--- title: "RGB vs YUV: Why Video Separates Luminance and Colour" url: https://postproduction.studio/en/image/rgb-vs-yuv language: en description: "Understand why video separates luminance from colour using Y'CbCr: luma and chroma separation, component video, colour differences, and their role in compression and grading." updated: 2026-10-25 --- # RGB vs YUV: Why Video Separates Luminance and Colour Level:Intermediate Section: Image 7 min ## TL;DR · Key takeaways * •Sensors and screens natively work in RGB (Red, Green, Blue), but transmitting video in pure RGB requires massive bandwidth that is incredibly inefficient. * •Human vision is heavily skewed towards detecting brightness (luminance) details over colour (chrominance) details, thanks to the rods and cones in our retinas. * •To exploit this, video signals are converted into Y'CbCr (often loosely called YUV). "Y'" is the luma (brightness), "Cb" is the blue difference, and "Cr" is the red difference. * •By separating the brightness from the colour, engineers can aggressively compress the colour channels (chroma subsampling) without the human eye noticing the degradation. * •Understanding the difference between Luma (Y') and true Luminance (Y) is critical for colorists, as mathematical approximations during RGB-to-Y'CbCr conversions can shift perceived brightness of saturated colours. Every digital camera sensor captures light using a Bayer pattern of Red, Green, and Blue photosites. Every computer monitor and television display emits light using Red, Green, and Blue pixels. Yet, between the moment the image is captured and the moment it is displayed, the video signal is almost never stored or transmitted as RGB. Instead, it is transformed into a radically different mathematical model: Y'CbCr (commonly referred to as YUV or Component Video). This transformation is not a legacy accident; it is the fundamental cornerstone of modern digital video compression. By understanding why and how video separates the brightness of an image from its colour, post-production professionals can unlock the mysteries behind chroma subsampling, properly configure their delivery exports, and avoid catastrophic colour shifts when moving between grading suites and broadcast deliverables. ## Table of Contents * [1. From sensor to screen](#section-1) * [2. Why the eye changes everything](#section-2) * [3. Building Y'CbCr](#section-3) * [4. Component video](#section-4) * [5. YUV / Y'CbCr / Y'PbPr vocabulary](#section-5) * [6. Luma is not luminance](#section-6) * [7. Why it matters in post-production](#section-7) ## 1. From sensor to screen At the hardware level, all light is processed in RGB. A digital cinema camera uses a sensor covered in microscopic red, green, and blue filters to measure the intensity of incoming light across the visible spectrum. When you look at an OLED or LCD monitor, you are looking at microscopic red, green, and blue sub-pixels lighting up in varying intensities to create the illusion of thousands of colours. If the capture device and the display device both operate in RGB, the most logical transmission method would be an uncompressed RGB signal. In fact, high-end computer graphics and VFX workflows (like EXR sequences) do exactly this. However, storing a 4K video file as uncompressed 10-bit RGB requires astronomical data rates—often exceeding gigabytes per second. To make broadcast, streaming, and practical editing possible, engineers had to find a way to compress the image without making it look terrible. ## 2. Why the eye changes everything The solution to the data problem was found not in computer science, but in human biology. The human retina is lined with two types of photoreceptor cells: rods and cones. Rods are highly sensitive to brightness and spatial detail, meaning they are incredibly good at seeing texture, sharp edges, and contrast (black and white). Cones are responsible for seeing colour, but they are far less numerous and less sensitive to fine spatial detail. If you take a high-resolution black-and-white photograph, your eye sees every sharp detail perfectly. If you take that same photograph, blur it slightly, but tint it with colour, your brain still perceives it as a sharp, high-quality image because the black-and-white contrast provides all the structural detail. Engineers realized that if they could separate the brightness of a video signal from its colour, they could compress the colour data aggressively while leaving the brightness data untouched, and the human eye would simply not notice the missing colour resolution. ## 3. Building Y'CbCr To achieve this separation, the RGB signal must be mathematically transformed into a new colour space called Y'CbCr. \*\*Y' (Luma):\*\* The Y' channel represents the brightness of the image. It is calculated by mixing specific percentages of the Red, Green, and Blue channels together to form a black-and-white image. Because the human eye is most sensitive to green light, the Y' channel is heavily weighted towards the green signal (often ~70% green, ~20% red, and ~10% blue). \*\*Cb (Chroma Blue):\*\* Once the brightness (Y') is extracted, the system needs a way to transmit the colour. Instead of sending raw blue data, it sends the difference between the blue signal and the luma signal (B - Y'). \*\*Cr (Chroma Red):\*\* Similarly, the system sends the difference between the red signal and the luma signal (R - Y'). You might notice there is no "Chroma Green" (Cg). This is a mathematical trick. Because the display device knows the total brightness (Y'), the amount of Blue (Cb), and the amount of Red (Cr), it can simply use basic algebra to deduce exactly how much Green must be present to make the math work out. By only transmitting three channels (Y', Cb, Cr) instead