View Full Version : 7D always 1 to 1.5 overexposed.
I'm still getting grips with the new body. I went to a wedding a few weeks back and a few days out with the kids. In all cases, I had to reduce the exposure of all the RAWs by at least 1 in dpp. Usually 1.33.
I can't find any settings that are causing this. Any ideas?
on your camera screen too bright or your monitor?? calibration on your monitor ok? also check your histogram on your cam as you shoot, that should guide you as well
You haven't knocked the exposure compensation have you? Or left it set from a previous shoot?
How are you shooting? I usually find I underexpose by about 1/3 or 2/3 stop in nearly all situations from what the camera's meter says is 'right', even on Matrix...
How are you shooting? I usually find I underexpose by about 1/3 or 2/3 stop in nearly all situations from what the camera's meter says is 'right', even on Matrix...
I do exactly the same, shooting at 0 EV always seems too light to me.
Tight Fart
08-05-2010, 19:04
Big white dress confusing the exposure?
phototuition
08-05-2010, 19:55
Big white chief says big white dress cause underexposure.
trencheel303
08-05-2010, 20:19
I do exactly the same, shooting at 0 EV always seems too light to me.
Yep, me too. regardless of whether the histogram is 'correct', I know for a fact that a photograph on 0 EV makes the situation look brighter than it really was.
But isn't it better to 'shoot to the right' and reduce it in post processing of the RAW file?
Freester
08-05-2010, 20:43
Metering mode?
How are you shooting Av, Tv? Manual and just getting the meter to the middle?
on your camera screen too bright or your monitor?? calibration on your monitor ok? also check your histogram on your cam as you shoot, that should guide you as well
On both the camera and Computer. Histograms are biased to the right. Pictures from the 400d AND 50D look fine.
You haven't knocked the exposure compensation have you? Or left it set from a previous shoot?
That was my first thought....but nope its blob on. I remember checking the 50D exposure on the LCD and on the main screen and seeing they were different. I checked the 7D and nope both are set to 0.
How are you shooting? I usually find I underexpose by about 1/3 or 2/3 stop in nearly all situations from what the camera's meter says is 'right', even on Matrix...
Usually AV, Evaluative Metering. I then trust the camera to get the exposure right.:bonk:
Big white dress confusing the exposure?
Big white chief says big white dress cause underexposure.
What about kite flying?, Pics of my boy in his pram?
Metering mode?
How are you shooting Av, Tv? Manual and just getting the meter to the middle?
The problems began in AV mode, Evaluative metering, Pic in AE program seam better. :bang:
Some examples:
http://farm5.static.flickr.com/4053/4589667331_061e335d21_o.jpg
http://farm5.static.flickr.com/4005/4589666651_2b63e82298_o.jpg
http://farm5.static.flickr.com/4030/4589665969_5e0a0f6821_o.jpg
http://farm5.static.flickr.com/4026/4590285662_2eacff8ab0_o.jpg
trencheel303
09-05-2010, 02:12
But isn't it better to 'shoot to the right' and reduce it in post processing of the RAW file?
Going through hundreds of RAW files isn't my idea of fun after a day out... I only ETTR if I am at high ISO.
...
Some examples:
...
Those exposures are fine, as expected shooting most of them in Auto as well. If you want more control over the final outcome, then start shooting more in Av, etc.
Those exposures are fine, as expected shooting most of them in Auto as well. If you want more control over the final outcome, then start shooting more in Av, etc.
I was with you all the way until you said "Av" at the end there. Did you mean Manual?
Autoexposure is not very clever, left to its own devices. It assumes the subject/scene is a middle tone unless you tell it different. In tricky lighting it can easily make assumptions that don't work well for the scene as a whole. Sometimes (most times) that requires evaluation of the subject/scene by the photographer and some input from him/her to tell the camera what it needs to know when coming up with its best guess at the exposure. If your subject/scene is darker than "middle grey" then you need to tell the camera. Ditto for subjects/scenes that are brighter than "middle grey".
In general terms, evaluative metering is a fickle beast and can catch out the unwary. While the metering does look at the whole scene, the exposure is influenced most strongly by the tones at the point where you focused, less so by the tones surrounding that focus point, and much less so elsewhere in the frame. In other words, if you focus on a dark subject, or dark area, then the camera will brighten up the exposure. If you focus on paler tones then the exposure may come out darker than you like.
As far as evaluative metering on the 7D is concerned, I have read that with its colour aware meter the 7D can expand the area which it considers to be the subject, by including areas of the scene of similar tone and distance to your subject (focused point) to have equal importance in influencing the exposure.
If you are going to do things like shoot at the shaded side of your subject while including brightly sunlit clouds in the scene it is more than likely you will need to step in and take some control. If you focus on a groom dressed in black, with a sunlit bride right beside him, you might well find the bride to be overexposed. The Exposure Compensation dial is there to be used. You can't simply point the camera at anything and everything and expect a perfect result without any input on your part. The EC dial is one tool you can/should use. The histogram and blinking highlight warnings after you've taken the shot are a couple of others.
If you have scenes of high dynamic range, or where nailing exposure accurately is tricky, then shooting raw and using good raw editing software will help you retrieve some image detail that might be lost forever when shooting JPEG. However, it is still important to try to get a perfect exposure in the first place, rather than always relying on raw as a safety net to bail you out.
There is a general rule (guideline) for digital photography, especially when shooting raw, which is to expose for the highlights and develop (edit) for the shadows. To do that well requires input from the photographer. If you just leave things to autoexposure, especially in evaluative metering mode, your results will tend to end up exposed for the subject (possibly wrongly) and bugger anything else in the scene.
My prefered approach, when I want to get my exposures spot on, is to use manual exposure and to spot meter off something that allows me to set my exposure very precisely. If I have time and need then I may well fire off a test shot and verify that my exposure setting is good before moving on. If I need to make an adjustment then I do. I'm not infallible either.
