integer encoding with values scaled from 0 to 255 and a gamma-corrected for perceptual uniformity. Adobe RGB (1998), introduced by Systems to accommodate wider color gamuts suitable for production, employs similar 8-bit or 16-bit integer encoding but significantly expands the reproducible colors over , particularly in cyan-greens, while maintaining compatibility with standard RGB pipelines. For professional photography requiring maximal color fidelity, ProPhoto RGB—developed by as an output-referred space—supports large gamuts exceeding Adobe RGB, typically encoded in 16-bit integer or floating-point formats to capture subtle tonal variations without clipping. Quantization converts continuous or normalized values to discrete integers for storage. In linear RGB spaces, for a linear intensity I normalized to [0, 1], the quantized value is \text{RGB\_value} = \round(I \times (2^n - 1)), where n is bits per . However, in gamma-encoded spaces like , linear intensities are first transformed using a (e.g., approximate gamma of 2.2) to perceptual values V, then quantized as \text{RGB\_value} = \round(V \times (2^n - 1)); this ensures even perceptual distribution as per encoding standards. Binary integer formats dominate for standard dynamic range (SDR) images, using fixed-point representation in the 0–255 scale for 8 bpc, while (HDR) applications employ floating-point encoding to handle values exceeding [0, 1], such as in files, which support 16-bit half-precision or 32-bit single-precision floats per for RGB data, enabling over 1,000 steps per f-stop in . File storage of RGB data must account for byte to maintain portability across architectures. Little-endian , common in x86 systems, stores the least significant byte first (e.g., for 16-bit channels, low byte precedes high byte), as seen in formats like ; big-endian reverses this, placing the most significant byte first, which is specified in headers for versatile formats like to allow cross-platform decoding without .

Geometric Modeling

The RGB color model is geometrically conceptualized as a unit in three-dimensional Cartesian , with orthogonal axes representing the normalized intensities of the (R), green (G), and blue (B) primary components, each ranging from 0 to 1. The at the (0,0,0) corresponds to , signifying the absence of , while the opposing at (1,1,1) denotes , the maximum additive combination of the primaries. This cubic representation facilitates intuitive visualization of color mixtures, where any point within the cube defines a unique color through vector addition of the basis vectors along each axis. As a over the reals, \mathbb{R}^3, the RGB model treats colors as position vectors, enabling linear algebraic operations for color computation and manipulation. For instance, , or lerping, between two colors \mathbf{A} = (R_A, G_A, B_A) and \mathbf{B} = (R_B, G_B, B_B) produces a of intermediate colors along the connecting them, parameterized as: \mathbf{C}(t) = (1-t)\mathbf{A} + t\mathbf{B}, \quad t \in [0,1]. This operation yields smooth gradients essential for rendering transitions in graphics and imaging, leveraging the model's inherent derived from additive mixing. The reproducible color in RGB forms a polyhedral volume, which, when transformed to the CIE tristimulus via a linear , approximates a bounded by the three primary chromaticities and the . This tetrahedral structure encapsulates the subset of visible colors achievable by varying R, G, and B intensities within [0,1], excluding negative or super-unity values that lie outside the cube. Out-of- handling, such as clipping, projects such colors onto the nearest gamut boundary to ensure device-reproducible outputs without introducing invalid tristimulus values. Despite its mathematical elegance, the RGB space suffers from perceptual non-uniformity, where geometric distances do not align with human visual sensitivity, prompting conversions to uniform spaces like CIELAB for psychovisually accurate analysis. The Euclidean distance metric, \|\mathbf{C_1} - \mathbf{C_2}\| = \sqrt{(R_1 - R_2)^2 + (G_1 - G_2)^2 + (B_1 - B_2)^2}, quantifies linear separation but over- or underestimates perceived differences, as human color perception follows non-Euclidean geometries influenced by cone responses and adaptation. Gamut comparisons across RGB variants rely on volume computations via tetrahedral tessellation in CIE XYZ, summing signed volumes of sub-tetrahedra to assess coverage and overlap relative to the full visible spectrum.

