Diffraction Limit

Not arguing because you agree - indeed, the tone of my voice didn't get conveyed well there!

The rest of this should all be vocalised in a laid-back and light conversational tone...

I'd disagree with the smaller format always being better. It's usually lighter, which is good, but I still see noise and/or noise reductions artifacts at base ISO on the GH3 which are simply not there on the Canon FF. Not that I can't get great shots with the GH3...

As to diffraction, in the real World I find having a mental view of apertures not to go below without careful consideration to be a useful concept in m43 (but less so in FF due to the wider range of apertures available, well unless m43 users are shooting bokehramas, but after trying it I find that really needs FF). You could see it as a limit of some sort...

BTW I suspect you'll have significant difficulty spotting diffraction blur in anything FF at f2.8 or less for real-World photographs. It's a scientific point, but not really important to people taking photographs, IMHO.

Finally I personally really think when people (generally, not meaning you in particular, I just have a bugbear here, so thought I'd say where I'm coming from) are talking about sharpness and blur using a CoC (or whatever) that assumes a certain size print viewed at a certain distance by someone with a certain standard of eyesight they shouldn't generalise, as, for example, someone who has done a big crop to zoom will see something different.

GeorgianBay1939 said:

GeorgianBay1939 wrote:

GeorgianBay1939 said:

Anders W wrote:

GeorgianBay1939 said:

Steen Bay wrote:

GeorgianBay1939 said:

Anders W wrote:

GeorgianBay1939 said:

LTZ470 wrote:

GeorgianBay1939 said:

Steen Bay wrote:

GeorgianBay1939 said:

LTZ470 wrote:

I learn by "concrete proof in the pudding"...there is none for the FZ200 being diffraction limited at f/2.8...

Click to expand...

But if it was (for example in the center of the image), then that would actually be a good thing! If a lens is "diffraction limited" (has its peak resolution) wide open, then that just means that it's a very good lens without any significant aberrations, where the resolution mostly will be limited by the unavoidable diffraction.

Click to expand...

IF or CONCRETE...I'll stick with concrete...and the fact that there is no concrete proof...

Click to expand...

So, if Panasonic came out with a successor to the FZ200 and the succesor were the same camera in every way except that it had a new lens that was "diffraction limited" already at f/2.8, which version would you prefer: the old one or the new one?

Click to expand...

Though, maybe Panasonic shouldn't use the term "diffraction limited" in the advertising. ;-)

Click to expand...

Why not? "Limited" seems to be a rather viable term in marketing speak. Certain Pentax and Oly lenses come to mind for example. And "diffraction" sounds rather advanced, doesn't it. Besides, Canon is already using the term in its lens marketing. Diffractive optics. Great stuff. ;-)

http://www.usa.canon.com/cusa/consumer/standard_display/Lens_Advantage_Perf

Click to expand...

I know that you are having some fun here Anders. But I went to the site and scrolled down to Diffractive Optics . I first thought this was marketing of interference coatings, but when looking at the illustration


Diffractive optics: http://www.usa.canon.com/cusa/consumer/standard_display/Lens_Advantage_Perf#f

... and reading this:

"Canon has also developed a triple-layer type DO lens that uses an advanced diffractive grating to deliver excellent performance, with superior control of color fringing."

That looks like concentric diffraction gratings, like a diffracting (instead of refracting) Fresnel lens.

This will warm the cockles of the hearts of those who believe the ideology that in-lens corrections are superior to software corrections.

Now back to the other form of diffraction, your regularly scheduled program ....

Click to expand...

Hi Tom,

I was just playing around with the words. I am well aware of what diffractive optics à la Canon actually is. If I hadn't already been interested in that from a more serious point of view, I wouldn't have remembered the label, and wouldn't even have come to think of the little word play it allowed in the present context. ;-)

Great Bustard said:

Great Bustard wrote:

Great Bustard said:

Too many relate the effects of diffraction to the size of the pixel as opposed to the proportion of the photo the Airy Disk spans.

Click to expand...

Click to expand...

As you may have read in my previous post (and take this slightly light-heartedly, I'm not trying to start an argument, just add a point)...

Too many assume the final photo will be the whole sensor area... and use that to make absolute judgements. (Sometimes without stating they are making that assumption.)

If you crop to zoom, as any wildlife, or especially astro, photographer will do a chunk of time, the diffraction to pixel size can be quite important, if nothing else is limiting resolution at that point. (Also noise, but that's another story.)

Jon555 said:

Dr_Jon wrote:

If you crop to zoom, as any wildlife, or especially astro,

Click to expand...

