Conundrum: exposure compensation for rise?

[Leigh Perry]

I have made a couple of attempts to shoot an ultra-wide angle shot with lots of front rise. My first attempt was wildly underexposed -- perhaps by 2-3 stops. This wasn’t a metering error. On the second attempt, I bracketed +1 stop, +2 stops, and +2.5 stops. I got the shot with the +2.5 exposure, but don’t understand the reasons.

The lens is a Schneider 38XL using roughly a 6x9 crop out of a 6x12 frame. I used a centre filter. The lens just ran out of image circle at the top of the photograph.

I suspect the cause must be some exposure compensation requirement for rise-type movement, perhaps exacerbated by the extreme hot-spot characteristics of the 38mm and the use of the image circle extremity. I’ve never seen any details of such compensation though, while there is plenty of mention of bellows-extension-related compensation.

Can anyone shed any light on the cause? If it helps, the shot is here (http://www.breakpoint.com.au/lpsite/temp/612-039-06-Harbour-bridge-u.jpg).

[Bruce Watson]

I've never used a center filter before, but it would seem to work by putting a neutral density filter over the center, and feathering this density out to the edges as required for the lens in question. IOW, I think you have to have exposure compensation for the center filter. I don't think it's a problem with using movements.

[David A. Goldfarb]

Front rise per se doesn't require compensation, but if there is falloff with an ultrawide lens, then you need to take account for it. If you're using a center ND filter, then you shouldn't have falloff, but you need to add the filter factor for the amount of density at the center of the image circle.

[Bob Salomon]

Yo use a center filter the lens must be stopped down at least 2 stops for the filter to have any effect. Then you need to add the filter factor of 1 1/3 to 1 1/2 stops. Was this done? Lastly 35 aand 38mm lenses do not allow much movement. The 35mm 120° Apo Grandagon allows 20mm of rise and 15mm of shift with a 6x9cm in horizontal orientation.The 38mm isn't much different in coverage.

[Brian Ellis]

Here's a highly untechnical explanation. When you use a wide angle lens the light from the outer portion of the image circle has a greater distance to travel and travels at a steeper angle to the film than the light from the center portion of the image circle, which can lead to under-exposure of the portions of the photograph at the outer edge of the image circle (which you already know). When you combine a wide angle lens with a lot of front rise you're exacerbating this situation, i.e. you're moving part of the image circle even farther from the film than it already was, thus leading to even greater potential underexposure. That's my understanding of the situation anyhow. If it's wrong please don't correct me because it makes sense to me and I'd like to keep it that way without being confused by the actual facts if they differ from my understanding.

[Henry Ambrose]

I think you are asking about the darkening at the very top edge of the frame, as the rest of the frame seems good? Assuming that you used the appropriate compensation for the center filter, it looks like to me that you simply ran out of image circle.

[Michael S. Briggs]

Both Schneider and Rodenstock publish curves of the illumination vs angle off-axis for their lenses. The curves typically follow closely the theoretically expected cosine to the third, or cosine to the fourth, functions expected. I've explained this theory in previous postings. From these curves, or from the theory, one could calculate exposure compensations when using rise. But this isn't applicable to your photo since you used a center filter.

I know of only one illumination curve published by a manufacturer for the case that a center filter IS used. The Schneider instruction brochure that accompanies their center filters shows the relative illumination vs distance from the axis for the 90 mm f5.6 Super-Angulon, both with and without center filter. The curves show that the center filter helps greatly, but does NOT result in uniform illumination. For example, by the corners of a 5x7 film, there is about 2 stops falloff without a center filter and 0.9 stops with the center filter.

Scheider explains this in the brochure: "In the interest of not extending the exposure time excessively and of fully exploiting the exposure latitude of the negative layers, no attempt was made at achieving full compensation of the brightness differential."

I think probably all manufacturers of center filters follow this approach of not full correcting the uneven illumination -- the recommended exposure corrections are about the same from all manufacturers.

