Posted by Steve Balke on July 13, 2012 at 11:26 AM
This morning, after struggling to obtain the minimum iso at correct exposure in accord with our recent discussion on minimizing noise....and watching birds fly away, I developed an approach that seems to work quite well. I'm posting it here to invite comments. It involves auto iso.
Here is the procedure that I used this morning:
i. Set manual mode on my 1D Mk IV
ii. Set ss to overcome my motion and the movement of the bird.
iii. Set aperture to obtain the depth of field that I want
iv. Set iso on auto.
v. Set exposure bracketing on +/- one stop.
vi. When shooting, if the iso selected is higher than I like, I increase aperture.
Note the advantages of the above approach:
i. practical for birds that don't like to sit around;
ii. allows one image to be obtained as exposed to the right in addition to one "correctly" exposed and one underexposed.
iii. these three images can be used for hdr or at least for combining images to recapture blown highlights. Each image is at a different ss, not different aperture.
iv. the photographer is still very much in control (e.g. ss, aperture and definition of the bracket for exposure bracketing are all set by us). Only the most awkward variable, iso, is set automatically. Reference to the in-camera meter as I take the shot can verify that the camera is setting iso to obtain "correct" exposure.
v. I like to use spot metering to focus attention on the bird.
There is an excellent article on the use of auto iso at:
http://www.dpreview.com/articles/0206161199/an-in-depth-discussion-of-m-auto-iso-for-canon-slrs
The main issue I see is the bracketing Steve... It has pluses and minuses... but I see more minuses. A few notes :
-- Expose to the Right (ETTR) doesn't typically refer to exposing some arbitrary value over the metered exposure. Instead, it refers to exposing just far enough to the right to avoid clipping highlights to get the maximum signal to noise ratio (SNR). Setting an arbitrary 1 stop over-exposure is likely to cause blown-out whites or fail to maximize the SNR.
-- For birds that don't like to sit around, bracketing isn't very practical in that merging mutliple images is isn't likely to be easy, and will be impossible if panning. Plus, if you want to capture behavior and like a high frame rate you prevent capturing that sequence that might nail the pose you want.
-- The underexposed image in the bracket isn't likely to be all that useful, unless you have blown highlights at the "correct" exposure that you want.
Aside from that, I think auto-ISO is great in situation where I have a something like you stated - I want a high enough shutter speed for movement, and to pick DOF by aperture. That's really the ideal for auto ISO, otherwise I do tend to manually keep the ISO at the lowest setting (or limit the auto to the lowest couple of settings).
And for your last numbered comment ("v"), I'm not sure what you mean. All spot metering does is use a narrow range of brightness values at the center of the frame for metering. If you've got a bright bird, it will create a darker exposure than other modes, and a dark bird will create a brighter exposure. The same thing could be accomplished in other ways, but spot metering is useful. It doesn't do anything to emphasize a particular subject necessarily.
Thank you Jack and David. David, regarding your comments:
i. I'm not sure why you state that bracketing is not practical as a means of attaining ETTR. You just need to decide how much above the "suggested exposure" of the camera you wish to place your bracket. You can use trial and error looking at the histogram too if you wish. How is that different from your normal procedure for finding ETTR conditions? I bet that you have a much better chance of obtaining an ETTR shot immediately without trial and error using the procedure.
ii. Bracketing would likely not help with the bird. However, it can help with the perch and possibly the bg. That relates to the spot metering comment too. If I base my exposure on the bird only and manage to blow the perch, for example, I can go to the underexposed image and use the perch from that. I use spot metering because I want to be sure that the bird is properly exposed.
Posted by Steve Balke on July 13, 2012 at 10:00 PM
By the way, when I say a bird that doesn't like to sit around I mean that the bird perches, but not for long. I'm not visualizing a bird in flight shot for this example. And another thing....bracketing can help with the bird. If bracketing is used to obtain ETTR then it is the bird image that is so exposed.
Posted by David Bostedo on July 13, 2012 at 10:21 PM
I'm not sure why you state that bracketing is not practical as a means of attaining ETTR. You just need to decide how much above the "suggested exposure" of the camera you wish to place your bracket.
I was more referring to your "+/- one stop" phrase. The best ETTR is not likely to be 1-stop in many cases. As you said, it takes some adjustment.
Bracketing would likely not help with the bird. However, it can help with the perch and possibly the bg.
Absolutely, but if you can leave the camera completely still like that, I call that a bird that DOES like to sit around I was, as you mentioned, thinking about birds that are actually in motion.
