4
\$\begingroup\$

Recently I experimented with using shutter-priority ("Tv" on my Canon DSLR) for images shot from my kayak, because movement is a constant factor in that situation (even if there's no wind, there's always a bit of current pushing me along). I started off with Auto ISO but quickly realized I needed to set the ISO higher to get enough depth of field, since the camera was otherwise going to use the lowest ISO available. In reviewing the images, I'm puzzled about how the camera selected the aperture. Over and over again, I see two images in a row, taken in the same light with almost the same composition, having different apertures - sometimes wildly different. Here's an example, with 1 1/3 F-stops difference, both shot at 100 ISO and 1/200 sec.

F/5.6
1/200 sec, F/5.6, ISO 100

F/9
1/200 sec, F/9, ISO 100

What is causing such a difference?

Metering was "partial" (Canon) for all of them. Light conditions were bright with a few clouds but mostly clear. I am used to using a dSLR in a kayak - the difference this time was I was trying TV (shutter-priority) rather than on M or AV as I've used in the past.

\$\endgroup\$
8
  • 2
    \$\begingroup\$ Two different images shot at the same shutter speed, same ISO, and "wildly different" apertures means, two "wildly different" exposures. What was the camera's light-metering algorithm looking at in the different pictures? I'm imagining kayaking on a sunny day, and I can picture quite a range of dark and bright subjects. \$\endgroup\$ Oct 1, 2019 at 16:02
  • 1
    \$\begingroup\$ I have a hard time imagining a situation in a kayak where I can both take pictures with a DSLR and have to worry about the current speed. Also, if you are on water you can have moving sun reflections and some of them can hit exposure metering spots in one shot and not in the other. \$\endgroup\$
    – xenoid
    Oct 1, 2019 at 19:56
  • 2
    \$\begingroup\$ Possible duplicate of How do DSLRs figure out what aperture to select in P mode? \$\endgroup\$
    – mattdm
    Oct 1, 2019 at 20:17
  • \$\begingroup\$ Other question is about P rather than S (or Tv, depending on brand), but fundamentally the same. \$\endgroup\$
    – mattdm
    Oct 1, 2019 at 20:17
  • \$\begingroup\$ The other question about P mode includes other variables as well: At the very leat Tv and quite possibly ISO as well. Much of the answers deal with how the camera selects which variable to prioritize over the other(s). With manually selected ISO, what Av gets selected in Tv mode is pretty much only about metering. \$\endgroup\$
    – Michael C
    Oct 2, 2019 at 7:08

2 Answers 2

3
\$\begingroup\$

The camera can be set to lock the exposure in auto modes using different buttons, but a system default is to lock in the exposure at the time you hold the shutter release half way down. So, you'd get to where you want to take a shot, go halfway down, the autofocus and exposure would lock in, and then you'd take the shot.

Since you were using partial metering in a scene that was dramatically different between shaded area and unshaded, my guess is that your bobbing up and down in a kayak caused you to meter mostly in the shade for one frame and mostly in the sun for the other.

You probably locked the exposure while the meter was located here on this frame:

enter image description here

and probably locked in while metering here on the other:

enter image description here

Subtle differences for sure, but enough to cause a big exposure calculation change.

For a scene like this, it's most important not to blow the highlights. If you shot RAW, it actually appears that your brighter shot may be the better of the two in the end, as you can tone down the exposure in post (assuming the highlights are salvageable).

If you only get one shot at the image (like kayaking down a river), then I'd tend to use evaluative or center weighted average, making sure to meter off averages in the scene, not the brightest or darkest points.

Also, there are custom functions that allow you to decouple the exposure lock and autofocus from the shutter release. I'm personally a fan of back-button focus and the shutter release used as exposure lock, but there are other options. I'd advise doing this so that you can figure out where to point your camera to get a good meter reading for the scene using a mode you are comfortable with, lock that in, and then recompose and focus and shoot.

\$\endgroup\$
4
  • \$\begingroup\$ With default settings, metering only locks with a half press on EOS cameras when in Evaluative metering mode. It does not lock in Spot, Partial, or Center-weighted average with a shutter half press unless the custom menu option for shutter half-press is changed from 'AE/AF Start' to 'AE Lock' . \$\endgroup\$
    – Michael C
    Oct 4, 2019 at 2:41
  • \$\begingroup\$ Actually I always use back-button focus - have done so for years. So I don't think I locked in the exposure accidentally. However, I do appreciate the info about partial metering and advice to use one of the others instead - will do so in future. Thanks. \$\endgroup\$
    – LMacB
    Oct 4, 2019 at 15:28
  • 1
    \$\begingroup\$ @LMacB If you have the shutter half-press option changed from option 0: 'AE/AF Start' to option 1:'AE Lock/AF' in order to enable back button AF, then you are locking exposure every time you half press the shutter in any metering mode. If you do not wish to lock exposure with a shutter button half-press when using back button AF, then instead select option 3:'AE/AF, no AE lock' \$\endgroup\$
    – Michael C
    Oct 9, 2019 at 0:46
  • \$\begingroup\$ Thanks for this tip, Michael - I did indeed have the lock on, unintentionally. I have now changed this to AE/AF, no AE lock. \$\endgroup\$
    – LMacB
    Oct 15, 2019 at 15:01
2
\$\begingroup\$

How does a camera determine aperture when in shutter-priority mode?

