18/04/2014

Decoder April 2014 - Field Curvature


It seems almost impossible, but this image was taken with a 28mm lens on a "full frame" camera at f3.5 yet has virtually everything in perfect focus, it is the unusual field curvature characteristics of the lens used that makes this possible!


Field Curvature,  A Tricky Bastard


Up front let me say "field caurvature has nothing to do with images where straight lines in the frame are bowed in or out as you move towards of the corners", that would be distortion. The name field curvature unfortunately leads many folk to think that field curvature and distortion are the same thing!


Instead of photos as I would usually use to explain ideas and techniques I have drawn up some diagrams to explain the concepts, actual photos at web resolution are simply not viable for demonstrating what we are looking at.


With a perfect lens the plane of focus would be a perfectly flat plane parallel to the cameras sensor. Specifically you would be able to focus on a flat object which is parallel with the sensor plane and it would appear sharp from edge to edge, corner to corner. Some lenses can indeed get close to this, generally macro lenses which have to work in such a way if they are to succeed in their intended purpose, which generally includes photographing flat artwork and similar subjects. In most cases the plane of focus for most general purpose lenses is anything but "planar".
(You may now be able to understand what Zeiss mean when they add the "Planar" moniker to some of their lenses.)




Flat field rendering, in other words no field curvature.


Simply put, field curvature is where a lens does not focus planar objects with equal clarity across the width of the subjects area, in other words the centre may be sharply focused at say 5 metres but at the edges of the frame objects at 5 metres are "out of focus" but objects at say 10 meters may actually be in focus. Field curvature can work in both directions, you could have the centre in focus at 5 metres and the edges in focus as say 2 metres.

The quick and easy to see give-away of field curvature is that should you focus on a flat field object, for example a brick wall at the wide open aperture only the middle point where you focused will ever appear sharply defined.

Digital capture is far less accommodating of field curvature than film was, principally because the resolution levels are so much higher thus any variation in the actual plain of focus now stands out in sharp contrast.

MTF charts, which are often used to judge a lenses performance quite readily show the effects of field curvature, but I must caution you that using an MTF chart alone as a means of judging lens performance is no where near as instructive of real world performance as you might think, unless of course you intend to use your new lens solely for the purposes of photographing flat field objects, like brick walls.

Field curvature may cause the focus plain curve away from the camera at the edges or it may come closer to the camera at the edges and the degree of curvature can in some cases be very strong or quite weak.  Some lenses produce a complicated form of curvature where  it may curve outwards as you move away from the centre and then move inwards again at the corners of the image, this is known as wave type field curvature and is a true devil of an issue to deal with or use creatively.




Uneven field curvature, this is very difficult if not impossible to use creatively.


The effect of field curvature cannot be underestimated, it makes for some very unpredictable results and may lead you to make totally wrong assumptions about your lenses performance. Probably even more frustrating, field curvature makes a complete mockery of depth of field scales and calculations which of course are created on the assumption that the lens is planar in its depth of field rendering.

But wait, there's more problems in store, field curvature for zooms varies with focal length   which makes it quite impossible to use the classic "focus long the reframe wide" techniques accurately with many lenses, this is especially the case with many wide angle zoom lenses. I have one lens where the field curvature radically changes from rear biased to flat and then forward biased with just a shift in focal length of 3 mm!

And yet more....field curvature is not normally consistent at all focused distances, it could well be quite flat at close up distances but become more curved as you move towards infinity, some macro lenses perform exactly in this way, or course it could work in the opposite way as well.  Field curvature may also mess with infinity performance, I have found that sometimes a lens may have to be focused beyond or well in front of infinity at wider apertures to give correct overall infinity focus due to field curvature anomalies.

Many lenses that are regarded as having truly bad edge definition are in reality very strongly effected by field curvature. In some cases the lens may have to be focused at just a couple meters or so to bring the extreme edges into focus, but because everyone focuses for the centre or there abouts the assumption is the dreadful corner resolution is simply an issue bad lens design. Strangely enough if you actually take a lot of these so called bad lenses, focus at a point just in from the corner and have a good look you will find surprising amounts of actual resolution. 


Rearward tending field curvature


It is normally possible to actually get a shot with sharp corners from lenses with even strong field curvature, you just have to completely rethink your focusing methods and work outside the optical box.

These days mirrorless cameras with good quality EVFs can help enormously as if you have the time when shooting say landscapes, you can zoom in on several parts of the image to find out if everything is sharp and readjust the focusing point accordingly to compensate.

