Same thing happens with those old alarm clocks with LED 7 segment displays.
I think it’s the difference between rods and cones on the retina. Rods are more sensitive to light but do not see colour. That’s what you use in the dark. I suspect they are also slower to react to light (probably because they need to gather enough over time to make an image). So since the lighter text is bright enough to be seen by the cones, the reaction time is quicker so it looks like it’s moving independently.
Anyway that’s my theory and I’m sticking to it. :P
That similar effect happens with some 7-segment displays because they actually flicker very fast (either didn’t bother to add an AC rectifier if it’s wall power, or it uses PWM to regulate the brightness). You can see it if you point a camera at them. The visual cortex makes it appear like it’s a continuous light but that illusion is destroyed as soon as the light moves. (Pretty sure this is the same effect which makes a moving image on CRT monitors appear smoother at low frame rate than LCD monitors.)
I don’t know for sure but I don’t think there’s any significant reaction time difference between rods and cones in different light levels. There is a difference in how long they take to adjust to different light conditions (IIRC rods take longer to adjust to darkness but can achieve much higher sensitivity in darkness).
I think the effect produced by the strobing due to DC rectification is markedly different. That produced a gapping when moved (since the LED is off during some parts of the movement and on for others). The effect when there’s a somewhat brighter light source around a darkened room is very different where the lighter source seems to move independently of the object it is attached to, with no stuttering/strobe effect present.
The independent movement effect is much more likely due to response time differences. I thought I’d take a look and see. This https://www.nature.com/articles/s41598-018-34073-8 article by nature does mention there is a response time difference. However the vast majority of the paper seems to be about the effects on cone response time due to lack of calcium and other cone specific testing.
This https://pmc.ncbi.nlm.nih.gov/articles/PMC2063471/ article seems to imply that there’s a 20ms lag for rod reaction time. Where they tested people’s reactions to both kinds of stimuli and measured the response times.
Hmm, interesting. Then I don’t think I’ve ever seen that myself, I thought you were talking about the strobe effect, that’s the only thing I associate with 7-segment displays (or other lights) in dark environments. I’ll have to experiment with it sometime :P
Also, I guess I stand corrected about the response time difference!
It worked best with alarm clocks because the outline of the clock was likely 100% only picked up by rods, but the display is visible clearly by the cones. So if you picked it up and moved it, the display would seem to move independently of the case it was part of.
I think the second link does suggest the difference between response times varies between people. So, maybe it’s just more noticeable by some?
Same thing happens with those old alarm clocks with LED 7 segment displays.
I think it’s the difference between rods and cones on the retina. Rods are more sensitive to light but do not see colour. That’s what you use in the dark. I suspect they are also slower to react to light (probably because they need to gather enough over time to make an image). So since the lighter text is bright enough to be seen by the cones, the reaction time is quicker so it looks like it’s moving independently.
Anyway that’s my theory and I’m sticking to it. :P
That similar effect happens with some 7-segment displays because they actually flicker very fast (either didn’t bother to add an AC rectifier if it’s wall power, or it uses PWM to regulate the brightness). You can see it if you point a camera at them. The visual cortex makes it appear like it’s a continuous light but that illusion is destroyed as soon as the light moves. (Pretty sure this is the same effect which makes a moving image on CRT monitors appear smoother at low frame rate than LCD monitors.)
I don’t know for sure but I don’t think there’s any significant reaction time difference between rods and cones in different light levels. There is a difference in how long they take to adjust to different light conditions (IIRC rods take longer to adjust to darkness but can achieve much higher sensitivity in darkness).
I think the effect produced by the strobing due to DC rectification is markedly different. That produced a gapping when moved (since the LED is off during some parts of the movement and on for others). The effect when there’s a somewhat brighter light source around a darkened room is very different where the lighter source seems to move independently of the object it is attached to, with no stuttering/strobe effect present.
The independent movement effect is much more likely due to response time differences. I thought I’d take a look and see. This https://www.nature.com/articles/s41598-018-34073-8 article by nature does mention there is a response time difference. However the vast majority of the paper seems to be about the effects on cone response time due to lack of calcium and other cone specific testing.
This https://pmc.ncbi.nlm.nih.gov/articles/PMC2063471/ article seems to imply that there’s a 20ms lag for rod reaction time. Where they tested people’s reactions to both kinds of stimuli and measured the response times.
Hmm, interesting. Then I don’t think I’ve ever seen that myself, I thought you were talking about the strobe effect, that’s the only thing I associate with 7-segment displays (or other lights) in dark environments. I’ll have to experiment with it sometime :P
Also, I guess I stand corrected about the response time difference!
It worked best with alarm clocks because the outline of the clock was likely 100% only picked up by rods, but the display is visible clearly by the cones. So if you picked it up and moved it, the display would seem to move independently of the case it was part of.
I think the second link does suggest the difference between response times varies between people. So, maybe it’s just more noticeable by some?