Reaching diffraction

FM Forum Rules
Canon SLRs, primes, and zooms lenses reviews

FM Forums \| Canon-mount SLRs \| Join Upload & Sell	Search Used
1 2 end

Archive 2009 · Reaching diffraction

pahrens
Offline
Upload & Sell: Off

p.1 #1 · Reaching diffraction

I'm trying to understand something here and wondering if someone can help me clarify.

Now that the 7D has 18mp and it means diffraction will be reached sooner, does this mean we cannot stop down to get great depth of field without suffering IQ loss?

If this is the case would that make it a worse landscape camera than something with a larger pixel pitch?

Nov 20, 2009 at 01:52 PM

Fred Tedsen
Offline
Upload & Sell: Off

p.1 #2 · Reaching diffraction

pahrens wrote:
I'm trying to understand something here and wondering if someone can help me clarify.

Now that the 7D has 18mp and it means diffraction will be reached sooner,

No.

does this mean we cannot stop down to get great depth of field without suffering IQ loss?

No

If this is the case would that make it a worse landscape camera than something with a larger pixel pitch?

No. The bottom line is that diffraction doesn't start any sooner on the 7D. You may be able to discern it earlier than say a 20D at 100% magnification, but that's because the 7D has greater resolution.

Nov 20, 2009 at 02:51 PM

cameron12x
Offline
Upload & Sell: On

p.1 #3 · Reaching diffraction

Like everything else in life, "it just depends..." Diffraction is a very interesting discussion topic.

The 18mp sensor on the 7D opens up a number of IQ tradeoffs. Whether you notice diffraction depends on your final print size or whether you do 100% pixel-peeping of your files.

Will you notice it sooner on 4x6" prints? Probably not. What about on A3 and larger sized prints? Or on a large, high-resolution monitor? Well, that all depends... There are two things to consider: image-level diffraction, and pixel-level diffraction.

Image-level diffraction is based on the sensor form-factor (the 7D is a crop camera with a smaller sensor than a FF camera). Pixel-level diffraction is based on the resolving power (pixel density) of the sensor itself. Given the same lens focal length and apertures, diffraction will visually appear sooner with a crop camera than with a full-frame camera, and will also manifest itself visually sooner (at the pixel-level) with a more densely packed sensor (such as the 7D).

DOF is a focusing characterstic which has an arbitrary CoC (Circle of Confusion) resolution metric underlying it. It basically says that, for a given focusing range, the image will be "resolved" at this basic CoC detail level. So, for small prints, the CoC is usually sufficiently large enough such that absolute pixel-level sharpness (and any diffraction effects at that level) will not be visually meaningful for a small print. The larger the print that you make, however, the more important CoC becomes, and the sooner that you'll likely start to see diffraction effects.

These concepts apply not only to the 7D, but also to all cameras. But since the 7D currently offers more pixel-level resolution than any other DSLR crop sensor, its effects are often-times seen sooner when pixel-peeping at 100% magnification.

Diffraction, as it manifests itself at the "pixel-level," may very well be a major reason why so many people have commented subjectively that "7D images appear soft."

Consider the tradeoffs when you are shooting. What is the target output medium? Is it a large print, or a small web image? Or something in between? How much DOF do you need?

The 7D can certainly do landscape photography, and do it well.

However, IF for a given landscape you require maximum sharpness throughout the entire intended DOF, then that does open up a discussion of whether the 7D would be the right tool for the job. Depending on your shooting requirements and the landscape subject matter, a full-frame camera with a smaller pixel-pitch could give you better results. Keep in mind, however, that the full-frame form factor also provides less DOF at a given aperture and focal length than a crop sensor.

Diffraction effects also vary from lens-to-lens, dependent upon their design limitations and compromises. So, also keep in mind the lens selections available for each vendor, and for each sensor form-factor (FF or crop).

In short, there are many things to consider, and it's not a simple answer. I will be keeping my 7D, but I am also looking forward to the full-frame successor to the 5D2, when it arrives perhaps a year from now.

Buyer beware. Your results may vary.

Edited on Nov 20, 2009 at 03:50 PM · View previous versions

Nov 20, 2009 at 02:53 PM

pahrens
Offline
Upload & Sell: Off

p.1 #4 · Reaching diffraction

So given equal print sizes the 7D would produce a better result than another camera with a less dense sensor at higher apertures?