Applications and Extensions

Web and Graphic Design

In web and , the RGB color model forms the foundation for specifying and manipulating colors through standardized syntax in Cascading Style Sheets (CSS). The rgb() function allows designers to define colors by providing , , and component values, typically as integers from 0 to 255 or percentages from 0% to 100%, such as rgb(255, 0, 0) for pure . This syntax originated in CSS Level 1, recommended by the (W3C) in 1996, enabling precise control over element colors like text and backgrounds. Complementing this, hexadecimal notation serves as a compact , using formats like #FF0000 for the same or the abbreviated #F00, where each pair of digits represents the of one RGB . Web standards have entrenched as the default for RGB specifications to ensure consistent rendering across devices. Adopted by the W3C in alignment with the 1996 proposal from and , sRGB provides a standardized for , minimizing discrepancies in color display. In technologies like (SVG), all colors are defined in sRGB, supporting RGB values in attributes for fills, strokes, and gradients to create resolution-independent visuals. Similarly, the Canvas API leverages RGB-based CSS colors for dynamic pixel manipulation, allowing to draw and edit images by setting properties like fillStyle to RGB or values. Graphic design tools integrate RGB as a core output for color selection and workflow adaptation. Color pickers in applications like display and export RGB values alongside other formats, enabling designers to sample hues from images or palettes and apply them directly to digital assets. When transitioning designs from (often CMYK-based) to formats, gamut mapping algorithms adjust out-of-gamut RGB colors to fit constraints, preserving visual intent by clipping or compressing vibrant tones that cannot be reproduced on screens. Evolutions in CSS standards have expanded RGB capabilities beyond to accommodate modern displays. The CSS Color Module Level 4, advancing since the and reaching Candidate Recommendation status in 2022 (with ongoing drafts as of 2025), introduces the color() function for wider gamuts, such as Display P3, allowing specifications like color(display-p3 1 0 0) for enhanced reds on compatible . For accessibility, (WCAG) rely on RGB-derived to compute , ensuring readable text by requiring at least 4.5:1 for normal text and 7:1 for large text. is calculated from component values using the formula L = 0.2126 \times R + 0.7152 \times G + 0.0722 \times B, where R, G, and B are linear RGB values, then applied in the contrast ratio \frac{L_1 + 0.05}{L_2 + 0.05} (with L_1 as the brighter ).

Color Management Practices

Color management practices in the RGB color model involve standardized systems to ensure color consistency across diverse devices and digital workflows, addressing variations in how RGB values are interpreted and rendered. These practices rely on device-independent color spaces like CIE XYZ as intermediaries for transformations, enabling accurate reproduction of RGB-defined colors from creation to output. By characterizing devices through profiles and applying transforms, discrepancies arising from hardware differences, such as phosphors or printer inks, are minimized. Central to these practices are International Color Consortium (ICC) profiles, which embed characterizations of RGB color spaces to link devices in a workflow. An ICC profile describes the transformation between a device's RGB values and a device-independent space, such as CIE XYZ, allowing software to adjust colors for accurate viewing or printing. Version 4 (v4) profiles, introduced in December 2001, enhanced support for perceptual rendering intents and parametric curves, improving handling of wide-gamut RGB spaces beyond sRGB. These profiles are embedded in image files or documents to preserve color intent, facilitating seamless transfer between applications and platforms. In professional workflows, RGB working spaces like Adobe RGB serve as standardized environments for editing and proofing in applications such as . Proofing simulates output conditions by applying the target device's profile to the RGB data, allowing creators to preview and adjust colors before final export. Soft-proofing, a non-destructive on-screen, uses the to render a gamut warning for out-of-gamut colors and applies rendering intents like perceptual or relative colorimetric to approximate the final appearance. Key standards underpin these practices, including Apple's ColorSync system, released in 1993 as an operating system-level for on Macintosh platforms. ColorSync integrates profiles to automate RGB-to-device transforms, supporting consistent color from input scanners to output printers. For open-source environments, the Little CMS library provides a lightweight engine for ICC-based RGB transformations, emphasizing accuracy and performance in cross-platform applications like image editors and browsers. Despite these advancements, challenges persist, such as illuminant metamerism, where RGB-defined colors match under one light source (e.g., D65 daylight) but appear