EOS 20D's evaluation of exposure in sunny conditions

[david-nicolas]

My EOS 20D seems to evaluate exposure wrongly in very bright, sunny conditions:

it overexposes the shot. Taking account of that, I find that manually correcting

exposure by 1 stop most often makes for a well-exposed image in very bright,

sunny conditions. What do you think about that? Is it normal? Or is there a

defect in my camera?

[lovcom_photo]

How are you gaging over exposure? By looking at the histogram? And if so, are you seeing blown highlights?

[at_nick]

At first I felt my Rebel XT did the same, and most of my friends with 20Ds and 30Ds felt that they overexpose by about 1-stop in bright light.

 

I got to where I could predict it, and adjust for it by turning down the exposure compensation by 1-stop.

 

But lately I've been working on biasing my exposures so that the histogram is snugged up to the right, but not to the point that the highlights are blown, and then I adjust the brightness in post-production. I do this because digital sensors record most information at the bright end of the scale. Check out this article on �exposing to the right� �

 

http://www.luminous-landscape.com/tutorials/expose-right.shtml

 

So now that I�m exposing to the right, and making adjustments in post production, I find that my Rebel XT only overexposes by 1/3 stop, on average.

[picturesque]

It's normal. You can also try changing the ambient metering type to averaging rather than evaluative. In some conditions, not all, averaging does not blow out highlights in contrasty sunlit scenes, but both still tend to blow them out if left uncompensated to the minus side.

[david-nicolas]

Thanks for your answers. How much the shot is overexposed is difficult to assess exactly. But when I made no correction before taking the shot, I had a raw image with which it was very hard, if not impossible, to get satisfactory results: too much was blown up. Now that I correct 1 stop or 2 thirds of a stop, I get nice colors from my raw image, often with little to change.

[PuppyDigs]

Do a "sunny F16" meter test. That is, meter a gray card (or any medium toned object, e.g.,

northern clear blue sky) near midday. Set ISO 100, AV mode at F16 and your resulting

shutter speed should be 1/100 (or nearest displayed value, e.g., 1/90 or 1/125). If your

meter passes, it is operating correctly.

 

However, no current digital camera can record the full dynamic range of a sunny day

landscape. Even if the mids and shadowns are perfect, the brightest highs will blow out.

Slide film was exactly the same (print film could take in the whole enchilada). I find my

10D, 20D and 5D all need -1/3 to 1/2 EC to keep the highlights from sunny day scenes

from blowing out. Also, if you shoot RAW you'll have at least a stop more headroom than

JPEG.

 

Oddly, cloudy day scenes usually need +1/3 to 1/2 EC to look right.