4 same model iPhones...
The article goes on to talk about identical component tolerances (typically you buy 1%, or 5%, or 10%, tolerance components, meaning there's a measured +/-% of variation within all the components of the same model/specs for a given tolerance rating) and how it may be the culprit.
"We suspect that the reason for these fluctuations lies in the tolerances of the components built into the phones. These are tolerances which the user does not notice in everyday use but which in direct comparisons have a great effect on the displayed illuminance values."
"Even when using the same app in structurally identical smartphones different measuring results are achieved.
Therefore apps are unfortunately not really of any great assistance for measuring illuminance and not even any use to obtain a general idea of the illuminance value. On the contrary: They lead the user in the completely wrong direction."
Do luxmeter apps replace a professional meter? DIAL has tested how they perform compared to a calibrated class A illuminance meter.
www.dial.de
Not trying to beat a dead horse, but I figured I'd just throw up a bit more of the data to help give a better picture. Idk this company or their credibility, ect, and they obviously haven't tested every make and model with every app. But their conclusion is technically sound imo and if the data is as consistently inconsistent across other makes and apps then I'd suggest not putting a ton of merit in what the app reads back??
I'd be curious to see some
pictures of members cellphone apps & corded luxmeters side by side and measuring lux at various heights ect. See how well they maintain proportionality. If the % difference between the 2 device's readings changes drastically with varying height/intensity then it would support the websites findings. But if they seemed to scale at the same rate, then I'd wonder how accurate the lux meters are, as well as the websites analysis. Without a known reference value to compare the readings to, then idt we'd actually be able to tell where the discrepancy is found (within corded lux meters or within the websites analysis), but there may be ways of teasing it out if it so happens that the 2 devices measurements seem to maintain a ~static % difference (if you like to go down the rabbit hole lol).