Yes indeed field curvature is a strange beast indeed and it is probably responsible for a good number of otherwise decent lenses being canned as useless simply because the photographer had no real idea of what was going on and how they might work around it.

Of all the lens types, macro lenses are the least likely to give you grief, but generally long telephotos are quite good also in the FC department largely because their field of view is much narrower.  The wider the maximum aperture of the lens the more likely it is to have significant field curvature unless the maker has gone to enormous trouble and expense to correct it with additional lens elements.  As you might expect, when you combine significant field curvature and very shallow depth of field such as when shooting fast primes at wide apertures the rendered results are often pretty dismal, in fact nailing good focus is often something of a lottery in these instances.

So far the news sounds all bad, and it is if you are shooting flat targets like test charts or art works (in which case you really do need a proper macro lens), but field curvature can actually be used to your advantage if you know something about the curvature characteristics of the lenses you are using. And trust me on this, pretty much every lens you own will have some FC, so you need to test for it!




Forward tending field curvature 


Consider the Possibilities

Lets say you want to take a photo of a street scene, you have a subject in the middle and buildings and other interesting details running along each side of the image into the distance. Now imagine you want the person in focus and separated from the distant background by having a shallow depth of field, but you actually want the detail moving away from the camera along the edges to be in sharp focus. 

If the focus area was planar the above is an impossible situation, the requirements are mutually exclusive, but if the lens has strong field curvature where as the plane of focus moves rearwards away from the centre of the image you could achieve just such an effect, even a wide aperture.

If the edges focus closer towards the camera of course the effect would be the opposite of what you would need and it would work against you making the edges far blurrier than you’d expect, but then again that might be a creative option as well providing far shallower depth of field for any given aperture than you might expect.

Lets take another scenario, imagine you want to photograph a relatively close object but you want to extend the depth of field at the bottom of the image back towards the camera, now with a view camera we could uses tilts to do such a thing, but think about this. If the lens has strong inwards field curvature it will allow you to gain an extra degree of depth of field for any given aperture for the bottom part of the frame, kind of like getting a tilt effect without the tilt!
Many of the difficulties of field curvature can be counteracted by not focusing in the middle of the frame but rather choosing a point somewhat out from dead centre, this will of course put the critical focus at a different point but importantly it may make your overall focus better.

Naturally like most things optical, stopping down mitigates the effects of field curvature, with some lenses the wider apertures may well be rendered practically useless by FC as for all but a few types of shots and you will be forced to stop way down to get acceptable overall focus despite the DOF scales telling you it should not be needed.




Infinity focus for rearward tending field curvature often works better if actual focus point is moved back towards the camera somewhat and allowing DOF to do its work.


So how can you test for field curvature?
The methodology of testing for field curvature is quite simple , it's the finding of suitable target subjects that will mess with your head. 

The ideal test target would be a flat area of ground covered in large coarse which gets larger as it becomes more distant but is evenly spaced and the area needs to be wide enough to fill the view of your widest focal length lens, got one of those out back? I thought not!

So what can we practically use?

You are going to need different targets for different focused distances, here is what I suggest and use. 

Break it down into three separate target scenes, close up, intermediate and infinity.

For the close up test you will need a large flat surface with strong textural content, a course textured rug would be a good choice. Place the carpet on a flat surface outdoors and then place your camera on a tripod which is lowered down to around 45 cm off the ground and angle your camera so it is looking along the depth of the rug. Make sure the carpet if filling your entire frame right out to the edges and corners, if it doesn’t, rearrange your set up till it does, wide angles lenses will cause the most consternation as always. 

Now, place an object on the carpet to use as the central target, something with lots of fine detail (I use an old circuit boards out of a computer). This object needs to be parallel with the cameras sensor, thus I tilt the circuit board a little.

Now all you need to do is focus (manually) on the central subject and take a series of shots at differing apertures and focal lengths (if it's a zoom). The shots taken at the widest aperture will be the most telling, but the others will also be useful, for example you may find at certain apertures the field curvature changes or even reverses.

You should be able to see if the focus curves back beyond or in front of the central focused point, at some point depth of field will likely be enough to counteract the effect, which by the way is a handy thing to know.

