I just finished reading Peterson’s Understanding Exposure - a great reference. Peterson recommends shooting an additional shoot at -2/3 EV saying it frequently produces a better photo (better contrast and saturation) than what the meter recommends. Has anyone found this to be true with a D90? I can possibly see this in high contrast situations, but would this be recommended for other situations? Terry
#1. "RE: shooting at -2/3 EV" In response to Reply # 0 Thu 03-Jan-13 09:34 PM by jef_dk4
An interesting related story is that when I had the d90 with a tamron 17-50 f2.8 I had always set exposure -1/3ev to achieve optimum results. That was required though only with that specific lens. With every other lens I ever tried on that body, I didn't need exposure fix, unless I wanted to achieve something on a specific occasion. So I guess it has to do with the camera/lens combination, I wouldn't just blindly set it unless I figured it is needed for a specific lens.
#2. "RE: shooting at -2/3 EV" In response to Reply # 0
Seattle, WA, US
>I just finished reading Peterson’s Understanding Exposure - a >great reference. Peterson recommends shooting an additional >shoot at -2/3 EV saying it frequently produces a better photo >(better contrast and saturation) than what the meter >recommends.
I remember there being some film that could benefit from a small amount of constant underexposure, but certainly not digital.
-2/3 EV for digital as a genera setting is just going to introduce lots of noise when you try to fix the underexposure in post processing.
If you are shooting JPG and want more contrast and saturation in the "just out of camera" JPGs, just set your camera to do that in the picture controls.
---------+---------+---------+---------+ Joseph K Seattle, WA, USA
#3. "RE: shooting at -2/3 EV" In response to Reply # 0
Cedar Park, US
Here's the extended version... hope you can use the experience...
I'd have to pretty much agree with Peterson. Since purchasing my D90 I've made over 35,000 shots. Early on I used Nikon's default exposure settings regardless of the lens and found that Nikon's algorithm constantly produced over-exposed images both RAW and JPG. Although images were only 1 to 2 EV over, I found it frustrating when comparing the same subject with Canon's default algorithm. Also, I've found this same to be true with other Nikon cameras. It's much easier for me to recover data from under-exposed images than from over-exposed. Over-exposure literally obliterates data. Since then on my D90, I've gone primarily to full manual control for critical work regardless of the lens. While Nikon's algorithm, in my opinion, is probably okay for point-and-shoot (snapshots) pictures, it's simply not the optimum for critical photography. There's a huge difference between intentionally constructed photographs and snapshots. As my skills mature, I find that I become more critical of technique. Funny that I find that when I'm just shooting non-critical stuff, I still use full-auto....crazy, aye?
Regarding post processing of critical stuff, I've found that noise is virtually a non-issue when shooting at ISO under 800 althought I've noticed that there are a couple of "hot" pixels in the CMOS. They can be adjusted out using higher-end photo editing software. JPEGs are great for many images such as time-lapse work but NEF's are invaluable for critical work. I normally capture both and tease out the details at post. Digital storage management is a whole different matter.
Regardles of the situation, a camara doesn't see what our eyes/brains see. Don't be afraid to learn what works for you. Most of all --learn to enjoy your work.
#4. "RE: shooting at -2/3 EV" In response to Reply # 0
To a great extent, the "correct" exposure is a matter of personal taste. Whilst it's wise to avoid the loss of detail that comes with too much overexposure, I would not recommend applying a fixed amount of compensation for all types of shot.
The metering systems in Nikon SLR's (including the D90) are generally very good, and get pretty close in most circumstances.
My advice would be to learn how your camera meter responds to different lighting conditions, and apply compensation if and when it seems necessary
#5. "RE: shooting at -2/3 EV" In response to Reply # 4
Thanks for all your input. So far I haven't seen the need to underexpose in all situations as Peterson suggested, but I'll bracket for comparison. Nothing beats experience with your equipment. Thanks again. Terry
#6. "RE: shooting at -2/3 EV" In response to Reply # 5 Fri 04-Jan-13 05:34 PM by quenton8
The statement in your extended quote that shots were generally 1 to 2 ev over tends to stretch credibility. 1/3 to 2/3 ev over "maybe".
As others have pointed out, a lot may depend on personal taste, final use etc. If I were only ever going to view on my computer monitor, I might want a 1/3 or 2/3 under exposure -- but I tend to print, so I prefer what my D90 provides.
If the "1 to 2 ev" overexposure is real, then I would want to compare to another D90 and/or take it to nikon to be checked out.
