Example: a photo of a woman sitting on a bench. If you zoom in at 100% and can count each and every hair on her head, you have a high resolving lens. If the photo looks crisp and sharp with plenty of contrast between the elements of the bench and other small elements of the photo at a normal viewing size, you have a lens with high micro contrast. A lens can be both but not necessarily
Bifurcator wrote:
And isn't lens resolution also a function of sharpness? We say a lens which can resolve small details has good micro-contrrast and we've already agreed that micro-contrast and sharpness are synonymous.
That's an odd way to put it, because technically there is no property like "sharpness", technically, there is only resolution and contrast, both of which can be measured. Lenses can have low contrast but high resolution (of course some contrast is needed to resolve detail at all, and you have to draw a line to decide at which contrast level you still consider a structure to be "resolved"), or they can have high contrast (in coarser structures) but low resolution (meaning they don't resolve finer structures at all).
"Sharpness" is a subjective perception resulting from the contrast levels at different frequencies, so the word shouldn't be used to describe the measurable properties from which it is (subjectively) derived.
Guys (Edward, Bif), I think you need to re-read the paper "How to Read MTF Curves" by Hubert Nasse.
edwardkaraa wrote:
Micro-contrast and sharpness are synonymous in my opinion. When a lens has high micro contrast, it means that it is a sharp lens. Micro-contrast corresponds roughly to the 10 lpmm and 20 lpmm on Zeiss MTF tables. A high number indicates that the lens gives a very good illusion of sharpness while the 40 lpmm indicates the resolving power of very fine details.
This isn't quite right. Resolution just means that two image elements (line-pairs being the most simple ones) have sufficient contrast to be distinguishable from each other. Because we're talking about small image elements, we're always talking about micro-contrast and not global contrast. See the bottom half of page 16 of the Zeiss paper.
You are redefining the term micro-contrast to correlate it to 10 and 20 lp/mm but that depends on the sensor size, viewing size and distance so it's kind of arbitrary. You couldn't use those same numbers for MFT or large format lenses for example. The vertical scale on the Zeiss MTF charts simply shows the microcontrast at different resolutions (10/20/40 lp/mm) across the image's diagonal.
The 10 lpmm and 20 lpmm lines show the contrast on coarse and medium detail while the 40 lpmm line shows the contrast on fine details. Of course we are talking about 35mm format here, so no confusion. The higher the contrast on coarse and medium detail, the better the apparent sharpness. The higher the contrast on fine detail, the higher the resolving power.
edwardkaraa wrote:
The 10 lpmm and 20 lpmm lines show the contrast on coarse and medium detail while the 40 lpmm line shows the contrast on fine details.
It still doesn't make sense to associate 10 or 20 lp/mm with micro-contrast and 40 lp/mm not. The sharpness of 40 lp/mm is defined by the micro-contrast just like any other number of lp/mm (unless you're talking about a sensor that's only a few millimeters across).
AhamB wrote:
It still doesn't make sense to associate 10 or 20 lp/mm with micro-contrast and 40 lp/mm not. The sharpness of 40 lp/mm is defined by the micro-contrast just like any other number of lp/mm (unless you're talking about a sensor that's only a few millimeters across).
In this case what is your definition of micro-contrast? For me, it is the ability of a lens to produce crisp and sharp results when you view the image as a whole. This sharpness is due to the high contrast between the finest details in the image. This fine detail corresponds to the 10 to 20 lpmm lines. The 40 lpmm is too small to affect the contrast unless you blow up the image to large sizes. Also, in MTF charts, you can always see a pattern of highest contrast at 10, followed by 20, then 40 lpmm. In many lenses, the 40 lpmm is considerably lower than the other two, and very rarely does it reach excellent levels even in Zeiss and Leica lenses.
"Incidentally, it is not the case that, when designing a lens, a decision has to be made between high resolving power and good contrast rendition; both are possible
for lenses with good correction.
But what does ‘contrast rendition’ actually mean? We must not forget that when we talk about ‘contrast’ we always mean micro contrast, i.e. structures, which we can just about see or just cannot see with the naked eye, for example on a slide. But if we photograph a chessboard, for example, so that it fills the format, the contrast between the black and the white squares has nothing to do with this."