The hidden assumption here is that you are lens-limited, i.e. you are using the same lens with two different sensor sizes and then need to crop to the same framing. This may be important for astro, but relatively little wildlife photography uses long lenses at f/stops where diffraction is significant (vs. other issues like focus, motion blur, atmospherics, etc.)

Great Bustard said:

Great Bustard wrote:

Great Bustard said:

Dr_Jon wrote:

Click to expand...

... the point where "diffraction is going to be really start killing the sharpness" depends heavily on the display size of the photo, the viewing distance, the visual acuity of the viewer, the other sources of blur in the photo, and the QT (quality threshold) of the viewer.

Click to expand...

When "diffraction really starts killing the sharpness", that effect (as you point out) exists in the spatial frequency response at the front-end in the lens-system (not the numerous other points along the processing path). That we look at the product of all individual "stages" is quite correct.

Great Bustard said:

Great Bustard said:

More pixels, all else equal, will *always* resolve more detail.
- True, although a chunk of the time it will be to so small a degree you don't care, the rest of the system needs to be in the ball-park. The number of times a friend's D800 out-resolves my 5DmkII are less than you'd think as the pixel effect gets lost in other factors. (He has consumery long lenses, for example)

Click to expand...

Goes along with what I said above -- allow me to paraphrase for the new paragraph:

Well, the point where "it will be so small a degree you don't care" depends heavily on the display size of the photo, the viewing distance, the visual acuity of the viewer, the other sources of blur in the photo, and the QT (quality threshold) of the viewer.

Click to expand...

Actually, an assessment of possible "detail" (in comparing lens-camera systems, here assuming a lens-system free from optical aberrations, and with zero focus-error and camera-motion) depends upon the ratio of the integrals of the composite spatial frequency (MTF) responses involving the following "front-end" parameters: Wavelength, F-Number, Optical Low-pass Filtering, Photosite aperture/shape/pitch. When that ratio of integrals closely approaches unity, "the game is over" where it comes to meaningful quantitative statements about "more" detail.

Whether or not numerous subsequent transfer-functions in the image processing-chain (may indeed) further limit spatial frequency (MTF) response in such a case then becomes essentially irrelevant (if and when "signal bandwidth" is fading into upcoming "noise-floors" of the "front-end").

Great Bustard said:

... That's not to say that there aren't differences in resolution still, as well as other differences, but just that, in my opinion, they are trivial by that point.

Click to expand...

Trivial seems an aptly descriptive term in such cases.

Great Bustard said:

When people start "caring about diffraction" depends, as I said, on a host of other factors, namely the display size of the photo, the viewing distance, the visual acuity of the viewer, the other sources of blur in the photo, and the QT (quality threshold) of the viewer.

Click to expand...

Not to mention the (even more directly relevant) "front-end" considerations surrounding the complicating (MTF) effects of lens-system optical aberrations, focus-errors, and camera-motions.

DM ...

Great Bustard said:

Great Bustard wrote:

Too many relate the effects of diffraction to the size of the pixel as opposed to the proportion of the photo the Airy Disk spans.

Click to expand...

The proportion of the an entire image-frame dimension is indeed important. However, the results amounts only to a simple scaling of the horizontal X-axis of the spatial frequency (MTF) function.

On the other hand, if/when the upper spatial frequency magnitudes of the lens-system MTF (itself) begin to attenuate the upper spatial frequency magnitudes of the composite MTF of lens/filtering/photosite system, we are talking about differences in the fundamental shape of that composite MTF response - something that is distinctly different from simple x-axis scalings due to Sensor Size.

Okay, fair cop, I should have just left it at crop-to-zoom, rather than giving hasty examples. My point is really about hidden assumptions which have been used a lot in this forum in this type of discussion and implied to be fully general.

Jon555 said:

Dr_Jon wrote:

Jon555 said:

Great Bustard wrote:

Jon555 said:

Too many relate the effects of diffraction to the size of the pixel as opposed to the proportion of the photo the Airy Disk spans.

Click to expand...

Click to expand...

As you may have read in my previous post (and take this slightly light-heartedly, I'm not trying to start an argument, just add a point)...

Click to expand...

I take it personally. As far as I'm concerned, you may as well have called my mother...

:-D

Jon555 said:

Too many assume the final photo will be the whole sensor area... and use that to make absolute judgements. (Sometimes without stating they are making that assumption.)

Click to expand...

I think it's fair to say that we are comparing on the basis of same perspective, framing, and display size. That is, we are not shooting 200mm on two different formats and cropping the larger format to the same framing, displaying one photo at 4x6 inches and the other at 30x40 inches. If any such thing were the case, then I would expect someone to explicitly mention that they were focal length or magnification limited with such and such system, and explain why they are comparing photos at different enlargements.

