[arpaagent]

I don't see why not. Perhaps there's something I'm missing about the term resolution.

I thought the general guide was: higher pixel density = lower image quality.So a K100 with 1.7 MP/cm² pixel density should give better image quality than a K200 at 2.8 MP/cm² pixel density at say a 3x5" 100% crop.

Or am I confusing terms again,
Brian

Lower megapixel cameras don't show the vices of lenses as much as higher megapixel cameras. Sure the IQ from a 6MP camera will probably be almost as good as a 10MP image downscaled to 6MP, and in most cases this is fine unless you need to print large. But for comparing lenses, the differences in their resolving capabilities may not be noticeable with only 6MP. You could still notice other lens attributes such as contrast or purple fringing, but pure resolution may be hard to measure with only 6MP.

[FHPhotographer]

Lower megapixel cameras don't show the vices of lenses as much as higher megapixel cameras. Sure the IQ from a 6MP camera will probably be almost as good as a 10MP image downscaled to 6MP, and in most cases this is fine unless you need to print large. But for comparing lenses, the differences in their resolving capabilities may not be noticeable with only 6MP. You could still notice other lens attributes such as contrast or purple fringing, but pure resolution may be hard to measure with only 6MP.

What is "pure resolution"?

Your post reiterates what audiobomber said in his/her post. If I'm reading it right. you're saying 10mp is better for lens comparison because it's better -- it's that way because it's that way -- but there's no support for why the same size APS-C sensor with lower pixel density would produce more resolution.

I'm not saying it does or does not work the way you claim, but I'd like to not just take it on faith without some reasonable explanation... if anybody wants to give me one,
Brian

[Lowell Goudge]

What is "pure resolution"?

Your post reiterates what audiobomber said in his/her post. If I'm reading it right. you're saying 10mp is better for lens comparison because it's better -- it's that way because it's that way -- but there's no support for why the same size APS-C sensor with lower pixel density would produce more resolution.

I'm not saying it does or does not work the way you claim, but I'd like to not just take it on faith without some reasonable explanation... if anybody wants to give me one,
Brian

I believe what he is saying is that perhaps a 6 MP camera can't resolve to teh limit of the lens, therefore the quality of a shot taken with 6MP even though excellent, is meaningless.

Since the sensors have not yet reached lens resolution, I guess he does not believe the shoot out is of any value.

That is his opinion. he can contribute or not at his choice, but he does admit, that ultimate resolution is not the only contributor to the image. In the end, the overall of the image is what counts, not the absolute resolution.

[audiobomber]

I don't see why not. Perhaps there's something I'm missing about the term resolution.

Resolution is a measure of the ability to resolve fine details. A 6mp camera cannot resolve as much detail as a high quality lens, so testing lenses with a 6mp camera will show you colour rendition, contrast, bokeh and distortion, etc, but it won't show you any differences in resolution. You can't see differences between resolution in screen-sized web photos, but you will see a difference in comparing crops of a 6, 10 and 14.6mp camera, and you will see it in large prints. There was some question about whether to use 6 or 10mps of resolution for this test. I'm saying it would be better to use 10. If resolution doesn't matter, then what's the argument for not allowing 14mp images?

I thought the general guide was: higher pixel density = lower image quality.So a K100 with 1.7 MP/cm² pixel density should give better image quality than a K200 at 2.8 MP/cm² pixel density at say a 3x5" 100% crop.

Or am I confusing terms again

The general rule is; higher pixel density = higher noise at high ISO.
Image quality is a complex term describing a lot of variables that are unrelated or partly related to pixel density. Even the higher pixel density correlation to noise is not set in stone. The K20D for example has better noise performance than the K100D, even though pixel density is higher, because it uses a different sensor technology (CCD vs CMOS).

---

Last edited by audiobomber; 12-15-2008 at 08:20 PM.

[audiobomber]

I believe what he is saying is that perhaps a 6 MP camera can't resolve to teh limit of the lens, therefore the quality of a shot taken with 6MP even though excellent, is meaningless.

As per my last post, there are many things a 6mp camera will show, but it won't show outright resolution of a high quality lens.

Since the sensors have not yet reached lens resolution, I guess he does not believe the shoot out is of any value.

I doubt that what you say is true about the camera not being able to show lens resolution. I can clearly rank the resolution of my three primes with the K20D.

That is his opinion. he can contribute or not at his choice, but he does admit, that ultimate resolution is not the only contributor to the image. In the end, the overall of the image is what counts, not the absolute resolution.

I agree with you, resolution is only one indicator of image quality, but it is a significant comparator for a many people. You asked a couple of times whether the test should be conducted at 6 or 10mp. I expressed my opionion. Sorry if you didn't like my answer.

