h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

When walking down the street after dark, I've noticed that the glow from a TV screen though a window always appears blue. However, TVs are actually full colour or black & white.

Can anyone explain this phenomenom?

BTW "They're here!" is not a valid answer!

I'm not sure i can remember the answer in full but if this is wrong someone else is bound to correct it.(H2G2 is friendly like that.)

The light from a TV is produced from an electron beam striking a phospher coating (the effect is similar to a flouresent tube.)
The spectrum of light emmitted from a TV includes ultra violet (and even Xrays) hence it has more components towards the blue end of the spectrum and appears to give a blue cast when viewed against a background of conventional lighting.

Also TV programmes made outdoors or in a studio under imitation daylight also contain a higher proportion of blue just as daylight does.

Additionally, if, like much of the UK, your area happens to use orange (low pressure sodium?) street lighting, light from a TV will appear even more blue-ish than it does when compared to standard domestic lighting.

Alternatively, I certainly find that if I'm out walking in the hills by starlight or moonlight, there's a definite impression of blueness. Maybe the blue-sensitive cells in the eye are slightly better than the red or green ones in marginal lighting, or, when it's so dark that all 3 kinds of colour detector (cones) stop working, maybe the brain 'tints' the picture from the remaining black-and-white detection system (rods).

If I remember correctly, that's right -- it's the same reason that 16-bit color in computer monitors is normally stored in 5-6-5 format. 5 Bits go to "red" shades, 6 to "green" (the type the eye is most sensitive to), and 5 to "blue" (the second most sensitive). When seeing a dim enough light, the most likely colors to fire are green and blue, leading to the impression of a bluish glow. The same is true for night vision: true night vision seems to have a blue color for many people even though it's supposedly the black-and-white "rods" firing and not the "cones", due to there being just enough to touch off some blue and green receptors.

I wonder what happens at intermediate light levels - when colours are clearly visible, but the light intensity isn't sufficient to saturate the rods. Do signals from the rods get mixed in to the brain's calculation of colours, or are colours purely formed from the cone cells. What is the spectral sensitivity curve for rods actually like?

Also, I was wondering to what extent does the brain 'invent' colour information, even when there isn't _any_ obvious incoming data to work from. When playing with some hyperbright red LEDs which give roughly single-coloured light, I noticed that if I properly blacked out a room at night, and used light from a LEDs to illuminate a wallful of 12x8 colour photographs, very quickly after the initial sensation of 'everything's all just shades of red', I began to 'see' green where I knew there was grass, and blue where I knew there was sky. White areas appeared different from red, even when the intensity seemed the same (as far as I could tell). It didn't feel like I was consciously remembering what colours should be where.
Even if the red light stimulated both rods, and red-sensitive cones (perhaps with lesser stimulation of green cones), surely, any given area of photograph could only be reflecting somewhere between 0% and 100% of the red light falling on it. A dark red area should have emmitted exactly the same light as a paler green, but it didn't feel like that. My few black-and-white pictures didn't appear to be at all coloured.