Did you use Schneider's IIa center filter? Schneider recommends 2 stops exposure compenstation for this filter. I have a different Schneider center filter and find that I need 1/3 to 1/2 more compensation than recommended by Schneider -- perhaps their are manufacturing variations, or perhaps I just judge differently than Schneider.

As Bob stated, the Schneider brochure also recommends stopping the lens down at least two stops. This is to get past vignetting and to something close to the best relative illumination possible according to the lens design. In the case of the Super-Angulon, that would be cosine to the third performance.

There are several aspects that aren't clear from your question. When you say that you applied +1 to +2.5 stops, if that from your normal metering, or in addition to the +2 stops needed when you use the center filter. Are you using transparency or negative film?

Using the lens to, or just past, its coverage limits, means that parts of the film are still receiving substantially less light than the on-axis location. With a negative film, you might want to apply additional exposure to correctly expose the extreme off-axis areas, allowing overexposure of the center. A negative film will work better because it has much more exposure latitude than a transparency film, particularly on the over-exposure side.

The extreme contrast of the scence you photographed is exacerbating the effect.

Some additional good advice from Schneider: "It is just as well to remember, that, whenever wide-angle lenses are used, high contrast negative material, under-exposure and strong contrast should, as far as possible, be avoided." Transparency films are more to be avoided than high-contrast negative films.

[Bob Salomon]

"I think probably all manufacturers of center filters follow this approach of not full correcting the uneven illumination -- the recommended exposure corrections are about the same from all manufacturers."

Heliopan was making 2 different density center filters. The standard 0.34 density and a 0.9 density. The 0.9 did result in a more even image but required a 3 stop factor. Due to the loss of 3 stops and the need to stop down at least 2 stops this was not a very popular filter and was discontinued. However, depending on the size required, we still do have some left in stock.

[David A. Goldfarb]

You can sometimes use this falloff to your advantage by composing around it. Use rise, fall, and shift to put the hotspot over the subject of the image or to darken the sky in a landscape. I think Adams mentions this technique in passing in _The Camera_.

[Leigh Perry]

Thanks for the responses.

You use a center filter the lens must be stopped down at least 2 stops for the filter to have any effect. Then you need to add the filter factor of 1 1/3 to 1 1/2 stops. Was this done?
Yes. I used the 38XL's centre filter and added the requisite two stops. The exposure compensation was in addition to the centre filter's 2 stops. I don't believe it is a metering error.

Brian's non-technical explanation matches my theory regarding lens hot-spot / light fall-off.

[Leigh Perry]

I think you are asking about the darkening at the very top edge of the frame, as the rest of the frame seems good? Assuming that you used the appropriate compensation for the center filter, it looks like to me that you simply ran out of image circle.

That is what happened. It was the way I planned the shot: to shoot as far up under the bridge while keeping the buildings straight. But that's not the phenomenon I was wondering about. I wanted to understand and quantify the need to radically overexpose in this setup.

[Dan Fromm]

Please tell us how you metered.

Thanks,

Dan

[Leigh Perry]

Please tell us how you metered.

I spot metered, setting the brightest part of the sky (lower right) at +1.5 stops relative to grey. But, on my previous botched attempt, when I used exactly the same metering approach, all was grossly underexposed, including the brightest part of the sky that I had metered.

Following up Michaels helpful comments regarding medium... I shot Provia 100F, but what you see is pretty much as it looks on the lightbox -- perhaps a touch darker. I used the Photoshop CS highlight tool to subdue the building highlights a little. I consider the final exposure to be pretty good, and Provia's latitude is enough to meet my purposes.

Apart from the specular highlights of the city buildings, which I expected, I'm happy with the exposure. I'd just like to be equipped with a bit more science if another shot like this rises.

[Michael S. Briggs]

The most scientific thing that you can do is to remember that at max rise, using your metering techniques, you should add another 2.5 stops of exposure. You have scientifically determined this in an experimental manner.