Here is one example of mine where I used manual exposure to lock in an exposure that was correct for the lighting conditions at the time. It has the histogram pegged very precisely at the right hand edge, with an exposure that is about as good as one could hope for. If I had tried to follow this bird in flight as it flew from a background of bright sky to one of shaded trees, using autoexposure, my exposures would have been all over the place, and many would have been disappointing to say the least. With manual exposure I was in charge and got the results I wanted from the first press of the shutter button.
http://lh4.ggpht.com/_4_R8TkwT74w/S-GCBeGlOHI/AAAAAAABSJI/9Pc6mqjJ01k/s800/20100505_153422_.JPG
If I had used autoexposure for this shot I couldn't even tell you what the result might have been as I have no idea how the camera would evaluate the scene. By the time I found out it might well be too late.
Here's another example. I'd say this would be a tricky scene for autoexposure to get just right. By setting my exposure manually I had no difficulty in getting things dialed in perfectly....
http://lh4.ggpht.com/_4_R8TkwT74w/S-ZMNM51HFI/AAAAAAABSRg/1tWLFzeAb_w/s800/20100509_064608_.JPG
malla1962
09-05-2010, 07:10
Those look ok to me ant to be honest I normally over expose all my shots by 1/3 on all my cameras.
Freester
09-05-2010, 07:10
Hi Brock,
Those images look fine - at least nothing wrong with the camera. Tim has given a very detailed explanation above. Without echoing what he says basically all of the examples you posted have a large dynamic range. Bright skies, but with blacks etc.
In these situations it can be difficult to dial in the correct exposure and the camera metering can get confused. At the end of the day the evaluative meter is just and average and so can get skewed quite easily.
You can either spot meter and shoot manual to get correct exposure on the subject. Or perhaps just take a shot, have a look at it and the histogram / blown highlights, then dial in some Exposure Compensation to compensate.
p.s. taking the wedding photo as an example, it is clear that the scene, especially the most important part(s) is brightly sunlit in early afternoon light. There is a handy rule of thumb for setting exposure when shooting in sunny conditions, called the "Sunny 16 Rule". In such conditions the rule says that with your aperture at f/16 your shutter speed should be the reciprocal of your ISO e.g. at 100 ISO you would shoot at 1/100. At 200 ISO you would shoot at 1/200 and so on. If I had been shooting this scene I would have set a manual exposure equivalent to a sunny 16 exposure, at least initially, although not at f/16, obviously. Something like f/5.6 at 1/800 would have been a better starting point. If I found that exposure to be slightly off then I would have adjusted slightly and then left it alone until the lighting conditions altered.
The exposure for the wedding shot is 1/640, f/4.5, 100 ISO. f/4.5 is 3.7 stops brighter than f/16 so we would expect a shutter speed 3.7 stops dimmer than 1/100 to compensate. Unfortunately, 1/640 is only 2.7 stops darker than 1/100, so overall, based on a sunny 16 standard exposure, the shot is too bright by about 1 stop. I'm not sure it's quite that bad, but clearly it is overexposed. Personally I would not blame the camera for that. It's a tricky scene to expose correctly, with lots of DR, and it needs the photographer to take control for best results. If this had been shot raw then there should have been quite a bit of opportunity to recover much of the highlight detail, although I think not so easily in DPP. Lightroom/ACR would almost certainly have been able to do a pretty good job of recovery.
Tdodd, that was quite a read and thanks for the time. The shots are shot in raw so I was able to corrent the 1stop error on all the shots from the day. Looking back at the set, there is a higher dr than i'd appreciated.
I'm still not confident to move into manual exposure yet. I've not had the 7d long so I'm still getting used to navigating through it. I still struggle with remembering iso, dof, af selection and metering modes in the time from seeing a shot to taking it before the moment has gone.
Your comment on evaluative metering are very surprising, and goes slightly against what I was taught by canon experience seminars. It does make sense. It sounds more like a slight centre weighted metering than full frame metering
out of interest, for that eagle shot, did you anticipate the bird flying there and set up prior, or are you saying you metered, manually set up and shot the image in a few moments.
jerry12953
09-05-2010, 09:19
Strangely, I find the same thing with most subject matter and my 5D Mk2.
1/3 to 2/3 stop over-exposure with evaluative metering...
As an ex-velvia user I'm used to exposing for the highlights and I usually do. The exposure compensation dial is my friend, and the develop sliders in Lightroom are then all my friends too!
But Tim, I thought the rule was to expose for the shadows and deal with the highlights in post processing, rather than the other way round?
Brocks, I'd say your pics nos 1 and 4 ARE over-exposed but it looks to me as if Lightroom could deal with them successfully. The other two..... well they look fine to me. But if you have Lightroom (or PS) you can adjust the blacks for more "saturation", and/or you could maybe turn the highlights down a little in no2. But that's really personal preference.
Brocks, for the "eagle" shot (I believe it's actually a kite) I had my exposure set up well in advance, and fired off many shots with identical exposures and all very good, if I may say so. I accomplished this by spot metering off the brightest part of the sky at +3 stops and locking the exposure there manually. That meant that regardless of whether the bird had a background of 100% sky (it was a grey day with fairly even cloud cover) or against a backdrop of trees, the exposure for the bird remained constant. As the sky was so flat it meant that the light did not change and so there was no need for my exposure to change either.