Personally, I'd prefer to meter the scene, spot or matrix depending on what I thought was appropriate, then tweak the exposure either manually or with exposure compensation to get to the actual exposure I want for all of the exposures. In a lot of cases with changing light, subjects, or moving subjects I don't bother trying to get too caught up in perfect ETTR, as much as trying to make sure I don't underexpose. But that's me, and if you find that the bracketing helps you in your workflow, great.
Posted by Keith Bauer on July 13, 2012 at 11:32 PM
I use spot metering because I want to be sure that the bird is properly
Steve: That statement is great if the bird you are photographing reflects 18% of the light. Reflective meters are designed and calibrated to assume that 18% of the light is reflected.
If you are spot metering off of a white bird, your meter will underexpose. If the white bird is in the sun, it will under expose by even more, something like 2 stops. On the other hand, if you are photographing a black bird, your meter will again be fooled, and this time it will over expose the bird. If it is something like a jet black Raven, it will be over exposed by 2 stops or so.
Assuming that spot metering will generate the right exposure for the subject is simply wrong. It all depends on the subject. Try it. Spot meter off of a white piece of paper in the sun, and see what the exposure looks like. Then spot meter off of a pure black subject and see what the exposure looks like. Both will tend to be gray, rather than the proper color.
Posted by Steve Balke on July 14, 2012 at 12:24 AM
Keith, I'm well aware of the "grey snow" problem associated with the light metering. All of the light metering modes (including evaluative metering) have the same deficiency. Which is better depends on whether our bird is closer to 18% grey tone or whether it would be better to include bird and surroundings. I'd certainly be willing to experiment if I was running into trouble. I've been using spot metering for many bird images and have had quite good luck with it. At any rate, the important thing to note in the context of the posted topic is that the auto-iso procedure detailed above is flexible enough to allow us to select our metering mode. That is not to say that this method is perfect: judging from the cited article, using auto iso with flash is not a good idea. Also, if light is very bright it may drive iso to 100 and still provide over-exposure. So, we need to be careful if we see the method providing either very low or what we consider too high iso values. I sincerely hope that you and others will try this method, Keith. I was really impressed this morning with how easy it was to use. I'd be very interested in knowing if others discover the same thing.
Posted by Brian Gatlin on July 14, 2012 at 3:06 PM
This doesn't directly relate to using auto-ISO (which I don't like), but I absolutely hate the idea of bracketing +/-1 stops for birds. This approach more or less guarantees that 67% of your shots will be worse than they could have been if the exposure was simply set correctly for all three shots. When subtle movements of the head can create all the difference in which shot in a burst is best, or one shot in a burst captures the perfect moment of action, it would surely be a shame for that shot to be a full stop underexposed (or over).
Very interesting response, Brian. I find your reaction to auto-iso curious. The first time I tried it was with aperture and with ss priority. In those modes I found it confusing, too computerized and unpredictable. However, with this manual method it's doing what I would want to do, only much faster.
Regarding bracketing: First of all, please note that as discussed above, you don't need to go as far as 1 stop in each direction. I really like the idea of exposing to the right. However, often when I try it I blow the whites and the shot is ruined. I'm sure that many of you are better than I am at it. However, there is always an element of risk and often there is no time for trial and error. So, on the one hand you risk having any acceptable shot at all (without bracketing) while with the other you have a backup (a properly exposed shot that is not ettr). At the 10 frames/s of the 1D Mk IV often I will have two frames with that perfect moment of action on them too... I guess this boils down to personal preference and the great thing about photography is that we can each do it our own way. I'm not seeking converts but I do love to try to explore and assess new ideas. A useful auto-iso method is particularly interesting because it is a complete break with the film technology. Thank you for your input, Brian.
Posted by Brian Gatlin on July 14, 2012 at 5:31 PM
regarding the auto-ISO subject:
I prefer to shoot AV, choosing my aperture, choosing my ISO, and letting the camera work out the SS (while adjusting the ISO upward if I'm not getting enough SS).
This preference may be based on the fact that I shoot a 7D, and I do not want the ISO jumping above 640 or perhaps 800. If forced to push the limits on ISO or SS, I'll opt for SS, knowing that I can often get good results at surprisingly slow SSs but that if I go too high on the ISO I virtually never like the result. If the 7D offered the opportunity to set a range of acceptable ISOs for AutoISO to choose from, I'd consider using it. Likewise, if, like you, I shot a 1DIV which has impressive high ISO capability, I'd also be willing to try it out. But with my current setup it doesn't work for me.