Jon555 said:

If you crop to zoom, as any wildlife, or especially astro, photographer will do a chunk of time, the diffraction to pixel size can be quite important, if nothing else is limiting resolution at that point. (Also noise, but that's another story.)

Click to expand...

Well, as I said, this is a specific scenario, and should be explicitly mentioned if it is the case. Otherwise, as I said, the natural, and implicit, assumption is same perspective, framing, and display size.

Jon555 said:

Dr_Jon wrote:

Click to expand...

Jon555 said:

Okay, fair cop, I should have just left it at crop-to-zoom,

Click to expand...

If you are crop-to-zoom such that the same lens, same f-stop, and same sensor area is used to get the same framing, then pixel density and sensor noise design are the distinguishing differences to the final image. Everything else will be the same (DOF, diffraction, shot noise, etc.) so likely you'd want to optimize for maximum pixel density.

If you are crop-to-zoom using different lenses and different sensor areas, then it's impossible to make any general statement as there are too many variables.

Jon555 said:

Dr_Jon wrote:

Not arguing because you agree - indeed, the tone of my voice didn't get conveyed well there!

Click to expand...

Them's fightin' words, Mister! Draw! :-D

Jon555 said:

The rest of this should all be vocalised in a laid-back and light conversational tone...

I'd disagree with the smaller format always being better.

Click to expand...

"Almost always better for the same DOF and shutter speed" -- very important qualifications for my statement.

Jon555 said:

It's usually lighter, which is good, but I still see noise and/or noise reductions artifacts at base ISO on the GH3 which are simply not there on the Canon FF. Not that I can't get great shots with the GH3...

Click to expand...

The advantages of the larger sensor system come from the use of more shallow DOFs and/or longer shutter speeds for a given DOF at base ISO.

Jon555 said:

As to diffraction, in the real World I find having a mental view of apertures not to go below without careful consideration to be a useful concept in m43 (but less so in FF due to the wider range of apertures available, well unless m43 users are shooting bokehramas, but after trying it I find that really needs FF). You could see it as a limit of some sort...

Click to expand...

Myself, it's incredibly rare for me not to be able to get the DOF I wanted for a scene by f/8 on FF (F/4 on mFT), so it's a non-issue for me. But, there are those who can never seem to get enough DOF, so it is an issue for them.

Jon555 said:

BTW I suspect you'll have significant difficulty spotting diffraction blur in anything FF at f2.8 or less for real-World photographs. It's a scientific point, but not really important to people taking photographs, IMHO.

Click to expand...

Keeping in mind, of course, that diffraction is merely one of many sources of blur, and the effect of f/2.8 differs in proportion to the size of the sensor.

Jon555 said:

Finally I personally really think when people (generally, not meaning you in particular, I just have a bugbear here, so thought I'd say where I'm coming from) are talking about sharpness and blur using a CoC (or whatever) that assumes a certain size print viewed at a certain distance by someone with a certain standard of eyesight they shouldn't generalise, as, for example, someone who has done a big crop to zoom will see something different.

Click to expand...

Well, the CoC used for online DOF calculators typically presumes viewing an 8x10 inch photo from 10 inches (or is it 12 inches?) away with 20-20 vision.

Great Bustard said:

Great Bustard wrote:

... As far as I'm concerned, you may as well have called my mother...

:-D

Click to expand...

Sometimes our mothers may have "padded their pixels" a bit in order to procreate utter geniuses.

BTW (Jessica and Megan ask) - Will you be in the running in the soon upcoming mayoral election ?

The OP is correct, diffraction is simply that. Everything else constant except changing the f number of the lens, diffraction just limits the information we can capture.

Imagine the aperture of the lens as a filter/hole and photons of light as balls. For a diffraction limited f-stop, larger photons/balls such as the wavelengths of RED gets filtered and can't get through this hole. So, what gets through that hole is always constant regardless of what catches/absorbs it (such as a sensor format or number of pixels).

If we have a large basket(such as a pixel) but two small green balls can fit in, then we are resolution limited. If we have very small baskets (small pixels) but one large red ball comes in, the large ball is sliced into pieces to fit in those baskets. What I'm trying to say here, smaller pixels don't add detail to a diffraction limited capture, it's just over-sampled.

Yet, why do people say that smaller pixels(high res sensors) is bad due in terms of diffraction? Simple, you just lose efficiency due to diminishing returns. Efficiency of image file sizes. We still get more detail on higher resolution sensors but for very little gain.

I would like to see a test on this subject since there are two Nikon DSLRs perfect for this, a D3s and a D800.