I had already positioned myself so that the brightest part of the sky was balancing the light in favour of the front of the bird, rather than contributing to a backlighting effect, leaving the bird in its own shadow. This gave me the best result possible under the circumstances. There is a thread over on the BirdForum where I have provided more examples and tried to explain in detail everything I do and why....
http://www.birdforum.net/showthread.php?t=165102
Coming back to evaluative metering, you can easily demonstrate how exposure alters depending on what kind of tones you are focusing on. Simply find a scene - I'm sure your back garden will do, or even indoors - or set one up where, without changing your composition, you can place different focus points over different areas of tone. Focus with a point over a bright tone and take a shot. Now, without moving the camera at all, pick a focus point that is placed over a dark tone. Focus once more and take the shot. Even though nothing in the scene has changed, or the lighting upon it, you will find you get completely different exposure results. It's actually like centre weighted metering, but instead of remaining weighted to the centre it becomes weighted to your selected focus point.
Like I say, this can catch people out, and some old hands prefer to avoid using evaluative metering and instead use centre weighted average, which gives them more predictable results, and ones for which they can more easily anticipate how much EC to dial in.
Here, I've shot the example I was talking about. Here are two images, both shot in Av mode with 0 EC and depending on the lens I focused on the exposure has altered significantly. That is Evaluative metering doing its thing....
http://lh5.ggpht.com/_4_R8TkwT74w/S-aPQV4NrPI/AAAAAAABST0/GNYzzrE8o-0/s800/20100509_113158_.JPG
http://lh4.ggpht.com/_4_R8TkwT74w/S-aMPghHlkI/AAAAAAABSTY/b2NpP2Q82Qk/s800/20100509_111905_.JPG
The only thing I altered between shots was the focus point. The camera has decided that the shutter speed should differ from 1/60 when focused on the dark lens to 1/320 when focused on the light lens. That's a difference of 2.3 stops. As far as I am concerned that is just fracking bonkers, if you're not expecting it, and it's a good example of why I prefer to lock my exposure manually. Same scene, same lighting, should be the same exposure, but not if you let autoexposure serve you up with surprises. It's no fault of the camera; that is how it is designed to work; but you do need to know how it works if you want to get the best from it.
astral16v
09-05-2010, 10:40
:clap:
jerry12953
09-05-2010, 10:57
An excellent demonstration of something you wouldn't expect at all. Is it in the manual, I wonder?
Fortunately I've never worked out how to use the outer focusing points!
It's not explained in the manual as far as I can tell from a search of the PDF. There is one slightly cryptic line in the specifications listed at the back, where it says that Evaluative metering is "linkable to any AF point" without any explanation of what that means. That seems to imply that linking the metering to a focus point is optional, which as far as I know it is not, unless you switch to manual focusing so that there is no active focus point.
There is a discussion about the topic, including input from Chuck Westfall, here....
http://www.openphotographyforums.com/forums/showthread.php?t=8526
That discussion took place before the 7D was launched and, as I stated earlier, I think the 7D's metering system has integrated focusing data with metering data more closely, so the bias may be stronger (or not) in the 7D than other Canon cameras. It's all a bit mysterious.
I've also read somewhere that metering does not cover the entire image area, so stuff happening at the edges of the frame will have no influence at all.
jerry12953
09-05-2010, 11:54
Perhaps its unique to the 7D then.....
petersmart
09-05-2010, 12:11
TBH on my monitor the images look fine - detail almost everywhere and no really badly burnt out highlights.
The ones with a large expanxe of sky may be giving the appearance of being too light just because of the bright sky.
You could probably knock the exposure down by 1/2 stop and then I think they'd be spot on.
Perhaps its unique to the 7D then.....
No. Like I said, those discussion took place prior to the release of the 7D and pertain to evaluative metering throughout the lineup. I am simply speculating that the 7D might have different algorithms, and with a new metering and AF system that doesn't seem impossible.
Here are results from my 5D2 with a similar test setup. Once again, the only change I made between exposures was to the active focus point. The camera adjusted the exposure settings from 1/50 to 1/125 under its own steam. Nothing to do with me....
http://lh6.ggpht.com/_4_R8TkwT74w/S-anIfu-ubI/AAAAAAABSUs/aNmcHwIsQWE/s800/20100509_131254_.JPG
http://lh4.ggpht.com/_4_R8TkwT74w/S-anI7Ru9xI/AAAAAAABSUw/kA8WzcXUQn4/s800/20100509_131334_.JPG
Perhaps with a less contrived scene the differences would not be so obvious in typical shooting, but when your exposures don't turn out as you might expect you may find that this behaviour is the reason.
astral16v
09-05-2010, 12:27
http://www.imaging-resource.com/PRODS/E7D/E7DA5.HTM
Canon 7D
Quote ‘Although Evaluative metering is linked to the active AF point (whether automatically or manually selected), Spot metering is fixed to the center of the viewfinder’
Manual also says to Use AE lock when the area of focus is to be different from the exposure metering area suggesting that linked is the default. Page 106 in manual
Have you looked at the Auto lighting optimizer function on page 75 as well
TBH on my monitor the images look fine - detail almost everywhere and no really badly burnt out highlights.
The ones with a large expanxe of sky may be giving the appearance of being too light just because of the bright sky.
I'm surprised you say that, because I can see plenty of burned out areas and the histogram and clipping warnings confirm the problem....
http://lh4.ggpht.com/_4_R8TkwT74w/S-aqE3pup_I/AAAAAAABSU0/LjkcUvuu_E8/s800/20100509_132449_.JPG
http://lh5.ggpht.com/_4_R8TkwT74w/S-aqFOEq3lI/AAAAAAABSU4/k2mmnqYtqQ0/s800/20100509_132524_.JPG
I agree with Tim on this - all those images looked way too bright to me...
maybe 1 - 1.5 stops over, apart from the kid in the push-chair which was only maybe 1/3-2/3 stop over...
I was with you all the way until you said "Av" at the end there. Did you mean Manual?...[/IMG]
I said Av as the OP had shot some of those in auto, and some in Av. So I suggested he stick to using Av all the time if they were his preferred options. It's going to be better than auto if you know how to use it after all.
