Thread: High ISO, noise, megapixels and etc.

As always...do whatever works for you...don't believe every theory that you read on the internet.

I alway find this conversation interesting. It is fascinating to hear people talk about image quality, at least in terms of the type of photography used creatively, in absolute terms. Image quality in applied photography comes down to "making pleasing images." Or, in other words, it is entirely subjective. I think the problem come from the term "quality" as we also associate the idea of "high" and "low" with it like it had some real meaning.

Image quality is the result of a process. Image quality is the flavor imparted to an image based on choices the photographer makes. Part of that is the camera and optics, but there is a lot more than that--exposure, processing, handling, focus, framing, and a whole of other factors that affect the final outcome. But you can make great image from cell phones, Holgas, 8x10 view cameras, wet plates, pinhole cameras and any kind of imaging device. If used well (this is the operator), the image quality will always be high. Making high-quality images has nothing to do with the equipment, but the skill of the operator.

If image quality were actually a technical, objective factor, then most of the worlds great photographs would be bad. Simply taking things like dynamic range and resolving power, then the work by a photographer like Henri Cartier Bresson would be poor as the DR of his film was low, his optics soft, and his exposure not very good in a purely technical sense. The power of his images is not simply a technical criteria. He had a process and he mastered it. Having more DR would not have made any of his work "better."

Nide to have you back in the mix, Hikari. I've missed your posts.

Gordon

Originally Posted by Hikari

I alway find this conversation interesting. It is fascinating to hear people talk about image quality, at least in terms of the type of photography used creatively, in absolute terms. Image quality in applied photography comes down to "making pleasing images." Or, in other words, it is entirely subjective. I think the problem come from the term "quality" as we also associate the idea of "high" and "low" with it like it had some real meaning.

Image quality is the result of a process. Image quality is the flavor imparted to an image based on choices the photographer makes. Part of that is the camera and optics, but there is a lot more than that--exposure, processing, handling, focus, framing, and a whole of other factors that affect the final outcome. But you can make great image from cell phones, Holgas, 8x10 view cameras, wet plates, pinhole cameras and any kind of imaging device. If used well (this is the operator), the image quality will always be high. Making high-quality images has nothing to do with the equipment, but the skill of the operator.

If image quality were actually a technical, objective factor, then most of the worlds great photographs would be bad. Simply taking things like dynamic range and resolving power, then the work by a photographer like Henri Cartier Bresson would be poor as the DR of his film was low, his optics soft, and his exposure not very good in a purely technical sense. The power of his images is not simply a technical criteria. He had a process and he mastered it. Having more DR would not have made any of his work "better."

I've looked around the internet and did not find any consensus for a singular definition for "Image Quality". Hikari, you choose to use the most subjective definition of Image Quality I could find. I think most photographers in this digital age, tend to use Image Quality as a general description for objective and measurable elements of an image, resolution, noise and dynamic range.

I used the term 'Image Impact' to describe the subjective elements of an image and 'Image Quality' to describe the objective/measurable elements.

Gary

As much as I like the message in Hikari's post, I have to agree with Gary. "Image Quality" or IQ as I assume most people here use it is meant of what the sensor (or camera) is capable of in a lab, without factoring in the skill level of the photographer.

The quality of the photograph is a separate conversation. Any to say that HCB's photographs wouldn't be better with more dynamic range is pure conjecture. One thing is for sure. He learned his tools because he was constantly out there shooting. He wasn't hanging out discussing the relative merits of different cameras and constantly upgrading.

Originally Posted by Gary

I've looked around the internet and did not find any consensus for a singular definition for "Image Quality". Hikari, you choose to use the most subjective definition of Image Quality I could find. I think most photographers in this digital age, tend to use Image Quality as a general description for objective and measurable elements of an image, resolution, noise and dynamic range.

I used the term 'Image Impact' to describe the subjective elements of an image and 'Image Quality' to describe the objective/measurable elements.

