Atomic official guide to AV cables, cabling and tech

Back in the days of the cathode ray tube, analogue made sense. Take a source video, break it up into red, green and blue, and modulate a pulse for each signal. The greater the voltage of any one part of the signal, the brighter that particular phosphorous dot would shine. Include an inverted-polarity section in one of the signals to tell the display that the rest of the signal was meant for a new line (or, with slightly more complexity, a new frame) and you have moving picture right there on your TV set!

Sounds easy, yes?

When dealing with the limited bandwidths that were available when colour displays were first introduced, however, there simply wasn’t room for separate signals for red, green and blue. What we ended up with instead was ‘chroma’ and ‘luma’ (or chrominance and luminance for all you syllablephiles). An entire article could be written on these two little words alone (and, indeed, many have been), but to put it as simply as possible: luma is brightness, chroma is colour.

Luma is a signal value derived from a combination of the red, green and blue values, which when combined give us a black-and-white picture. The information is there and, crucially, is made up from red, green and blue values, weighted in such a manner that this purely black-and-white picture would appear ‘natural’ to a human eye more sensitive to blue than to red, and red than green. It is commonly abbreviated, for reasons arcane and mystic, as Y’ (a raw, unweighted combination of these signals would be Y, more of which later).

How the eyes respond to colour. Ponder this and you'll be able and extrapolate why black and white images are normalised differently when you compress colour information into less bandwidth.

Chroma is a signal comprised of two further signals, U and V. Take your original blue signal value and subtract the Y’ value, and you have your U. Do the same with Red and you have your V. Since the Red, Green and Blue values are encoded in the Luma, and the Red and Blue values are encoded in the Chroma, a bit of the kind of subtraction magic that analogue circuits can perform so well will give you the green. In practice, of course, it’s quite a bit more complicated than that, but without wanting to fill this article full of numbers and algebra, the upshot of it is that in theory the initial R, G and B values can be extracted from the Y’, U and V signals.

Composite vs S-video vs Component
Most everyone reading this article will have watched a Y’UV signal at some point. Composite is well-known to everyone, that single little yellow plug that tends to give appalling picture quality. The reason for this is that a composite cable squeezes the Y’ luma and UV chroma into the one cable. As the chroma signal interferes with the luma, the ever-common ‘dot crawl’ creeps in. S-video attempted to alleviate this problem of interference by separating out the luma and chroma signals into different wires (hence the four pins on the end of the one cable – a signal and a ground each for luma and chroma).

Both composite and S-video suffered from a terminal bandwidth problem, however. With most of the available bandwidth given up to the more-important luma data (which could still be used to generate a black and white picture), the colour resolution suffered dramatically. Component was an attempt to rectify this by further splitting the signal. To start with, since black and white TV is well and truly dead, component video was able to get rid of that pesky apostrophe, giving equal weighting to the red, green and blue potions of the luma, which was now represented by a Y. It then split the UV signals into separate Y-minus-blue and Y-minus-red cables, resulting in the YPbPr and YCbCr colour spaces that provide far greater resolution and colour accuracy.

FRAPS footage of HL2 powered CS, as viewed with Adobe Premiere Pro 2's YCbCr scope

...and again, this time as viewed with an RGB scope

PbPr and CbCr scale their signals differently to UV to achieve a ‘natural’ looking picture in the absence of a weighted luma value, but use the same principle – luminance subtracted from blue and red signal values. The difference between PbPr and CbCr is almost entirely irrelevant, with the CbCr originating from component’s original use in professional equipment where it needed to deal with digital signals, before filtering down to the technically analogue-only PbPr. These days the terms are often used interchangeably and unless you have very old component equipment can be safely ignored.