Spot metering off a white object - exposure compensation value?

[rogerb]

If you spot meter on an 100% white area, the camera will adjust its exposure to render it 18%grey. How many stops of compensation will I need to apply to make that white area ALMOST perfectly white (but not blown out)?

I would guess that it will depend on the dynamic range of the camera. If the dynamic range of the camera is supposedly 10 stops, does that mean that a +5 stop compensation will do it? That is a lot more than the "common wisdom" number of 2 - 3 stops.

I would appreciate some input on this subject, thanks.

[Colin Southern]

If you spot meter on an 100% white area, the camera will adjust its exposure to render it 18%grey. How many stops of compensation will I need to apply to make that white area ALMOST perfectly white (but not blown out)?

I would guess that it will depend on the dynamic range of the camera. If the dynamic range of the camera is supposedly 10 stops, does that mean that a +5 stop compensation will do it? That is a lot more than the "common wisdom" number of 2 - 3 stops.

I would appreciate some input on this subject, thanks.

Hi Roger,

Short answer: You'll need an EC of approx +2

Long answer: The difference in reflectivity between a black and white objects is only 4 stops, and 18% gray is 1/2 way between them. The rest of the dynamic range (often around 12 in total for many modern cameras) allows "headroom" for shadow detail and backlighting (although most monitors won't display more than 5 or 6 stops at best, so you usually have to compress the cameras range).

Does this help?

[rogerb]

Hi Collin.
At first I was surprised to learn that there is only about a 4-stop variation in reflected intensity between black and white, but then again the so-called "18% grey" (actually should be closer to 12 - 13%) is supposedly the mid-point in tonal reflection, so that would make sense. Assuming black results in more than zero reflection, 4 stops would make white in the 20% - 30% range which is pretty well in line.
I think I may conduct some experiments on my own to verify it but I suspect now that you are right on.
Thanks for responding.
Roger

[Raycer]

Hi Roger,
I think Colin gave a simplified answer to complicated questions. Yes you are correct that most camera now can record 10 to 12 stops of data in raw. But how the camera/software converts jpg can greatly vary the apparent dynamic range.

For high key set up, the white background is about 2 stops over the subject.

Looks like there's going to be a 12% vs 18% debate going here. I personally think its pretty useless for such a debate. When I use spot metering, I normally put it on my subject and select where I want the subject on the histogram. Because I don't use a grey card for metering, I don't care what spec manufacturers calibrate the cameras to.

Looking forward to see your test results.
Cheers

[rogerb]

Thanks Raycer.
I had intended to follow the "actually should be closer to 12 - 13%" with a "?"
I shoot raw exclusively and do the conversion in ACR. The more I read on the subject, the more I realize that there may not be a "one factor fits all", since each camera model will have its own interpretation of "grey". I have read in more than one article that the histogram shown on the lcd is derived from the jpg equivalent of the image (presumably to apply the non-linear correction). However I'm more concerned with determining when the logic from which the raw file is derived reaches saturation (electrically, not photographically!) and the 14-bit per channel output hits the top.
As far as my own experimentation, I'm concerned that my results may vary considerably as I expose my "black" and "white" targets to different kinds of light. In the end, I may just settle on "+2 stops" and be done with it!

Any thoughts on this?

[Colin Southern]

At first I was surprised to learn that there is only about a 4-stop variation in reflected intensity between black and white.

The easy way to verify it is to set the camera to spot metering (preferably manual mode) - set ut to read -2EV pointing to a black object ... and watch it go to +2EV when you point at a white object.

the so-called "18% grey" (actually should be closer to 12 - 13%) is supposedly the mid-point in tonal reflection, so that would make sense.

My understanding is that modern camera meters are calibrated to an ANSI standard, which equates to around 13% - however - I'm also told that the difference between 13% and 18% is only about 1/3 of a stop.

I think I may conduct some experiments on my own to verify it but I suspect now that you are right on.

I'd guarantee it

[Colin Southern]

I think Colin gave a simplified answer to complicated questions. Yes you are correct that most camera now can record 10 to 12 stops of data in raw. But how the camera/software converts jpg can greatly vary the apparent dynamic range.

What's jpg?