Exposing Sky vs. Foreground

[Suzumushi]

I've noticed that with landscapes, I keep getting a sky full of blown highlights. Even in the evening, with exposure compensation down at -2 I managed to blow out the sky in one case. However, I noticed that this only happens when I focus on the foreground; if I focus on the sky, it exposes the sky properly but leaves the foreground in shadow.

Ok, so the camera is trying to expose whichever area I focus on as best it can, at the expense of the other area. If it has to make a sacrifice somewhere, I guess it makes sense to do it this way.

My question is, short of using a graduated filter, is there a way to autofocus on the foreground but still expose for the sky? Does this have something to do with "metering?" I may not mind deep shadows in the foreground as much as white blotches in the sky. Or can I get a wide enough depth of field with a small aperture to focus on the sky and still get the foreground sharp?

 

[nynfortoo]

My question is, short of using a graduated filter, is there a way to autofocus on the foreground but still expose for the sky? Does this have something to do with "metering?" I may not mind deep shadows in the foreground as much as white blotches in the sky. Or can I get a wide enough depth of field with a small aperture to focus on the sky and still get the foreground sharp?

Your camera is metering, and subsequently exposing for, whatever you're focusing on. Try looking into your Exposure Lock function—or switch to manual mode (presumably you're in auto/semi-auto mode), meter off whichever, then recompose.

Focus and exposure are separate issues.

Have a read of this to learn about metering, and this to learn about Depth of Field.

 

[tirediron]

An age-old question... of course the best way is to use graduated ND filters (A couple of Cokin 'P' size, holder and apaptor ring will weigh in at under $100, so it's a pretty economical way too). If however, you can't or don't want to purchase them, then your best bet is going to be to expose for the highlights, in this case, the sky.

What I would suggest is with your camera set to 'Manual', and your exposure meter set to either 'Centre-weighted' or 'Spot', meter the sky (The term "metering" simply means the use of a light-meter, wether hand-held or in-camera to determine the required exposure).

From there, take a series of test exposures, gradually increasing the exposure, and watching your histogram. Once you get to the point where you're seeing blown highlights/clipping, then that's as far as you're going to go with the sky. Hopefully that will give you enough exposure latitude that you can brighten up the foreground in post.

Another alternative is to do an HDR merge; make a number of exposures of the scene, starting at the darkest and working ot the lightest (in 1-2 stop increments) and then use software such as Photoshop, PSP, or other to put them together, giving you an image with a vastly greater dynamic range than could be captured at once.

Bear in mind that anytime you have a white, cloudy sky, it's going to be very difficult, if not impossible to get good foreground expsoure in one image.

With respect to your question on DoF (which is totally unrelated to the one of exposure), it depends on how near the foreground is that you want sharp, what kind of lens you're using, and what apeture you can stop down to.

Hope that helps,
~John

 

[Suzumushi]

"meter off whichever, then recompose."

This is what I have been doing, but as far as I can tell it only meters when I autofocus. Is it ok to focus on the sky but then shoot the foreground? It seemed like that should leave the foreground at least slightly out of focus unless everything is very far away. Especially in the evening when I need to use a wider aperture. That's why it seemed like the two issues were related. I guess the distances would usually be large enough that changing the angle of the camera a bit won't have a noticeable effect? I suppose to really do it properly I should use a narrower aperture and a tripod. Until then I'll check out those links and try the different metering settings.

 

[djrichie28]

Sounds like your meter is set to 'spot' If you can change it to 'matrix' then you should be in business. Spot metering usually meters only on what focus box you are using to focus. But that depends on what camera you are using.

 

[nynfortoo]

"meter off whichever, then recompose."

This is what I have been doing, but as far as I can tell it only meters when I autofocus. Is it ok to focus on the sky but then shoot the foreground? It seemed like that should leave the foreground at least slightly out of focus unless everything is very far away. Especially in the evening when I need to use a wider aperture. That's why it seemed like the two issues were related. I guess the distances would usually be large enough that changing the angle of the camera a bit won't have a noticeable effect? I suppose to really do it properly I should use a narrower aperture and a tripod. Until then I'll check out those links and try the different metering settings.

You don't have to meter off your focus point, even if your camera requires a half-pressed shutter to actually meter something.

Check out your manual for 'Exposure Lock'. This will allow you to get your metering done by focusing on something, lock the exposure to that part of the scene, then recompose without the settings changing; you can then focus on whatever you want without messing with your exposure.

 

[Suzumushi]

I've been using the default evaluative metering, I tried the others just to see what would happen and they tended to make it worse. Of course, I didn't really know how they worked or where to point the camera. I think the exposure lock answers my question... I'll check the manual.

 

[Mav]

Cheap $40 or less screw in grad ND filter. Problem solved. And yes, it sounds like your camera is set in center weighted or spot metering mode. If you flip it back to Matrix it should give you about the same exposure either way.

