PDA

View Full Version : Light falloff on wide angle lenses.



AnselAdamsX
10-Oct-2006, 11:53
I've been having trouble getting the correct exposure in my photos. I thought it was just me but now I'm wondering if it is due to light falloff on my wide angle lenses. I have a Nikkor 90mm f/8 and a 120mm f/8. And I have been using extreme rise. Enough that I have some vignetting int 2 corners. I resently tried some macro shots with a Schneider 210mm f/8 and it looked like the exposure was correct. How much light falloff can you expect from a wide angle lens? 1 stop? And how far from the center of the lens does it begin to become noticable?

Thanks
Chris

Nick_3536
10-Oct-2006, 11:57
What format? The Nikon 120mm covers 8x10 basically. Are you pushing that on 4x5?

Brian Ellis
10-Oct-2006, 12:26
"How much light falloff can you expect from a wide angle lens"

There's no fixed rule, it depends on the size of the lens' image circle and the type and extent of the movements you're using and probably other things that don't come immediately to mind. However, I think the key in your situation is the extreme front rise. When you use any front movements you're moving the center of the lens away from the center of the film. So instead of having the heart of the image circle centered around the film you're moving the portions of the image circle that are closer to its edge into the film and using them to form a part of the image. Obviously the more you move the lens the more of the areas near the edge of the image circle are used to form the image. Since it's those areas that exhibit the greatest amount of light fall-off, when those areas are used to make the image you increase the risk of vignetting.

You can greatly minimize the problem with tilts and swings by using the back rather than the front since with tilt and swing of the back you aren't changing the relationship between the center of the lens and the center of the film. However, that doesn't work with rise, fall, and shift since even when using the back for them you're changing the position of the center of the lens relative to the center of the film.

Ron Marshall
10-Oct-2006, 13:12
Here is a link to a page on this site about center-filters:

http://www.largeformatphotography.info/filters.html

Ole Tjugen
10-Oct-2006, 14:03
There are basically two kinds of wide-angle lenses, which can be somewhat siply summed up as "big ones" and "small ones". "Big ones" are those with big oversize front and rear elements, like the Super-Angulon. The "small ones" are the ones which look more "normal" - like Angulon. The Super-Symmar XL is "small" even if it has a large front element,since the rear element is small (I use Schneider lenses as an example, for the simple reason that they're the only ones I don't have to look up!).

Very simply the "small ones" have a light fall-off corresponding to cos^4, the "big ones cos^3. So at 45 degrees from the optical axis the "small ones" will transmit (cos(45))^4, or 0.25 times the on-axis illumination - two stops loss. A "Big one" transmits half a stop more.

Using extreme shifts cam well make this noticeable even when using a 120mm on 4x5". The only way to improve this significantly is with center filters.

AnselAdamsX
10-Oct-2006, 14:04
I'm using it on 4x5. I thought it only covered 5x7. I'm using a monorail and I have 45mm front rise and 55mm rear fall. I believe I only have 80mm of shift on a vertical shot before I reach the edge of the image circle. The light falloff on the 120 is not as bad because of the larger image circle and not needing as much rise if I can stand far enough back to use that lens. I also have been using a circular polarizer. I will stop using it to see if it is causing some vignetting. I don't see a problem with the 90mm if I'm only using a little swing. I just been taking pictures of buildings mostly.
What format? The Nikon 120mm covers 8x10 basically. Are you pushing that on 4x5?

Nick_3536
10-Oct-2006, 14:28
I think at F/22 the 120mm is rated at 310mm of coverage by Nikon. You can check the Nikon website. It's not a very wide lens on 4x5.

To the best of my knowledge polarizers cause problems on wide lens but maybe I'm remembering that wrong.

AnselAdamsX
11-Oct-2006, 08:37
So how do you determine where you will see 0.5 stop of light fall-off for a given lens on a given format? So on a 90mm lens on 4x5 format I should expect to see 1 stop at 33 degrees and 0.5 stop at 22 degrees? I attached a graph of the function.
There are basically two kinds of wide-angle lenses, which can be somewhat siply summed up as "big ones" and "small ones". "Big ones" are those with big oversize front and rear elements, like the Super-Angulon. The "small ones" are the ones which look more "normal" - like Angulon. The Super-Symmar XL is "small" even if it has a large front element,since the rear element is small (I use Schneider lenses as an example, for the simple reason that they're the only ones I don't have to look up!).

Very simply the "small ones" have a light fall-off corresponding to cos^4, the "big ones cos^3. So at 45 degrees from the optical axis the "small ones" will transmit (cos(45))^4, or 0.25 times the on-axis illumination - two stops loss. A "Big one" transmits half a stop more.

Using extreme shifts cam well make this noticeable even when using a 120mm on 4x5". The only way to improve this significantly is with center filters.

Ron Marshall
11-Oct-2006, 09:39
So how do you determine where you will see 0.5 stop of light fall-off for a given lens on a given format? So on a 90mm lens on 4x5 format I should expect to see 1 stop at 33 degrees and 0.5 stop at 22 degrees? I attached a graph of the function.

The distance from the center of the image to the point at which light fall-off is equal to 1 stop, (using 33 degrees, and focal length of 90mm and infinity focus) is equal to:

90mm * tangent(33) = 58.44mm

The circle inside which falloff is less than 1 stop is double this, or 116.9mm, or about 76% of the image circle required for 4x5 film.

Michael S. Briggs
12-Oct-2006, 08:55
So how do you determine where you will see 0.5 stop of light fall-off for a given lens on a given format? So on a 90mm lens on 4x5 format I should expect to see 1 stop at 33 degrees and 0.5 stop at 22 degrees? I attached a graph of the function.

cos^4 33 degress = 0.5, so 1 stop. But the Nikkor-SW series uses the large, negative outer elements just like the Super-Angulon series, and thus the tilting pupil trick to improve off-axis illumination. So cosine to the fourth isn't the right equation to predict the falloff for the 90 mm lens that started this question. Probably cosine to the third power is a good approximation. This only changes the angle to 37 degrees.

The differences become larger further off-axis. Frequently photographers will not notice a 1 stop change in illumination, so the performance further off-axis is more interesting. There are posts where people say that some lens has "no falloff" when that isn't true for any non-fisheye lens that I know of.

Rodenstock and Schneider publish graphs of the illumination on their datasheets. But even though Fuji and Nikon don't do the same, we can judge that the performance of lenses of similar design are similar.

Are you using transparency or negative film? Negative film has more exposure range and so could cope with this better. Since the lattitude of negative is on the side of over exposure, trying giving extra exposure to get sufficient exposure to the corners.

There are previous discussions in the archives, e.g., http://www.largeformatphotography.info/forum/showthread.php?t=17295, http://www.largeformatphotography.info/forum/showthread.php?t=17205 and elsewhere, http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=005gK2