It measures the amount of light reflected from various areas of the scene and calculates a desired exposure value based upon the metering mode used. It then selects an aperture value that, combined with the shutter duration and ISO setting you have selected, will result in the desired exposure value.

If an appropriate aperture value is available, the camera will use that aperture, along with your selected ISO and Tv, to take the picture.

If an aperture value needed to give the desired exposure value is not available , one of several things may happen depending upon the camera in question and the state of certain user selectable settings.

For instance, with your selection of ISO 100 and shutter duration of 1/1000 second, if the camera wants an aperture of f/1.4 but your lens' maximum aperture is f/3.5, the following scenarios are possible:

  • The camera will take the photo at ISO 100, 1/1000, f/3.5 and the exposure will be two and two-thirds (2 2/3) stops darker than desired.
  • If "Safety Shift", a feature offered by most Canon DSLRs, is enabled and "Tv" is selected as the first shifted value, the camera might take the photo at ISO 100, 1/160, f/3.5. The camera has reduced your selected Tv by two and two-thirds (2 2/3) stops to give the indicated exposure value.
  • If "Safety Shift", a feature offered by most Canon DSLRs, is enabled and "ISO" is selected as the first shifted value, the camera might take the photo at ISO 640, 1/1000, f/3.5. The camera has increased your selected ISO by two and two-thirds (2 2/3) stops to give the indicated exposure value.

In reviewing the images, I'm puzzled about how the camera selected the aperture. Over and over again, I see two images in a row, taken in the same light with almost the same composition, having different apertures - sometimes wildly different. Why is this?

Different metering modes measure different parts of the scene to determine proper exposure. Here are the typical metering mode choices for many Canon cameras:

enter image description here

The gray areas in each "viewfinder" are what is used to measure how much light is in a scene. With center-weighted averaging, the darker areas are given more weight than the lighter areas. For 'Spot' and 'Partial' metering the active areas shown in grey are given the same weight and the rest of the frame is ignored.

'Evaluative' is a bit different. In this metering mode, the camera attempts to identify what type of scene the camera is being pointed at and then adjust metering based on the result. The readings over the entire light meter are compared to a "library" of different scenarios and the instructions for the one that best matches the current scene are used.

In early implementations using monochromatic light meters back during the film days, the first evaluative metering systems (Nikon calls in "Matrix" metering, other makers have different names for it) were fairly simplistic. For instance, if the upper 1/3 of the scene was very bright and the lower 2/3 was not as bright, the camera might be programmed to identify this as a "landscape" scene and expose for the lower two-thirds (2/3) while allowing the sky in the upper one-third (1/3) to blow out. If the upper two-thirds (2/3) was very bright and the lower one-third 91/3) was darker, the camera might be programmed to expose for the sky instead, and leave the landscape in the lower one-third (1/3) very dark. This would have been based on an assumption by the programmers that the part of the scene the photographer wished to expose properly was the part that took up more of the frame.

With modern color based light meters (in DSLRs with an RGB+IR meter, this is a low resolution color imaging sensor located in the prism area of the viewfinder - with mirrorless cameras it is the main imaging sensor) and an extensive "library" of different scenarios, cameras are getting better and better at "guessing" what kind of scene is in the frame and what the photographer wants to be "properly" exposed as a mid-tone.

With 'Partial' metering only about 10% of the total field of view is measured for brightness. The percentage can be a little lower or a little higher depending on the specific camera model, but it is rarely less than about 6-7% and rarely more than about 12-14% with Canon cameras. Whatever is in the rest of the frame is ignored by the light meter, only the center circle is metered.

  • If the camera is pointed at a dark part of the scene, the exposure will be very bright because the camera is attempting to make the dark area a mid-tone (often called "medium gray", even when shooting color images) in the resulting photograph.
  • If the camera is pointed at a very bright part of the scene, the exposure will be very dark because the camera is attempting to make the very bright area a mid-tone in the resulting photograph.
  • If the camera is pointed at a medium bright object, the exposure will be somewhere in between because the camera is attempting to make the medium bright object a mid-tone in the resulting photograph.

When taking pictures on the water with the sun shining brightly, reflections of the sun off the water can dramatically affect metering. As the water moves and the angles between the sun, the water, and the camera are constantly changing the amount of brightness in any particular small area can vary wildly.

