Cropped sensors better for landscape work. Huh?

[Brev00]

But, if one changes only the focal length, accepting the field of view as it is in either image, the depth of field will be larger with the wider focal length. The thing is, when I use my ultra wide, I am taking pics using its unique qualities and am not trying to get closer to simulate a 24mm image. Once it is on my camera, just as when I put a macro lens on, I am in a different world using every available mm possible. So, within that world, I am getting more depth of field than if I had a longer lens on--in the final image without trying to equalize them in terms of field of view. I am just changing one variable--the focal length. The image will be different with a wider frame. There is no framing I do not want. I am recommending that Dan get an ultra wide and see things from that perspective. Explore a different way of framing a scene. I am not recommending that he use it to somehow get the same 24 mm framing of his shot. I am sure your point is true in theory but I think my suggestion to try an ultra wide still holds.

[NorthernFocus]

Back to the original post, being new to this site I was quite surprised to read that someone was concerned about going with smaller aperture than f/8 because of diffraction. I have done landscapes with both cropped and full sensors and never saw diffraction as a significant factor even at f/22 with normal viewing conditions.....

In the OP I didn't exactly say I was "concerned" about diffraction. Simply pointing out that it could be avoided while still retaining adequate DOF to accomplish the desired outcome.

With regard to never having seen effects of diffraction before, the anti-aliasing filter in most cameras softens the image more than the effects of diffraction. So you wouldn't likely notice it. But the D7100 offers a new level of potential sharpness because it doesn't have an AA filter. So why not take advantage of the fact?

Unfortunately, as Colin pointed out, choice of aperture is often dictated by more than just desired DOF. So realistically when shooting landscape I'll rarely get the opportunity to take full advantage of the potential sharpness achievable by the D7100 sensor. Heck I might as well be shooting Canon

[NorthernFocus]

...I am recommending that Dan get an ultra wide and see things from that perspective. Explore a different way of framing a scene....

That's another whole topic. When I went back and looked at my prior landscape work I wondered at why 22-24mm (35mm FF equiv) seemed to be a "floor" for my shooting. Particularly since the two landscape lenses I've owned have been a 17-55 and 16-85. The conclusion I came to is that is about the limit where the FOV starts to go beyond my "normal" perceived vision. In other words, I have trouble visualizing how the scene will be rendered in the image. So learning to shoot truly WA scenes is on my to do list.

[Polar01]

I use a 16-35mm on a D600, I myself find that I shoot a lot at the 16-18mm with the camera in portrait format for pans. Just one shot, takes it is around 24mm, I think that is has a lot to do with what I find pleasing to my eye. Most of the time I will bring the camera to my eye, zoom in and out, more around do some more zooming until the composition look good then shoot, looking down at the lens I will often see that it is around the 24mm mark. So I think that whatever xxxmm we use to get the FOV depends a lot on what our brain is happy with or use to.

Cheers: Allan

[Glenn NK]

And would be wonderful unless there is something 8 ft away that you want included in the frame and want to be in focus all the way to the horizon

Another benefit of APS-C not yet mentioned is the fact that only the center portion of the image is captured where the lens is sharpest and has the least distortion. Another one of those little things, each of which add to better overall IQ...

You don't have to focus right on an object to have it appear acceptably sharp - that's what DOF is all about.

I agree about the APS-C sensor having that advantage - I still use my ancient 30D with very good results, but the difference in DOF between the FF and Crop is not a lot.

[Brev00]

That's another whole topic. When I went back and looked at my prior landscape work I wondered at why 22-24mm (35mm FF equiv) seemed to be a "floor" for my shooting. Particularly since the two landscape lenses I've owned have been a 17-55 and 16-85. The conclusion I came to is that is about the limit where the FOV starts to go beyond my "normal" perceived vision. In other words, I have trouble visualizing how the scene will be rendered in the image. So learning to shoot truly WA scenes is on my to do list.

That is why you have not shot very wide in the past. You have been using lenses that allow you to stay in your comfort zone. When I got my 12-24, I kept it on my D90 for a couple of weeks and shot with it full time. I had a trip to Hawaii in a few weeks so I wanted to get used to it as quickly as possible. The more I used it, the more I liked the wider fl's. It took a while, though. It did seem strange at first. For the most part, I got it to shoot in the 17-24 range--backing up my midrange where the 17-70 has its worst distortion. So, I can take landscape shots with some of the same focal lengths I already had but without needing much in the way of corrections to get level horizons and straight lines. So, it might just take a certain level of commitment to finally enjoy this type of lens. It helps to have a goal. Just getting wide for no real reason might lead to a quick return.

