I am generally wary of using AI tools for old-TV stuff as the risk of flooding Youtube with even more shitty mocks is definitely real but in the interests of science I tried telling Claude to recreate Test Card F from the technical specification published in EID & George Hersee's documentation rather than just finding an existing image and cleaning it up. It didn't do a bad job. as you can see below, but it rather pedantically found a couple of discrepancies from the original documentation and what became more familiar on air. The original version did include a full set of colour bars (white through to black) at the very top of the image above the blue castellations and which were presumably intended to be more or less hidden in the VBI but available to check with a vectorscope. So what we ended up with here eventually is a rather half-assed compromise where the castellations are thinner and the rest of the space is taken up by the bars (which I had changed to omit the white bar to match the on-air look that came later). The point of the arrow marks the top line as the top of the bars, when in the original it was at the top of the castellations with the bars above. I originally left the image out but later supplied it with a copy and told it to match the image parameters with those of the testcard as a whole, zoom and crop to fit and - most importantly! - to place the "X" at the exact centre of the image. Being basically a simulation of a pedantic nerd it got this right. Anyway, here it is. Don't reuse it anywhere else as I don't want yet another crappy version of TC-F muddying the historical record. Still, it's quite nice to be able to create the testcard from a Python script, right?