different under another due to spectral mismatches in device responses. This phenomenon complicates verification in mixed lighting environments and requires spectral profiling beyond standard RGB characterizations. Additionally, embedding ICC profiles in formats like JPEG and PDF ensures portability but demands careful handling to avoid corruption or incompatibility; JPEG supports profile markers in the APP2 segment, while PDF allows output intents at the document level for RGB content. A fundamental transform in RGB color management converts linear RGB values to CIE XYZ using a 3x3 M, as defined for specific spaces like : \begin{pmatrix} X \\ Y \\ Z \end{pmatrix} = \begin{pmatrix} 0.4124564 & 0.3575761 & 0.1804375 \\ 0.2126729 & 0.7151522 & 0.0721750 \\ 0.0193339 & 0.1191920 & 0.9503041 \end{pmatrix} \begin{pmatrix} R_{\text{lin}} \\ G_{\text{lin}} \\ B_{\text{lin}} \end{pmatrix} This , derived from sRGB primaries and D65 , linearizes encoded RGB (via ) before applying the transform to achieve device-independent representation. The inverse enables XYZ-to-RGB conversion, essential for profile-based rendering. The RGB color model serves as the foundation for several luminance-chrominance color spaces, which separate the brightness () component from the (color) components to facilitate efficient processing in video and imaging applications. One prominent example is the color space, derived from RGB for video compression, as defined in the BT.601 standard originally established in 1982. The conversion equations for this space, applied to gamma-corrected RGB values (R', G', B') in the range [0, 1], are: Y' = 0.299 R' + 0.587 G' + 0.114 B' C_b' = -0.169 R' - 0.331 G' + 0.500 B' + 0.5 C_r' = 0.500 R' - 0.419 G' - 0.081 B' + 0.5 These coefficients weight the RGB contributions to luminance based on human visual sensitivity, with green having the highest impact, while the chrominance components Cb' and Cr' represent blue-luminance and red-luminance differences, respectively. YUV, an earlier analog predecessor, follows a similar structure but uses scaled versions of these differences for broadcast television signals. The RGB model connects to device-independent through linear transformations to the CIE 1931 color space, which provides an absolute, illuminant-independent representation of colors based on human vision. For common RGB variants like , the transformation from linear RGB values (after de-gamma correction) to uses a 3x3 derived from the primaries and : \begin{pmatrix} X \\ Y \\ Z \end{pmatrix} = \begin{pmatrix} 0.4124564 & 0.3575761 & 0.1804375 \\ 0.2126729 & 0.7151522 & 0.0721750 \\ 0.0193339 & 0.1191920 & 0.9503041 \end{pmatrix} \begin{pmatrix} R \\ G \\ B \end{pmatrix} This matrix maps RGB primaries to the tristimulus values, often requiring imaginary (negative) coordinates for some primaries to achieve full coverage, ensuring RGB can represent the entire visible when transformed. 's illuminant independence contrasts with RGB's device-specific nature, making it ideal for cross-device color matching. Perceptual uniform spaces like CIELAB further derive from via nonlinear transformations, enabling metric-based color differences (ΔE) independent of viewing conditions. Conversions from RGB to subtractive spaces like CMYK for typically route through an XYZ intermediate to maintain colorimetric accuracy, as direct mappings would amplify device dependencies. In ICC color management frameworks, RGB profiles transform to the profile connection space (PCS) in , which then maps to CMYK profiles tailored to printer inks and . This intermediate step preserves perceptual intent, such as relative colorimetric rendering. These links enable key applications, particularly in video compression where YUV/YCbCr reduces bandwidth by subsampling chrominance signals, as in the 4:2:2 format that halves color data per line compared to full RGB () while retaining near-full detail for human . This subsampling cuts transmission by approximately one-third in standards. Modern codecs like HEVC (H.265) build on these RGB-derived formats, supporting YCbCr with optional RGB coding for higher fidelity in workflows. A core limitation of RGB in these conversions is its inherent device-dependency, tied to specific primaries and white points, which can lead to mismatches or metamerism under different illuminants, unlike XYZ's standardized, observer-based foundation that ensures consistency across environments.

References

  1. [1]
    A Standard Default Color Space for the Internet - sRGB - W3C
    Definition: A color space is a model for representing color numerically in terms of three or more coordinates. e.g. The RGB color space represents colors in ...
  2. [2]
    Additive & Subtractive Color Models - DINFOS Pavilion
    Oct 1, 2024 · The additive color model describes how light produces color. The additive colors are red, green and blue, or RGB.
  3. [3]
    [PDF] Chapter 7 ADDITIVE COLOR MIXING - cs.wisc.edu
    Additive color mixing works for mixing lights. The rules of additive color mixing apply to color TV, computer monitors, projectors, stage and disco lights.