I think he is staying within the realm of the data as it is measured in the image plane. When you talk about micro-contrast in an enlarged image then it gets more complicated because display size and viewing distance become factors. Maybe it's better to use a word like edge contrast (though I'm not sure how that term is defined), something that scales with the display size. You're either talking about an enlargement or the lens projection but it can't be both, IMO.
Most people talk about resolution and contrast, so if I understand it correctly, Dr. Nasse is using "microcontrast" interchangeably with "resolution"?
To my limited mind, there is a continuum:
On one end, there is resolution, the ability to resolve the finest detail at the pixel level. But at regular viewing size we are not very sensitive to this.
On the other end, there is global contrast
And somewhere in between, there is microcontrast/edge contrast, the ability to give clear edge definition and texture at normal viewing condition. This is what I think most Zeiss lenses are being optimized for.
Lens that have high resolution but low microcontrast that I can think of:
Canon 55/1.2, here's Edwin Put's description of the lens despite it's high resolution:
The snag is that micro contrast is quite low and even the 10 lp/mm pattern has soft edges. In practical shooting you see this as a quite soft image when looking at finer details. You see the detail in the mage, but it lacks the sharp delineation of outlines that give the picture transparancy and sparkle.
aleksanderpolo wrote:
And somewhere in between, there is microcontrast/edge contrast, the ability to give clear edge definition and texture at normal viewing condition. This is what I think most Zeiss lenses are being optimized for.
Exactly!!!! And that's the reason why I am a Zeiss fanatic!
Most people talk about resolution and contrast, so if I understand it correctly, Dr. Nasse is using "microcontrast" interchangeably with "resolution"?
It seems so. And this is what I was wanting to point out - or get clarification on if otherwise.
To my limited mind, there is a continuum:
On one end, there is resolution, the ability to resolve the finest detail at the pixel level. But at regular viewing size we are not very sensitive to this.
Well, we can't easily point to the individual pixels and say: There it is! but IMO this is one of the more critical (and enjoyable) qualities of a lens. And when scaling down using "For Smooth Gradients" the level of that attribute stays with the image from the original 6000x4000 all the way down to about 1,500x1000 past where it begins to degrade and is just about totally gone by 800x600. But again I equate the three terms as being synonymous: sharpness, micro-contrast, and resolution.
On the other end, there is global contrast
And somewhere in between, there is microcontrast/edge contrast, the ability to give clear edge definition and texture at normal viewing condition. This is what I think most Zeiss lenses are being optimized for.
OK, now I'm confused. I can understand what "global contrast" is. Where one lens will render the darks and lights of Dr. Nasse's chess board appropriately we can say it has good global contrast and where another lens renders the darks too light and/or the lights too dark we can say the lens has poor global contrast. And of course a lens that make the darks too dark and/or the lights too light we can say the lens is too contrasty.
But you lost me with the idea of there being something in-between and separate from the others. Although I can see how the two (micro and global contrast) work together.
Lens that have high resolution but low microcontrast that I can think of:
Canon 55/1.2, here's Edwin Put's description of the lens despite it's high resolution:
"The snag is that micro contrast is quite low and even the 10 lp/mm pattern has soft edges. In practical shooting you see this as a quite soft image when looking at finer details. You see the detail in the mage, but it lacks the sharp delineation of outlines that give the picture transparancy and sparkle."
There are no image samples so I have to chalk up that page as just so much useless blather. Unfortunately. Also his terminology is way over on the romantic side making most of his descriptions seemingly nonsensical.
I have the Canon 58/1.2 and there is a good amount of lens bloom at wide apertures which cause halos in hi contrast (large structure) areas and seemingly wash the micro-contrast out in fine structure areas.
Bif, I don't know. I am just trying to reconcile with the fact that, there are some lens that have better test resolution than others, but doesn't seem to have the clear edge definition. Some examples in mind:
CV 50/1.5 vs ZM 50/2, the CV have better test resolution, but to my eyes ZM have better edge definition
Samyang 85/1.4 vs Planar 85/1.4, similar story
Just as our eyes respond to color variation best when it is in the mid-tone region (the S curve), I think there is also a spatial frequency that we are most sensitive to, and that of course depends on the viewing size of the image. Scale it down to very small size, we will only notice global contrast. View it at 1:1, we can only see the resolution difference. View it at a normal viewing angle (30-40 degree?), then it would be some frequency between the two ends?
