How good is Evaluative Metering?

Discussion in 'Canon EOS' started by dillan k, Jan 7, 2006.

  1. I'm a slide shooter with an Elan 7, and I've never been brave enough
    to rely upon evaluative metering. I use partial metering almost
    exclusively. I'm ok at it, but I still have to learn to trust myself
    more. I usually take a long time to meter my subject. I'd like to give
    the evaluative metering a try. Before I waste a roll of film, what can
    any of you say about it, both with and without flash (420ex). Thanks
    for your help!
  2. It's much more of a pain switching between metering modes on an Elan 7 than on a Rebel (just push * to get partial metering to compare with evaluative that is otherwise standard in P, Av and Tv modes) or Elan II (lever). However, you will learn most by comparing scenes for yourself. Don't forget to explore the metering CF8 option.

    So far as flash is concerned metering of the flash exposure is quite independent of the ambient exposure. Ambient metering for flash depends on the exposure mode and scene brightness relative to ISO. You would do well to read the flash bible here:
  3. great for neg's

    so-so for chromes

    crappy for dslr's

    The only way to learn exposure is trial and error. Bracketing will speed the learning process.

  4. ky2


    It's pretty good as long as your target image follows the 18% gray scheme.
  5. Really depends on what you shoot. But you should be able to get an idea by first metering yourself on partial, note the parameters, then switch to evaluative metering and see if the parameters are close to your own. If so, good; if not, you can expose two shots and see whether you or the evaluative is better. You have to experience for yourself when it's okay to use evaluative, and not to worry too much about film cost. For myself, if time allow, I like to meter myself. I don't shoot enough flash chrome to comment on that.
  6. I only use slide film these days, and my camera is the EOS 1V. I find evaluative metering to be correct for around 95% of the time. I shoot landscapes and travel.

    Sometimes, just for the sake of it, and to confirm, I compare readings from evaluative to spot, and they agree.
  7. I've owned and used an Elan2, Elan7, EOS3 and several others. Eval metering was excellent on all of them. 95% of my chromes were properly exposed.

    Eval metering pretty much sucks on the DSLR's I've used, including my 300D. If I shoot JPEG, half the shots are incorrectly exposed, specially when using flash. I only shoot raw, and it needs exposure compensation almost 100% of the time. No big. Raw works.
  8. >>The only way to learn exposure is trial and error. <<

    I should say not. Exposure is a science and its srtistic application can be learned systematically. The best tool, IMO, is the Zone System, which was specifically invented as a learning tool for new photographers. To see how it applies to digital imaging and relates to histograms, check this site:

    Then Google a bit for the Zone System and learn how to expose intelligently.
  9. When Evaluative Metering first came out with the Canon 650 (about the same time as the similar Nikon Matrix Metering), the photography magazines published detailed descriptions about how it worked. I still have my copy of the July 87 "Modern Photography" that had a stripdown test of the Canon 650.

    Evaluative Metering is basically a "knowledge base" system. When you meter a scene, the system measures the light levels of each of the sectors then converts the readings into a mathmatic expression using a proprietary algorithm. It uses the activated focusing point to determine where the subject is as part of the calculation (which is why EM locks exposure with focusing).

    It then searches its database of algorithms produced from tens of thousands of correctly exposed photographs (Nikon claimed 90,000 photographs for its matrix metering). The photographs include scenes with as many different lighting schemes as possible. When it finds a match, or the closest possible, it sets the exposure.

    It can normally identify things like intentional backlighting and correctly underexpose the subject by a stop...but this presumes the photographer intended a "conventional" backlit scene. If the photographer wanted the subject fully silhouetted, the camera would not know that.

    Also, like all meters, the EM system still presumes the subject should be medium gray. Therefore, if the subject is some other tonal level, the photographer should compensate for the subject difference--but not the lighting.

    For example, if the subject was a backlighted white stallion, the camera would correctly identify the lighting situation, but would assume the subject was an 18 percent gray that should be underexposed a stop to preserve the backlighted effect.

    If the horse would normally be Zone 7, and properly exposed for the backlighting at Zone 6, the camera would place it on Zone 4. Because the subject is two stops lighter than Zone 5, the photographer should dial in a +2 stop compensation for it.
  10. I mostly shot negatives, with an Elan 7E and an Elan II before that. Occasionally, I'd shoot slide film. I found evaluative metering on the 7 a bit better than on the II, but both were pretty good for most situations. The only situation which typically caused problems was a backlit subject, and neither one handled those very well. But, knowing that, I could fix it by adding some exposure compensation.
  11. Canon has not published the details of how evaluative metering works, so I consider it to be
    a black box. It can provide decent enough exposures much of the time, but it isn't the same
    as a human who knows what they want from a scene.
  12. Your answers are all helpful, and the over-all response is more of less what I expected. I am usually pretty good at judging the metering, but sometimes I tend to second guess myself. I've never really used the evaluative metering, and I just wanted some opinions. Thanks everyone!
  13. On the Elan 7, the first press of the exposure lock button locks the exposure. The second press (and all subsequent presses) re-meters the scene with partial center metering.
    Try this:
    1. Set the camera in evaluative metering, aperture priority.
    2. Compose on a subject that is very bright, but have something dark that fills the partial metering circle.
    3. Half-press the shutter to meter the scene.
    4. Press expsoure lock once, shutter speed should remain the same.
    5. Press exposure lock again. On the second press the shutter speed should change.
    • Take a couple of readings and you can decide what exposure you would choose and compare to the intial evaluative metering reading.
  14. That isn't the way my camera meters, or as described in the manual. AE lock is supposed to operate as follows:

    In Evaluative metering exposure will be biassed to a manually selected focus point (or one chosen by eye control), or to the point that achieves focus with all points active. The centre point is used with evaluative metering if you use MF. In Partial or centre weighted metering the exposure is taken round the centre point at all times except if CF8 is set to 1 with partial metering, in which case the selected (manually or by eye control) point is used.

    Simple tests show that the bias can be quite strong in evaluative, but not quite as strong as using partial metering. I used a wide angle lens with just a single lamp in my living room toward the right of the frame switched on. Using AF on the lamp in evaluative gave 1/2 stop more exposure than partially metering from the lamp. Focussing elsewhere in the scene results in much longer exposures - up to 2 stops more. This is with the same framing (using 17mm focal length).

Share This Page