Personally I like to shoot in manual with spot metering all the time. But it doesn't matter, so long as you control the outcome it doesn't matter how you get there does it.
Anyway, the thread seems to have moved on a bit since then.
Av is auto. It is one of several autoexposure modes, including Av, Tv and P.
On cameras with a basic exposure zone, offering modes such as Portrait, Sports, Landscape and so on, those are all autoexposure modes as well. The camera is making the exposure calculations, not the photographer. I call that "autoexposure".
Tim, your example blows everything the experiance seminars taught me out the water. Thanks for taking the time to show it so so well
I never really noticed the issue with the 400 or 50d.
Arkady, I agree hence the question. It's better to get the shot right rather than sorting dozens of images out on the pc. It's been an education!
I like to think of the green sqaure as auto, p,av,tv and ca as semi auto and m as hells fury! :)
Av is auto. It is one of several autoexposure modes, including Av, Tv and P.
On cameras with a basic exposure zone, offering modes such as Portrait, Sports, Landscape and so on, those are all autoexposure modes as well. The camera is making the exposure calculations, not the photographer. I call that "autoexposure".
OK, to be pedantic, substitute 'auto' in my posts, with 'full auto'. ;)
I was just going off what the Canon put in the EXIF, that being;
"Exposure Mode:Auto"
But then you already knew that.
James, I didn't already know that. In fact, I would be surprised if Canon actually encodes the word "Auto" into the EXIF. There is probably a simple numeric value that gets interpreted by the EXIF viewer and displayed however that EXIF viewer sees fit. Certainly I've seen images shot in Program mode where the EXIF says "Normal". What the frack is "Normal" supposed to mean.
Anyway, as far as I see things, either you shoot with manual exposure or something else. Well surely, if it's not manual exposure it must follow that it is autoexposure. I was not trying to be funny/clever. I genuinely thought that you meant that manual exposure would be the solution and had simply mistyped, not least because even if you were to switch from "auto" to Av it wouldn't make a scrap of difference to the outcome if you were still relying on the same metering pattern to determine the exposure.
When I shoot in manual mode, which is nearly all the time, I use spot metering and choose carefully what to meter from within the scene (or perhaps not even within the scene at all) and what value to aim for on the meter itself. I evaluate the scene and I decide how to meter it and how to set exposure.
e.g. to hold detail in the white clouds I would spot meter off the clouds themselves and set an exposure to place the meter needle at +3. That might leave my subject/foreground underexposed, but at least I would have captured enough raw data from which to craft my final image. In such shooting situations it might actually be better to use an ND grad to tame the sky, but that's taking things to another level of refinement. Alternatively I might meter off the green grass at -2/3, or off my own palm at +1 1/3. In any event, I would not get fooled by weird metering patterns, such as evaluative, doing funky things, and I would be very much in charge of how to manage the dynamic range in the scene. For a simple sunny day, with my subjects sunlit, such as the wedding party, I would have my exposure locked manually into a variant of Sunny 16, and I would not blow the whites in the bride's dress or men's shirts. If I did see some odd results then a small tweak would have me sorted and good to carry on shooting.
The settings I've described above are a good starting point for preserving highlight details, say feathers on a swan, when shooting in bright conditions. If the subject/scene has no important highlight information then I may well increase my exposure in order to capture more tonal range and reduce shadow noise.
It is results such as those that started this thread that have put me off relying on autoexposure. It is often not predictable enough and thus not reliable enough, if you seek perfection. If you are more interested in the scene content rather than absolute technical perfection, and many are, then autoexposure (any mode) will usually give you a result you are happy with. If you want your results to improve technically then you can't necessarily leave everything to the camera to get right, without input from you.
I've been shooting with DSLRs for almost four years and the best thing I ever did was to move away from autoexposure (in the broad sense) and start metering precisely and setting my exposures manually. There are, without question, times when autoexposure (in the broad sense) is the right way to go, but I don't find that I encounter those times very often at all.
For those who prefer to use autoexposure (any auto mode), let's say you are shooting motorsport and panning vehicles as they come past. You can pretty much fill the frame. First you have a white car, then black, then yellow, then silver, then red, then green and blue, then yellow and red, then black and white. And so on. Which metering mode would you choose? What, if anything, would you do with your exposure compensation? How do you think your exposures would turn out? How consistent do you think the tones of the track and surrounding grass (or whatever) would be? How would you ensure that you didn't blow out your whites/brights?
If you don't shoot motorsport then how about horses - and you have to expose for white horses, black horses, patchy horses, possibly against a tricky sky. How would you meter? How would you control/adjust as horses of differing tones came past?
If it's wedding parties, let's say you first shoot the bride, then the groom, then the bride and groom together, then the groom and all his mates, then the bride and all hers. Which metering mode would you use and would you adjust EC to compensate for the differing tones parading in front of you? Would you expect to nail exposure first time for all the shots? Would the foreground and background remain the same for every shot? Would the skin tones remain the same from shot to shot?
Manual exposure is very powerful when your lighting conditions remain constant, because the tonal content of the scene can be changed freely without affecting the exposure. However, it can become a pain if the lighting is changing frequently. Conversely, autoexposure is very powerful when your subject and scene remain constant and it can react quickly and automatically to changing light levels. It is my experience that the subject/scene typically changes more rapidly than the lighting does, and for my purposes that makes manual exposure the better choice, more often than not.
trencheel303
10-05-2010, 09:47
I love reading your posts Tim - I have learned a lot in this thread!
hsuffyan
10-05-2010, 11:03
as a recent convert to spot metering and manual, I wholeheartedly agree with tim (who is always spot on IMO). I find that I am so much more confident in my exposures and tackling difficult scenes, and also allowed me to think more about each shot and get more involved with understanding the subject (which for me is people).
