Sorry for the dumb question, but how come sometimes when taking pictures of or recording lighting, pink lines appear in the photo/video that you can't see with your eyes?

Example:

If the lamp is on magnetic gear / on electronic with failing input filtering capacitor, it flickers with double the AC power frequency (100 or 120 Hz)

In the times when the current is going, the light is coming from the mercury arc (green / blue) and the phosphor (mix of pink/orange + green + blue components). In the times when the current is crossing zero, the arc is out and the light is from the afterglow of the phosphor alone

The phosphor by itself may be composed to have extra of the pink/orange component to balance the green and blue coming from the arc, and still more if it in a warm white lamp

Additionally it is possible that out of the 3 components in the phosphor mix, the pink/orange have the longest afterglow. This you can check - see what color the lamp does glow in the dark after it is switched off

When you shoot the pic the digital camera is scanning the image it sees. If at some point during the scan the arc strikes or goes out, in part of the pic the lamp and everything under its light will appear with the color of the arc + phosphor, and in part of the pic with the color of the phosphor afterglow

The pink background is as well a response of most sensors to an IR radiation. Normally there shouldbe a filter in the camera (very frequently directly on the sensor), but when the IR is more intense, it may scatter it (even only in one direction - depends on the exact structure of the filter layer).
Soifthefilter scatters theIR in just one direction, you get the pink haze smeared from the lamp.

Note, the fluorescents generate IR as the wasteradiation when electron in the phosphor return to theirbase levels (only some transitions belong to the visible, many others generate just an IR).

@Medved: I've seen this effect in any type of lamp with my camera, regardless to the amount of IR emitted. This is usually caused in cameras with CCD sensor. Didn't seen this happens in cameras with CMOS sensor.

The concerned IR isthe shorter wavelength range (close tothe visible), it is emitted by practically all light sources (include the white LED's - there it comes from the phosphor).
The CMOS uses different pixel sensor concept to distinguish the colors, the effect is highly related to the filter concept (used mainly by the CCD's; sensor sub-pixels have "band-pass" filters utilizing the interference to block the passage of the light. As the IR has about twice the wavelength compare to the blue and it is not that far from the red pass band, it excites both, hence the pink color)


CMOS could use the same filter concept as CCD, but frequently use a bit different concept for color component separation: Three junctions stacked underneath each other (N+, Pwell, Nwell, P-substrate), where the doping gradient is different in each junction (the sharpest in the N+/Pwell, smallest in the Nwell/P-substrate). Thelight then goes firstthrough the first junction and some part ofthe spectrum is trapped there to form the current. The remaining parts then go deeper, the Pwell/Nwell junction catching the green, so only the rest remaining spectrum passes to the last Nwell/Psubstrate junction. That means the colors are separated directly by the photon energy, so no wavelength multiplication/division could occure, so the IR just affect the closest part, so the red only and not the blue.

So no wonder, it uses to be not prone to the smeared pink stripes.