Hidden overexposure...

kelly1

Hello folks,

My recent purchase of a light meter has got me thinking about in-camera metering and how it can get it so wrong. At first I thought the lightmeter was underexposing but when I look at the images close up I can see how much details is preserved in high-lights on a "textural" level. When I say highlights, I'm not referring to large blown out white areas like sky. I mean details on any surface whether it's black or white.

Take for instance an smooth surface e.g. skin. On the exposure scale, the skin overall might be a stop above mid-grey but when you look at the exposure close up, there will be peaks on the skin which reflect more light than the "valleys" for want of a better word. So I think it's very easy to overexpose these almost microscopic details and in this way you loose details of the skin texture. Does this make any sense?

To give another example, I took 2 exposures of a tree in the evening sun and I shot one with a light meter reading and the other with the camera in Av mode. The one metered with the lightmeter show far more detail on the branches compared with the shot done in Av mode. The surfaces directly facing the camera were overexposed and hence details was lost even though
the histogram curve was placed in the middle and showed no sign of blowing out.

Anyone notice anything similar?

pixbyjohn

http://www.bythom.com/graycards.htm

kelly1

pixbyjohn wrote: »

http://www.bythom.com/graycards.htm

Interesting article John. It actually occurred to me that manufacturers deliberately over-expose with "amatueur" lenses attached and expose correctly with pro lenses which would increase saturation, detail and contrast. I rarely see the camera underexpose!

kelly1

Another 18/12% article:

http://www.richardhess.com/photo/18no.htm

CabanSail

Kelly - Are you talking about detail on film exposures here or digital? If it's digital is it jpeg or RAW files?

kelly1

CabanSail wrote: »

Kelly - Are you talking about detail on film exposures here or digital? If it's digital is it jpeg or RAW files?

Digital RAW. I'll have to post some samples of what I mean.

zerohamster

Mostly I would say the exposure diversities are from the camera's different methods of metering compared to say the handheld unit spot metering(?), are you using the same type of metering on camera as the light meter?

Different lenses at the same settings can let more or less light through than others because of the light transmission values of the lenses which are measured in T-stops rather than F-stops which the camera compensates with ISO for (without telling you).
This could have a factor in how the camera meters compared to the handheld meter but Im not sure.

ditpaintball

Are you using spot metering, average metering, evaluation metering with the camera etc, all these setting will change on what the camera is trying to meter for. If you hold the meter at the tree branches and get an exact reading for that area, then you should expose for that. Where as the camera might not be able to detect such a fine an area.

Generally it is know that on some camera sensors, that using +1/3 exposure compensation is needed as camera firmware can be a little bit off, depending on the camera.

If in doubt you can always over expose a 1\3 or 1\2 a stop as digital sensors are more sensitive to bright light so you can record more details. I.e. you can recover a blown out image better than you can recover an underexposed image by the same amount.... generally.

FoxT

I have not tried this myself, but it would be an interesting experiment to do this:

Take a photo of an evenly lit (with colour balanced source) white surface, and print it. Then you will have a gray card that matches the camera/lens combination perfectly ( assuming printer is calibrated etc) . The reason I havent done this is its all a bit too much like hard work....especially when you have a histogram on the back of the camera!

kelly1

After a bit more testing, it seems the problem is I'm just a poor judge of the number of stops above or below mid-grey for a given subject. But I do think in general cameras tend to over expose. Maybe it's just Canon.

I just don't trust the camera meter any more.