Cine Meter II’s color metering exploits iOS’s built-in color measurement API, available
starting with iOS 8.
I found in testing that the reported temperature values tended to be low, so I added a calibration step for correction.
How accurate is it? Overall, Cine Meter II is about as accurate as other color meters: its readings are within the same general ranges as those reported by “real” color meters. For continuous-spectrum lights such as tungsten or sunlight, Cine Meter II (and the other meters) all agree very closely. For LEDs, fluorescents, and other light sources with unusual spectra, no two color meters agree—but Cine Meter still gives readings close to the average of the other meters’ readings.
Here are comparisons of Cine Meter II, both uncalibrated and calibrated, with color meters from Minolta, Sekonic, and Asensetek. There’s background info on these other meters at DVInfo.net, along with a discussion of why color metering is difficult and subject to disagreement between meters.
In all tests below, I show several “real” color meters’ readings, then Cine Meter II “raw” values without calibration, then the calibrated values. Where “average” is shown, it’s the average of all the “real” meter readings. “Trusted average” is the average of the Asenstek Lighting Passport Essence, the Sekonic C-500 in digital mode, and the Sekonic C-700 in digital mode (since the digital modes are supposed to match the color sensitivity of a digital camera).
These tests were the first for the “grayGains” calibration method (four different calibration methods were tested, and grayGains gave the best results):
In these tests the devices running Cine Meter II were all calibrated using the "Halogen spot" light, so they all show exactly the same correlated color temperature under that source.
The following tests were done with the Cine Meter II devices calibrated under a different tungsten light (a tungsten reflector lamp in my hallway, not listed), using the final code for the release:
Results are similar to test #2: calibrated Cine Meter II CCT readings, in general, are close to the average or trusted average of the “real” meters, and almost always within the bounds of the highest and lowest CCTs.
Note that I don’t have readings from the C-500 for Halogen + 30G (a full plusgreen gel over the halogen light) as the meter said the light level was too low.
For tint, the different numerical readouts of the meters were all converted to CC (Wratten) numbers for comparison. Some meters aren’t shown: the Lighting Passport Essence doesn’t report tint (it’s in the meter’s data file as a Du’v’ value, but I didn’t know how to extract it when I ran these tests), and the C-500 again reported “low light” in the Halogen + 30G test.
Cine Meter II has a tendency to see more magenta under certain lighting conditions than the other meters. My 5500K Kino-Flo tubes are very old and they do radiate a magenta light; arguably Cine Meter II gives me better guidance than the other meters with these old tubes. Under overcast daylight, the other meters are probably more accurate for tint; for the Litepanels Micro (a 1st-generation one with poor color rendering) I find that a 1/4 plusgreen gel makes it look better, as Cine Meter II suggests.
I didn’t trust my own judgment as to the value of these readings (I'm obviously biased, since I wrote the code) so I ran a two-month beta test with 16 cinematographers in the Americas, Europe, Australia, and Hong Kong. All of them reported that Cine Meter II was no worse than their other color meters, and that it was just as useful on scouts and for balancing light colors during setup—and they unanimously said I should release it. So I did.
I found in testing that the reported temperature values tended to be low, so I added a calibration step for correction.
How accurate is it? Overall, Cine Meter II is about as accurate as other color meters: its readings are within the same general ranges as those reported by “real” color meters. For continuous-spectrum lights such as tungsten or sunlight, Cine Meter II (and the other meters) all agree very closely. For LEDs, fluorescents, and other light sources with unusual spectra, no two color meters agree—but Cine Meter still gives readings close to the average of the other meters’ readings.
Here are comparisons of Cine Meter II, both uncalibrated and calibrated, with color meters from Minolta, Sekonic, and Asensetek. There’s background info on these other meters at DVInfo.net, along with a discussion of why color metering is difficult and subject to disagreement between meters.
In all tests below, I show several “real” color meters’ readings, then Cine Meter II “raw” values without calibration, then the calibrated values. Where “average” is shown, it’s the average of all the “real” meter readings. “Trusted average” is the average of the Asenstek Lighting Passport Essence, the Sekonic C-500 in digital mode, and the Sekonic C-700 in digital mode (since the digital modes are supposed to match the color sensitivity of a digital camera).
These tests were the first for the “grayGains” calibration method (four different calibration methods were tested, and grayGains gave the best results):
In these tests the devices running Cine Meter II were all calibrated using the "Halogen spot" light, so they all show exactly the same correlated color temperature under that source.
The following tests were done with the Cine Meter II devices calibrated under a different tungsten light (a tungsten reflector lamp in my hallway, not listed), using the final code for the release:
Results are similar to test #2: calibrated Cine Meter II CCT readings, in general, are close to the average or trusted average of the “real” meters, and almost always within the bounds of the highest and lowest CCTs.
Note that I don’t have readings from the C-500 for Halogen + 30G (a full plusgreen gel over the halogen light) as the meter said the light level was too low.
For tint, the different numerical readouts of the meters were all converted to CC (Wratten) numbers for comparison. Some meters aren’t shown: the Lighting Passport Essence doesn’t report tint (it’s in the meter’s data file as a Du’v’ value, but I didn’t know how to extract it when I ran these tests), and the C-500 again reported “low light” in the Halogen + 30G test.
Cine Meter II has a tendency to see more magenta under certain lighting conditions than the other meters. My 5500K Kino-Flo tubes are very old and they do radiate a magenta light; arguably Cine Meter II gives me better guidance than the other meters with these old tubes. Under overcast daylight, the other meters are probably more accurate for tint; for the Litepanels Micro (a 1st-generation one with poor color rendering) I find that a 1/4 plusgreen gel makes it look better, as Cine Meter II suggests.
I didn’t trust my own judgment as to the value of these readings (I'm obviously biased, since I wrote the code) so I ran a two-month beta test with 16 cinematographers in the Americas, Europe, Australia, and Hong Kong. All of them reported that Cine Meter II was no worse than their other color meters, and that it was just as useful on scouts and for balancing light colors during setup—and they unanimously said I should release it. So I did.
© 2015–2018 Adam J. Wilt. Last updated 2018-05-24