What can account for this overexposure?

Discussion in 'Beginner Questions' started by adan|1, Jul 26, 2009.

  1. I was shooting with a Canon Rebel DSLR and the pictures came out completely overexposed. I tried both automatic and manual settings and still encountered the same problem. Checked the ISO setting, cleaned the lense and no improvement. I'd appreciate any advice as to what can be causing this so that I can try to fix it before taking it to the shop. Thanks.
    Here is an example:
     
  2. SCL

    SCL

    What type of metering were you using and what was the (or were the) points being metered?. Manual settings should have cleared it up unless you only set focus to manual, not exposure. Cleaning the lens has nothing to do with the exposure unless it is filthy so badly that you diminish the light entering the system by 50 % (1 f stop).
     
  3. The EXIF data shows ISO 100, 1/160th @ f/9. That would be about right for bright hazy conditions. If the sky was clear the EV would be closer to 15 or 16, in which case you'd be overexposed by one or two stops.
    It's difficult to tell from your photo but judging from the shadows, reflections in the water and what appears to be fog overhead, I'd guess EV 15 would have been about right. But it appears to be a fairly difficult scene to meter.
    Each camera system implements automatic metering a little differently. With my Nikon D2H in matrix metering mode and a wide angle lens I would have anticipated a slight underexposure. Based on experience I would have metered the foreground separately from the sky and water, then estimated how much detail I'd want to preserve in the washed out sky. So I'd have added +1/3 or +2/3 EV.
    The only way to determine how to make these adjustments for your own camera is to set up some testing scenarios and see how the camera responds. Or, in tricky situations, bracket your shots and evaluate the EXIF data later while comparing the results. If the results are consistently satisfactory at the recommended exposure, then use that. If the results consistently favor plus or minus exposure compensation for a similar situation, use that.
    A good baseline starting point is green grass. In any daylight situation - sunny or overcast - green grass tends to be comparable to an 18% gray card, and the meter reading should be very close to what you'd get with an incident meter reading. Try that. JPEGs straight from the camera should be very close to properly exposed. Lock in the exposure, recompose and take other photos in the same location without changing the exposure. Just mundane stuff will do - backyard, houses, whatever. As long as the lighting is the same and the exposure settings don't change the results should be acceptable in a JPEG straight from the camera. If the exposures are way off then, yes, it's possible there is a problem with the camera. But don't jump to that conclusion before running a few tests.
     
  4. William Michael

    William Michael Moderator Staff Member

    "a Canon Rebel DSLR and the pictures came out completely overexposed. I tried both automatic and manual settings and still encountered the same problem"
    If the camera is not faulty, then irrespective of the MODE used to take the image which is set by the Mode Dial – i.e. Green Rectangle Full Auto or “M” Manual: In this scene it is most likely that the METERING MODE was set:
    1. EVALUATIVE (Selected Automatically in FULL AUTO and all BASIC ZONE MODE and cannot be changed) or
    2. CENTRE WEIGHTED AVERAGE and was also no correct exposure compensation applied for the scene for CWA Metering.
    For this scene with your camera, I would selected PARTIAL METERING and taken the exposure reading with the centre of the frame at the two points shown by the PURPLE SQUARES and then recomposed the image.
    My guess is that there would be just a little over 1 stop difference if you took two readings, with the viewfinder centred at the two purple squares – the left hand (grass area) being the brighter: I would have taken about 1/3 stop under that brighter reading, for a one shot exposure.
    A good (technically better) method would be to bracket, in third stops, through those two meter readings.

    It would be good idea to research and test the three metering selection you have available to you because, as mentioned that scene seems difficult to measure: IMO the camera was confused by a combination of:
    1. the haze / fog.
    2. the parched grass.
    3. the direction of the light.

    Assuming the image is a full frame crop, your view through the viewfinder would have been as below.
    WW
    00U3BV-158725584.JPG
     
  5. Did you have any filters on the lens? At what elevation were you approximatly?
    Ed
     
  6. Have you checked to see if you've inadvertantly set exposure compensation to something other than '0', especially a positive value ?
     
  7. I AM WITH DB COOPER- CHECK THE EXPOSURE COMPENSATION. the ec see if it is set to anything other than the middle or zero.
     
  8. Metering from the camera and ec was set at 0. No filter used and elevation was 85m or 280 ft. The area is called Crystal Springs and here is a link with more info:

    http://en.wikipedia.org/wiki/Crystal_Springs_Reservoir
    I'm going back to retest later this week. Thanks guys.
     