Understandable, Brian. In my case I am accustomed to leaving iso set at 1250, taking advantage of the low noise capabilities of the 1DIV. I use ss priority because the maximum aperture that I can set is f/8 with this lens system. If I use aperture priority I find that very often the camera dives into a low ss and I end up with a blurry hand held shot. With ss priority the problem has been that quite often low light meant that I ended up underexposing even at an iso of 1250. I learned lately that it would generally be better to raise the iso to permit correct in-camera exposure rather than raise exposure using software later. Also, at the other extreme, for bright light, I was needlessly using a high iso and accepting a very small aperture by fixing iso at 1250. Based upon my very brief experience with the auto iso method, in using the auto-iso method I don't try to constrain the iso automatically. It is very easy to adjust aperture to accommodate what I consider too high an iso or too low an iso. [Please see above for what is meant by "too low" an iso.]
I think either method probably works well, since you have one automatic exposure variable, and one manually changed exposure variable, and one fixed exposure variable... I.e. :
Method 1 (Brian) : Aperture is fixed, Shutter is automatic, ISO is manually adjusted
Method 2 (Steve) : Shutter is fixed, ISO is automatic, Aperture is manually adjusted
The automatic variable will provide a limited range of exposures that the camera can play with. And you have a variable that you can freely change manually to slide the range. I tend to shoot like Method 1, because I generally care least about shutter speed, and want to force lower ISOs - it's a matter of control. Shooting like Method 2 might lead to the same exact result, especially where you need the shutter speed and are perfectly willing to raise ISO.
However, with Method 2, you are adjusting the exposure range via aperture, which in addition to changing exposure, changes DOF. For Method 1, you are adjusting the exposure range via ISO, which leaves DOF the same.
I suppose preferences in these cases all depend on how your brain works and the results you get.
Thank you, David. I think you are right. However, there are more angles to this too: Yesterday evening I was further evaluating the auto iso method and as the sun went down iso headed up to the maximum value and, of course, the constraint on aperture meant that eventually it could not be increased further and iso stayed at max. It then forced me to lower ss....and look around for some kind of support for the camera so auto-iso would come down from the maximum value. Myself, I like that progression from controlling aperture to controlling ss. (I have a better feel for those two variables than I do for iso.) Another aspect: Using the auto-iso method, I could just as easily have decided to adjust ss first followed by aperture. The ease with which either aperture or ss can be adjusted is another aspect that I like about the method.
Discussion
13 repliesHere is the procedure that I used this morning:
i. Set manual mode on my 1D Mk IV
ii. Set ss to overcome my motion and the movement of the bird.
iii. Set aperture to obtain the depth of field that I want
iv. Set iso on auto.
v. Set exposure bracketing on +/- one stop.
vi. When shooting, if the iso selected is higher than I like, I increase aperture.
Note the advantages of the above approach:
i. practical for birds that don't like to sit around;
ii. allows one image to be obtained as exposed to the right in addition to one "correctly" exposed and one underexposed.
iii. these three images can be used for hdr or at least for combining images to recapture blown highlights. Each image is at a different ss, not different aperture.
iv. the photographer is still very much in control (e.g. ss, aperture and definition of the bracket for exposure bracketing are all set by us). Only the most awkward variable, iso, is set automatically. Reference to the in-camera meter as I take the shot can verify that the camera is setting iso to obtain "correct" exposure.
v. I like to use spot metering to focus attention on the bird.
There is an excellent article on the use of auto iso at:
http://www.dpreview.com/articles/0206161199/an-in-depth-discussion-of-m-auto-iso-for-canon-slrs
All comments welcome.
-- Expose to the Right (ETTR) doesn't typically refer to exposing some arbitrary value over the metered exposure. Instead, it refers to exposing just far enough to the right to avoid clipping highlights to get the maximum signal to noise ratio (SNR). Setting an arbitrary 1 stop over-exposure is likely to cause blown-out whites or fail to maximize the SNR.
-- For birds that don't like to sit around, bracketing isn't very practical in that merging mutliple images is isn't likely to be easy, and will be impossible if panning. Plus, if you want to capture behavior and like a high frame rate you prevent capturing that sequence that might nail the pose you want.
-- The underexposed image in the bracket isn't likely to be all that useful, unless you have blown highlights at the "correct" exposure that you want.
Aside from that, I think auto-ISO is great in situation where I have a something like you stated - I want a high enough shutter speed for movement, and to pick DOF by aperture. That's really the ideal for auto ISO, otherwise I do tend to manually keep the ISO at the lowest setting (or limit the auto to the lowest couple of settings).