Nov 20, 2009 at 02:54 PM

pahrens
Offline
Upload & Sell: Off

p.1 #5 · Reaching diffraction

I must have posted at the same time as you Cameron. I'll be printing on canvas and would like to use the largest sizes possible. For some landscape and macro shots I'd like a lot of depth of field.

Reason I'm asking this is I'm weighing up to stay with Canon and get a 7D or 5D, or move over to Nikon and get a D700. It's such a tradeoff, Canon has the 5D with a great sensor, Nikon has the D700 with great build, speed and focus (Not to mention the 14-24 2.8).

Thanks everyone.

Nov 20, 2009 at 03:01 PM

cameron12x
Offline
Upload & Sell: On

p.1 #6 · Reaching diffraction

pahrens wrote:
So given equal print sizes the 7D would produce a better result than another camera with a less dense sensor at higher apertures?

No, it depends on the print size. There would be no visible difference at 4x6, for example.

For larger print sizes it also depends, as noise may also start to become more visible. But a properly exposed and processed 7D file (shot with a good lens) should do very, very well at large print sizes.

pahrens wrote:
I must have posted at the same time as you Cameron. I'll be printing on canvas and would like to use the largest sizes possible. For some landscape and macro shots I'd like a lot of depth of field.

Reason I'm asking this is I'm weighing up to stay with Canon and get a 7D or 5D, or move over to Nikon and get a D700. It's such a tradeoff, Canon has the 5D with a great sensor, Nikon has the D700 with great build, speed and focus (Not to mention the 14-24 2.8).

Thanks everyone.

I was updating my post at the same time that you made yours. I hope my post helps.

There is probably no simple answer (or camera/lens combination) which will cover all of your shooting requirements, but diffraction is a very interesting general discussion topic, however.

Edited on Nov 20, 2009 at 04:12 PM · View previous versions

Nov 20, 2009 at 03:39 PM

AGeoJO
Offline
Dedicated FM
Upload & Sell: On

p.1 #7 · Reaching diffraction

pahrens wrote:
I must have posted at the same time as you Cameron. I'll be printing on canvas and would like to use the largest sizes possible. For some landscape and macro shots I'd like a lot of depth of field.

Reason I'm asking this is I'm weighing up to stay with Canon and get a 7D or 5D, or move over to Nikon and get a D700. It's such a tradeoff, Canon has the 5D with a great sensor, Nikon has the D700 with great build, speed and focus (Not to mention the 14-24 2.8).

Thanks everyone.

There is practically no universal answer to your situation. The more choices we have the more complicated the soluation. On one hand, you want to get an extended DOF for landscape and macro. Plus, it sounds like you want larger prints. In that respect, a 5D Mark II will probably your best bet but in the same breath, you mentioned speed and focus which leans the balance more towards a 7D. Take your pick or get a 1Ds Mark II or III as a single solution. With respect to D700 - yes, it is a great camera but resolution-wise, it is on the short end. Now, we may consider a D3x, which comes up closer to your ideal camera. Oh, yeah the Nikkor 14-24 is a great lens but it can also be adopted to be used on a Canon body. Life is full of comprimises, huh?

Nov 20, 2009 at 04:08 PM

gdanmitchell
Offline
Dedicated FM
Upload & Sell: Off

p.1 #8 · Reaching diffraction

The subject of diffraction relative to increasing photosite density (e.g. - "more megapixels") is one that often creates a lot of confusion.

In a nutshell, if you move from a 12 MP cropped sensor body to a 18 MP cropped sensor body the "amount of diffraction blur" is...

... exactly the same.

The statement that "diffraction will be reached sooner" doesn't make any sense, and this is not a concern when you compare various MP flavors of cropped sensor bodies.

An example: Make a photograph with your 12 MP cropped sensor body using some lens (let's make it a 50mm lens at f/8) and then make an identical photograph with the same lens/aperture but using a cropped sensor body an 18 MP sensor. Now take both images through your normal workflow and make prints of the same size - let's say 12 x 18 inches. You will have precisely the same "amount" of diffraction blur in each print.

Now, if you happen to have a really high resolution lens that can exceed the resolution capability of the 12 MP sensor and you focus and shoot very carefully such that lens/sensor resolution is the limiting factor, you might (in some very limited cases) get slightly better resolution from the higher photosite density sensor if you open up a bit and the lens you are using performs exceptionally well at a slightly larger aperture.