  4. [4]
    [PDF] The RYB, CMYK, RGB, HSV and Hexadecimal Color Models
    The RGB model is also used for recording colors in digital cameras, including still image and video cameras. The RGB model is an additive system.
  5. [5]
    3.1 Color – Geographic Information Technologies
    The RGB model combines light beams of the primary hues of red, green, and blue to yield additive secondary hues of magenta, cyan, and yellow. Although there is ...
  6. [6]
    [PDF] Parameter values for the HDTV standards for production and ... - ITU
    ITU-R BT.709-6. 1. RECOMMENDATION ITU-R BT.709-6. Parameter values for the HDTV1 standards for production and international programme exchange. (Question ITU-R ...
  7. [7]
    BT.709 - INTERNATIONAL COLOR CONSORTIUM
    BT.709. Reference Name: ITU-R BT.709. Definition: ITU-R BT.709-6: Parameter values for the HDTV standards for production and international programme exchange.
  8. [8]
    [PDF] Lecture 17: Color
    Grassmann was intrigued by the idea that two colors could mix to produce a different color, such as red+blue=purple. He showed that when you have an equation ...Missing: primary | Show results with:primary
  9. [9]
    Color mixture - Scholarpedia
    Jan 12, 2011 · The rules, or laws, of additive color mixture were defined in 1853 by mathematician Hermann Günther Grassmann (mathematically formalized by ...
  10. [10]
    Additive versus Subtractive Color Models | X-Rite Blog
    Aug 10, 2018 · To effectively manage color, you need to know the difference between additive and subtractive color models.Rgb--Additive Primaries · Is Rgb Additive Or... · Cmy And Cmyk--Subtractive...
  11. [11]
    Additive vs. Subtractive Color Models - HunterLab
    Dec 27, 2023 · The difference between additive vs. subtractive color is in the ways color is created.
  12. [12]
    II. The Bakerian Lecture. On the theory of light and colours - Journals
    The Bakerian Lecture. On the theory of light and colours. Thomas Young.Missing: primary | Show results with:primary<|control11|><|separator|>
  13. [13]
    Thomas Young and the Trichromatic Theory of Colour Vision
    Thomas Young published no original experiments on colour mixing or on the perception of colour. Although he did enter into the spirit of the Royal Institution ...
  14. [14]
    The evolution of concepts of color vision - PMC - PubMed Central
    This argument was finally resolved by Thomas Young, who noted that trichromatic color matching was consistent with a continuous spectrum if there were just ...
  15. [15]
    Hermann von Helmholtz • LITFL • Medical Eponym Library
    Aug 10, 2025 · In the mid-1850s, Helmholtz refined this into a testable physiological theory. The retina contains three classes of photoreceptors (cones) ...Biographical Timeline · Medical Eponyms · Helmholtz And The...
  16. [16]
    Zur Theorie der Farbenmischung - Grassmann - Wiley Online Library
    This paper, 'Zur Theorie der Farbenmischung', is by Professor H. Grassmann, published in 1853.
  17. [17]
    How did a colour scientist come up with linear algebra? - Medium
    Apr 13, 2024 · In this seminal work, Grassmann developed a mathematical framework for colour mixing and introduced the concept of the colour triangle based on the idea of ...
  18. [18]
    XVIII.—Experiments on Colour, as perceived by the Eye, with ...
    Jan 17, 2013 · A method by which every variety of visible colour may be exhibited to the eye in such a form as to admit of accurate comparison.<|control11|><|separator|>
  19. [19]
    IV. On the theory of compound colours, and the relations ... - Journals
    On the theory of compound colours, and the relations of the colours of the spectrum. James Clerk Maxwell.
  20. [20]
    A short history of colour photography
    Jul 7, 2020 · Nearly 200 years later, in 1861, a young Scottish physicist, James Clerk Maxwell, conducted an experiment to show that all colours can be made ...
  21. [21]
    Deriving the cone fundamentals: a subspace intersection method
    König & Dieterici developed a method for measuring the colour confusions of dichromats to estimate the cone fundamentals [18]. The data and theory have been ...Missing: 1880s | Show results with:1880s
  22. [22]
    Color Spaces – color models, LMS, CIE XYZ, RGB ... - RP Photonics
    The primary colors of CIE XYZ color space (e.g. with X = 1, Y = Z = 0) are imaginary colors, i.e., they cannot be physically realized.