  9. William Michael

    William Michael Moderator Staff Member

    D.B. Gary,
    When this camera is in Automatic (Green Rectangle) Exposure Correction is not functional.
    Nor is E.C. functional any of the BASIC ZONES.
    If the camera is set to "M" (Manual Mode) the E.C. is not functional.
    If the original question means that the pictures were taken in any of the modes mentioned above, it is not possible that E.C is the cause of this problem.
    Adan,
    Whilst you have confirmed where the E.C. was set, as I mentioned earlier, it would be valuable for you to check what Metering Mode your camera is set to, when you are in "M" Manual Mode.
    Note: as I also mentioned earlier, if you are in Automatic Mode (Green Rectangle) your camera automatically selects "Evaluative Metering" - note it also selects Evaluative Metering for ALL of the BASIC ZONE Modes, as I am still of the opinion, the most likely cause of the problem is the selection of the Metering Mode - unless the camera is faulty.
    Truly, on most Landscape shots I can easily get Canon's Evaluative Metering to go up to 1 stop over (compared to using a selective Centre Weighted Average on a green grassed area).
    Just for fun this morning, we had a little mist over the Bay - Evaluative went 1 and 1/3 overexposure compared to a Grey Card meter reading. The sunrise was side light into the misty bay with tree surrounds - and that was with my 5D - the 400D and 450D meters are a little more suspect in this regard, IMO.
    WW
     
  10. Evaluative went 1 and 1/3 overexposure compared to a Grey Card meter reading.​
    That's interesting info. I've noticed my D2H tends toward underexposure when the sky is misty white. It's one of those situations that can fool matrix metering.
    As handy as complex metering modes, auto-exposure, etc., can be, sometimes it's best to do it the old fashioned way.
     
  11. William Michael

    William Michael Moderator Staff Member

    Lex, for detail / clarity: I understand Nikon's Matrix Metering and I also would expect the scene in question to be underexposed, if anything with a Nikon. So too, underexposure is the assumed case when using Evaluative Metering with a Canon 400D / 450D . . . but IME, Evaluative Metering can give an overexposure, also, even if on the face of it seems odd, in this particular case

    It seems to me, sometimes the camera believes there is a backlit central subject, when there is not - but that is only my guess - as I do not use Evaluative Metering very much at all as I find it the most difficult to predict in odd lighting circumstances or scenes.

    The Metering Mode might or might not be the reason for this over exposure in this particular scene, but IMO the most logical route of action is to rule it in, or out, and if necessary proceed to diagnose further.

    WW
     
  12. William,
    "If the original question means that the pictures were taken in any of the modes mentioned above, it is not possible that E.C is the cause of this problem."​
    What if it doesn't?
    "...the most logical route of action is to rule it in, or out, and if necessary proceed to diagnose further."​

    Exactly.
     
  13. William Michael

    William Michael Moderator Staff Member

    D.B.
    Feather ruffling is not my style. The written word is difficult to convey nuance, and writing styles differ.
    I too was always fully aware that the original question might lack precise technical information such that the word "Automatic" and "Manual" may NOT refer to the "Green Rectangle" and "M" mode on the MAIN DIAL. I was also aware that others might be thinking this way too.
    My use of technical language (as used in the Canon Users' Manual) was to refine / define the problem. I usually write and read literally and technically – but am aware enough to perceive that my comments might have been an irritant to you - that was not intended.
    My reference to your post was only making statements of fact which would be useful to solving the original question, to provide information about how the Rebel Camera functions, irrespective of whatever SHOOTING MODES were actually selected on the main dial. If you misinterpreted my intent, then please Chilax: I interpreted from the beginning we were both just attempting to assist, not compete.
    However, in that spirit of working together, and also furthering knowledge generally, for your information and consideration, below is an extract of the FULL EXIF DATA of the image originally posted.
    Please note: I interrogated these FULL EXIF data, before posting my first response, so therefore was fully aware of this information when posted my second and I directed some comments to yourself and your post – but I thought it would be impolite of me to come straight out and state that checking Exposure Compensation would be “simply irrelevant” for the image under question.
    -- EXIF-Data ---------------------------
    Exif : 9=Directory 9 entries
    Make : Canon
    Model : Canon EOS DIGITAL REBEL
    Exposure Time : 1/160=1/160 sec [1/160.00]
    F Number : 90/10=F9.0
    ISO Speed : 100
    Metering Mode : 5=MultiSegment/Pattern [Canon = Evaluative Metering]
    Focal Length : 36/1=36.00 mm
    Exposure Mode : 0=Auto
    White Balance : 0=Auto
    Scene Capture Type : 0=Normal
    Focus Mode : 0=One-Shot
    Canon-0x0001-0x0008 : 0
    Easy Shoot : 0=Full Auto [Canon = Full Auto - i.e. Green Rectangle]
    Metering Mode : 0=Default
    AF Point : 0=n/a
    Exposure Mode : 0=Easy Shooting
    -- EXIF Summary -(long) ----------------
    I therefore originally concluded that the image posted was captured using Full Auto (Green Rectangle) – (unless the EXIF had been tampered with), however, I thought it would serve very little purpose to just blast that information against your EC theory.
    I trust that explains my previous posts and the rationale and scrutiny behind what I wrote and why I wrote it.
    Regards,
    WW
     

Share This Page