Sigma 18-200mm f3.5-6.3 "OS" DC Review

I feel the IQ & colors are very good for a lens of it's price range, and I am happy with mine so far for my intended uses. Here are some full size crops as shot, nothing done except the crop. The pail was the tacal point, at 200 mm from about 45 feet at F 6.3. The flower shot is at 51 mm F 5.0 from about 10 feet away.

Byte size: ZERO | Content warning: NOT AN IMAGE

Byte size: ZERO | Content warning: NOT AN IMAGE

The colors all look right to my eyes considering they haven't had any adjustments of any kind.

OK, as I mentioned before, this is the real trick. I'll tell you what I know about it and maybe it'll help a little. This is going to be a bit long, so my apologies. Please bear with me (and those of you who really know this stuff, please chime in and correct anything I get wrong).

I'll talk about this in terms of your camera's metering system. There are other ways to meter the light in a scene, but your camera's meter is always available. In the simplest terms, your camera's metering system will tell you what your camera thinks the exposure values (combination of shutter and aperture) need to be in order to properly expose the scene. The problem is that your camera isn't always right. Understanding how the camera meters the scene is a huge component of getting good exposures in situations where the camera itself may not compute the exposure values properly.

The first thing you need to know is that camera metering systems in general (and those in the Canon EOS line in particular) are calibrated relative to middle gray. Middle gray is just that: a gray with pixel values halfway between pure black and pure white.

So you need to see what middle gray looks like on your computer's monitor. To do that, take a look at the Wikipedia article on the Zone System here: http://en.wikipedia.or​g/wiki/Zone_system. About a third into the article, you'll find some images showing a gray scale.

Zone V is middle gray, with an RGB (Red, Green, Blue) value of (127, 127, 127) with pure white being (255, 255, 255).

For most metering modes, the intensity of middle gray is the intensity your camera's meter will try to get the exposure in the area being metered to match.

Want to know what middle gray looks like in "real life"? The simplest way to do that would be to get a gray card. You can pick one up from any decent camera shop. They're pretty cheap, on the order of five dollars. They're also useful, because you can use them not just to meter a standard exposure, but also to set your white balance in environments where none of your camera's white balance settings will quite do the trick.

If you know what middle gray looks like in real life and on your computer monitor, then you'll know roughly what intensity the camera is trying to make the average metered part of the image look like.


Internally, there are two steps to the camera's metering process. The first is the actual measurement of the incoming light. The second is the interpretation of that light. Both are relevant to the entire process.

Your camera has multiple metering modes. They are:

• Evaluative (or "matrix")
  
• Center weighted average
  
• Partial

Some cameras, such as the 30D, also include "spot" metering mode, which is basically the same as partial except that the size of the metered area is smaller.

Evaluative and center weighted average mode both measure the amount of light coming into the entire frame. Your camera has an array of light sensors scattered across the frame. You can't see them, of course, but they're there. Each one measures the amount of light hitting its particular section. In the case of your camera, there are 35 sensors in a 7 horizontal by 5 vertical grid that overlays the entire frame (if the manual is anything to go by, at least). Both evaluative and center weighted average metering modes make use of all of those sensors. They just differ in terms of how they interpret them.

Partial metering mode ignores most of those sensors except for the center cluster, which occupies about 9% of the frame. I'm not sure exactly which sensors it uses but the results are what matter: the center area gets metered and the rest gets ignored.


The biggest difference between the modes is in the interpretation.

Evaluative mode on many cameras is something of a "magic" mode. In general what it does is average the scene and give weight to the focus point(s). But from what I understand, it also tries to be "smart" about its interpretation, in that (if I remember right from discussions elsewhere) it tries to detect and compensate for backlit subjects. While that means it's less likely you'll have to use exposure compensation to get the exposure to come out right, it also means you have to experiment with it more to really understand it. In principle, it's not necessarily as predictable as, e.g., center-weighted average, although in the case of our cameras it might be. It's also not really officially documented in any way, so how it works under the covers can change radically as the manufacturer refines it. That's a good thing if you just want to use it without learning how to compensate for its shortcomings, but will mean a little more work on your part when they change it, because you'll have to re-learn its shortcomings all over again, through experimentation.

Center-weighted average performs an average of the entire scene and gives weight to the center, as the name suggests. It always behaves the same way no matter what you're shooting, so in that respect it's easier to predict. What you'll need experience for is to know how much bias it gives to the center, in order to know how much exposure compensation to use.

Partial metering, I believe, performs a straight average across the area it's metering, i.e. the center 9% of the scene. The area metered is roughly the circle with its centerpoint on the center focus point and the edge touching the top and bottom focus points.

And spot metering (which you find on the 30D, for instance) averages the center 3.5% of the scene.


All that's nice and all, but how do you use it? For that, we have to talk about your camera's exposure modes and about exposure compensation. First, exposure modes.

You have 4 main exposure modes (all of them being "creative" modes): Program (P), Shutter Priority (Tv), Aperture Priority (Av) and Manual (M).