Gary

Gary, the internet is the best source of information. Even Thomas Jefferson is quoted stating, "A man's words are the most reasonable when posted in the thread."

My background in photography is technical. One of the first things you learn is that when Kodak and those folks where building emulsions and processes, they way they determined the right combinations were to make images and decide if they looked good. Then the reversed engineered the process to find the criteria that produced "the most pleasing" image. I an not making up the term. In my old company, Konica Minolta Photo Imaging, they did the same thing. They actually took pictures and looked at them to judge the output. All of this is purely subjective. It is also why the colors from different camera companies are not the same. If the were some objective measure, all cameras would basically have the same output--that would not be hard are the filters on the sensors are essentially the same and could be mixed the same by by every manufacturer essentially achieving the same result. How many photographers post that they prefer so-and-so manufacturer's colors? It is all subjective coming down to what someone finds "pleasing."

But with most professions, the workings are not usually public knowledge. And you know this. Remember the glamour of being a photographer? it is just as everyone says it is. I certainly know what most photographer say on internet forums. And there is certainly a lot of misinformation posted.

Originally Posted by Luke

As much as I like the message in Hikari's post, I have to agree with Gary. "Image Quality" or IQ as I assume most people here use it is meant of what the sensor (or camera) is capable of in a lab, without factoring in the skill level of the photographer.

I am not "most people." I simply thought to broaden the conversation to add my view.

Any to say that HCB's photographs wouldn't be better with more dynamic range is pure conjecture.

Actually, it is not. More DR would make it different--better would be a personal judgement. HCB, like many other photographers, as part of his skill learned how his system saw the world. He could visualize the results (Adams used "previsualization," which even he later said was a daft term). So it is more than likely that he anticipated what the final image would be because of the limits of the technology and factored that into its creation. As you pointed out, HBC was no gear head. After a while working as a photographer you realize that chasing technical qualities ultimately does not really add to your work--there are still wonderful photographers like Mary Ellen Mark that prefer film systems.

I don't mind folks referring to IQ as Dr, resolution, etc., but it is not really very valuable and mostly comes from a simple idea that more is better. Unfortunately, it is not true. Everyone raves about the images from the Foveon sensor. Part of the great look is the narrower DR. "More" is not "better." And when you rave about IQ of a sensor, what more could it be than a subjective opinion? No one understands what, for example, 13.5EV DR really means in actual use.

Originally Posted by Hikari

Gary, the internet is the best source of information. Even Thomas Jefferson is quoted stating, "A man's words are the most reasonable when posted in the thread."

My background in photography is technical. One of the first things you learn is that when Kodak and those folks where building emulsions and processes, they way they determined the right combinations were to make images and decide if they looked good. Then the reversed engineered the process to find the criteria that produced "the most pleasing" image. I an not making up the term. In my old company, Konica Minolta Photo Imaging, they did the same thing. They actually took pictures and looked at them to judge the output. All of this is purely subjective. It is also why the colors from different camera companies are not the same. If the were some objective measure, all cameras would basically have the same output--that would not be hard are the filters on the sensors are essentially the same and could be mixed the same by by every manufacturer essentially achieving the same result. How many photographers post that they prefer so-and-so manufacturer's colors? It is all subjective coming down to what someone finds "pleasing."

But with most professions, the workings are not usually public knowledge. And you know this. Remember the glamour of being a photographer? it is just as everyone says it is. I certainly know what most photographer say on internet forums. And there is certainly a lot of misinformation posted.

Jefferson also said that ... in the marketplace of free ideas, the truth will prevail ... or was that Milton ...

I understand where you're coming from. Some feel that the pinhole lens (sans glass) is the baseline to judge IQ. But, (the big but), I think most/some/a lot of us judge IQ purely on an objective, measurable basis. And, (the big and), while using IQ in this manner may not be technically correct or historically true, it certainly has merit as a working definition for measurable qualities of a sensor and lens.

