Canon EOS 40D - exposure metering calibration test

A number of years ago (when I was just a kid, perhaps only 65 or so) , Chuck Westfall of Canon US described a "quick-and-dirty" test that could be used to see of the exposure metering "calibration" of a Canon EOS camera was roughly in line with Canon's intent.

In the test, with the camera set for a "benign" processing style, a metered shot was taken of a uniform-luminance, neutral test target.

The resulting image was examined in Photoshop with the gray gamma 2.2 color space in effect. Ideally, the average K value of the image should be 55.

K is the sole coordinate of the gray gamma 1.8 and gray gamma 2.2 color spaces. Its symbol is evocative of "K" used to mean "black" in the CMYK color space, and it indeed increases for "darker" colors. It is zero for reference white, and 100 for black.

In gray gamma 2.5, a K value of 55 suggests a luminance of 17.3% of saturation.

Just for kicks, I ran the "Westfall test" on my trusty Canon EOS 40D.

The average sRGB color of a modest-sized region in the center of the frame was 114, 115, 113 (the test target was indeed not exactly neutral). That suggests a luminance of 17.3% of saturation!

Photoshop, with the gray gamma 2.2 color space in effect, reported the color of that region to be K = 55.

Better lucky than good, Carla always says.

We note that for an automatic exposure system following the ISO standard, fed an "exposure index" that is the ISO SOS for the sensor system (not the ISO speed), the expected average color of a metered shot of a uniform luminance neutral target would be 18.1% of saturation. The Canon "bogey" is about 0.07 stop lower than that.

And Chuck, if you're lurking: thanks for everything.

Best regards,

Doug

Today I did a basic "automatic exposure metering calibration test" on my Canon PowerShot G16 camera. This is a small, small-sensor camera with sophisticated user controls, generally in the "Canon way".

In the test, I took test shots of a nominally uniform luminance, fairly neutral test target (hey - the wall in front of me as I sit at my computer - that way I don't have to get out of my chair to do this). These snots were taken with three different metering "patterns" offered by the G16: center-weighted average (CWA), evaluative, and spot.

I had commented before that with my Canon EOS 40D, the results of such a test were very nearly the same for the different metering patterns (not that being consistent is "correct" - that is a rather complicated matter beyond the scope of this note).

In the case of the G16:

• There was a significant difference between the results for the different metering patterns.

• In all cases, the average exposure result was substantially greater (in normalized luminance) than with the 40D.

Here are the results for the three metering patterns. For each, I first give the normalized image luminance (in percent of "saturation"), followed by the degree (in stops) that this varies from the nominal value suggested some years ago by Canon as normal for EOS digital cameras (and very closely followed by my EOS 40D), 17.3%, and then compared to the value we would theoretically expect (following my usual trail through the various ISO standards), 18.1%

• Center-weighted average (CWA): 20.4%, +0.24 stops (vs 17.3%), +0.17 stops (vs. 18.1%)
• Evaluative: 27.1, +0.65 stops, +0.58 stops
• Spot: 24.3% +0.49 stops, +0.42 stops

Interesting.

Best regards,

Doug