[BrianR]

I have a hard time with a score like 4 megapixels with a zoom. Surely at some point in the zoom range it is better and at certain apertures.

I haven't look much at their lens testing, but it looks like the same problem as the one number 'sensor score'- they have test data across the zoom and aperture range (but apparently not focusing distance) and are distilling it down (in some not totally specified way) to one easy to consume number. You can see more details:

Tests and reviews for the lens Tamron AF18-200mm F/3.5-6.3 XR Di-II LD ASPHERICAL (IF) MACRO Pentax mounted on Pentax K-5 IIs Measurements - DxOMark
Tests and reviews for the lens Tamron AF18-200mm F/3.5-6.3 XR Di-II LD ASPHERICAL (IF) MACRO Pentax mounted on Pentax K-3 Measurements - DxOMark

What I find a little odd is how (under the 'sharpness' measurement tab) the lens on the k5iis seems to have a bizarre local sharpness performance peak at 85mm when stopped way down but the one on the k-3 has a local peak at 50mm (also when stopped way down). Tests on other bodies don't seem to have this peak.

---

Last edited by BrianR; 02-25-2015 at 08:57 AM.

[wildman]

Definitely the weakest part of the DxO logic is the inference that their standards mean something in terms of image production.

If the DXO score does not explicitly imply image quality what is it for?
After all in photography, more than in most areas, perception is reality.

Also an inverse case can be made that after a certain point, take your pick, the DXO score becomes meaningless as there is no real perceptual improvement after a certain point on the high end given the way photographer's actually use a lens in the real world.

[audiobomber]

In other words does a "horrible" DXO score necessarily give you, at least technically, a horrible image?

The sharpness score tells you about sharpness. Since when is sharpness the determinant of a great image? Is it even in the top five? Sharpness is a useful technical aspect of a lens.

Furthermore, the sharpness score is a single number intended to generally describe the sharpness of a zoom at all apertures, at every focal length and across the frame. No single number can tell you everything about the lens, and DXOMark says so, quite plainly as I quoted above.

[normhead]

If the DXO score does not explicitly imply image quality what is it for?
After all in photography, more than in most areas, perception is reality.

It's to help sell their software, what? You think they are some kind of philanthropists?

Or as I continually point out... what DxO and most people on this site refer to as Image Quality (as in resolution etc.) has never been linked to people image enjoyment factor. Which is really all that matters.

https://www.pentaxforums.com/forums/38-photographic-technique/203024-who-took...hotograph.html

Look at the Cezanne. It is very low res.... you can't even read the guy's cards.... technical image quality is not the same as the value of the photo, in any way related to people's enjoyment of an image.

The image below was taken in low light, with a consumer grade ƒ5.6 lens and probably isn't functionally more than 4 or5 MP. But it got over 750 likes on Facebook. It is so far from my best technical image it isn't funny, but it's my most "liked" image ever.



I keep saying, the things that make good photographs are not the numbers, but the content.
https://www.pentaxforums.com/forums/38-photographic-technique/203024-who-took...hotograph.html

There is absolutely no human value to a perfect high res, low noise, perfect exposure, of a bland subject. The values that make photographs humanly meaningful, are not technical. I wish I could beat folks over the head with a baseball bat, until they get it.

And no, a more high res, higher dynamic range, higher anything image of a compelling, low res, low anything image would not necessarily be more compelling. Technical improvement in a numbers related environment does not necessarily make for a more enjoyable image when looking at how a human values that image. That is perhaps the biggest lie perpetrated by camera companies in their drive for more sales and more profits.

The simple fact is, if you can't take a compelling image with a 6 MP camera, you can't take a compelling image with a 51 Mp camera. If you can't take a compelling image with APS-c, you can't take a compelling image with FF or MF or 4/3. The format of the camera is not going to save you. The resolution is not going to save you, and the technical specs are not going to save you.

---

Last edited by normhead; 02-25-2015 at 03:05 PM.

[audiobomber]

I have a hard time with a score like 4 megapixels with a zoom. Surely at some point in the zoom range it is better and at certain apertures. I couldn't give the DA 55-300 a single score. It is pretty decent at 200mm and f5.6, but at 300mm needs to be at f8 minimum for good results. And when DXO Mark is giving numbers that are well below the predicted numbers, what does that mean?

It doesn't really make sense that a lens will score 4 megapixels on a K5 (16 megapixel sensor) and then score 5 megapixels on a K3 (24 megapixel sensor). Why should the lens get any better moving to a sensor with higher pixel density?

My experience with the K5 II versus K3 is that there is some improvement in resolution, as long as you shoot in your lens's sweet spot, shoot low iso, and have good technique. But it certainly isn't earth shattering. As to whether or not a "20 percent" resolution increase would translate to an increase in printing size, I have no idea, but I don't think it would be a huge difference.

See my reply to Wildman about the single number. You can view the performance in detail by aperture and focal length by clicking on the Measurements tab and the sub-tabs.

Regarding a higher score on the K-3 than a K-5, it does make sense. The effect of an AA filter is easy to understand, right? The point of a blur filter is to blur the image, IOW, reduce the resolution. So the K-3 and K-5 IIs will have higher resolution. Then there are more pixels in the K-3 image, which magnifies the image more on a computer screen or in a larger print. If the lens is not at the limit of its resolution on the K-5 IIs, the K-3's extra magnification will show greater detail.

