Underexposure question

I'm currently using a 300D (and have a 30D on the way - yay!) and I'm a newbie in the photography world (mainly shot on auto for the last 5 years). Now I mainly use AV mode letting the camera pick the speed and am dabbling with M mode. However, the majority of my shots come out underexposed. Why is this and what can I do to fix it? I've tried adjusting the ISO which does help in some cases but also gives me more noise in the pic than I would care for. I try to avoid using the built-in flash indoors alot because I'm not equiped or experienced enough to avoid a shadow caused by it and outdoors because my subject wouldn't be affected anyway (I think). I don't know if I've given enough info but any help would be appreciated.

When the shots come out underexposed, what does the exposure indicator in the viewfinder show? Does the camera think it's exposed correctly? If the camera thinks it's OK but your eye still says its underexposed, you can either use some exposure compensation or switch to manual and adjust your shutter speed until the camera says you're a stop overexposed. You'll have to play with it to get a feel for it. Or maybe you accidentally set exposure compensation to force it to underexpose and it's doing exactly that.

Can you post a shot, with the embedded EXIF information intact? You might inadvertantly have a negative value EC set into the camera, which would tell it to underexpose!

If you fill the viewfinder with just your lawn, does it come out exposed correctly? A lawn is about 18% tonality, similar to a gray card (there are variations in lawn brightness, but it is close enough without a calibrated target in the viewfinder!)

If you aim your meter at a mid day blue sky, if your ISO is set to 100, does the camera indicate 1/100 f/16 as the right exposure to use?

Here's one I did the other day. This is straight out of the camera.

I'll try your grass and sky suggestions if we can ever get a break from the rain.

eta: I don't know why it dosn't give the shooting mode but it was AV.

^ actually it's not that badly exposed.. It's possibly 1/2 a stop or so under which can be easily fixed in post processing..

yes, I have recently tried to do the exposure compensation and that does help improve things. I guess I was wondering why the camera doesn't "get it right" without that. Or is that because every capture is subjective and situational? - no one hard and fast rule?

I can see your camera was set to ISO 400, auto exposure with no EC bias, and the camera used 1/320 f/3.5 ambient light only. Nothing seems to be set in a manner which would mistakenly cause underexposure.

I believe you are experiencing what is known as 'subject failure'...the skin is +1EV brighter than an 18% gray card. So it is trying to convert your infant's skin to 18% gray tone, thereby underexposing by -1EV and leading to the result that you posted. The remedy would be to deliberately dial in EC = +1, and your infant's skin would be correctly captured at its inherent brightness.

Let me use this as a teaching point to everyone reading this thread...a meter tries to turn everything that it sees to a mean 18%.
It tries to take the black cat in a coal mine and render the scene to 18% gray -- overexposing the scene.
It tries to take the winter bride in her white gown outside in the snow to 18% gray -- unexderexposing the scene.
Therefore it is essential that the photographer engage his/her brain, and alter the exposure accordingly in interpreting the inherent brightness of the subject, via use of EC (when in one of the auto modes whether it be P or Tv or Av) And switching the camera from Evaluative to Center weighted to Partial or Spot does NOT bias the reading, it merely controls the area of the frame used to decide what to turn 18% gray!

Well put Wilt..

yogestee wrote in post #8468323
Well put Wilt..

I could have put it better!

The PALM of the hand, regardless of racial background (or status of one's tan from exposure to sunlight) is +1EV from 18% gray. Since the OP's infant is a white M/F newborn, the +1EV description works for that subject's inherent brightness, too!

daystar wrote in post #8468234
yes, I have recently tried to do the exposure compensation and that does help improve things. I guess I was wondering why the camera doesn't "get it right" without that. Or is that because every capture is subjective and situational? - no one hard and fast rule?

Yes, every situation is different.

In a nutshell, this is the problem. The correct exposure for any situation is dependent on the amount of light falling on the subject (the incident light) which your camera is not equipped to measure.

Your camera measures the amount of light that bounces back off the subject (the reflected light) and uses that to make a best guess about the correct exposure. The problem is, the brighter the subject the more light it reflects back. So if you take a picture of a polar pear in a snow field the camera will think it is a heck of a lot brighter out than it is. You end up with an underexposed shot.

I think this is what happened in this case, the white patches on the blanket plus the tone of the infants skin made for a bright scene. About +2/3 stop exposure compensation would have fixed the shot.