For the middle distance shots I use a graveled garden area to the side of my home, the gravel is multi toned and about 12mm in size, in other words perfect for the job! In my case the area is wide enough and deep enough to accommodate all my lenses, if you look around I am sure you find a location that gives a similar target, an area of rough tarmac could be pressed into service (try an avoid being run over of course). I set the camera up so its about 60 cm off the ground and place a target 1/3 of the way in from the start of the gravel, basically this represents shot you would commonly take like group portraits, photos of largish objects ( cars/motorbikes) etc   As before take a sequence of shots at all apertures and focal lengths.

Again it should be easy to see how the field curvature works, and in the scheme of things this is probably the most important of the three tests as it represents where field curvature is most likely to cause you grief in real world shooting.

Finally the infinity test. You will need a distant landscape which has some distinct near elements as well, a good example would be a field which is bordered to the rear by a row of hills. Here the testing is a little different. 

Set the camera up so the line of the tops of the hills runs through the centre of the frame. We are trying to determine if when the centre of the image is focused at infinity the edges of the frame are also focused at infinity, beyond infinity (yes beyond infinity is possible) or much closer towards the camera.

Should you want to be even more demanding you could hold the camera (perferably on a tripod) so it is angled thus having the hills or mountains run diagonally across the frame.

To carry this out take the shots at the wider apertures only, say maximum aperture and up to f7.1, there is little point in shooting distant landscapes at small apertures despite the fact that many folk do.  

First focus at infinity and take a sequence using the available apertures , then focus beyond infinity if possible (some lenses do actually go beyond infinity) and repeat, then come back from infinity a bit and repeat again. You will need to have a good close look at the results zoomed in to 100 or 200% on screen to see how it worked out, the differences in this case will be subtle.

In the infinity case, a lens that focuses too close on the edges with infinity in focus at the centre will have to be either focused just beyond infinity (if this is possible) or always used stopped down enough to cover the discrepancy with DOF. 

If the lens focuses too far out at the edges then you need to slightly bring the central point of focus back towards the camera a little, you probably won't need to stop down an extra in this case as DOF will likely cover your needs.. 

Infinity focus issues can be a real pain for landscape photographers and especially those who make really big prints, as always, the problems are more likely accentuated for wide aperture lenses and wide angles optics.

If when testing your lens you find the field curvature is quite lopsided , meaning perhaps one side focuses too close and the other too far, then you likely have a de-centered lens, in this case replacement or repair is needed. De-centered lenses are very common, even amongst the more expensive brands, some brands though seem to have a dreadful reputation with de-centering, low end Sigma lenses are a case in point,  where some lenses are gems and others dogs, and by that I mean nasty pit bull terrier type dogs, and no I can’t tell you why.

If you're really clever you may actually be able to even use a de-centrered lens creatively, so long as you know which way the curvature is actually running in the frame, but generally de-centered lenses are in no mans land.

Dropping a lens can cause de-centering as it can misalign the elements, I have one such lens I keep for demos, a little minolta 35-70 f4. When I bought it, it was a gem, sharp everywhere with just the smallest amount of stopping down, but clumsily I dropped it on a concrete floor and now it is hopelessly de-centered, is it fixable, probably, but I have two others.

The point is, if buying a second hand lens, don’t ever buy one that has a dented filter thread, chips on the body or any other bent bits, there is a good chance it has an internal problem.

Its easy to see that having intimate knowledge of your lenses field curvature can allow you to use it far more creatively, and next time you look at the edge performance of a lenses you may wonder, is it just poor edge definition or is the problem really field curvature?








A word to the wise:
Many techie photographers love to use DOF scales to help calculate out precise DOF, especially for studio work and other demanding applications, like images to be enlarged to big prints.  The sad truth however is that for many lenses these DOF scales are utterly useless and at the best misleading. Why?

Field curvature means that most lenses do not behave in a textbook manner anywhere in the frame except in the very centre making your DOF calculations null and void. Additionally not many lenses these days have sufficiently refined distance markings to set the lens to anyway, and even if they do there is no guarantee they are accurately placed on the lens.

An additional issue with DOF scales is that "your definition of acceptable sharpness" and the calculation tools definition may not gel, in fact cameras such as the Nikon D800 and the Sony A7r really give us cause to re-evaluate the whole idea of sharp.

And a Second Word to the Wise
So you might think that examining your image for detail in review on the camera will tell you what you need to know about focus, clarity and field curvature, Oh if only that were the case. If you shoot in RAW only, be aware most cameras will create a low resolution version of the file internally so you can review it , the problem is that as you zoom in it will lose clarity even though the file could be a sharp as a tack. So leave your decisions till you get to the computer and a 100% zoomed in view of the RAW files.

No comments:

Post a Comment