Bifurcator wrote:
It seems so. And this is what I was wanting to point out - or get clarification on if otherwise.
They are not interchangeable. Resolution just means two image features (lines usually) being resolved at a certain minimum contrast value.
MTF however can be said to be contrast at a certain resolution (actually the percentage of the original object contrast that is transferred to the image). Because we're always talking about small features in the image it's valid to call this microcontrast.
There is no point where microcontrast suddenly turns into "resolution", but for lens discussion purposes it's fine to talk about microcontrast as the contrast in structures just larger than the contrast in the very finest detail (which we usually just call sharpness).
After moving to m43, I bought the much maligned Oly 17 little pancake prime. I like the size and FL so I use it for most my shooting even though the lens is not nearly the sharpest I own. Occasionally, with a large print I notice a little bit of corner softness, then I frame the print and never think it matters... you have to really look close and know what you're looking for to even see it. Most people don't look at photos that way anyway.
Yes, this is a gear forum and most of us have problems ignoring the tiniest defects, but the point in the end is a photograph, so we should stay focused on that.
Once, at a gallery I was showing at, some crusty old Leica user showed up with a loupe. When I saw hime examining each print though the loupe I laughed and asked him what he was doing.
He said, "I always look at photos with a loupe first, if they don't look good there I don't bother looking at the photo."
I told him that he was probably missing a lot of good photographs then.
It's pretty easy to look at a photo and admire it (or not) based only on the art. That's not anything to do with this discussion; that's a whole'nother thing.
Bifurcator wrote:
It's pretty easy to look at a photo and admire it (or not) based only on the art. That's not anything to do with this discussion; that's a whole'nother thing.
i thought that was kinda the point of the discussion. how much does sharpness really matter? or does it matter at all?
incidentally, the answer is no, not for art.
if you want to sell stock photography or impress people on forums, then the answer is usually yes.
I think for sports photography, product photography, commercial shooting, travel shooting, fashion shooting, and most journalism, sharpness counts.
A lot of other things matter first but its the extreme exception and not the rule where something like iPhone photos are used.
As retina style displays become more common, I think you'll only see a greater demand in high quality images. People aren't going to tolerate looking at blurry pixelated images in their high quality tablets and computer displays while reading their favorite digitized magazines and newspapers.
I think for everything sharpness counts - a lot! For your art you can always blur an image but you can't get details out of an image where none exist. It's MUCH better to have and use the superior tool and then form your art instead of using an inferior tool and be stuck with just one result.
retina style displays are actually making low res images look better in my opinion due to superior interpolation and the fact that they don't make the image any larger in pixel independent size.
This is a matter of degrees. At one point -- a long time ago -- AA said that most "modern" lenses were "good enough" to take "good photos." This was more than 25 years ago I think, lenses are now much better.
When you can barely see any difference, if any, in sharpness when viewing a print, it is hard to support your claim that, "sharpness counts - a lot!"
Yes, better is better, but chasing the "best" gear is rarely the most productive way to get your best photos.
Bifurcator wrote:
I think for everything sharpness counts - a lot! For your art you can always blur an image but you can't get details out of an image where none exist. It's MUCH better to have and use the superior tool and then form your art instead of using an inferior tool and be stuck with just one result.
retina style displays are actually making low res images look better in my opinion due to superior interpolation and the fact that they don't make the image any larger in pixel independent size.
I don't know, I used my work ipad 2 to look at an image then saw it on my personal iPad 3 and a 800x600 image taking up the full screen looked horrrible on the iPad 3.
I can understand what "global contrast" is. Where one lens will render the darks and lights of Dr. Nasse's chess board appropriately we can say it has good global contrast and where another lens renders the darks too light and/or the lights too dark we can say the lens has poor global contrast. And of course a lens that make the darks too dark and/or the lights too light we can say the lens is too contrasty.