[NorthernFocus]

That is why you have not shot very wide in the past. You have been using lenses that allow you to stay in your comfort zone.....It helps to have a goal...

So much to shoot. So little time... I'm saving the challenges of WA landscapes for retirement. The only thing I've really spent any time on is moving water (which is why I'm rarely concerned with the horizon). For now I've got enough challenges chasing critters around. In the meantime landscapes are pretty much photos of opportunity or vacation shots

[FlyingSquirrel]

I'm late to this, and frankly haven't the time to read the whole thread. Might have been mentioned, but you have the additional benefit of crop sensor in that the central area of the optics are being used to render the image, i.e. the image from the outer regions of the optics are not captured (they are "cropped" if you will) thus you are using the area of the optic which is sharper and has less defects/abberation/softness/vignetting

[Colin Southern]

I'm late to this, and frankly haven't the time to read the whole thread. Might have been mentioned, but you have the additional benefit of crop sensor in that the central area of the optics are being used to render the image, i.e. the image from the outer regions of the optics are not captured (they are "cropped" if you will) thus you are using the area of the optic which is sharper and has less defects/abberation/softness/vignetting

Only if it's a lens designed for a full-frame camera, and even then, the difference is usually minimal.

[AB26]

you have the additional benefit of crop sensor in that the central area of the optics are being used to render the image, i.e. the image from the outer regions of the optics are not captured

Hi Matt,

It seems this is a general misconception about crop frame.

There are two different tipes of lenses for DSLR cameras. You get crop frame lenses for crop frame bodies and full frame lenses for full frame bodies. Full frame lenses are much more expensive than crop frame lenses, for the simple reason, they are built to a much higher standard, using more material than crop frame lenses.

Crop frame lenses are built for amateurs and are much more affordable. These lenses have a much smaller image circle than a FF lens. Look at specifications for a similar FF lens and a crop frame lens, the angle of view differs. Example: Nikkor 18-200mm DX lens at 200mm AOV = 8 degrees. Nikkor 70-200mm FX lens at 200mm AOV = 12.2 degrees. The image circle of a FX lens is much bigger than the image circle of a DX lens.

A crop frame lens on a crop frame body has no advantage over a FF lens on a FF camera, in fact, the FF lens is in most cases far superior to the crop frame lens. The only time a lens has the advantage of only using the “best part” of the image circle, on a crop frame body, is when a FF lens is fitted to a crop frame body. Such a combination should not render any vignetting even when shot wide open. On the contrary, there are FF lenses rendering no vignetting on FF bodies even when shot wide open.

[FlyingSquirrel]

Only if it's a lens designed for a full-frame camera, and even then, the difference is usually minimal.

True, only on a full-frame lens, or what I simply refer to as a lens. And I suppose the difference COULD be minimal, depending on the quality of the lens; but also since we are talking about landscape shots and therefore generally wider angle lenses, those are the type of lenses that tend to have the most aberrations and softness at edges. But that is just my general observations, and surely the differences will vary for different lenses and bodies.

[FlyingSquirrel]

The only time a lens has the advantage of only using the “best part” of the image circle, on a crop frame body, is when a FF lens is fitted to a crop frame body.

That is precisely the case that I was referring to. Also see my comment in the post above in response to Colin.

I honestly rarely even think about 'crop frame' vs 'full frame' when it comes to lenses, since all of my lenses are 'full frame' lenses if we are talking about it, except for my 10-22 crop sensor lens, which, true to your comment, is inferior in quality.

[Colin Southern]

True, only on a full-frame lens, or what I simply refer to as a lens. And I suppose the difference COULD be minimal, depending on the quality of the lens; but also since we are talking about landscape shots and therefore generally wider angle lenses, those are the type of lenses that tend to have the most aberrations and softness at edges. But that is just my general observations, and surely the differences will vary for different lenses and bodies.

To be honest, I think it's one of those areas where the theory and practice are different; I'm sure the lab rats could come up with 100% crops to show a difference, but I'd challenge them to show a difference in any whole image either at normal internet sizes or normal print sizes.

