I’ve tested the following settings:
- Screen, max brightness, bright cold ambient lights (5000K)
- Screen, max brightness, no ambient lights
- Screen, lowest brightness, no ambient lights
- Screen, max brightness, bright warm lights (2200K)
- Paper, bright cold lights (5000K), directly illuminating paper
- Paper, low cold lights (5000K), directly illuminating paper
- Paper, warm bright lights (2200K), directly illuminating paper
- Paper, warm low lights (2200K), directly illuminating paper
For measurements I used the 3 times testing I generally using for cm measurements.
- My hypothesis seems to be correct. In low light settings it’s easier to detect the edge of blur. It was easiest to detect on screen with lowest brightness. But on paper even on my lowest lamp setting it was brighter than the screen with lowest brightness. So I’m not sure what would be the result if they would have been equal.
- Interestingly warm light also has positive effect on edge of blur detection. Even when it’s just ambient light, which I’ve really not expected.
- Ambient light brightness seems to have no substantial effect when measuring on screen. Though total darkness with bright screen seems to strain the eyes (no suprise here though). Ambient light color seems to have effect though, as I mentioned.
- Doing so much measurement prooved to be a mistake. About halfway my eyes (especially my right eye) started to get strained. I have to aborted the last measurement, right eye was totally tired at that point, I was practically not able to see clearly with it.
- I taped the paper on my computer screen to have similar setup, but this resulted in a white - black edge near the text I used for measuring. I’m not sure if this had any effect or not, I was halfway through paper measuring when I realized the problem, so did not want to change.
- The paper and screen result not comparable anyhow. The brightness obviously was not matched, and the text font and size was not the same (though similar). But this was expected.
- Lowest average variance in the measurements was in the “screen, darkness, low screen brightness” settings. If you check the raw data, the “screen, bright warm lights” shows lower variance, but I suspect in the former case there were a measurement error for the left eye. Without that it would be superior (both left eye and both eye showed superior variance in measurement, only one data point is a big outlier for the left eye). Though warm bright ambient light is close to it, which is really interesting for me.
If anyone is interested here is the raw data, and the remarks I’ve written during testing:
- I’m not sure how valid my paper measurements at all though. My eyes was tired when I’ve reached that point. So I don’t think I can make any conclusion there, nor compare the measurement on paper and on screen. Most likely I will do a separate paper testing and get the best of both test result to do a third test to compare them.
- For me lower light settings on screen results in better detection of edge of blur and as a result lower variance. They also results lower cm measurements. I may switch to measure like this in the future, but currently I’m not sure psychologically I’m ready to see lower numbers I already have to bear through that during the summer when I’ve stricted up my measurement. A possible compromise is to use warm light, which also shows lower variance, though I’m not sure why. More testing need for this.
- When I redo the paper tests I will also include one where the light is as low as possible to still be able to read the letters.
Edit: the “warm” light was 2200K, not 2700K