Luma and colour differences can be computed from R’, G’, and B’ through a 3× 3 matrix multiplication.

Y’PBPR

Y’CBCR

Y’UV

Y’IQ

Colour difference coding (chroma)

In component video, the three components necessary to convey colour information are transmitted separately. Rather than conveying R’G’B’ directly, the relatively poor colour acuity of vision is exploited to reduce data capacity accorded to the colour information, while maintaining full luma detail. First, luma is formed according to Equation 12.1 (or for HD, Equation 12.2). Then, two colour difference signals based upon gammacorrected B’ minus luma and R’ minus luma, B’-Y’ and

R’-Y’, are formed by “matrixing.” Finally, subsampling (filtering) reduces detail in the colour difference (or chroma) components, as I will outline on page 127. Subsampling incurs no loss in sharpness at any reasonable viewing distance.

In component analog video, B’-Y’ and R’-Y’ were scaled to form colour difference signals denoted PB and PR, which were then analog lowpass filtered (horizontally) to about half the luma bandwidth. Y’, PB, and PR each have unity excursion (i.e., 0 to 1, ±0.5, and ±0.5). In computing, these components are commonly denoted U and V.

In component digital video (including M-JPEG, MPEG, and H.264), B’-Y’ and R’-Y’ are scaled to form CB and CR components, which can then be subsampled by digital filtering denoted 4:2:2 or 4:2:0; I will describe the subsampling in a moment.

In composite NTSC or PAL video, B’-Y’ and R’-Y’ were scaled to form U and V components. U and V were then lowpass filtered and combined into a modulated chroma component, C. Luma was then summed with modulated chroma to produce the composite NTSC or PAL signal. Scaling of U and V was arranged so that the excursion of the composite signal (Y’+C) was constrained to the range -1⁄3 to +4⁄3 of the unity excursion of luma. The historical U and V components are now obsolete, and today UV denotes Y’PBPR (see above).

Composite NTSC video was standardized in 1953 based upon I and Q components that were essentially U and V components rotated 33° and axis-exchanged. Excess detail was supposed to be removed from the Q component so as to improve colour quality. The scheme never achieved significant deployment; I and Q components have been completely obsolete for many decades.

4:2:0

ITU-T Rec. H.261 is a 1990s-vintage videoconferencing standard.

The use of 4 as the numerical basis for subsampling notation is a historical reference to sampling at roughly four times the NTSC colour subcarrier frequency. The 4fSC rate was already in use for composite digital video.

Figure 12.3 Chroma subsampling notation indicates, in the first digit, the luma horizontal sampling reference. The second digit specifies the horizontal subsampling of CB and CR with respect to luma. The third digit originally specified the horizontal subsampling of CR. The notation developed without anticipating vertical subsampling; a third digit of zero now denotes 2:1 vertical subsampling of both CB and CR.

This scheme is used in JPEG/JFIF, MPEG-2, and H.264. CB and CR are each subsampled by a factor of 2 horizontally and a factor of 2 vertically. The 12 bytes of R’G’B’ are reduced to 6. CB and CR are effectively centered vertically halfway between image rows. There are three variants of 4:2:0, having different vertical and horizontal siting. In MPEG-2, CB and CR are cosited horizontally. In JPEG/JFIF, H.261, and MPEG-1, CB and CR are sited interstitially, halfway between alternate luma samples. In 4:2:0 DV, CB and CR alternate line by line. Figure 12.2 overleaf summarizes the various schemes.

Subsampling effects 1.5:1 or 2:1 lossy compression. However, in studio terminology, subsampled video is referred to as uncompressed: The word compression is reserved for techniques such as JPEG, M-JPEG, MPEG, or H.264 that use transform coding (DCT or wavelets).

Chroma subsampling notation

At the outset of digital video, subsampling notation was logical; unfortunately, technology outgrew the notation. In Figure 12.3 below, I strive to clarify today’s nomenclature. Despite appearances, the notation doesn’t specify a ratio! The first digit originally specified luma sample rate relative to 33⁄8 MHz; the leading digit is now relative to the sample rate in use. The initial digit is typically 4, since all common chroma ratios are small powers of two – 4, 2, or 1. (3:1:1 subsampling was commercialized in an HD production system – Sony’s HDCAM – and 3 appeared as the leading digit. HDCAM has been superseded by HDCAM SR, which uses 4:2:2 or 4:4:4 subsampling.) Some people use 4:0:0 to denote greyscale (“monochrome”).

Luma horizontal sampling reference (originally, luma fS as multiple of 3 3⁄8 MHz)

CB and CR horizontal factor (relative to first digit)

Same as second digit;

or zero, indicating CB and CR are subsampled 2:1 vertically

4:2:2:4If present, same as luma digit; indicates

alpha (key) component

Figure 12.2 Subsampling schemes are summarized here. C indicates a [CB, CR] sample pair when located at the same site; otherwise (as in the DV schemes) individual CB and CR notations indicate the centers of the respective chroma samples. Y’ indicates the center of a luma sample. The schemes in the left column are progressive. The schemes in the right column are interlaced; there, solid letters indicate top field samples and shaded letters indicate bottom field samples.