[chauncey]

Do be honest, I think it's one of those areas where the theory and practice are different; I'm sure the lab rats could come up with 100% crops to show a difference, but I'd challenge them to show a difference in any whole image either at normal internet sizes or normal print sizes.

All that is true Colin, but...those lab rats are the ones that tell me which gear to buy, the gear that doesn't show problems when blown up to 200% on my monitor. I want my photography imperfections to be based on the inadequate operator skills rather than the fault of my gear.

[Manfred M]

All that is true Colin, but...those lab rats are the ones that tell me which gear to buy, the gear that doesn't show problems when blown up to 200% on my monitor. I want my photography imperfections to be based on the inadequate operator skills rather than the fault of my gear.

I would have to disagree; the lab rats measure what is easy to measure in a lab, rather than things that matter in most real life photography. I rarely shoot under laboratory conditions, so the results that they get are quoting are something that I generally cannot reproduce. I am also suspcious that they are testing "cherry picked" samples from the manufacturer / distributor so getting a piece of equipment that close to optimal in real life; well probably not. Besides, write a bad review and you will get cut off of your supply of equipment and there goes your career...

The other issue is that they never cover things that are really important to a real life photographer; things like robustness of the equipment, backwards compatability with older gear, etc.

Reviews; I hardly ever look at them because I have found them to be meaningless.

[NorthernFocus]

Matt brings up a good point regarding the cropped sensor using the center of the lens. Distortion, resolution, and CA are all much worse at the edges. Vignetting is also, by definition, an edge phenomenon. One can take advantage of the fact and can get quite good results on a crop body with inexpensive full frame lenses that perform poorly on full frame bodies. I'd recommend checking out the Photozone database. They provide test data on a lot of lenses on both full and cropped frame bodies and the differences can be pretty dramatic.

Regarding diffraction not being a real world issue, for most people that is likely the case due to the camera bodies they are shooting. However, for anyone shooting one of the new bodies with no AA filter, it will show up and possibly quite noticeably. Currently to my knowledge the Nikon D7100, D800E, and one of the new Sony bodies that just came out are the only ones on the market with no AA filtering.

Diffraction was considered more of an issue back in the film days. When digital came along AA filters were introduced which intentionally soften the image to counter moire effects. I've not seen any supporting data but some suggest that AA filter softening exceeds the effects of diffraction. Now with the higher resolution sensors moire patterns are less of an issue and in the examples I mentioned the AA filter has been removed. So diffraction re-enters as a real issue, assuming one is interested in optimizing the optical path and resultant IQ.

For anyone interested in supporting data you can pull up the DxOMark lens database, pick any given Nikon mount lens, and compare the test results for sharpness on D7100 and D5200 bodies. Both bodies share the same APS-C 24MP sensor, one with and one without AA filtering. The difference in sharpness is substantial.

For those of you who don't shoot the subject cameras, it may be true that diffraction is not a real world consideration. But in the context of CIC being a learning forum, this topic may warrant further thought. Particularly if AA filters are eliminated in more camera bodies in the future.

[ajohnw]

To much to read all posts but I may be able to add a couple of things that should clear up any confusion and if not known best to bear in mind. Google will explain some terms.

Web info on diffraction - usually in black and white and all pixels register that. The colour pixel resolution size of a camera is a 4 pixel square and talking about that in terms of diffraction isn't entirely sensible as debayering to generate the image is a multi way interpolating process. Generally two green and one blue and one red pixel making things even more difficult to get hard answers. In practice the presence of diffraction resolution loss will largely be determined by what is in front of the camera and even what size it's finally viewed at. What people do not realise about diffraction limits is that contrast levels are reduced to less than 10% of what goes into the camera. That I am afraid is a physical fact. Actually with the lens testing used these days I feel mentioning it is a bit dubious as there is very little black and white to actually measure at this level. There are charts on this page for real lens testing - in black and white - that show what 10% contrast looks like.

http://www.normankoren.com/Tutorials/MTF5.html

Depth of field. The figures are based on the size of focus blur usually called the size of the circles of confusion. An image is considered to be made of of a huge number of individual points all merged together.. A lens produces it's smallest circle of confusion at one point, the one it's focused to. It will gradually get larger each side of that. This is an idealised way of looking at this but it's true for a perfect lens. When a calculator is used it assumes a certain maximum sized circle of confusion based on sensor size. Say a full frame size is 1 a camera with a crop factor of 2 would use 0.5. The actual size is calculated on the basis of producing a 12x8in print that can be viewed from 10in without loss of focus being visible to an unaided eye. So smaller sensors need smaller circles of confusion as they have to be enlarged more. If for some reason the settings for a shot produces circles of confusion that are twice as big as this all it means is that it would best be viewed from 20in. If you want to LOL realise that the Japanese introduced larger maximum circles of confusion than some as soon as they started exporting cameras! They used them to set the spacings of depth of field markings on older lenses.