  • The brightest reflections are so much brighter than anything else in the scene that only the very bright reflections will have any details when partial metering measures a bright reflection. The rest of the frame will be near black.
  • On the other hand, if there are no bright reflections with in the partial metering area, the exposure will be much brighter and most of the frame will appear properly exposed, but the details of bright reflections in other parts of the scene will be "blown out".

Update following addition of example image to the question:

The two examples added to the question show center areas that are roughly the same brightness in the actual scene. It appears exposure was locked prior to the camera being moved to the final composition used.

  • Exposure was locked for the first example when the center of the viewfinder was pointed at a darker area in the scene.
  • Exposure was locked for the second example when the center of the viewfinder was pointed at a slightly brighter area in the scene.

With default settings, exposure is only locked by a half-press of the shutter button when EOS cameras are in 'Evaluative' metering mode. In 'Spot', 'Partial', and 'Center-weighted average' metering modes, by default exposure is not set until the moment the shutter button is fully pressed and the image is taken. But there are custom settings which change this.

There are three possibilities for locking exposure prior to fully pressing the shutter button in 'Partial' metering mode:

  • Using Manual exposure mode to set ISO, Tv, and Av manually.
  • Pressing the 'AE Lock' Button on the back of the camera. It's the button marked with a six pointed asterisk (*) that shows up as a five point superscript star when we type it here.

The 'AE Lock' button by default is set to lock exposure until the photo is taken or metering times out, typically about 5 seconds after you take your finger off the shutter button.

With some EOS camera models, the 'AE Lock' button (or any other button mapped to perform the 'AE Lock' function) can be modified to act as a "toggle switch" that locks exposure when first pressed and does not unlock exposure again until the 'AE Lock' button is pressed a second time, regardless of how many frames are captured or how long metering has been inactive. Be very careful if you select this option while shooting in a "semi-automatic" exposure mode (P, TV, or Av)! If you press it once and leave it on, every picture you take will use the initial exposure settings that you locked in, even if the camera goes into 'standby mode' for several minutes. Turning the camera off with the 'On/Off' switch will clear any locked exposure values.

  • Changing the behavior of the shutter half-press using a custom setting. The exact description of the menu item will vary by your specific camera model, which you have not disclosed, but it will either be named something like 'Shutter/AE Lock Button' for some older/lower tier models or it will be under the 'Shutter button half-press' section with models that have a 'Custom Controls' section.

For cameras that use the 'Shutter/AE Lock Button' menu option, the following from the EOS Rebel T3i/EOS 600D Instruction Manual (p.256) is typical:

enter image description here

Under each of the options, the behavior described to the left of the slash (/) is assigned to the shutter half press and the behavior described to the right of the slash (/) is assigned to the 'AE Lock' button. Exposure will only lock upon a shutter half-press in 'Partial' metering mode if '1: AE Lock/AF' is selected. Note: in such a case, half pressing the shutter button will not initiate AF, you'll need to press the 'AE Lock' button that has been remapped to be an 'AF ON' button to use autofocus.

For EOS cameras that use the 'Custom Controls' menu, the following from the EOS 7D Instruction manual is typical:

enter image description here

The camera will only lock exposure in 'Partial' metering mode upon a shutter half-press if the third option in Step 3, 'AE Lock' represented by the six starred asterisk, is selected.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Thanks - I really appreciate the detail you've provided re: metering modes, especially partial metering. Maybe that was the determining factor for me. I don't think I realized how little of the scene it was metering. I may opt for center-weighted instead, next time I'm in a similar situation. \$\endgroup\$
    – LMacB
    Oct 3, 2019 at 15:35
  • \$\begingroup\$ Evaluative is usually pretty good in such situations unless a bright reflection is directly in the center of the frame. If the light is not changing, I'd meter with evaluative, double check the histogram (which always reflects the full image), and set ISO, Tv, and Av manually. \$\endgroup\$
    – Michael C
    Oct 4, 2019 at 2:27
  • \$\begingroup\$ It's also possible that I accidentally locked the exposure, as you explained. On a bright day out on the water, it's hard to see details in the viewfinder like the asterisk that comes on when exposure is locked - I can usually manage to see the histogram but sometimes even that is a challenge to see in the glare. \$\endgroup\$
    – LMacB
    Oct 8, 2019 at 18:52
  • \$\begingroup\$ A big thank you, Michael C and Hueco, for the detail and attention you put into answering my question - very much appreciated. I probably won't be out in the kayak again until next spring, so I've packed that camera away for the winter (it's my second camera - I won't risk the main one when I'm on the water). But I changed its metering mode before putting it away, so that I won't forget to do this once I get back in my kayak. And I'll try to remember to take care not to accidentally set the exposure lock in future. \$\endgroup\$
    – LMacB
    Oct 8, 2019 at 18:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.