Manual mode means you have full control of both the aperture and the shutter. You can change one or both, fully independently of each other.

Program mode picks a combination of shutter speed and aperture that it thinks will work best. You can change the combination it uses by pressing the shutter button halfway and then turning the front dial. But the combination it uses will always produce the same exposure, unless you use exposure compensation.

Shutter priority means you manually select the shutter speed you want the camera to use, and the camera will select the aperture based on what the meter tells it. You can move the front dial to change the shutter speed.

Aperture priority means you manually select the aperture you want the camera to use, and the camera will select the shutter speed (based, of course, on what the meter tells it). Moving the front dial changes the selected aperture.


Exposure compensation is how you change the actual exposure the camera will take relative to what the meter is telling it to use. You use this when you're in either Program, Shutter, or Aperture mode and know that the camera will overexpose or underexpose the shot if left on its own. If you know that the camera would underexpose the shot, then you need to add positive exposure compensation, which means moving the meter's needle to the right of center. Conversely, if you know that it would overexpose the shot, then you add negative exposure compensation, which means moving the meter's needle to the left of center.

Manual mode, of course, doesn't listen to the meter because it doesn't automatically adjust anything. Instead, it will show you what the meter thinks of your exposure parameters. If it thinks you're going to underexpose the shot, it will show the needle to the left of center, and if it thinks you're going to overexpose then it'll show the needle to the right of center. So if you want to manually set your exposure parameters to get an exposure that is "correct" according to the meter, just adjust the shutter and/or the aperture (depending on what and how you're trying to shoot) until the needle is centered. In manual mode, "exposure compensation" merely means manually changing aperture and/or shutter so that the actual exposure is above or below the meter's idea by however much is appropriate (and if it's within 2 stops of what the meter says it should be then you'll be able to see it based on the position of the meter's needle).


And now, finally, we get to the heart of the matter: how do you tell how much exposure compensation you're going to need?

That's where experience comes in. As you know, the meter is going to give you an exposure value that will make the intensity of the metered area average out to Zone V gray (how exactly it computes that average depends on the metering mode). If what you're metering is actually dark (like, say, a black car) then it will cause the shot to be overexposed, and if what you're metering is actually bright (like, say, a white car) then it will cause the shot to be underexposed.

Different photographers have different approaches to this particular problem. I can only tell you about mine, and keep in mind that I'm relatively new to this as well. My methods may change over time as I gain experience, but for now I've found something that works reasonably well for me (depending on who you ask! ).

For me, the first step was knowing what middle gray looks like and how many stops of difference there is between that and the other objects I might want to meter. I use spot metering almost exclusively because it makes metering a specific object in the scene easier than any other metering mode. In your case, the closest you can get to that is partial metering mode. That may be enough.

The easiest way I've found to "calibrate my eye" in this regard is to shoot objects in the same room that my computer is in. The first thing I'd do is take a shot of my gray card and pull it up on my screen. In a reasonable amount of ambient light, the image of my gray card on the screen should look very much like my gray card does in that ambient light. This will tell me, roughly, whether or not my monitor needs to be calibrated. If you're using a color calibration system like the Pantone system then you probably won't need to calibrate your monitor. In any case, it'll tell you what Zone V gray looks like on your monitor and in the environment you'll be using for the rest of the shots in your room.

Then I'll start shooting objects in my room. I'll take a guess at how much expousure compensation to use, take the shot, and after doing this a few times, pull the resulting shots up on the computer. Then I can see whether what I see on the monitor matches what I see in the room and adjust my idea of how many stops of compensation any given shade I see requires.

I'll perform the same experiments outside as well, to get an idea of whether or not my judgment works in that environment (it seems to, more or less).

The more you shoot like this, the better your exposures will be because you'll be better at assessing the brightness of an object, how that should relate to the rest of the shot, and how much compensation to use to get it to look right.


I would expect you can do roughly the same thing with the full-frame metering modes, too, but the problem is that they're averaging the entire scene, so you have to know from experience how much of an effect each part of the scene is going to have on the meter's assessment of the scene. This may take you a long time, or it may be something that you just "get" quickly. One thing that'll help: if the scene is composed mostly of relatively bright objects and you want them to show up as being bright in the final picture, then you're going to have to add some exposure (since the meter will try to make it look like middle gray). If the scene is predominantly dark then you'll have to compensate the other way. Only experience will tell you how much compensation you'll need.

One more thing that might end up helping, but which I haven't tried myself yet. Your camera's metering system doesn't know a thing about colors, only intensities. That means, for practical purposes, it sees "black and white". Which means it might prove instructive to shoot in black and white, and pay particular attention to how objects which are already relatively colorless (i.e., white, black, or some shade of gray) look both in real life and in your shots. If you're shooting in RAW then you can get the best of both worlds, because even if you take a picture with a black and white picture style (which will show up on the camera's screen as a black and white picture), the RAW image will retain all the color information, and you can see what the picture looks like in color and black and white.

Anyway, I hope this makes at least a little sense, and I certainly hope that it's not riddled with errors!