Bottom line: There is no "diffraction blur liability" when moving to a camera with increased photosite density on a given size of sensor. The higher photosite density sensor could provide a very small improvement in image resolution at slightly larger apertures if your lens and technique are up the task.

In the context of your landscape question:

1. If you shoot at smaller apertures (say f/8) there is no disadvantage in terms of diffraction blur in using the higher photosite density camera.

2. If you shoot at larger apertures with excellent lenses and very careful technique there may be a very small improvement in resolution on the 18 MP body, but it is almost certainly entirely insignificant.

3. If you really want to improve resolution for landscape photography you want to consider a larger sensor. There are several reasons for this and diffraction blur is one of them. A diffraction blur of some absolute "size" on the sensor surface of a full frame sensor comprises a smaller relative portion of the overall larger frame. (But even here the differences in resolution are unlikely to be significant at all unless you regularly make fairly large and carefully produced prints.)

Dan

BTW: People often make a fundamental mistake when comparing 100% magnification crops on the screen. If you do my test outlined above in which the "amount of diffraction" from the 12 MP and 18 MP prints will be identical, you might notice that the 100% on-screen image of from the 18 MP camera looks softer. Of course it does, but this doesn't mean that the image is less soft! If that sounds bizarre, just think it through a bit more carefully. If you look at, say, a 200 x 200 pixel section of your 18 MP image at 100% you are looking "more closely" at a smaller portion of the image than when you look at a 100% 200 x 200 section from the 12 MP camera. This is conceptually similar to looking at one negative with a 10X loupe and then looking at the next one with a 12X loupe!

pahrens wrote:
I'm trying to understand something here and wondering if someone can help me clarify.

Now that the 7D has 18mp and it means diffraction will be reached sooner, does this mean we cannot stop down to get great depth of field without suffering IQ loss?

If this is the case would that make it a worse landscape camera than something with a larger pixel pitch?

Nov 20, 2009 at 04:19 PM

abqnmusa
Offline
Upload & Sell: On

p.1 #9 · Reaching diffraction

I see no evidence using the 7D that diffraction is reached sooner then the 40D crop I had before. I see no loss of image sharpness at F8, F10. Diffraction does not seem to have any effect until past F11.

Nov 20, 2009 at 04:39 PM

cameron12x
Offline
Upload & Sell: On

p.1 #10 · Reaching diffraction

There is also a very interesting (and lengthy) discussion thread on the topic of diffraction and DLA here:

For those worried about diffraction...

Its ultimate effect on IQ will vary, depending on your shooting needs, display needs, and printing needs.

Nov 20, 2009 at 05:11 PM

cameron12x
Offline
Upload & Sell: On

p.1 #11 · Reaching diffraction

Here is a landscape image I shot with my 7D and the Canon EF-S 10-22mm zoom.

I didn't have any neutral density filters with me at the time so I was forced to shoot at f/16 in order to get a slow enough shutter speed to get the feathery water effect. With a neutral density filter, I probably would have selected an aperture closer to f/8 for this image, since the foreground objects were not insanely close and DOF requirements were not severe.

I have a 16x20" print on the way to me in the mail. So, I'll soon be able to see first-hand how a decent-sized print (with some post-processing applied, including sharpening) holds up to diffraction effects at f/16 for this camera/lens combination.

Canon EOS 7D 22 mm f/16.0 1/2 sec 100 ISO 0.0 EV

Nov 20, 2009 at 05:34 PM

vachss
Offline
Upload & Sell: Off

p.1 #12 · Reaching diffraction

Diffraction produces a fundamental trade off between resolved detail and depth of field. Regardless of the sensor the best you can ever do is get a depth of field (DOF) on the order of DOF ~ (dx)**2 / lambda, where dx is the resolved feature size in the image and lambda the wavelength of light. In other words, as you resolve ever finer features you have ever shallower depth of field. The only way around this is to image stack or do some other clever processing to generate detail not present in the original image.

What high pixel density sensors do is allow us to reach this limit at somewhat larger apertures and smaller feature sizes if we so choose, but the limit remains the same.

Nov 20, 2009 at 05:43 PM

Grant808
Offline
Upload & Sell: Off

p.1 #13 · Reaching diffraction

Good points from everyone so far! One factor that may increase the effect of diffraction is the gapless microlenses. The more discrete data gathered from older sensors with larger gaps are just less likely to show the diffraction, which is always there for a given lens and settings.