IQ is a relatively new term to me. When I was working news, I do not remember lenses being judged by lines or images by dots. The higher the ASA, the greater the grain is all one had to remember. The human element of exposure, development technique (agitation) and chemicals had as much, if not greater, impact on grain than film type. 90%+ of photo journalists used Nikon and back then Nikon only made one level of lenses, the best they could. The 50mm came in a couple of flavors of F-stop's, but the rest of the focal lengths only had one lens in the line-up.

Gary

Originally Posted by Gary

I think most/some/a lot of us judge IQ purely on an objective, measurable basis.

And that is fine, if people really understood those numbers. If you have a sensor with a DR of 13EV, would that be enough for a picture of your garden right now? That of course depends if you or your camera gets the exposure right. And if you lens is high or low contrast, that changes that again.

Now, I use the numbers too, but as a simple gauge. EV13.5 and EV14, is that difference something really to worry about? Is the difference in the EV coming from the shadows or the highlights (it does make a difference). I imagine you do something I do, I learn how my camera sees. I recognize the situation where I need to change my exposure and decide whether I lose the shadows or highlights or lose them for effect. I also understand how far I can push my processing.The DR values has nothing to do with it, at least after I buy my camera.

I have also had folks write that this objective data like DxO mark scores are wrong--they have taken an picture with such and such camera and they "know" it is wrong! So much for objectivity.

I remember when the D800 was released. Folks were saying the lenses would not have enough resolution for the sensor. It would be a disaster! But why? Objectively, it is not hard to calculate that the D800 sensor has a resolution of about 100 l/mm. Optically speaking, that is not a tough target. But you still get folks saying that the lenses have to be made of Kriptonite and forges by elves in Mordor. And here is the kicker. I get an optical report that states my lens can resolve 200 l/mm. But it only does that with a high-contrast target. I go out on an overcast day and my lens is resolving less.

People look at noise, but at 100%. At 100%, if it looks the same, then the sensors are judged the same. But a given amount of noise on a 12MP sensor is not really the same as the same noise on a 24MP sensor.

We both have done a great deal of photography without the need for the technical minutia. I certainly look at that stuff, but like many of the engineers I knew, they are indicators of problems, not of "goodness."

I can agree that objectives numbers are good to have, but you need to know what their relevance is. If you have an incomplete picture, the following false conclusion can be made (which are also widely published over the internet):

Print size is limited by pixel resolution

DoF decreases with pixel resolution

Diffraction limit is based on pixel pitch

All of these are false. Not because there is not a number associated with these ideas, but rather people think there is an objective, absolute frame in which to judge them. The problem is they need to judged based on a subjective, relative one--which has always been true.

The stress on the numbers is far too great. How many people actually choose a camera based on pretty pictures they see on the net rather than DxOmark scores? Pretty picture sell more cameras. Camera ultimately are sold because the buyer thinks they are in someway cool. So much for objectivity.

I don't think we are really disagreeing. I think in many ways we chose and use gear the same way. I just find the emphasis on measured values is not very useful if people cannot place them in context. And when placed in context, they really are not very important at all.

Originally Posted by Hikari

... All of these are false. ...

Whew ... I am glad you added that. My head was spinning for a while there. (Doesn't take much to spin my head.)

I don't think we are arguing. In fact my most over-used, ad nauseam, words on these forums are "significant" and "the only thing that matters is the final image". In my opening post I stated I feel compelled to respond but really don't know how I should respond. What your posts have crystalized for me is that with the equipment I now possess, "IQ", (the measurable side of IQ), and increases in measurable IQ really doesn't matter much to me. Any significant increases or betterment of "IQ" (the subjective, total, all-encompassing impact of the image), will be operator driven not hardware driven.

I had a Canon 1D-X, but I got rid of it. My final images from the $7,000 Canon were not significantly better than with the OM-D. I shot the same with the Canon as I did with the Olympus, same subjects, same focal lengths, et cetera. The Canon was not significantly more capable of capturing and delivering my vision than the Olympus. (Except for sports/action stuff.)

Gary