One question for Tim: How do you expose +3 in camera? Because doesnt the in camera meter just go upto +2? Or you expose to +2 and then manually dial in 2 or 3 increments (depending how your camera is setup) to take you to +3?
The 7D and all 1 series bodies have a 3 stop meter, so it is a piece of cake to meter at +3. You are absolutely right that for bodies with a+/-2 meter you first have to get the meter to +2 and then add another 3 (or 2) clicks to brighten by one more stop.
It is a bit of a pain when you need to recheck, if the light changes, because you have to come back down to +2 again and then go back up 1 again. With the 7D and 1 series you just glance at the meter and confirm or adjust your settings instantly.
Perhaps the best/quickest way to deal with the 2 stop cameras is by changing ISO. After all, you pick an aperture to give the DOF you wish, and a shutter speed to give you creative control of motion, and you pick whatever ISO is necessary to allow you to use that combination. So, let's say you want to shoot f/4 at 1/250, for example, and to get the meter to +2 you need to shoot at 200 ISO. The quickest way to get to +3 is simply to bump the ISO by 1 stop. You don't really want to be adjusting your creative controls if they are where you want/need them. ISO is usually the floating "don't care" variable, although obviously as the ISO starts to climb you might choose to compromise a little on shutter speed or aperture to avoid going too high.
I only ever use full stop ISOs - 100, 200, 400, 800 etc. - because the intermediate ones are faked and serve no useful purpose when shooting raw. In fact thay can do more harm than good. So I can just click back and forth between ISOs with a single click (of course I have to press the ISO button first).
Your post has also raised another point worth mentioning. As discussed, you can manually dial in any exposure you like, placing the meter needle anywhere you want, including off the scale. When you shoot with autoexposure and a camera with only a 2 stop meter you can't usually dial in exposure compensation of more than +/-2, so if it is your aim to meter and expose at +3 you are SOL (http://onlineslangdictionary.com/definition+of/sol) in any of the autoexposure modes. That places them at a serious disadvantage when trying to meter and expose for the highlights, which is a technique I use a lot of the time.
hsuffyan
10-05-2010, 12:43
another very useful post. I never thought of treating ISO as my floating variable, I used to get so hung up about noise, when it is easily the most 'correctable' of controls. Much more so than an out of focus or motion blur.
its a shame Canon didnt give the 3 stop meter to the 5D MKII
trencheel303
10-05-2010, 12:44
Tim, not sure what you mean that the intermediate ISOs are fakes - surely ISO sensitivity as a whole is 'fake' anyway, as all you're doing is upping the gain on the sensor. Would you be able to explain in laymens terms to me how intermediate ISOs are any more fake than the full stop increments?
The full stop ISOs are generated using an analogue amplifier to boost the signal, a bit like an aerial booster might for your TV. But that amplifier is only designed to have "gain" settings that hit the full stops. In order to get an ISO such as 125, 250, 500 and so on the analogue gain used within the camera will only get you as far as 100, 200, 400, etc.. All that happens within the camera to make the image 1/3 stop brighter is a mathematical multiplication of each pixel brightness value by a factor of 5/4.
What that means is that the camera is doing a software boost of the exposure, much as you might do yourself in your raw editing software. Therefore this is not a "real" amplification of the signal. It is a mathematical fake. So what actually happens is that, for example, you think you are shooting at 500 ISO, but secretly the camera is only shooting at 400 ISO and underexposing your shot by 1/3 stop. The camera then covers up the underexposure with a little bit of mathematical massage. As with any underexposed image, the software boost may show a little more noise in the image. What is equally bad is that you may have a perfectly good base exposure but the exposure boost in camera may send your highlights into clipping, when the real exposure was not clipped at all.
In other words, any ISO 1/3 above a full stop ISO is an underexposed shot that might lead to more visible noise and/or might lead to highlight clipping. If you'd underexposed for yourself by sticking to 400 ISO you might end up with more noise, but at least you would have an extra 1/3 stop of highlight headroom to play with.
A similar thing happens when you shoot at 160, 320, 640, 1250 ISO and so on, except this time the camera shoots at the ISO above and then mathematically reduces the exposure by a factor of 5/4 to achieve the faked ISOs. This in camera exposure reduction may make noise less obvious, which people think is good (and I suppose it is) but bearing in mind the initial capture was originally 1/3 brighter, you may actually clip your highlights initially and then the camera will pull them down so you don't notice. You could blow crucial detail in that brightest 1/3 stop and simply not be aware until it was too late.
In my opinion it is better to make your own decisions about how high to push your luck with the exposure and then you decide what to do with the raw file in your raw software. Why have the camera muck about with the data and potentially FUBAR it before you get your hands on it? OK, it's only 1/3 stop we're talking about here, but if you seriously want to obtain the best image capture you can then 1/3 stop matters.
For much the same reasons I have equal distaste for HTP. It is an in camera underexposure by one full stop and then a sneaky patch up of the underexposed file within the camera, or DPP for raw files. If you shoot to JPEG I concede HTP may have some small merit. If you shoot to raw then it has no value that I can see. Anyone can make their own choice whether to underexpose by 1 stop, 2/3 stop 1/3 stop manually, and then can process the file as they wish later, knowing they are in full control of the original raw data and squeezing the best from it. But underexposure is a bad thing, so why do it at all, especially intentionally?
Don't even get me started on ALO. Fortunately it doesn't screw with raw data, but it can certainly lead to disappointing results in JPEG (and some very pleasing ones too), and won't do anything to help those who are trying to learn more about photography. Using ALO is a bit like saying to the camera - "You get on with it. You fix my files for me. I don't want to get involved." You won't learn a thing if you have ALO try to cover up your mistakes. It's like all those old negatives that had exposures all over the map, but the print lab sorted them out for you so you never knew how bad they really were.
