Flicker is a catch-all phrase that we use for changes in light output over time. And depending on the frequency of the change, we call it different things.
If the light is changing in output up to 80 Hertz, which means it’s going on and off—or from a high level to a low level—80 times a second, that is called “visible flicker.” And then when we see the light oscillating in output from about 80 Hertz to 2000 Hertz, and it makes moving objects appear to be bright or dim as they move, that is called the “stroboscopic effect.” There’s a third effect called the “phantom array effect” or called “ghosting effect” that also occurs from 80 Hertz up to several thousand Hertz. The “phantom array effect” is due to your eye movement and it turns a flickering light source or the object it’s lighting into a series of dots or lines or repeated images across your visual field, rather like rabbit tracks in the snow. You may have noticed this from some automobile tail lights when driving at night.
We have two technical terms to keep in mind. One is “temporal light modulation,” and that’s basically the stimulus, the waveform that illustrates the change in light output over time. And then we have a term called “temporal light artifacts,” and that refers to the response to that stimulus. So are you responding by seeing the flicker, or being distracted by it, or developing a headache, or having a seizure — those kinds of things. That is the temporal light artifact.
Flicker has been a problem for decades with all electric light sources. Even incandescent lights can exhibit a small amount of flicker, but usually it’s such a low change in the amplitude of the light that it’s not noticeable. But fluorescent lighting for example, especially when it was operated with magnetic ballasts, had a pronounced light output variation, about 120 Hertz of flicker. That was known to cause some degradation in task performance, and people who worked in offices with magnetically ballasted fluorescent had higher rates of headaches and other problems like malaise. The same thing was true with high-intensity discharge sources, so either metal halide or high-pressure sodium all had fairly large output changes, about 120 times per second.
When fluorescent and metal halide lamps and even some high-pressure sodium lamps were driven by high-frequency electronic ballasts, that got rid of the flicker. Or at least it minimized it to a very small level, and the frequency was increased to a point where you wouldn’t notice it. Your eye and brain system can’t pick up changes that fast.
Then came LEDs. Now, the complication with LEDs is that unlike incandescent filaments, or fluorescent tubes that have a glowing phosphor on the inside, or arc tubes from metal halide that glow and continue to glow even when there’s a slight change in the voltage delivered to them, LEDs don’t have any persistence. So, the incandescent lamp, because it stays hot even though there’s a point in time when it’s not getting any voltage because of the AC waveform, it’s still delivering light because it’s still hot. Same thing with the fluorescent lamp; it still has some persistence in it due to the phosphor.
But LEDs don’t have phosphors that operate like that. So LEDs are basically on or off. There is no persistence to keep them on. But LEDs usually have drivers connected with them. And these drivers, these electronic devices, are designed in different ways, and some electronic drivers are able to maintain the output of the LED over time through various techniques. And if that’s done, then the LED doesn’t exhibit flicker, or exhibits far less flicker. But some drivers don’t support the output of the LED over time. So LEDs, if they’re not paired with a good driver, or if they’re not paired with a driver that is designed to respond to the output of a dimmer for example, those LEDs can flicker, and they can flicker as much as 100 percent, from on to full off.
Flicker—we know how to measure it. We know that sometimes it can be causing medical problems. But what we don’t know is what the thresholds are for those different kinds of reactions. We need research to tell us what are these thresholds? What’s acceptable for what kind of population? If you’re designing the lighting for schools, and you know that perhaps one or two percent of the population of the school may be autistic children, we really don’t know what ranges of waveforms are going to be problematic for them, and which are not likely to have any effect at all. So although we’re moving towards having some set standards for different applications, we really don’t know the medical implications, and that’s where we need the most research right now.