[Steinback]

If I do wind up shooting the same setups on film and digital, in some ways I would prefer to use up somewhere close to a full roll. I shoot film seldom enough that I keep that camera unloaded and generally run a full roll at a time through when I do use it. For the shoot out I was thinking of buying a fresh roll of whatever we agree on and then running it through the camera over the course of a few days.

As far as resolution and resolving power I wouldn't worry about it too much. With a number of different people shooting different (but similar) setups at different locations with different camera and lens combinations it's not like we are doing scientifically rigorous testing.

I'm looking at this a more of an interesting experiment that will show how several similar lenses perform in every day shooting with as many controlled variables as we can reasonably agree on. If everyone records (to the best of their ability) things like shutter speed, aperture, camera model and settings (possibly default settings for digital), and lens model then we have some idea of what is different between the photos submitted.

The end result is sort of a database or reference that can be used to get some idea of how these lenses compare. If I like the look of somebody's Tamron photos under similar conditions to my Vivitar maybe I'll go out looking for one. If I decide that another lens is really close to what I'm already using then maybe I'll give it a pass. Who knows, maybe we'll discover that half the lenses compared are indistinguishable in real world use, or somebody will show up with an unusual lens that's a real winner (and its value on the marketplace suddenly triples).

[Blue]

Here is what I have to shoot with:

K20D, K200D or film on one of several bodies. To do 6 mp, I'll have to get a mount for the Tamron and shoot it on the Nikon D70s. I'd really rather not have to mess with that because it won't be able to meter and I'd end up using the histogram as a crutch. The only hand held meter I have at the moment is 60 years old.

[Class A]

6mp is not enough resolution for lens comparisons.

Sorry, but I don't buy this at all. You once did an experiment to that effect but admitted yourself that it didn't meet scientific standards.

If you do the math, i.e., demonstrate that the resolution of a 6MP sensor is not sufficient to distinguish between the kit lens and a high quality prime, then I'll buy your argument. I won't buy it on the basis of your own quick test.

I see distinct differences between different lenses with my 6MP camera. Please show me that I'm wrong with proper arguments, but please don't bash 6MP sensors based on an experiment that you've performed, the outcome of which was questioned by several people.

Check Do Sensors “Outresolve” Lenses? again. Depending on the wavelength under consideration, you'll need a lens that has its optimal sharpness at apertures below f/11 or f/8 in order to outresolve a 7 MP sensor. Good on you if you have one. And, as you have said yourself, you can still judge other important IQ aspect, such as contrast, even when the lens has higher resolution than the sensor.

I thought the general guide was: [U]higher pixel density = lower image quality.

The general rule is; higher pixel density = higher noise at high ISO.

This is not true. High pixel density means you'll have higher noise per pixel. But with respect to the whole image higher density is even better since it leads to finer grained noise. With respect to noise, sensor size matters, but not pixel density.

[Lowell Goudge]

I agree with you, resolution is only one indicator of image quality, but it is a significant comparator for a many people. You asked a couple of times whether the test should be conducted at 6 or 10mp. I expressed my opionion. Sorry if you didn't like my answer.

The issue of MP is not whether the test should be conducted at 6 or 10, but I don't want to exclude people, including my self that don't have a K20.

As for ultimate resolution of the lens, while it may be a significant facotr for pixle peepers, it is not going to be the only basis for a test.

Note also, that this was intended to be a fun challenge not a technical bashing issue.

[audiobomber]

Here is what I have to shoot with:

K20D, K200D or film on one of several bodies. To do 6 mp, I'll have to get a mount for the Tamron and shoot it on the Nikon D70s.

The K20 can be used to shoot 6 or 10mp jpegs. Or you can shoot in raw and downconvert in p-p.

[audiobomber]

The issue of MP is not whether the test should be conducted at 6 or 10, but I don't want to exclude people, including my self that don't have a K20.

Now you've confused me. Are you saying 6 and 10mp can be used, but not 14? I thought you were looking for one standard resolution.

As for ultimate resolution of the lens, while it may be a significant facotr for pixle peepers, it is not going to be the only basis for a test.

I'm sure you'd agree that I never said it would. But resolution is a factor, especially with a telephoto lens, because that type of photography often involves cropping.

Note also, that this was intended to be a fun challenge not a technical bashing issue.

I don't see that I've bashed anything or anyone. You talked about 6 or 10mp a couple of times. I thought you were looking for some input so I gave mine. If the idea of this thread is just for fun, great, it's your idea, it's your thread, make it universal and set the standard at 6mp, or don't have a standard resolution. Either way will maximize the number of people who can participate. But if the standard is 6 or 10 but not 14, then I don't understand what you're doing.