Just my opinion. :)
p.s. the above holds true for all Canon cameras, as far as I know, except the 1 series bodies. The 1 series bodies (at least prior to the 1D4) have a two stage analogue amplifier, but I am given to understand that the second stage amplifier is not terribly good and effectively renders intermediate ISOs on the 1 series bodies also of questionable value. I don't know how true that is, but to keep the operation of all my bodies consistent I only use full stop ISOs on my 1D3.
I have no information on the behaviours of cameras from other manufacturers.
p.p.s. ISO 50 is also fake. The sensor cannot shoot at 50 ISO. It is just in camera mathematics that pulls the file down by one stop from 100. By then, if your scene contains highlights they might already be lost. You may just as well shoot at 100 ISO in the first place and take control of any clipping/overexposure yourself.
trencheel303
10-05-2010, 13:27
That was one heck of an explanation, thanks a bunch for taking the time to write that out. Really helpful. I particularly enjoyed the paragraphs on ALO and HTP, particularly as I just recently turned them off in camera.
The Reverend JT
10-05-2010, 15:07
top forumming Tim - very helpful indeed.
ta
But Tim, I thought the rule was to expose for the shadows and deal with the highlights in post processing, rather than the other way round?
Jerry, sorry, I think I overlooked your question and filed to reply. When shooting raw there is a very powerful technique called "Expose to the Right" or ETTR. It means you set your exposure so that the histogram is nudged over towards the right hand edge, as close to clipping as possible, without actually clipping anything important in the scene. Specular highlights and light sources (street lamps, for example) are an exception. By exposing in such a way you capture as much image DATA as possible, while keeping the image data as far as possible above the noise floor in the shadow region. Think about shooting raw as capturing DATA (after all, it is just bits and bytes, not an actual image), not about taking pictures. You don't need your exposure to be "correct". You want it to record as much DATA as possible. You will make it look nice later on.
If you underexpose and then boost the exposure in post you will also boost the noise in the shadows. If you expose as brightly as possible, then reduce the exposure in post to make it look good, you will also lower the shadow noise until it is even less visible.
This approach would be standard for most types of raw shooting. This is what I mean by "exposing for the highlights". You place them very precisely to maximise dynamic range. The downside is that in scenes of high dynamic range the shadows have to make do with ending up wherever they may, but at least you did your best.
However, if your subject and interest actually lies more in the shadow region, such as a black dog, then you would indeed expose for the shadows, but in so doing you may have to wave goodbye to some highlight details as a consequence, and may even blow the sky altogether, for example. So it is a bit of a juggling act but experience should help guide you to optimising the exposure for the subject specifically and the scene as a whole.
You can read more on the whys and wherefores of ETTR here....
http://www.luminous-landscape.com/tutorials/expose-right.shtml
and here....
http://ronbigelow.com/articles/exposure/exposure.htm
My technique of exposing my highlights at +3 on the meter means I can easily place my highlights at the right hand edge of the histogram. It is a reliable and effortless way to accomplish that aim instantly. By locking in my exposure manually I can then feel free to compose however I like, zooming in or out and waving my camera about with not a care in the world about what the meter sees and thinks from that point on. If I find my exposure just a little off then a single click left or right will have me sorted and good to go. This is far easier than having to adjust EC every time the subject or scene changes, even just a little.
I haven't read through it, but I think there may be some interesting stuff about the benefits of raw, here - http://ronbigelow.com/articles/raw/raw.htm.
Firstly Tim, I’d like to thank you for the time and effort you have put in to these posts lately. I know I am among many who have learned quite a lot more about exposure in general and the 7D in particular (in the bird forum)
I have just started to try manual metering on static bird shots but have up to now “cheated” byfor example, setting exposure on a tree trunk bathed in sunlight then waited for the bird (treecreeper) to, hopefully, pop along. Whilst that may be ok for a very specific set of conditions and a specific bird, I would like to be able to set exposure manually and be ready for anything as it were.
This morning, I was stood in sunlight, not a cloud in the sky, and tried manual exposure. I took a test shot of a branch with blue sky behind and adjusted exposure accordingly. A blue tit popped along and after I took the shot and checked the histogram, the white cheek of the bird was blown. In this instance, should I have metered off the grass as you have previously documented? I also want to photograph terns down by the local river. These are obviously going to be the brightest part of the image. Given the same sunlit conditions and nice blue sky, should I meter off the grass beforehand? If there are nice white fluffy clouds, should I meter off the cloud at +3? Would this also be the same method for taking shots of planes at an airshow? I will need to set the shutter speed at 1/200th for prop blur so will need to adjust aperture and ISO appropriately.
Too many questions I know, but my head is buzzing at the mo and I am keen to crack manual exposure :bonk:
Nigel, I would not call it cheating if you are attempting to take charge and precisely set your exposures to suit the subject/scene/lighting. You can use any technique that works for you. I have my favourites, but that doesn't mean there are no other ways to tackle the challenge. My take would be....
If you have clear blue skies and strong sunshine somewhere behind you then one of the easiest manual settings to use is a standard "Sunny 16" exposure. Using such an exposure will usually preserve highlight detali in whites and bright colours.
A "Sunny 16" exposure is one where if you set your aperture to f/16 then your shutter speed should be the reciprocal of your ISO. In other words at 100 ISO your shutter speed would be 1/100. At 200 ISO your shutter speed would be 1/200 and so on.
However, f/16 is seldom a good choice of aperture, so it is usually better to open up wider, maybe to f/8, f/5.6, f/4 or whatever is best for your subject and then to increase the shutter speed to compensate. For example, for BIF you might try f/8, 400 ISO and 1/1600 in bright conditions. For perched birds you could lower the shutter speed and the ISO for better IQ, so long as you can keep the lens steady enough. If you have a fast, sharp prime lens and a good tripod and technique then you could shoot at something like f/5.6, 100 ISO, 1/800 for incredible IQ.