of four, bandwidth is saved immediately. ## 4. Component video This Y'CbCr signal is what engineers refer to as "Component Video". By successfully divorcing the structural detail (Y') from the colour tint (Cb and Cr), the door is opened to Chroma Subsampling (4:2:2, 4:2:0). The encoder can now safely discard 50% or 75% of the Cb and Cr pixels, knowing that the high-resolution Y' channel will preserve the sharpness of the image. When a streaming platform like Netflix delivers a 4K movie to your TV, it is transmitting a heavily compressed Y'CbCr 4:2:0 signal. Inside your television, a processor takes that Y'CbCr signal, performs a matrix multiplication to reverse the math, and converts it back into the RGB signal that the physical screen requires to light up its pixels. ## 5. YUV / Y'CbCr / Y'PbPr vocabulary The terminology surrounding this concept is notoriously confusing and often misused in the industry. \*\*YUV:\*\* Historically, YUV was the specific colour encoding system used for analog PAL television. Today, software developers and video engineers use "YUV" as a casual, generic umbrella term to describe any colour space that separates luma from chroma. When Premiere Pro or DaVinci Resolve refers to "YUV processing," they are technically referring to Y'CbCr. \*\*Y'PbPr:\*\* This refers to the analog version of component video. If you remember the red, green, and blue RCA cables used to plug a DVD player into an older HDTV, that was an analog Y'PbPr signal. \*\*Y'CbCr:\*\* This is the strictly correct term for the digital representation of component video. Whether you are dealing with an MP4 file, an Apple ProRes MOV, or a digital broadcast signal, you are dealing with Y'CbCr. ## 6. Luma is not luminance A critical distinction for high-end colourists and compression engineers is the difference between Luma (Y') and true Luminance (Y). True Luminance (Y) is a linear measurement of light intensity, calculated directly from linear RGB values. It is a physically accurate representation of brightness. Luma (Y'), denoted by the prime symbol ('), is calculated after the RGB signal has already had a non-linear gamma curve (like Rec.709) applied to it. Because the gamma curve warps the mathematical relationships between the colours, calculating the brightness of a gamma-encoded image results in a value that is not physically accurate. Why does this matter? When highly saturated colours (like a bright red neon sign) are converted from RGB into Y'CbCr using Luma equations, the mathematical approximations can cause the brightness of the red sign to shift slightly compared to its original RGB value. High-end colour management systems like ACES process in linear RGB to avoid these Y'CbCr rounding errors, only converting to Y'CbCr at the absolute final stage of delivery. ## 7. Why it matters in post-production Understanding the RGB to Y'CbCr conversion is essential because almost all standard video formats (H.264, HEVC, ProRes, DNxHR) operate in Y'CbCr. When you bring an RGB graphic (like a lower-third title generated in After Effects) into a Y'CbCr editing timeline, the software must perform a matrix conversion. If the software uses the wrong matrix (for example, applying a Rec.601 matrix to Rec.709 footage), the math will be wrong. This is the root cause of the infamous "gamma shift" or "washed out colours" bug that editors experience when exporting files for the web. Furthermore, if you are attempting to pull a green-screen key, your keying software is looking for colour differences. If the footage was shot in a heavily compressed Y'CbCr 4:2:0 format, the chroma (Cb/Cr) channels are physically blocky and low-resolution. The keyer will struggle to find a clean edge because the colour data has literally been averaged out over multiple pixels to save space. The separation of RGB into Y'CbCr is one of the most elegant and crucial inventions in the history of digital media. By exploiting the physiological weaknesses of the human eye, engineers created a system that allows for massive data reduction without destroying perceived image quality. While RGB remains the absolute truth for sensors, displays, and high-end VFX compositing, Y'CbCr is the practical reality of almost all video distribution. By understanding the math that bridges these two worlds, post-production professionals can ensure their colours remain accurate, their exports remain clean, and their workflows remain highly efficient. To fully grasp how this Y'CbCr signal is compressed, proceed to the article on 'Chroma subsampling and bit depth explained.' To understand the specific RGB primaries that feed into these equations, explore 'Color Spaces & Gamuts.' For a broader view of how these files are packaged, see 'Codecs and containers'. ## Related Articles [Advanced ### Chroma subsampling and bit depth explained Discover what happens to the Y'CbCr signal when chroma resolution is halved. Read article](https://postproduction.studio/en/formats/chroma-subsampling)[Fundamental ### Color Spaces & Gamuts: Rec.709, DCI-P3, Rec.2020 Learn about the primary RGB gamuts that the Y'CbCr signal must be derived from. Read article](https://postproduction.studio/en/image/color-spaces)[Advanced ### Video compression: principles and methods See how this color/luma separation enables all modern spatial compression. Read article](https://postproduction.studio/en/formats/video-compression) ## Key terms in this article * [Luminance](https://postproduction.studio/en/glossary/luminance) * [Rec.709](https://postproduction.studio/en/glossary/rec-709) * [Gamma](https://postproduction.studio/en/glossary/gamma) * [Resolution](https://postproduction.studio/en/glossary/resolution) * [ProRes](https://postproduction.studio/en/glossary/prores)