#7. "RE: shooting at -2/3 EV" In response to Reply # 6
St Petersburg, RU
So much depends on the scene, that a set rule is bound to lead to problems. What metering mode, what is it biased for, and where is the active focal point? A meter has to have some reference to compare to, and the standard is the mid point between highest tones and fully black, which is about 18% grey. That is what the camera sees as properly balanced between highest tones and pure black, so that is what the meter says is proper exposure. For many scenes it is. But looking at the scene with your eye, you can see if there is a bias for one end of the spectrum of light level, you know the camera will lead you astray in a predictable amount. For example take an shot of a bright beach, reflective water, bright blue sky. Well, looking at it you are going to assume the camera is going to try to under expose it so it can reach the 18% grey mid tone. It is going to be underexposed, you know it, so you dial in some +EC, maybe 1-2 ev and the image turns out over exposed according to the camera's sensibilities but it looks great with the white or light sand white or bright. Take a photo of a man in a black suit as the dominate element in the scene. It comes out looking grey, not black. But that would be the assumption if you keep in mind the camera is seeking to center between the extremes. So you intentionally underexpose the next frame by an ev or two and find the black suit has returned and given up its "greyness". Ever notice how dull and lifeless your photos of your ski trip were? The snow was a dingy grey in the prints. That is the meter trying to find balance in an unbalanced world, so everything was underexposed. With a little practice and stopping to look at the scene before putting the camera to your eye will tell you how you will need to trick the camera's meter into giving what you want instead of what it measures. You dial in a little +EC and the white snowy world comes out bright white just as your eye saw it. It is doing this as a very good thing, it knows your camera and media for display can't handle the full dynamic range of the scene, any scene so it tries to center the window of DR that it can handle in the middle of the range of the scene. For average scenes, that works very well. But it is surprising how many scenes are not average, the average scene is not average:>) The foregoing comments did not consider the mode of metering: spot, center-weighted and matrix. They are all used to your advantage when knowing which will allow you to get your desired results. Matrix is surprisingly smart and is suitable for most shooting but there are times when the scene or the specific subject would benefit from metering only on the subject as in the tight 1% of the frame under your active focal point. That can allow backgrounds to be badly exposed but your desired subject will be spot on most of the time. Center weighted considers the entire frame but gives a strong bias to the area under the selected focal point. Matrix considers the whole scene, its colors and contrast and distance information if supplied by the lens. So, for these reasons, setting a fixed bias might be making things worse and often increased noise and limits DR. Consider the scene before deciding whether to dial in compensation, but only do it for a reason that relates to the scene in front of you. Stan St Petersburg Russia
#8. "RE: shooting at -2/3 EV" In response to Reply # 7
Thanks Stan - very good advise especially about metering mode. I know from my early days with film that high contrast scenes (snow, beach/desert, etc.) that the -2/3 recommendation by Petersen was incorrect. I'm still surprised that he would make such a blanket statement like that. Thanks for all the replies. As a digital newbie, I really appreciate the input from such knowledgeable photographers. Terry
#9. "RE: shooting at -2/3 EV" In response to Reply # 8
I've always had a problem with my D90 overexposing. I would say 95 percent of my images are taken with -3/6 being "baked in" using the b4 fine tune optimal exposure menu setting. This gives me a zero (balanced) meter reading with -3/6 baked in. Maybe its just my camera, or some of these d90s, but even with this -1/2 I still usually have to dial in additional underexposure if there is any sunlit areas in a mixed lighting scene. I would say the majority of images i discard are due to blown out highlights that are unrecoverable (CNX2 and RAW). I continually struggle with overexposure even with this -1/2 stop baked in. Evenly lit scenes with lessor dynamic range of illumination are ok, when i get into medium to high contrast settings with sunlight/shade I usually will need to dial in additional minus exposure compensation. I have shot identical images with another semi-high end camera (non Nikon) that i have and do not have this problem with that camera.
#10. "RE: shooting at -2/3 EV" In response to Reply # 0
So glad to read these comments. I too have deleted several photos due to overexposure. As a matter of fact, my wife and I just returned from the Great Smokey Mountains and, quite honestly, in going through our photos since returning home, we deleted many from my D90 and kept the photos taken with her inexpensive Canon point & shoot which, generally speaking, yielded much better exposure. I'll try the -2/3 next time out. Thank you.
#13. "RE: shooting at -2/3 EV" In response to Reply # 9
St Petersburg, RU
Are the focus targets over exposed or the background away from the selected focal point? Even in Matrix metering there is a bias for the metering falling under the selected focus point. Can you post an image that is typical of your overexposure problem? If it is a high DR image, that greatly exceeds the capture range of the camera something has to give either lost detail in shadows or blowing hightlights and the camera is going to strive for the 18% grey, mid point between full black and full well saturation. Stan St Petersburg Russia
#14. "RE: shooting at -2/3 EV" In response to Reply # 9
>I've always had a problem with my D90 overexposing. I would >say 95 percent of my images are taken with -3/6 being >"baked in" using the b4 fine tune optimal exposure >menu setting. This gives me a zero (balanced) meter reading >with -3/6 baked in. Maybe its just my camera, or some of >these d90s, but even with this -1/2 I still usually have to >dial in additional underexposure if there is any sunlit areas >in a mixed lighting scene. I would say the majority of images >i discard are due to blown out highlights that are >unrecoverable (CNX2 and RAW). I continually struggle with >overexposure even with this -1/2 stop baked in. Evenly lit >scenes with lessor dynamic range of illumination are ok, when >i get into medium to high contrast settings with >sunlight/shade I usually will need to dial in additional minus >exposure compensation. I have shot identical images with >another semi-high end camera (non Nikon) that i have and do >not have this problem with that camera.