  23. [23]
    [PDF] Color measurement – the CIE color space - Datacolor
    The scientific foundation for color measurement is based on the existence of 3 different groups of signals (primary color stimuli blue, green and red) that are ...
  24. [24]
    History of the autochrome: The dawn of colour photography
    Jun 5, 2009 · Light, passing through the coloured starch grains, combines to recreate a full colour image of the original subject. How were autochromes made?Missing: RGB | Show results with:RGB
  25. [25]
    Photography - Kodak
    Introduced in 1935, KODACHROME Film became the first commercially successful amateur color film. Initially offered in 16 mm format for motion pictures, formats ...
  26. [26]
    The long-term development of three-color Kodachrome. An odyssey ...
    Mannes and Godowsky's collaboration led to the introduction of three-color Kodachrome in 1935. The researchers transitioned from additive to subtractive color ...
  27. [27]
    Color imaging array - US3971065A - Google Patents
    A sensing array for color imaging includes individual luminance- and chrominance-sensitive elements that are so intermixed that each type of element
  28. [28]
    Old Color Movie Processes - American WideScreen Museum
    The three color system was deemed virtually unworkable, and the registration in this photo would seem to bear that out. ... Natural Color Films - Additive Color.
  29. [29]
    Baird Mechanical Color System - Early Television Museum
    John Logie Baird demonstrated a color TV system, the first one that actually worked, in 1928. Baird used a Nipkow disk, with the disk divided into three ...Missing: demonstration RGB
  30. [30]
    [PDF] Analog TV Systems: Migration from Monochrome to Color
    – 1953 FCC approves RCA color TV system (compatible with B/W systems). (NTSC ... Why does NTSC color system uses the YIQ representation instead of the RGB?
  31. [31]
    The NTSC TV Set
    The shadow mask makes sure that the electron beams are aligned and strike the right phosphors. The beams of the electron gun must converge exactly at the holes ...
  32. [32]
    [PDF] Conventional Analog Television - An Introduction
    The color encoding principles for the PAL system are the same as those of the NTSC system -- with one minor difference. In the PAL system, the phase of the ...
  33. [33]
    Understanding Analog Video Signals
    Sep 18, 2002 · NTSC,4 PAL,5 and SECAM6 are the names of the broadcast video formats developed in the US, Germany, and France to encode color video and sound ...
  34. [34]
  35. [35]
    Famous Graphics Chips: EGA to VGA - IEEE Computer Society
    Nov 15, 2018 · IBM introduced the second-generation Enhanced Graphics Adapter (EGA) in 1984 that superseded and exceeded the capabilities of the CGA. The EGA ...
  36. [36]
    Famous Graphics Chips: IBM's VGA - IEEE Computer Society
    Mar 12, 2019 · The chip could support a planar mode: up to 16 colors (four, bit planes), and a packed-pixel mode: 256 colors (Mode 13h as it was commonly ...
  37. [37]
    S3 - DOS Days
    With a maximum of 4MB of VRAM, the fully VESA-compatible 86C928 chip offers true-color support for resolutions up to 1,024-by-768 and a maximum resolution of ...
  38. [38]
    The Apple Macintosh was first released 40 years ago - BBC
    Jan 23, 2024 · On 24 January 1984, the Apple Macintosh 128K personal ... The first Macintosh with a colour screen, the Macintosh II, only arrived in 1987.
  39. [39]
    Graphics pipeline - Win32 apps - Microsoft Learn
    Feb 23, 2022 · The Direct3D 11 programmable pipeline is designed for generating graphics for realtime gaming applications.Missing: history | Show results with:history
  40. [40]
    OpenGL - The Khronos Group
    From the very beginnings of OpenGL in 1992 to the release of ANARI 1.0, the timeline below charts the history of Khronos achievements throughout the years.
  41. [41]
    How the World's First GPU Leveled Up Gaming and Ignited the AI Era
    Oct 11, 2024 · In 2018, NVIDIA debuted GeForce RTX (20 Series) with RT Cores and Tensor Cores, designed specifically for real-time ray tracing and AI workloads ...Missing: RGB | Show results with:RGB
  42. [42]
    Chapter 4 • Color Management – Digital Photography for Graphic ...
    The advent of desktop publishing in the 1990s enabled users to connect various “plug-and-play” color peripherals, including cameras, monitors, and printers.
