Nadine/Rob B, ken + others-more sample exposures

Discussion in 'Mirrorless Digital Cameras' started by dennis osipiak, May 3, 2005.

  1. Two more photos to day, EXACT same light(overcast,no sun) and background identical, same clothing color. Photos taken 17 seconds apart,550EX both flash comp of 1 1/3. Shutter priority (TV). No camera comp (0) ISO 400. Focus points nearly identical. Obvious difference in quality. Canon EOS utility tells me the camera chose F22, 1/100th for the dark one, F10, 1/100th for the good one. Now thats an incredible difference in F stop. If I understand the math that is 4X diffrence in light. (double from F11 to F16, double again from F16 to F22).Can't understand the camera's divergent Fstop choices. Given all conditions are the same as stated above, do I accept defeat and pack this new camera off to Canon? Sorry for the repetive nature of this question but I want to exhaust all possibilites before taking action. As some have suggested I will try full manual, but comment on what you see so far. 2 photos below Dennis
  2. F10 ..100th photo TV
  3. 1 error..its a 580 EX flash.
  4. OK--you're right--that's a big difference. The only thing I can think of is that even though the lighting, background and clothing color are the same, the pose isn't. I'd be much more suspicious that something was wrong with the gear if it were the exact same kid in the exact same pose, etc., and the exposure varied by 2+ stops. Since the camera's/flash's metering is determined by an averaging of segments, it could be that it "thought" there was a lot more "important" light areas in the top photo (hence stopping down)--you did have the ambient metering set to matrix and not partial, right? If it was partial, there is a lot more white in the area that the center spot would be. Try using manual mode on the camera and E-TTl on the flash and see if you have big swings in flash exposure. Set something up at home so you don't have to experiment on a paying job.
  5. The pose is a good give-away. If you notice there's more white (especially center frame) on the first picture. That told the camera to measure the shirt as 18% gray. The second pose has more of the background in the center thereby overexposing the girl.
  6. I overlayed the two images in Photoshop. I also believe it is the pose.

    The only thing I can suggest is to re-read the Canon flash tutorial on There really is (that I've found at least) no better guide to flash photography with the Canon system than what is on that website.
  7. Dennis--what was your ambient metering mode--matrix, partial or center weighted? And what flash metering mode--matrix or averaging?
  8. If these photos were taken 17 seconds apart, then she must have
    been pretty darn busy in those 17 seconds. She painted her nails,
    took out an ear stud, and changed the garment she is wearing
    underneath her numbered uniform shirt -- all in 17 seconds!
  9. Thanks to all..
    I went partial metering to be more precise and to avoid the dark hemlock throwing off the average. I'll try going back to full evaluative metering to see how that works.
    Posing is slightly different, arm up with white shirts vs arm down, less white...might be some of the metering problem...but if this was an action shot I'm sure there wouldn't be time to consider that.(it wasn't obviously)
    Sam..I'm an easy going guy with a sense of humor..and I'll take the comment in jest as I'm sure it was offered. They may be two girls, with the same team, same uniform but they aren't the same kid. While I don't grind out shoot an entire league 17 seconds apart/kid I do sometimes get in a rythm for a few kids.
    More to do tommorrow,Friday, and weather permitting Saturday. I've gotten rained out out of my TBALL leagues two weeks in a row so far in Connecticut. Pray for sunshine so I can wrap this up.
  10. Thanks, Dennis. Somehow I had genuinely missed that it was a
    different player.

    I do have a technical theory to offer. I could be remembering
    this completely wrong, but I seem to recall that Canon's flash
    algorithm has an feature in which, if the ambient light is above
    a certain limit, it decides you must want fill-flash for
    catchlights, so it reduces the flash exposure. Could it be
    that the ambient light was very close to that cutoff and
    one exposure kicked in that particular feature and the other
    did not?

    My other thought was that the exposure algorithm is weighted
    for the acctive focus point. Obviously neither exposure is
    focused on the trees, but possibly one is focused on the face
    and the other on the uniform, which is white?
  11. << but if this was an action shot I'm sure there wouldn't be time to consider that. >>

    Which is why shooting in Manual mode is so powerful. Just pick your aperture for your DOF, pick your shutter speed for how much ambient light you want to see, be conscious of your focus point, and fire away with the flash on.
  12. Couple of things. First, Dennis if you had the ambient exposure meter set for partial, then that explains what happened with these two photos. Raised white arm fully in the center spot vs. shadowed white clothing, skin and part of background/bat in the other. If you continue to use the automated modes, I would definitely not use partial, but matrix. Even then, the camera meter can be fooled. As Rob said, in controllable situations, manual mode is best, possibly combined with an incident handheld meter or gray card with camera meter.

    For action shots, use AV or TV and matrix metering and get to know how it will react to different situations, but unless the light is changing rapidly, manual still works best. If shooting the game and it was basically overcast all the time, one incident reading and you would have been set. Also, I think it's best to shoot RAW in those situations so you can save an exposure which is not so great. It's amazing what you can do with a RAW file that you can't do with a jpeg.

    Sam--automatic fill flash starts about EV10 (?) I think. I don't think that will account for the two stop difference here, especially since auto fill flash affects the flash exposure, not the ambient.

    The 20D is not so biased to focusing points--it supposedly averages segment data, both for the ambient and for the flash, but it will still underexpose for lighter scenes and overexpose for darker scenes. For fill flash, I've found that I definitely need to pull the flash back (minus compensation) or it will overexpose.

    For the shot of the kid you show, I would have incident metered, set my camera to the reading (having previously compared how the camera meter reads in comparison to my meter), and set the flash 1.5 to 2 stops below whatever "home base" flash setting you normally use (maybe 1.5 or 1 since the kids are wearing white shirts). If using an automated mode, I would have used AV to try to throw the greenery out of focus in the background, watching the shutter speed so that it didn't fall slower than handholdable.

Share This Page