I had the same experience when I first got my D90, coming from a D50 with no exposure problems. I did some quick searching on the forum and ended up "baking-in" -3/6 exposure for the matrix metering reading only, not the other two. Everything has been correct since then. I can't say that I was smart enough at the time to check exposure in the focus spots, and since then I have just been able to forget about it and make the normal compensations for snow etc. that we do anyway.
There are enough of us that have done this that there must be something to it for general use.
#15. "RE: shooting at -2/3 EV" In response to Reply # 12 Fri 01-Feb-13 07:34 PM by limeyzen
In fairness to Mr. Peterson I think we may be misinterpreting what he says. To qoute in full from Understanding Exposure (section on exposure meters page 125) he says: "In addition, I do recommend taking another exposure at - 2/3 stop when shooting most any subject, since this often improves contrast and the overall color saturation of a given scene. This comparison shot will allow you a comparison example so that you can decide later which of the two you prefer. Don't be surprised if you often pick the - 2/3 exposure."
Seems to me that he is offering a suggestion for many situations for one's digestion and comparison not an absolute rule for everything! Of course we may disagree with his suggestion
#16. "RE: shooting at -2/3 EV" In response to Reply # 15 Thu 07-Feb-13 08:56 PM by km6xz
St Petersburg, RU
I have not read his book, although I often recommend it because I know it covers the basics well without math and in an upbeat enthusiastic style he is known for. So according to what he wrote, he was suggesting bracketing shots, which is a standard method used for decades, of have the bases covered.
The biggest problem with a rule is that it often makes it worse. What if the focus point of a scene in brighter than other detail. It Matrix will calculate overall scene level and then bias it with the spot under the focus point. That means the -2/3ev will be further underexposed because the focus point will bias the AF towards 18%grey at the selected point. If the focus point is darker than the rest of the scene, sure enough, the metering calculation will bias it towards higher exposure to increase exposure of the focus point closer to 18% grey. I have yet to see a D90 that was brought to me with that problem, actually have a problem. Knowing how the tone under the focus point impacts scene exposure clears up the problem. The D50 did not have matrix biased for the focus point. The philosophy behind the focuc point biasing is simple, the assumption is the subject focused on is the key element of the scene and by nudging exposure to get that subject right, the subject photo will be well exposed and higher tones might be blown out which are not the key elements. Something in scene has to be blown out or lost in shadows in any daylight sunny scene, due to the wider range of the scene than the range of a camera. If something is lost, the camera by design just makes sure it is not the subject that is the target of the shot. People notice and complain about highlights blown but often there is more lost on the dark side. Both lose data if the scene is wide enough in tone span and the subject is well exposed. The only way to have it all is to bracket and combine using HDR, or by exposing for the highest tones and bring up the shadows in posts if the ISO is low enough to prevent excess noise when pushing the shadows a stop or two. Or wait until a 18 stop camera is released in 10 years. Actually the technology is here now for that, some point and shoot cameras take a burst of frames all bracketed and create HDR on the fly. It is prone to artifacts and ghosting however.
Edit: forgot to include the most common fix used every day. Fill light. If there is backlit situations, a very common sort of scene, expose to the highest tone and fill the darker key subject with fill light either from a reflector or flash. I use flash on a large percentage of my outdoor daytime photos and it works great if the subject is close enough or there is enough power to over power the background. Here is an example. I saw a girl in the park laying in the leaves while her GF was shooting photos which were very dark in the face. The reason was that the sun was low and behind the girl so metering for the girl's face totally blew out the background and exposing for the overall scene left the girl's face in deep shadows so her own mother would not recognize her. They saw I had a camera and asked me for assistance. I was in back of the GF by 15 feet with a 70-200 2.8 and took one shot. I cheated, exposed for the high tones and use the SB900 for fill. The background was bright, the girl was in shadows directly in front of the setting sun, and the river in the background that also reflected more light from the back. As a coincidence, her camera was a D90 with 18-105vr. She was very frustrated and complained that faces were dark many times and guessed she needed a better camera. I assured her that she had a great tool that was just being applied to the task wrong. We both had the same backlit situation, but here is what mine looked like:
#18. "RE: shooting at -2/3 EV" In response to Reply # 0
The main reason behind this is that digital cameras can recover shadows a lot better than highlights, so it's better to err on the side of underexposure. That said, it depends a lot on which camera you're using. When shooting with my D40, I've found that I get the best results with -2/3 when shooting landscapes and 0 or -1/3 when shooting people. But with my D5100, the meter is usually spot-on, and I leave it at 0 unless I know my subject is really dark or really bright. I would imagine that the D90, being a pretty modern camera, would be pretty accurate, so leave it at 0 or maybe go down to -1/3 if necessary.
#19. "RE: shooting at -2/3 EV" In response to Reply # 9
San Diego, US
>I've always had a problem with my D90 overexposing. I would >say 95 percent of my images are taken with -3/6 being >"baked in" using the b4 fine tune optimal exposure >menu setting.
How do you go about baking in an exposure? I also have a problem of overexposing with my D90 and would like to dial it back for outdoor exposures that look pale and washed out compared to what i can actually see.