If the sun is not more or less behind you, or maybe there is a little haze, then you may need to increase the exposure slightly. Remember, this approach will be well suited to retaining bright highlight details. If your subject is of darker tones, especially something like a crow or blackbird, then you would be well advised to increase the exposure further, to capture extra detail in the dark tones. That is assuming that you will not blow out the sky, or anything else of importance to your composition, behind the bird.
If your subject is on or over water, or whiteish sand, or snow, or anything else that will reflect brightly and add more light to the subject then you may need to dial the exposure back a touch.
Here's an example where I would personally find the scene overall, and the subject, extremely hard to meter. The whites on the bird are not large enough to be spot metered and I'm certainly not skilled enough to be able to make a judgement about where to place the meter needle if metering off the water. The scene does have a high dynamic range, and will not fit within the range of the histogram. What I did observe is that the subject and scene is sunlit. The sun may not be doing much for the water, but it is certainly making the whites on the bird light up. The easiest option, at least to get things started, was to dial in an exposure equivalent to Sunny 16 and adjust if necessary. Here is the shot as a Sunny 16 capture, with no edits except white balance...
http://lh6.ggpht.com/_4_R8TkwT74w/S-qgXLFznQI/AAAAAAABSZM/aMT1DYX2tJs/s800/20100512_133406_.JPG
The darker areas do look a little dark, and would need some PP to bring them up, but the whites are awfully close to the right hand edge of the histogram. This exposure might have been improved by going a little brighter, but as a starting point it is very close.
So that's one good yardstick for shooting in bright, sunny conditions.
If the sun angle is not in your favour, or you have conditions other than bright sun, or even if you do have the sun on your side, there are lots of options for metering that will get you very close to the right ballpark. If you have bright, white clouds in your background, and expect them to form part of your scene (after cropping) then spot metering the clouds at +3 is a good starting point, if you shoot raw. However, you need to consider what sort of birds might be your subject, because some white feathers can reflect more brightly than clouds. Backlit clouds can be dazzling, especially relative to your subject. Front lit clouds may not be as bright as the feathers on a gull, swan, magpie and others. If your intended subject is likely to be less reflective (less bright) than the clouds then use the clouds as your +3 target. If the subject may be brighter than the clouds then hold a little in reserve, maybe metering the clouds at +2.3 instead of +3.
Basically, your aim is to capture the highlights in the scene, whether they are in the subject or other areas, as brightly as you can, without overexposing them, so save +3 for the brightest thing you expect the scene to contain.
Another option, which should give you an exposure that will also hold highlight details, is to hold your hand up, with the back of your hand facing the subject direction, and your palm facing back towards you, angled a little so as to "catch" the light from behind you rather than err towards glancing light or shade. You need to ensure your lens hood does not cast a shadow onto your palm!. For my palm, which is pale caucasian skin, setting the meter to +1.3 stops will give me an exposure that will secure my highlights. People with differing skin tones may need to aim for some other value. If I know the subject/scene will have bright content then I stick with +1.3. If I know my subject/scene does not have bright tones, and I do not expect bright tones to enter the scene, then I might aim to meter my palm at +2.3 so that I get a brighter/stronger/less noisy exposure. For palm reading to work you do need to make sure that you and your palm are in the same light as your subject/scene. If you are hiding in the shade of a building or tree while your subject is in the clear then palm reading won't work too well. Equally, if you are in the sunshine and your subject is in shade then the same argument applies.
Metering from grass is another option too. I find that with bright highlights in the scene that lush, green grass metered at -2/3 will give me a good exposure to secure the highlights. Once again, if my subject/scene contains nothing bright (of importance) then I may well increase the exposure by a stop or so. Don't forget, you might end up with something in the scene becoming overexposed, such as the sky, but you may be shooting in the knowledge that you plan to crop it out later, so hanging on to details there is not a concerns. Far better to capture better detail (and lower noise) in your subject than to worry about parts of the scene that are of no concern.
There is nothing stopping you metering off other things too, but you must carefully consider how bright those things are and whereabouts on the tonal scale they should be positioned. Think of it like this....
+3 on the meter is equivalent to (almost) pure white and this is where the brightest parts of the scene can/should be metered if they are indeed brightish/whiteish. The camera will be able to record detail at this level, such as texture in clothing, but it might be difficult to see. You will have the option, in your raw editor, to adjust the rightness if you need to, to reveal the texture concealed in the file. For safety, and if shooting to JPEG, then you can aim a little lower. I like to push to the edge, if I can.
+2 on the meter is just off white and anything there will hold plenty of tonal detail and texture.
+1 1/3 on the meter is where I will set an exposure from my own outstretched palm, if I have bright highlights I need to preserve. If there are no such highlights then I may increase the exposure by around 1 stop.
0 on the meter is just a sort of middle tone, and not very exciting or special. That is where the camera thinks things should be unless you tell it different. This is about the shade that you get from the inside material of a Lowepro camera bag. You'll probably find that concrete pavement is quite close to this tonal value too.
-2/3 on the meter is where I would place lush, green grass, if my scene had bright highlights I needed to preserve. In a scene with no such highlights I would meter grass around 1 stop brighter.
-2 on the meter is very dark but still contains good detail. If you were shooting a dark subject, such as a black horse, dog, cow, bird then you would be looking to expose the animal at around the -2 mark. After all, you don't want a "black" dog to turn out looking "grey", but at the same time you don't want the fur to be turned into a featureless black pool of nothingness. Even then, fur can have shadowy areas, and even a remarkably bright sheen to it, so you need to think about the location you are metering from and whether it should in fact be nearly completely black or actually quite a light black.