In matters of DOF and varying sensor formats on diffraction, I think it is a complicated yet futile discussion since we are lens limited anyway, by the smaller formats.

LTZ470 said:

LTZ470 wrote:

From a very credible source:

"Diffraction thus sets a fundamental resolution limit that is independent of the number of megapixels, or the size of the film format. It depends only on the f-number of your lens, and on the wavelength of light being imaged. "

--
--Really there is a God...and He loves you..
FlickR Photostream:
www.flickr.com/photos/46756347@N08/
Mr Ichiro Kitao, I support the call to upgrade the FZ50.
I will not only buy one but two no questions asked...

Click to expand...

Great Bustard said:

Great Bustard wrote:

Great Bustard said:

Anders W wrote:

Great Bustard said:

Great Bustard wrote:

So, where does that leave us? Oh yes -- it leaves us back where we started: you do not understand diffraction in the least.

Click to expand...

If that were the only thing he doesn't understand, the situation would be vastly improved. Regrettably it isn't. ;-)

Click to expand...

There is no sin in not understanding. The sin lies in refusing to learn. The greater sin is to be proud of the first sin.

Click to expand...

Let me guess: you hate the sin but love the sinner.

You're making me feel 'itchy'

Detail Man said:

Detail Man wrote:

Detail Man said:

Great Bustard wrote:

... As far as I'm concerned, you may as well have called my mother...

:-D

Click to expand...

Sometimes our mothers may have "padded their pixels" a bit in order to procreate utter geniuses.

Click to expand...

Unfortunately, here I am reaping the benefits of "au naturale". ;-)

Detail Man said:

BTW (Jessica and Megan ask) - Will you be in the running in the soon upcoming mayoral election ?

Click to expand...

Of course not. But I will be watching the mammarian race with interest. :-D

Wolf Hass said:

Wolf Hass wrote:

Wolf Hass said:

Great Bustard wrote:

Wolf Hass said:

Anders W wrote:

Wolf Hass said:

Great Bustard wrote:

So, where does that leave us? Oh yes -- it leaves us back where we started: you do not understand diffraction in the least.

Click to expand...

If that were the only thing he doesn't understand, the situation would be vastly improved. Regrettably it isn't. ;-)

Click to expand...

There is no sin in not understanding. The sin lies in refusing to learn. The greater sin is to be proud of the first sin.

Click to expand...

Let me guess: you hate the sin but love the sinner.

You're making me feel 'itchy'

Click to expand...

:-D

LTZ470 said:

LTZ470 wrote:

LTZ470 said:

Detail Man wrote:

LTZ470 said:

LTZ470 wrote:

LTZ470 said:

Detail Man wrote:

Whereas the phrase "Diffusion Limited" might be more applicable to debates, the phrase, "Phased Array Linear Dispersion Orthagonal Retro Rockets", always has had a rather sexy ring to me.

Click to expand...

Must I keep repeating:

"inebriated by the exuberance of their own verbosity"

Lol...

Click to expand...

Here's another of your exuberant "golden oldies", magister:

http://www.dpreview.com/forums/thread/3458992

Click to expand...

Aaahhh the golden oldies...those were the days...love it Dm...there is some variance somewhere...strange episode for sure at high ISO's it becomes more relevant...

But actually it is "irrelevant" as well, as I would never use either of these cams for anything over ISO 1250...if that...

Click to expand...

Yes, matters of accuracy and completeness, in addition to specific relevance, ultimately determine the utility, and the overall coherence, of the asking, as well as the answering, of technical questions.

If a little knowledge is dangerous, where is the man who has so much as to be out of danger ?
- Thomas Henry Huxley

Regards,

DM ...

Wolf Hass said:

Wolf Hass wrote:

Wolf Hass said:

Great Bustard wrote:

Wolf Hass said:

Anders W wrote:

Wolf Hass said:

Great Bustard wrote:

So, where does that leave us? Oh yes -- it leaves us back where we started: you do not understand diffraction in the least.

Click to expand...

If that were the only thing he doesn't understand, the situation would be vastly improved. Regrettably it isn't. ;-)

Click to expand...

There is no sin in not understanding. The sin lies in refusing to learn. The greater sin is to be proud of the first sin.

Click to expand...

Let me guess: you hate the sin but love the sinner.

You're making me feel 'itchy'

Click to expand...

Thus, a small "cottage industry" of techno-nannies emerged and thrived in the foral domains, pedantic inquisitors whose very souls have dedicated their efforts to mitigating the wanton self-misuse of "false senses of concreteness". An often thankless, but nevertheless sanctious, societal role that will surely (soon come Thunderdome) be of paramount importance, as future luminaries diligently search the ancient disk-drive scrolls, duly discounting any and all "bum steers" ...

zodiacfml said:

zodiacfml wrote:

The OP is correct, diffraction is simply that.