PS I don't expect to show the 55-300mm in these tests. I dropped my K20 the other day and it landed on the lens hood. The 55-300 is damaged, it's producing fuzzy images. Maybe someone else can oblige because I'd be very interested in the results.

[audiobomber]

As far as resolution and resolving power I wouldn't worry about it too much. With a number of different people shooting different (but similar) setups at different locations with different camera and lens combinations it's not like we are doing scientifically rigorous testing.

True, but I find that a photograph of a printed page is very telling of resolution. It would have to be the same page under similar conditions. It is feasible.

[audiobomber]

I see distinct differences between different lenses with my 6MP camera. Please show me that I'm wrong with proper arguments, but please don't bash 6MP sensors based on an experiment that you've performed, the outcome of which was questioned by several people.

Check Do Sensors “Outresolve” Lenses? again. Depending on the wavelength under consideration, you'll need a lens that has its optimal sharpness at apertures below f/11 or f/8 in order to outresolve a 7 MP sensor.

Lots of lenses reach optimal sharpness at F5.6. The FA 50mm reaches maxiimum center sharpness at F4.

I don't know why you think I'm bashing 6mp sensors. I love the IQ of my K100DS, but let's be realistic. If you're going to crop hard or print a large image, then 6mp will show its limitations. People on these boards complain that Photozone only uses a K10D for their resolution tests. Do you honestly think it would be fair of them to use a K100D?

This is not true. High pixel density means you'll have higher noise per pixel. But with respect to the whole image higher density is even better since it leads to finer grained noise. With respect to noise, sensor size matters, but not pixel density.

So you would say a K10D, with the same sensor and higher pixel density, produces better images than a K100D at high ISO. That's a controversial statement, isn't it?

The linked article below is pretty dense reading, but the conclusion at the end is simple enough. More pixels produce higher noise at higher ISO and larger sensors with more pixels produce higher quality images at low ISO.

"Bottom line: Among the important measures of image quality are signal-to-noise ratio of the capture process, and resolution. It was shown that for fixed sensor format, the light collection efficiency per unit area is essentially independent of pixel size, over a huge range of pixel sizes from 2 microns to over 8 microns, and is therefore independent of the number of megapixels. Noise performance per unit area was seen to be only weakly dependent on pixel size. The S/N ratio per unit area is much the same over a wide range of pixel sizes. There is an advantage to big pixels in low light (high ISO) applications, where read noise is an important detractor from image quality, and big pixels currently have lower read noise than aggregations of small pixels of equal area. For low ISO applications, the situation is reversed in current implementations -- if anything, smaller pixels perform somewhat better in terms of S/N ratio (while offering more resolution). A further exploration of these issues can be found on the supplemental page. Rather than having strong dependence on the pixel size, the noise performance instead depends quite strongly on sensor size -- bigger sensors yield higher quality images, by capturing more signal (photons).

The other main measure of image quality is the resolution in line pairs/picture height; it is by definition independent of the sensor size, and depends only on the megapixel count. The more megapixels, the more resolution, up to the limits imposed by the system's optics."

http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html

[arpaagent]

Enough with the discussion. We can do both 6MP and 10MP. If there are enough people involved, we can look at both, or simply downsample the 10MP images to 6MP if that's what we need to end up doing.

But all this theory isn't telling us anything about the comparison of the zoom lenses! Now lets pic some subjects! Here would be some good starters, generally things that can be found everywhere:

Trees (either close-up of trunks or full shot with lots of branches/leaves)
Architectural buildings (would have be same material, brick or stone or etc)
People! (Plenty of them to go around, but more subjective in terms of image quality so maybe not the best choice)
Statues (a little bit harder to find, but make great photographic subjects)

Other suggestions? Thoughts?

-Jim

[Steinback]

I had been trying to think of things that could be found anywhere in the world without too much trouble. It will almost becomes a photo scavenger hunt of sorts to find subjects.

These are some ideas:

A fire hydrant at 140mm zoom, mid focus distance
A traffic light at 200mm zoom, somewhere around infinity focus
A car with chrome on it (anybody want to specify body colour?) at 80mm zoom, mid focus range
A fork (or other utensil as available) at 80mm, closest focus
A drinking glass (half full? half empty?) in sunlight, 80mm, mid focus range
A person, head and shoulders shot (same person? same lighting?) at 80, 140 and 200mm with whatever focus you need to make it work

I like the idea of statues, but I also live down the street from a giant sculpture park.
I will post other subject ideas as I come up with them.