-3, for practical purposes is virtually black, and you won't see much in tones captured at -3 on the meter. In truth there is detail recorded there, and at levels below -3, but to see it you will need to brighten the shadow tones and that will start to reveal noise that you may prefer not to see. The higher your ISO, and/or the worse any underexposre, the worse that noise will appear.
So, when you choose to meter from a tree, that is all well and good, but different trees have different tones in their leaves, and in their bark, and those things themselves have shadows and highlights, so while you certainly can meter from them, you need to think about how dark they are relative to the rest of the scene. Metering a part of a tree at 0 might be perfectly correct, but it might be that -1 would be better or maybe +1, or maybe some other figure, depending on the tree itself and how the light is striking it.
Here is a typical example of a shot where I metered the sky at +3. Aesthetically it is overexposed, but my raw capture has recorded as much data as possible, without losing anything inportant to clipping, and it will take but a moment to adjust the aesthetic appearance of the image, without increasing noise or otherwise harming IQ.
http://lh6.ggpht.com/_4_R8TkwT74w/S-p0srIhTKI/AAAAAAABSYE/Sjr2EcxVzqc/s800/20100512_102659_.JPG
Here's another example (poor composition but a good example of what I want to explain, showing the case where the bird's feathers are in some areas brighter than the sky, and in this example are overexposed by around 2/3 stop. That in itself is not a problem, because (a) there is not much in the way of important detail there that needs to be faithfully recorded; (b) the raw headroom will almost certainly allow me to recover some detail in the blown out areas. The important thing to observe is that the sky itself is clearly not a bright white, and it would indeed have been risky to try to meter from this sky at +3. If you evaluate the exposure settings I used for this you will note that they are an exposure 1.7 stops brighter than Sunny 16. Now, given the hazy sky it would be reasonable to expect an exposure a little brighter than that required for full sunshine, and as you see this exposure is pushed around 0.7 stops too bright for these feathers, maybe 1 stop too bright for really dazzling feathers. The other 2/3 stop is for the hazy light. The reason I was set up like this, not to take account of white feathered birds, is because I had been shooting swifts (or sand martins or something like that) and wished to maximise the detail recorded for the darker tones of those birds and their shadowy undersides.
http://lh5.ggpht.com/_4_R8TkwT74w/S-qVjDXiP4I/AAAAAAABSYg/hvl-l9x4-nU/s800/20100512_124720_.JPG
So, even with guidelines to follow it is still important to pay close attention to the results you are achieving in practice and to be willing and ready to fine tune things if needed. On the back of the camera the histogram would almost certainly not make known the small amount of clipping in the feathers, but the blinking highlight warnings in the preview image would have been a useful alert that I was pushing my luck here.
As you may also note, I positioned myself so that the sun was working for me, rather than against me, so that my subject was well lit and the sky was not a luminous light source in the background. If I had been shooting in the opposite direction the birds would have been in their own shadows and the sky would have been much more difficult to control while achieving correct exposure for the birds. Even with the sun behind me, you may note that I have still caught some shadow under the wings, because the sun was so high in the sky. There are many things to consider when going out to shoot and the time of day can matter just as much as everything else. The most difficult thing to do well is to shoot dark, shadowy undersides when the sky is bright behind the bird. Blue skies are often not a problem, but hazy or white skies and backlit cloud can be.
Here's an example where the bird's underside was shaded, but the almost clear blue sky in the background made it easy enough to get a nicely balanced exposure. The exposure here was 1/2000 at f/5.6 and 400 ISO, being 2/3 stop brighter than sunny 16. It has had no exposure related adjustments other then to darken the blacks to create more contrast/punch...
http://lh4.ggpht.com/_4_R8TkwT74w/S-qdP6YCllI/AAAAAAABSZE/atZUGPN6Afs/s800/20100421_131603_5566_LR-2.jpg
Going back to the example of the blue tit with the blown cheek feathers, so long as nothing else was blown you can probably deduce that you were only out by maybe 1/3 stop or 2/3 stop, so a quick adjustment to your manual exposure to get you where you need to be will then leave you free to carry on shooting without worrying about remetering and adjusting repeatedly from that point. You can adjust your composition, taking in a little more sky, or a little less, for example, and your exposure will remain fixed just as you want it. If the light changes then you might need to re-evaluate your exposure setting.
Tim, thanks once again for a brilliant explanation of how to expose manually under different conditions.
Lots to take in but I will certainly be giving it a go with static birds, birds in flight and at an upcoming airshow. Should be fun :)
I'll need to get myself a quick ready reckoner of the sunny 16 rule and calculate the variable combinations to still maintain the equivalent of sunny 16
:cuckoo:
trencheel303
12-05-2010, 18:57
I'll need to get myself a quick ready reckoner of the sunny 16 rule and calculate the variable combinations to still maintain the equivalent of sunny 16
:cuckoo:
As will I, I'd better spend half a day working it out too as I am really bad with numbers.
madmackem
13-05-2010, 09:32
great reading and advice Tim, Added to my favourites cos theres too much too take in in one go!
Trixster
13-05-2010, 10:25
What the frack
Battlestar Galactica.
Is this thread a contender for thread of the year? I think so. :thankyou: Tim, awesome stuff.
Incredible bit of forum work here Tim, much appreciated. Dying to get out now and try it out, pity I'm in a grey office all day at work :(
pity I'm in a grey office all day at work :(
If it's 18% grey you shouldn't have a problem with exposure :D
If it's 18% grey you shouldn't have a problem with exposure :D
Quality :clap:
Here is a three page article which I think describes in a more digestable way, and with more useful examples, the points I have been making about maximising exposure by metering important highlights at +3 stops. If you didn't follow what I have been saying then this might help it sink in....
http://daystarvisions.com/Docs/Tuts/DCExp/pg1.html
vBulletin® v3.8.4, Copyright ©2000-2012, Jelsoft Enterprises Ltd.