Click to expand...

You make a compelling case, sir! ;-)

zodiacfml said:

Everything else constant except changing the f number of the lens, diffraction just limits the information we can capture.

Click to expand...

Yes. However, lens aberrations also limit that information, and each lens on each system has the balance point for maximum information capture at a different point (not to mention other sources of blur, such DOF, motion, and noise).

zodiacfml said:

Imagine the aperture of the lens as a filter/hole and photons of light as balls. For a diffraction limited f-stop, larger photons/balls such as the wavelengths of RED gets filtered and can't get through this hole. So, what gets through that hole is always constant regardless of what catches/absorbs it (such as a sensor format or number of pixels).

Click to expand...

All wavelengths get through the hole, it's just that the longer wavelengths smear out more than shorter wavelengths when passing through the aperture.

zodiacfml said:

If we have a large basket (such as a pixel) but two small green balls can fit in, then we are resolution limited. If we have very small baskets (small pixels) but one large red ball comes in, the large ball is sliced into pieces to fit in those baskets. What I'm trying to say here, smaller pixels don't add detail to a diffraction limited capture, it's just over-sampled.

Click to expand...

The photon will either be recorded by a specific pixel or not. The smaller the aperture, the wider the range of pixels that might absorb the photon. In other words, the sensor will record a specific position for the photon, but the smaller the aperture, the greater the chance that the wrong position is recorded.

That said, smaller pixels for a given sensor size (greater sampling) will still record a more accurate image, on average, than larger pixels. It's just that this greater accuracy becomes trivial by some point (in my opinion, that point is f/16 on mFT, f/32 on FF, f/5.6 on the FZ200, etc.).

zodiacfml said:

Yet, why do people say that smaller pixels(high res sensors) is bad due in terms of diffraction? Simple, you just lose efficiency due to diminishing returns. Efficiency of image file sizes. We still get more detail on higher resolution sensors but for very little gain.

Click to expand...

Yes. The blur of diffraction is obscured by the blur of larger pixels. Thus, we notice the effects of diffraction sooner with smaller pixels because we can. With larger pixels, we could not see the diffraction as early since the diffraction itself was absorbed by the blur of larger pixels.

zodiacfml said:

I would like to see a test on this subject since there are two Nikon DSLRs perfect for this, a D3s and a D800.

Click to expand...

Here's an excellent demonstration of pixel size vs aperture in regards to diffraction:

http://www.talknex.com/f2/what-does-diffraction-limited-mean-387/

zodiacfml said:

In matters of DOF and varying sensor formats on diffraction, I think it is a complicated yet futile discussion since we are lens limited anyway, by the smaller formats.

Click to expand...

Actually, smaller formats are more diffraction limited than larger formats, and larger formats are more lens limited than diffraction limited than smaller formats. This is where they oft touted "wide open advantage" that so many mFT and 4/3 users say without understanding the implications.

For example, let's consider a 45 / 1.8 on mFT vs an 85 / 1.8 on FF. Wide open at f/1.8 on mFT is equivalent to stopped down to f/3.5 on FF. So, while the 45 / 1.8 is "sharp wide open", peaking, at, say, f/2.8, the FF lens is not sharp until, say, f/2.8 and peaking at f/5.6.

So, in reality, the FF lens gets sharp a bit earlier than the mFT lens, but both peak at the same point due to diffraction.

Detail Man said:

Detail Man wrote:

If a little knowledge is dangerous, where is the man who has so much as to be out of danger ?
- Thomas Henry Huxley

Click to expand...

...that some are in more danger than others. ;-)

Great Bustard said:

Great Bustard wrote:

Great Bustard said:

Detail Man wrote:

Great Bustard said:

Great Bustard wrote:

... As far as I'm concerned, you may as well have called my mother...

:-D

Click to expand...

Sometimes our mothers may have "padded their pixels" a bit in order to procreate utter geniuses.

Click to expand...

Unfortunately, here I am reaping the benefits of "au naturale". ;-)

Click to expand...

Actual corporeal forms of "beauty" will never match that which our fertile imaginations do forge.

Great Bustard said:

Great Bustard said:

BTW (Jessica and Megan ask) - Will you be in the running in the soon upcoming mayoral election ?

Click to expand...

Of course not. But I will be watching the mammarian race with interest. :-D

Click to expand...

Some folks like to replace the letter "l" with the letter "r" in the term "election". I wonder if your fearful leader will perhaps "rise again" (in "Weineresque" brilliance) to re-enter the running ?