 

[Suzumushi]

It was definitely in evaluative meter mode. I imagine by focusing on the sky, much more of the image became bright so the evaluative metering exposed the sky better even though it was evaluating the whole image. I just checked the manual and apparently the exposure lock is what the * button is for. Next time I'll try that.

 

[tirediron]

Cheap $40 or less screw in grad ND filter.
<snipped>

Ehhh... allow me to dispute my learned colleague's recommendation if I may. IMO, cheap filters are often worse than no filter at all. As the very first element through which light passes, these should be the best quality you can afford. Buy used to save money, but always buy quality. With respect to the issue of screw-in grad NDs, I would suggest the Cokin-style gels since these allow you to place the transition wherever you want in the image, affording greater flexibility in composition. You can also stack them with the transitions at different levels giving you even greater range.

 

[Mav]

Ehhh... allow me to dispute my learned colleague's recommendation if I may. IMO, cheap filters are often worse than no filter at all. As the very first element through which light passes, these should be the best quality you can afford. Buy used to save money, but always buy quality.

Here's a recent photo taken with my 18-55VR kit lens and a "cheap" screw-in Tiffen 0.6 ND Grad. Do you consider Tiffen to be a cheap brand? How bout Hoya?

That's straight off the camera too (D80).


It's a flat piece of glass. It's not a curved "element" and is ridiculously easy to make well and cheaply. The comparison of screw-in filters to the front element of a lens is completely irrelevant. Is it bending or shaping the light at all? No. It's passing the light straight through to your lens, albeit attenuating it on the top half. So long as there are no irregularities in the filter glass as far as thickness variations or surface roughness or bubbles or anything and the mechanics hold the filter in good alignment, there isn't going to be any difference in image quality. I've shot with non-coated, single-coated, multi-coated, super-duper-pro multi-coated, and they've all worked fine. The only issue I have is occasional filter-induced ghosting issues at night from street lamps or other bright light sources against dark backgrounds, usually with a UV filter. No problem. I pop the filter off, take the shot, and then put it back on.

I guarantee you that this photo would have been far worse without a filter, even compared to the cheapest off-brand grad ND you can find on eBay (which I wouldn't recommend doing, stick with Hoya, Tiffen, Promaster, etc.) I would have had to shoot in RAW, do a lot of tweaking in photoshop later, but the only thing I like to do with my photos once they're on my computer is enjoy them and not tweak them.

Without the filter I had my choice of a properly exposed sky but a black beach, or a properly exposed beach but a white and blown-out sky. This is what Grad ND filters are for, and no you don't need to go with a clumsy Cokin system to get quality in a filter.

 

[tirediron]

Whoa Mav, whoa! When you said "Cheap" I meant the cheap nameless "brands" that you see on eBay and places like that. Tiffen, Hoya, B+W, Heliopan among others are very good quality, and even their economical lines are better than the "no-name".

I was also not extolling the virtues of Cokin for their quality, but rather there versatility. In my opinion, being able to move the transition zone anywhere [vertically] on the image is a plus that outweighs the admitted PIA of dealing with their holder.

It's a flat piece of glass. It's not a curved "element" and is ridiculously easy to make well and cheaply. The comparison of screw-in filters to the front element of a lens is completely irrelevant. Is it bending or shaping the light at all?

I do however choose to disagree with that statement. It is an element by virtue of the fact that it is now, for all intents and purposes a component of the lens; granted it's not shaped (or it shouldn't be) but it is the very first thing the light passes through to get to your sensor. I firmly believe that saving money on anything that gets between the lightsource and the sensor is a bad idea.

I guarantee you that this photo would have been far worse without a filter, even compared to the cheapest off-brand grad ND you can find on eBay

Remind me when I get home to send you a couple of the "Cokin compatible" G-NDs I bought through eBay a year or two ago; I think you'll really enjoy the light blue caste they add to your photos!

 

[Mav]

I've seen it suggested on this forum that you should put no less than a $150 filter on even the KIT LENS, which is probably one of the most ridiculous things I've ever seen posted here, so people have WIDELY varying perspectives on this. You and I are not too far off. When I say "cheap" I mean stick with the name brand stuff and don't go for the ebay crap, but don't feel like you have to put a super duper fancy shmancy ultra multi-coated pro filter on every lens you've got, and for every filter type you need.

 

[tirediron]

I've seen it suggested on this forum that you should put no less than a $150 filter on even the KIT LENS, which is probably one of the most ridiculous things I've ever seen posted here, so people have WIDELY varying perspectives on this. You and I are not too far off. When I say "cheap" I mean stick with the name brand stuff and don't go for the ebay crap, but don't feel like you have to put a super duper fancy shmancy ultra multi-coated pro filter on every lens you've got, and for every filter type you need.

On this we definitely agree (and apologies to the OP for highjacking his thread) but don't forget used! Filters are a great thing to buy used and you can save up to 75% of the new price, as long as they're scratch-free and not cracked, they're as good as new!