I look at the 7D as having the benefit of 'oversampling' or collecting more data than I really need in most cases. The good news for me is that I don't 'waste' the sharpness of my best lenses...and confirms for me that there is room to make use of even higher pixel densities. The bad news about higher pixel densities is that for many shots and lenses, you will just be getting more, but blurrier pixels.

From 500/4IS:

100% crop of above (f/5.6):

From this scene with Leica 19/2.8:

100% crops from the tree on the right. Focus was shy of infinity, so the f/4 and more open shots might be suffering from being out of focus.

f/4:

f/5.6:

f/8:

f/11:

f/16:

Nov 20, 2009 at 06:11 PM

alundeb
Online
Upload & Sell: Off

p.1 #14 · Reaching diffraction

Good examples, Grant. It is very difficult to see the effects of diffraction in real photos because there may not be microdetail in the subject, or it is very difficult to overcome other sharpness difiiculties.

Here is the 7D with EF-S 60 macro, apertures from 2.8 to 16. Amazingly (or not), the performance peaks at f/4 in the focal plane. And what a sweet lens the 60 is!

Bottom line: the 7D is capable of resolving some very fine detail, but we should disregard tests of 7D "sharpness" done at f11 - f16 as representative of what the 7D can do. You might even want to take a little care about f8 sometimes.

2012 © alundeb

Nov 20, 2009 at 06:23 PM

cameron12x
Offline
Upload & Sell: On

p.1 #15 · Reaching diffraction

alundeb wrote:
Bottom line: the 7D is capable of resolving some very fine detail, but we should disregard tests of 7D "sharpness" done at f11 - f16 as representative of what the 7D can do. You might even want to take a little care about f8 sometimes.

It's definitely a balancing act, for sure.

If you have large DOF requirements (but don't need to print big), then you would be more inclined to use the narrower apertures (f/11 -- f/16 range).

If your DOF requirements are smaller (but you need to print big) then you'd probably want to shoot at f/8 and wider.

There are thousands of permutations in between these two use cases. Each will require individual evaluation.

All come with caveat that you also need to test your lenses first in order to find the MTF sweet spot, as was nicely done above.

Nov 20, 2009 at 06:36 PM

pahrens
Offline
Upload & Sell: Off

p.1 #16 · Reaching diffraction

Wow, thank you so much everyone for taking so much time for the really in depth posts.

I think the ultimate solution here is to get a 5D MII and the 7D. It will be a pain not having them together in one body though. Or as some one suggested one of the 1Ds Mxx series.

I'd like to try out the Nikon D700 plus some of the great new lenses they have now, they never had them when I was with Nikon. I feel like it takes the best of both worlds but never quite as good in any one aspect.

Thanks again everyone.

Nov 20, 2009 at 10:53 PM

stanj
Offline
Dedicated FM
Upload & Sell: On

p.1 #17 · Reaching diffraction

cameron12x wrote:
If you have large DOF requirements (but don't need to print big), then you would be more inclined to use the narrower apertures (f/11 -- f/16 range).

Or you could be shooting at whatever aperture you please (the sweet spot) and use focus stacking. Not possible for moving subjects, but some things won't move.

Personally I have clearly seen issues of limited DOF, compensating for it, and resulting diffraction when shooting MF, as there to get anything in focus you have to be at f8 at least and if one was to stop down to say f32 it did get ugly really quick. So it's all relative

Nov 20, 2009 at 11:01 PM

pahrens
Offline
Upload & Sell: Off

p.1 #18 · Reaching diffraction

Just thinking about it, will the diffraction only become obvious from the dense sensors in prints larger than what the less pixel dense sensors will be able to print at?

Nov 20, 2009 at 11:10 PM

Grant808
Offline
Upload & Sell: Off

p.1 #19 · Reaching diffraction

pahrens wrote:
Just thinking about it, will the diffraction only become obvious from the dense sensors in prints larger than what the less pixel dense sensors will be able to print at?

Not really.
A billboard can be printed from a 300D or a 7D and you won't notice the difference in diffraction. With printing the native L-sized files, I believe you'll see pixelation intrude first for either body.

EDIT: On a 30' x 20' foot billboard it would be the difference of about 8 pixels per inch vs.15 in my example above and trying to discern detail at the 1/4" level.