There is no easy way to scientifically determine the exposure compensation from theory. If you weren't using a center filter, there are simple calculations that would give an answer. But you are using a center filter, and since the manufacturer has choosen to partially correct the illumination falloff with the center filter, one either needs data from the manufacturer, or to measure the center filter and do some more difficult calculations. The simplest scientific approach is to take photos and learn what compensation is needed.

[Dan Fromm]

Thanks for the explanation of how you metered.

The underside of the bridge, which I gather is where you wanted to capture detail, is many stops under the brightest part of the sky. Why did you expect that the exposure needed to get the sky right would also be right for the underside of the bridge? And why did you use a narrow latitude transparency film?

You might have been better off with a color negative film, since most of 'em have more latitude than color transparency films, and a lot of darkroom manipulation to get the print you want.

Cheers,

Dan

[Leigh Perry]

Dan, there were a couple of reasons for spot metering the way I did: one aesthetic and one prosaic. Mostly, I wanted to capture the rich twilight sky and let the rest of the levels fall where they landed. Secondly, that was the only place I could get a reading! Under the bridge and throughout most of the rest of the frame, my meter said 'Eu'. The city lights registered on the spotmeter, but I doubted that I could make a cogent metering decision based on that reading.

The final exposure was 4 minutes. Great thing about shooting LF -- closing the shutter while those Lavender Bay ferries cross through your shot!

I should say that I'm not unhappy with the final result. Aesthetically, I'm pleased with the level of under-bridge detail in comparison with the rest of the shot.

Thanks to all for your help & interest.

[Henry Ambrose]

I've never had to apply 2 stops (or any) correction for maximum rise with any lens. Sure the top edge of the frame will show some fall off as you come near to the lens's coverage but not the overall underexposure you claim.

Your subject matter is quite difficult and I think the wrong test to generate a definitive answer. I think you should shoot an evenly lit subject with max rise and see what really happens. Yes, a "blank wall test". Until I hear the results of this test I'm not believing in added exposure, as I have seen otherwise. That said there may be some oddity about the 38XL that I don't see with other super wide lenses (widest I have is a 58mm) but I think physics pretty much works the same for all of them.

Now for more - I wonder why you measured only the sky? I would have metered several parts of the scene to see if they all fell within the film's range. You may have done so but you have not said so. Also as Dan stated, if there was ever a good reason to use negative film this is one.

Last, what about reciprocity? Just how long was the exposure and did you adjust?

[Mark Sampson]

Another point- with dusk/twilight shots, there can be a difference in scene brightness between the time you meter and the time you actually expose- it's getting darker all the time. I'd repeat a full-rise shot in full daylight and then see if any exposure increase is needed. I'd bet not.

[Leigh Perry]

there can be a difference in scene brightness between the time you meter and the time you actually expose- it's getting darker all the time

Mark, that's true, but I take a lot of long shots in changing (mostly dawning) light and I make it a habit to spot meter throughout the duration of the composition. The meter reading was unchanged at the end of the 4 minutes, at least within the 0.5 stop display quantisation.

Henry, your incredulity matches mine :) I gave details of the exposure above, but it included roughly half a stop for reciprocity failure prior to the addition of the 2.5 stops.

[Leigh Perry]

FWIW I based the reciprocity correction on this empirical graph, which I found while researching the cause of my previous ignominious attempt:
http://home.earthlink.net/~kitathome/LunarLight/moonlight_gallery/technique/curve_3.gif(see here (http://home.earthlink.net/~kitathome/LunarLight/moonlight_gallery/technique/reciprocity.htm#provia2) for full details).

I erred on the generous side.

[Robert Lawrence - Fivebyfour Commercial Photography, UK]

It seems most likely from what you say about Eu as the indicated exposure on most of the scene that your meter is the source of the error and is insensitive and non-linear at this end of it's capability.

[Leigh Perry]

The lightmeter is a Sekonic 608. I frequently use it to meter predawn exposures in the range from 4-10 minutes, without problem. This is the only scenario where I have experienced the exposure problem, but then I don't often have a need for 38mm lens on 6x12 with maximum rise :)