The world famous Test Card "F" was originally produced by scanning a 35mm transparency in a dedicated Rank Cintel television slide scanner, which was located in the Central Apparatus Room at the BBC's Television Centre in Shepherds Bush, London. The slide actually had two carefully registered films in a plastic slide-mount with thin glass windows protecting the actual films. One film was colour for the picture and edge castellations, and the other was monochrome for the main design. This was to prevent any possibility of colour casts or shading on the neutral grey background. However, this construction demanded absolutely precise registration of the two films. There are a few specs of dust on the above scan, but we dare not try to open up the slide casing to clean it lest the registration of the two films be lost - others have made that mistake in the past! The image above is from one of these slides, scanned at the same time as the original transparency of Carole Hersee that was used for the new Test Cards "J" and "W". If you look closely, you can see that the red and blue colour blocks on the left encroach slightly on the white border, as do the blue castellations on the top. The slide also has rather more picture than is used: the arrow-heads in the middle of each side have small mirror-images beyond them. This was to aid the alignment of the slide-scanner, making it easy to adjust the height, width, and positions, to ensure the arrow-heads exactly reached the edges of the picture.
In 1984 an electronic version was produced, giving an absolutely stable and accurate size, a plain unshaded grey background, and needing no adjustments - focus or otherwise! While essentially the same design, there were a few subtle differences - a comparison of the two can be seen below, with the slide on the left, and the electronic version on the right:
Perhaps the most obvious difference is the lack of a station identification name in the letterbox at the bottom. This was added by a separate caption-keyer downstream of the test card generator itself, and the style changed several times in the life of the test card with the changing style of the Networks' identification logos. Two versions were thus available for BBC-1 and BBC-2, but a plain, unidentified version was also distributed internally for monitor line-ups in the studios. The caption letterbox itself was also a clean rectangle, unlike on the slide version where stubs from the vertical lines protrude into it like the remains of prison bars cut by a hacksaw! Just above that, the ident letter F changed from a plain, sans-serif font on the slide, to a serif'd one on the electronic card.
The central picture appears slightly different on the electronic card. Rather than going back to the original transparency as was done for the new test cards J and W (presumably because slide-scanning technology wasn't good enough back in 1984), the producers of the electronic test card F grabbed a frame of the test card generated from the slide, and cut that into the background pattern. However, in order to draw their own clean white circle around the picture, they had to zoom in a little to lose that on the original slide, and this resulted in Carole's hair almost touching the circle, whereas there is a clear gap on the slide. In the comparison above, the picture also looks desaturated (pale) and sat-up (brightness lifted). While some of that difference could be due to the different way the slide was scanned for the above picture, it was none-the-less generally felt that the picture in the electronic test card F was inferior to the original.
For most of its life, the slide test card had some electronic colour-bars switched over the top blue castellations, with the unfortunate side-effect that the top arrow-head was obliterated. However, this was a small price to pay for having an accurate signal which could be used to objectively measure colour saturation. When creating the electronic test card, the colour-bars were retained, but the arrow-head was cut over the top of the central magenta bar. The colours of the coloured blocks around the outside of the electronic test card are much brighter than those on the slide: being generated electronically, it was possible to accurately make them the maximum level allowed, and they thus matched the "95%" colour-bars at the top.
The image above of the electronic test card has a fat black border around it. This is because it is shown to the same scale as the slide on the left, but being generated electronically it didn't need, and indeed couldn't have, the "overscan" of the original slide.
There are some dots in the top and bottom steps of the greyscale to assist in setting brightness and contrast on televisions. These dots are circular on the original slide (see the image on the left, below), but became rectangular on the electronic card (below, right). (These images are double size, and have had the contrast enhanced to ease viewing this normally subtle feature.)
The frequency gratings changed from "square" on the slide to "sinusoidal" on the electronic card. The image on the left below is part of a high-resolution scan of a Test Card "F" slide, where the gratings can clearly be seen to be comprised of alternate stripes of white and black. Once filtered to TV bandwidths (see centre picture), the edges are softened, but the low-frequency (top) grating can still be seen to give a flat-topped signal. Square-waves are not very accurate for measuring frequency response, as they have large amounts of higher frequency harmonics. The electronic test card thus generated pure sinusoidal gratings (right-hand picture below). Even though these images are three times normal size, it is admittedly still rather hard to see the difference between the centre and right images. In fact, on the second grating down the centre image looks softer, but that is just caused by a slightly dirty slide giving low contrast - as can be seen from the far left image. This sinusoidal low-frequency grating gave the impression of being softer, or out of focus, by comparison to the old slide, and took a little getting used to!
The high-resolution image of the frequency gratings on the slide (above, left) also clearly shows that the frequency gratings were centred within the box available, which was not the case on the electronic card, which appears to have been aligned so that the falling edge of the signals were coincident in the centre of the box (see right-hand image above). Comparison of the centre and right images also shows that the horizontal white lines were narrower than the slide, being only five lines wide compared to seven on the slide.
The electronic test card was created in 1984, back in the days when the BBC micro and Sinclair Spectrum were the home computers of choice! Unlike today's computers, processing power, memory, and disc space were serious constraints. The test card was created using four Z-80 based computers ("Zelda's" for anyone within the Corp. who remembers them), and unfortunately the pressures of the constraints meant that the finished test card had a few minor errors.
In the top left-hand corner, the keying of the grating over the background grid has gone
slightly wrong in a number of places, as indicated by the arrows on the enlarged picture to the
Similarly, where the black/white/black lines meet the top and bottom of the circle isn't
The left and right arrow-heads are drooping down a little! As shown on the right, their bases are not symmetrical about the centre line.
Also, the anti-aliasing is not correct where the sloping edges of the arrow-head meet the vertical line. The slopes are about two pixels horizontally for one pixel vertically, so the two pairs of pixels (arrowed) where the slopes meet the anti-aliased edge of the vertical line should be slightly lighter. While it may look OK on this enlarged view, at normal size there appears to be notches cutting in to the vertical line where the arrow-head meets it.
Further, the edges of these arrow heads seem to have only been anti-aliased horizontally, and
not vertically as well. As a result they appear to have a much harder edge than their
equivalents on the top and bottom edges of the test card.
There is generally some minor vertical asymmetry about the pattern: the black-bordered
section of the background grill extends 34 lines above, but only 32 lines below; the
coloured blocks at the edge are also misplaced by the same two lines, and the other side blocks
are variously one or two lines different above compared to below the centre. Even the
centre circle extends one line higher above than below the centre!
The top arrow-head has a small grey zit just to the left of the arrow's tip. This
is very difficult to detect on a PAL coded signal, and was only became apparent when viewed
The second step down on the greyscale actually has two levels, 1% apart. This is
practically invisible to the naked eye, but can be seen on an oscilloscope as two closely spaced
lines. The image on the right has been enlarged, and the contrast considerably expanded to
demonstrate the effect.
Finally, the chrominance on the centre picture is slightly displaced to the right with
respect to the luminance. This can be seen on the enlarged picture to the right, on the
yellow of the doll's buttons.
Views and opinions expressed on this page are entirely those of the author, and not of the BBC or BBC Technology Ltd.