Nov 20, 2009 at 11:17 PM

AJSJones
Offline
Upload & Sell: On

p.1 #20 · Reaching diffraction

pahrens wrote:
Just thinking about it, will the diffraction only become obvious from the dense sensors in prints larger than what the less pixel dense sensors will be able to print at?
For those who expect to be able to make bigger prints from cameras because there are more MP (which usually means smaller pixels aka denser sensors) the effect of any kind of blur (diffraction OOF motion) will be enlarged more than for images from sensors with bigger pixels, so the effect of diffraction will be more noticeable. To your question : in a word, yes

Nov 20, 2009 at 11:42 PM

pahrens
Offline
Upload & Sell: Off

p.1 #21 · Reaching diffraction

So really it doesn't matter a whole lot that this sensor is so much denser than older models?

Nov 21, 2009 at 03:46 AM

Grant808
Offline
Upload & Sell: Off

p.1 #22 · Reaching diffraction

AJSJones wrote:
pahrens wrote:
Just thinking about it, will the diffraction only become obvious from the dense sensors in prints larger than what the less pixel dense sensors will be able to print at?

For those who expect to be able to make bigger prints from cameras because there are more MP (which usually means smaller pixels aka denser sensors) the effect of any kind of blur (diffraction OOF motion) will be enlarged more than for images from sensors with bigger pixels, so the effect of diffraction will be more noticeable. To your question : in a word, yes

Yeah, false expectations seem to be the problem here. Many don't seem to know that the real limit is still in the optical properties of the glass. The nice thing to me is that you can crop *hard* on the 7D files *if* you have a sharp to the pixel level capture.

pahrens wrote:
So really it doesn't matter a whole lot that this sensor is so much denser than older models?

It does to me. But I have lenses that capture detail to the 7D pixel level.

Guess what this is (100% crop):
This image is copyrighted by the owner

Shows some bayer artifacting...and no that's not the 'maze pattern' problem.

Nov 21, 2009 at 06:29 AM

gdanmitchell
Offline
Dedicated FM
Upload & Sell: Off

p.1 #23 · Reaching diffraction

Grant808 wrote:
Good points from everyone so far! One factor that may increase the effect of diffraction is the gapless microlenses. The more discrete data gathered from older sensors with larger gaps are just less likely to show the diffraction, which is always there for a given lens and settings.

I'm afraid that this is simply not going to be the case. Images don't line up with only vertical and horizontal components that neatly agree to not spill across photosite boundaries.

Dan

Edited on Nov 21, 2009 at 03:46 PM · View previous versions

Nov 21, 2009 at 06:48 AM

gdanmitchell
Offline
Dedicated FM
Upload & Sell: Off

p.1 #24 · Reaching diffraction

pahrens wrote:
So really it doesn't matter a whole lot that this sensor is so much denser than older models?

It might matter, but not in terms of diffraction blur.

For example, if you have a very high quality lens and shoot with extreme care you might get some tiny increment of increased system resolution if you select the optimum aperture. And you might get slightly better gradients with more, smaller photosites.

But the difference is really, really minor when it is visible at all. Unless you have very high quality lenses, shoot with extreme care from a tripod, and do very precise and careful post-processing, make very large prints, and then inspect them very closely...

... it is almost certain that you'll never notice any difference at all.

If you never print larger than letter size and/or you only post jpgs online or send them via email, this whole discussion is entirely pointless in terms of the quality of your photographs.

Nov 21, 2009 at 06:53 AM

Grant808
Offline
Upload & Sell: Off

p.1 #25 · Reaching diffraction

gdanmitchell wrote:
Grant808 wrote:
Good points from everyone so far! One factor that may increase the effect of diffraction is the gapless microlenses. The more discrete data gathered from older sensors with larger gaps are just less likely to show the diffraction, which is always there for a given lens and settings.

I'm afraid that this is simply not going to be the case. Images don't like up with only vertical and horizontal components that neatly agree to not spill across photosite boundaries.

Dan

Yeah, they don't 'like up' as you say.

Lining up has nothing to do with what I am saying. The lack of sampling via the gaps makes diffraction effects less likely to show.

Nov 21, 2009 at 08:02 AM


FM Forums \| Canon-mount SLRs \| Join Upload & Sell 1 2 end