  43. [43]
    [PDF] Rethinking Color Cameras - People | MIT CSAIL
    The most common alternating mosaic is the. Bayer pattern [1] shown in the top of Fig. 1. Full color images are then reconstructed from these samples through an ...
  44. [44]
    [PDF] HIGH-QUALITY LINEAR INTERPOLATION FOR DEMOSAICING OF ...
    This paper introduces a new interpolation technique for demo- saicing of color images produced by single-CCD digital cameras.
  45. [45]
    Flatbed Scanner - an overview | ScienceDirect Topics
    Flatbed scanners have a large glass plate on which the photograph is placed, face down, to be scanned. A linear CCD on a moveable bar, with a single row of ...
  46. [46]
    None
    ### Summary of Drum and Flatbed Scanners for RGB Color Capture
  47. [47]
    [PDF] AN1904 White Balance and Color Correction - NXP Semiconductors
    To accomplish this many cameras have white balance algorithms used to ... Illuminant. Color Channel Gain Settings (hex). Red. Green (of Red). Green (of ...
  48. [48]
    [PDF] Understanding Digital Raw Capture - Adobe
    Raw files also include some addi- tional metadata that raw converters need in order to process the raw capture into an RGB image. In addition to the grayscale ...
  49. [49]
    The Geometry of Noise in Color and Spectral Image Sensors - PMC
    Digital images are always affected by noise and the reduction of its impact is an active field of research. Noise due to random photon fall onto the sensor ...
  50. [50]
    Compressive recovery of smartphone RGB spectral sensitivity ... - NIH
    Apr 1, 2021 · We report a compressive sensing framework in the frequency domain for accurately predicting RGB spectral response functions only with several primary colors.
  51. [51]
    RGB Color Codes Chart 🎨 - RapidTables.com
    The red, green and blue use 8 bits each, which have integer values from 0 to 255. This makes 256*256*256=16777216 possible colors.HTML Color Picker · Color Wheel Tool · HTML color codes · Gold color
  52. [52]
    [PDF] EBU R 103 - Video Signal Tolerance in Digital Television Systems
    Currently nearly all TV studios are equipped with an SDI-based environment using 10-bit video signals with a Y′C′BC′R 4:2:2 sub-sampling representing the ...
  53. [53]
    [PDF] Adobe RGB (1998) Color Image Encoding
    May 5, 2005 · The Adobe® RGB (1998) color image encoding is defined by Adobe Systems to meet the demands for an RGB working space suited for print production.
  54. [54]
    [PDF] Using the ProPhoto RGB Profile in Adobe Photoshop v5.0 - Scarse
    Jun 26, 2000 · The objective of the ProPhoto RGB working space is to let you maintain and edit an image in a device-independent color space while displaying it ...
  55. [55]
    [PDF] How to interpret the sRGB color space (specified in IEC 61966-2-1 ...
    sRGB has primary chromaticity coordinates, a gamma of 2.2, a white point of 80 cd/m2, and a black point of 0.2 cd/m2. CIE 1931 XYZ values are scaled from 0.0 ...Missing: 1996 | Show results with:1996
  56. [56]
    Technical Introduction to OpenEXR
    Encoding images with one luminance and two chroma channels, rather than as RGB data, allows a simple but effective form of lossy data compression that is ...Missing: HDR | Show results with:HDR
  57. [57]
    7.3 Big Endian Versus Little Endian Byte Ordering
    Little Endian byte ordering is defined as follows: In a binary number consisting of multiple bytes (e.g., a 32-bit unsigned integer value, the Group Number, the ...
  58. [58]
    [PDF] RGB Color - UBC Computer Science
    • RGB color cube sits within CIE color space. • subset of perceivable colors. • scale, rotate, shear cube. 43. HSV Color Space. • more intuitive color space for ...
  59. [59]
    14. Color Systems and Artificial Illumination
    We are going to try to give you a geometric picture of color mixing in RGB and HSV by means of a tour of the RGB cube depicted in Figure The RGB color cube.
  60. [60]
    [PDF] Computer Graphics: Color II
    ▫ Example: RGB color space. □ Primaries: Red Green Blue. Blue. Cyan. RGB ... ▫ Linear interpolation between 2 RGB colours in RGB space: C. 0. = (r. 0. , g. 0.
  61. [61]
    [PDF] Lecture Notes 11 Introduction to Color Imaging • Color filter options ...
    Thus C can be expressed as a linear combination of three basis vectors. C = L ... ◦ Color correction: transforms the camera output to the color space.