When depth of field is considered as this thread started 2 things are going on. For the same view the cropped sensor uses a shorter focal length lens so it's depth of field is greater however the circle of confusion it has to use for the same sized final image has to be smaller so this part off sets the difference. In order to achieve exactly the same depth of field behaviour after that is accounted for the aperture needs to be changed. Oddly that relates to the crop factor as well. 1 stop if the crop factor is 2. So to just cover a landscape right at the limits a camera with a crop fact of 2 could use a 1 stop wider aperture for the same depth of field.

Anti aliasing filters slowly going is interesting. They are needed so the lenses themselves must be doing the job. People talk about spatial frequencies - so many line pairs per mm, one black one white. The frequency goes up as more and more are crammed in the mm. There is plenty of info on the web about what aliasing means but usually related to electronics but the same principle applies. At some point the frequency can go up to a point where the pixels on the sensor record aliasing effects rather than what is actually there. Lenses can only resolve so many line pairs per mm so can do more or less the same thing as an anti aliasing filter if the pixel density is high enough. Having looked at test results and read dpreview's coverage of the 2 types of D800 I'm not really convinced there is much difference really. I feel lenses make far more difference in practice.

John
-

[Manfred M]

Anti aliasing filters slowly going is interesting.

So far as I know, the medium format cameras. Mamiya, Hasselblad, Leica S and Phase One nor the smaller sensor Leica M series digital have never used AA filters, so I would suggest that this is more something we are seeing with volume manufacturers

[ajohnw]

So far as I know, the medium format cameras. Mamiya, Hasselblad, Leica S and Phase One nor the smaller sensor Leica M series digital have never used AA filters, so I would suggest that this is more something we are seeing with volume manufacturers

If I come completely clean Manfred I have wondered what they are about. I've spent rather lot of time working on aliasing problems and it's easy to miss some aspects. It could be that they are there really to get round problems encountered with debayering techniques that produce artefacts - sort of mush it up to help. In other words it's more of a software pixel colour sampling problem than strict aliasing. My comment was based round smaller pixels effectively allowing the lens to spread and mush the light. The wiki adds another point that would help with that aspect.

Optical anti-aliasing filter[edit]

In the case of optical image sampling, as by image sensors in digital cameras, the anti-aliasing filter is also known as an optical lowpass filter or blur filter or AA filter. The mathematics of sampling in two spatial dimensions is similar to the mathematics of time-domain sampling, but the filter implementation technologies are different. The typical implementation in digital cameras is two layers of birefringent material such as lithium niobate, which spreads each optical point into a cluster of four points.[1]

The choice of spot separation for such a filter involves a tradeoff among sharpness, aliasing, and fill factor (the ratio of the active refracting area of a microlens array to the total contiguous area occupied by the array). In a monochrome or three-CCD or Foveon X3 camera, the microlens array alone, if near 100% effective, can provide a significant anti-aliasing effect,[2] while in color filter array (CFA, e.g. Bayer filter) cameras, an additional filter is generally needed to reduce aliasing to an acceptable level

That suggest the RGB aspect is taken care of by evenly spreading light and that now lenses are doing it. I've seen comments about better software as well. The processing power available in cameras has increased too. But medium format - pass. Maybe the microlenses do it. On the other hand larger formats can have higher resolutions at a cost and it could be that the lenses or software have achieved the same thing for a long time.

The wike also mentions the Pentax that micro vibrates the sensor to do the same thing.

John
-

[Colin Southern]

I've spent rather lot of time working on aliasing problems and it's easy to miss some aspects.

I think that the aspect people are most likely to miss when working on aliasing "problems" is the big "picture". In terms of DSLRs, just forget about it and go take some photos!

It was Jay Maisel who said "too many people ruin their photos trying to save the pixels".