There are many ways to skin this cat. You can over time develop a sense for scene tones and how your camera meter will read it. You could also get an incident light meter and simply stop using your camera meter altogether (this is what I do). Finally, you can always take a test shot and look at the histogram.

"Tis my firm belief that PP was invented to cover the the mistakes that we make during the taking of the shot, and as Jurgen pointed out, this is an easy fix, providing you shoot in RAW.
It however, doesn't excuse the photographer from having a thorough understanding of how his camera functions.

Wilt wrote in post #8468243
I can see your camera was set to ISO 400, auto exposure with no EC bias, and the camera used 1/320 f/3.5 ambient light only. Nothing seems to be set in a manner which would mistakenly cause underexposure.

I believe you are experiencing what is known as 'subject failure'...the skin is +1EV brighter than an 18% gray card. So it is trying to convert your infant's skin to 18% gray tone, thereby underexposing by -1EV and leading to the result that you posted. The remedy would be to deliberately dial in EC = +1, and your infant's skin would be correctly captured at its inherent brightness.

Let me use this as a teaching point to everyone reading this thread...a meter tries to turn everything that it sees to a mean 18%.
It tries to take the black cat in a coal mine and render the scene to 18% gray -- overexposing the scene.
It tries to take the winter bride in her white gown outside in the snow to 18% gray -- unexderexposing the scene.
Therefore it is essential that the photographer engage his/her brain, and alter the exposure accordingly in interpreting the inherent brightness of the subject, via use of EC (when in one of the auto modes whether it be P or Tv or Av) And switching the camera from Evaluative to Center weighted to Partial or Spot does NOT bias the reading, it merely controls the area of the frame used to decide what to turn 18% gray!

Thank you, thank you! I think that's the best explanation I've gotten on the topic - the best because now I get it! Now I understand what the camera is trying to do. Now, Grasshopper will try to apply.

JeffreyG wrote in post #8468352
Yes, every situation is different.

In a nutshell, this is the problem. The correct exposure for any situation is dependent on the amount of light falling on the subject (the incident light) which your camera is not equipped to measure.

Your camera measures the amount of light that bounces back off the subject (the reflected light) and uses that to make a best guess about the correct exposure. The problem is, the brighter the subject the more light it reflects back. So if you take a picture of a polar pear in a snow field the camera will think it is a heck of a lot brighter out than it is. You end up with an underexposed shot.

I think this is what happened in this case, the white patches on the blanket plus the tone of the infants skin made for a bright scene. About +2/3 stop exposure compensation would have fixed the shot.

There are many ways to skin this cat. You can over time develop a sense for scene tones and how your camera meter will read it. You could also get an incident light meter and simply stop using your camera meter altogether (this is what I do). Finally, you can always take a test shot and look at the histogram.

And thank you - trying to understand how lighting effects the subject and the image has been challenging for me and I've got sooooo much to learn. I really appreciate the time taken to help me understand and (hopefully) improve.

chauncey wrote in post #8468361
"Tis my firm belief that PP was invented to cover the the mistakes that we make during the taking of the shot, and as Jurgen pointed out, this is an easy fix, providing you shoot in RAW.
It however, doesn't excuse the photographer from having a thorough understanding of how his camera functions.

^^^

It takes a very practiced eye to judge tonality accurately. And unless you have developed preciseness in the practiced eye, there will be small errors in exposure (1/3EV to 2/3EV error for example) that are best adjusted out in post processing. Most of us don't have exposure meters for brains, so we rely upon tools like incident meters or gray cards to reduce the error of our brains trying to interpret red or green or blue (or whatever hue) tonal intensities relative to 18% gray.

Wilt wrote in post #8468464
^^^

It takes a very practiced eye to judge tonality accurately. And unless you have developed preciseness in the practiced eye, there will be small errors in exposure (1/3EV to 2/3EV error for example) that are best adjusted out in post processing. Most of us don't have exposure meters for brains, so we rely upon tools like incident meters or gray cards to reduce the error of our brains trying to interpret red or green or blue (or whatever hue) tonal intensities relative to 18% gray.

I agree with that. The #1 reason that I started to use an incident meter is that I dislike going through 100-200 shots from an outing and having to adjust the exposure of every shot +/- 1/2 stop as was commonly the case when I was less experienced and shooting in Av mode.

Few people can look at a scene in the viewfinder and simply dial in exposure compensation based on the tones.