  62. [62]
    [PDF] REPORT ITU-R BT.2407-0 - Colour gamut conversion from ...
    This report addresses colour gamut conversion from ITU-R BT.2020 to BT.709, aiming for high-quality mapping in UHDTV to HDTV scenarios.
  63. [63]
    [PDF] Color Characterization and Modeling of a Scanner - Purdue e-Pubs
    Jul 25, 2008 · The gamut volume is calculated by a tetrahedral tesselation method in the device color space which is converted to the CIE LAB space and the ...
  64. [64]
    The non-Riemannian nature of perceptual color space - PNAS
    Color spaces can be directly related to the three cones: red, green, and blue, such as CIE RGB. However, most color spaces are mathematical constructs or models ...
  65. [65]
    [PDF] Colorimetry - Yuhao Zhu - Computer Science : University of Rochester
    ProPhoto RGB has a wide gamut, but. ~13% of is imaginary. Two of the three primaries are not physically realizable. • Q: Can you build a display that shows all ...Missing: limitations | Show results with:limitations<|control11|><|separator|>
  66. [66]
    [PDF] A Color Gamut Description Algorithm for Liquid Crystal Displays in ...
    ... gamut boundaries (CIE-a. ∗b∗ boundary and CIE-C. ∗L∗ boundary) are calculated which are often used in gamut mapping process. When our algorithm is ...<|control11|><|separator|>
  67. [67]
    Cascading Style Sheets, level 1 - W3C
    CSS1 is a simple style sheet mechanism that allows authors and readers to attach style (e.g. fonts, colors and spacing) to HTML documents.
  68. [68]
    CSS Color Module Level 3 - W3C
    Jan 18, 2022 · All RGB colors are specified in the sRGB color space (see [SRGB]). User agents may vary in the fidelity with which they represent these colors, ...
  69. [69]
  70. [70]
    Color profile - SVG 1.1 (Second Edition)
    All SVG colors are specified in the sRGB color space [SRGB]. At a minimum, SVG user agents shall conform to the color behavior requirements specified in the ...Color profile descriptions · The 'color-profile' element · The CSS @color-profile rule
  71. [71]
  72. [72]
    CSS Color Module Level 4 - W3C
    Apr 24, 2025 · This module describes CSS properties which allow authors to specify the foreground color and opacity of the text content of an element.
  73. [73]
    Understanding Success Criterion 1.4.3: Contrast (Minimum) | WAI
    A contrast ratio of 3:1 is the minimum level recommended by [ISO-9241-3] and [ANSI-HFES-100-1988] for standard text and vision. The 4.5:1 ratio is used in this ...
  74. [74]
    ICC.1 v4 Specification - INTERNATIONAL COLOR CONSORTIUM
    The ICC.1 v4 specification, updated in May 2022, includes a CICP tag for HDR and metadata, and removes ambiguities from previous versions.
  75. [75]
    Proofing colors in Photoshop - Adobe Help Center
    May 24, 2023 · When soft proofing is on, a check mark appears next to the Proof Colors command, and the name of the proof preset or profile appears at the top ...
  76. [76]
    Little Color Management System
    Open source small-footprint color management engine, with special focus on accuracy and speed. It fully supports all kind of V2 and V4 profiles.Color Management on Qt with... · Little CMS 2.17 released · Color engine
  77. [77]
    [PDF] ICC Technical Note Embedding ICC profiles
    Images in PDF files can also include an embedded profile within the image file, using the. Page 5. method of embedding defined by the image file format ...
  78. [78]
    Where Perception Meets Reality: The Science of Measuring Color
    Oct 1, 2022 · A consequence of this occurrence is color metamerism. Illuminant metamerism is the change in the color of an object because of a change in the ...Missing: management | Show results with:management
  79. [79]
    How to convert between sRGB and CIEXYZ - Image Engineering
    Aug 9, 2017 · To convert sRGB to CIE-XYZ, linearize sRGB data, then apply a 3x3 matrix. To convert back, apply a transformation matrix and then gamma.
  80. [80]
    None
    ### Summary of RGB to YCbCr Conversion Equations from ITU-R BT.601
  81. [81]
    [PDF] INTERNATIONAL STANDARD IEC 61966-2-1
    IEC 61966-2-1 is about multimedia colour management, specifically the default RGB colour space sRGB, and its transformation to sYCC.
  82. [82]