And for your last numbered comment ("v"), I'm not sure what you mean. All spot metering does is use a narrow range of brightness values at the center of the frame for metering. If you've got a bright bird, it will create a darker exposure than other modes, and a dark bird will create a brighter exposure. The same thing could be accomplished in other ways, but spot metering is useful. It doesn't do anything to emphasize a particular subject necessarily.
i. I'm not sure why you state that bracketing is not practical as a means of attaining ETTR. You just need to decide how much above the "suggested exposure" of the camera you wish to place your bracket. You can use trial and error looking at the histogram too if you wish. How is that different from your normal procedure for finding ETTR conditions? I bet that you have a much better chance of obtaining an ETTR shot immediately without trial and error using the procedure.
ii. Bracketing would likely not help with the bird. However, it can help with the perch and possibly the bg. That relates to the spot metering comment too. If I base my exposure on the bird only and manage to blow the perch, for example, I can go to the underexposed image and use the perch from that. I use spot metering because I want to be sure that the bird is properly exposed.
I was more referring to your "+/- one stop" phrase. The best ETTR is not likely to be 1-stop in many cases. As you said, it takes some adjustment.
Absolutely, but if you can leave the camera completely still like that, I call that a bird that DOES like to sit around
Personally, I'd prefer to meter the scene, spot or matrix depending on what I thought was appropriate, then tweak the exposure either manually or with exposure compensation to get to the actual exposure I want for all of the exposures. In a lot of cases with changing light, subjects, or moving subjects I don't bother trying to get too caught up in perfect ETTR, as much as trying to make sure I don't underexpose. But that's me, and if you find that the bracketing helps you in your workflow, great.
Steve: That statement is great if the bird you are photographing reflects 18% of the light. Reflective meters are designed and calibrated to assume that 18% of the light is reflected.
If you are spot metering off of a white bird, your meter will underexpose. If the white bird is in the sun, it will under expose by even more, something like 2 stops. On the other hand, if you are photographing a black bird, your meter will again be fooled, and this time it will over expose the bird. If it is something like a jet black Raven, it will be over exposed by 2 stops or so.
Assuming that spot metering will generate the right exposure for the subject is simply wrong. It all depends on the subject. Try it. Spot meter off of a white piece of paper in the sun, and see what the exposure looks like. Then spot meter off of a pure black subject and see what the exposure looks like. Both will tend to be gray, rather than the proper color.
Regarding bracketing: First of all, please note that as discussed above, you don't need to go as far as 1 stop in each direction. I really like the idea of exposing to the right. However, often when I try it I blow the whites and the shot is ruined. I'm sure that many of you are better than I am at it. However, there is always an element of risk and often there is no time for trial and error. So, on the one hand you risk having any acceptable shot at all (without bracketing) while with the other you have a backup (a properly exposed shot that is not ettr). At the 10 frames/s of the 1D Mk IV often I will have two frames with that perfect moment of action on them too... I guess this boils down to personal preference and the great thing about photography is that we can each do it our own way. I'm not seeking converts but I do love to try to explore and assess new ideas. A useful auto-iso method is particularly interesting because it is a complete break with the film technology. Thank you for your input, Brian.
I prefer to shoot AV, choosing my aperture, choosing my ISO, and letting the camera work out the SS (while adjusting the ISO upward if I'm not getting enough SS).
This preference may be based on the fact that I shoot a 7D, and I do not want the ISO jumping above 640 or perhaps 800. If forced to push the limits on ISO or SS, I'll opt for SS, knowing that I can often get good results at surprisingly slow SSs but that if I go too high on the ISO I virtually never like the result. If the 7D offered the opportunity to set a range of acceptable ISOs for AutoISO to choose from, I'd consider using it. Likewise, if, like you, I shot a 1DIV which has impressive high ISO capability, I'd also be willing to try it out. But with my current setup it doesn't work for me.
Method 1 (Brian) : Aperture is fixed, Shutter is automatic, ISO is manually adjusted
Method 2 (Steve) : Shutter is fixed, ISO is automatic, Aperture is manually adjusted
The automatic variable will provide a limited range of exposures that the camera can play with. And you have a variable that you can freely change manually to slide the range. I tend to shoot like Method 1, because I generally care least about shutter speed, and want to force lower ISOs - it's a matter of control. Shooting like Method 2 might lead to the same exact result, especially where you need the shutter speed and are perfectly willing to raise ISO.
However, with Method 2, you are adjusting the exposure range via aperture, which in addition to changing exposure, changes DOF. For Method 1, you are adjusting the exposure range via ISO, which leaves DOF the same.
I suppose preferences in these cases all depend on how your brain works and the results you get.