Regarding your last point, I believe sharpness is highly overrated as a comparator. It certainly matters to me a lot less than subject matter, composition, colour and contrast. Sharpness is valuable in some images and some types of photography and doesn't matter at all in a lot of images.

[osv]

The simple fact is, if you can't take a compelling image with a 6 MP camera, you can't take a compelling image with a 51 Mp camera. If you can't take a compelling image with APS-c, you can't take a compelling image with FF or MF or 4/3. The format of the camera is not going to save you. The resolution is not going to save you, and the technical specs are not going to save you.

and it's not going to save you from an out of focus image, it's not going to save you from an overexposed image, it's not going to save you from an underexposed image, blah blah blah

good grief, lol

[Dartmoor Dave]

The simple fact is, if you can't take a compelling image with a 6 MP camera, you can't take a compelling image with a 51 Mp camera. If you can't take a compelling image with APS-c, you can't take a compelling image with FF or MF or 4/3. The format of the camera is not going to save you. The resolution is not going to save you, and the technical specs are not going to save you.

So true. So true that somebody should carve it in stone and use it as a plinth for a Statue of the Unknown Photographer.

[Sagitta]

There is absolutely no human value to a perfect high res, low noise, perfect exposure, of a bland subject. The values that make photographs humanly meaningful, are not technical. I wish I could beat folks over the head with a baseball bat, until they get it.

That said, some photos that are downright boring today could evoke interest in 50 years.

That shot of those two beat up cars parked on the roadside would be meh today, but in 50 years someone might get excited over that shot of the vintage SmartCar parked next to that 2013 Hyundai.

You never know.

[normhead]

That said, some photos that are downright boring today could evoke interest in 50 years.

That shot of those two beat up cars parked on the roadside would be meh today, but in 50 years someone might get excited over that shot of the vintage SmartCar parked next to that 2013 Hyundai.

You never know.

I'm going with the "untestable hypothesis" on that one. At least untestable for me. Some of these young whipper snappers might be around in 50 years, but my odds aren't good.

[wildman]

Since when is sharpness the determinant of a great image?

When photography became a mass market.
At least to some extent sharp can be quantified and it's easy to sell numbers so 30 is better than 25 and 300hp is better than 250.

It's to help sell their software,

Touche'.

[audiobomber]

When photography became a mass market.
At least to some extent sharp can be quantified and it's easy to sell numbers so 30 is better than 25 and 300hp is better than 250.

Most importantly, numbers are useful in internet pissing contests.

[Nicolas06]

"Obviously" is not a scientific term. After such a statement the most obvious response is "prove it".

I'd like to see a proof that included the relative size of circles of confusional various subject distances of subject and back ground proving that you can match both DoF and have the same size circles of confusion using different formats. I can imagine it being both possible and impossible. So I need to see a proof.

If you are serious about that, you would study the laws of optics from witch the equivalence thing is taken from. You could try the resources on the net, a book on optics or the tool that compute the deph of field of you get for a given camera, focal lens, focussing distance and apperture and play with it. It is not like this is difficult to find.

I really means no offense but I will not spend my evening trying to convince you while the problem is you are not really interrested to change your mind but more into keeping your current point of view.

You could say it apply to me, perfectly true. As it would apply to the guys that made the formula that help design the lenses we use or for the online tools that compute the deph of field automatically for you. Your choice.

[normhead]

If you are serious about that, you would study the laws of optics from witch the equivalence we have is taken from.

I already studied optical physics etc. in a compulsory lens design course at Ryerson Polytechnical, first year photography, and am still uncertain... that's why i'd need to see a proof. And no I don't think it's worth doing.. but I wouldn't claim it's true unless I knew it to be true. When uncertain, don't muddy the waters.

If anyone wants to chip in... "is it possible to create the same size subject, depth of field and circles of confusion when shooting with different lenses on different formats" would be the discussion point. The only relevant images I've seen would suggest, not necessarily.

[Nicolas06]

The sharpness score tells you about sharpness. Since when is sharpness the determinant of a great image? Is it even in the top five? Sharpness is a useful technical aspect of a lens.

Furthermore, the sharpness score is a single number intended to generally describe the sharpness of a zoom at all apertures, at every focal length and across the frame. No single number can tell you everything about the lens, and DXOMark says so, quite plainly as I quoted above.

This at least give some indication that for example if you want to print large you'd want at least some sharpness level. And if you want to crop the image, you want some margin too. What K3 has given to me with great lenses is more margin for reframing. it is not essential all the time but it is usefull.

Thoses shoots of 18-200, if you crop them a bit too much, it will show. And the image of normhead with the 700 likes to me look lke uninterresting and not sharp even on the very small size he has given. I think this is more due to facebook terrible Jpeg compression than anything else. And I find it uninterresting because I'am not found of bird picture to begin with, the borders put on the photos and the colors rendering. But it is just me, art is not always universal.

But what the likes really means? That most would really prefer this photo on the living room than another one that is less liked? That they really feel it is better?

For me, it is because the subject is funny, a rare thing to have this kind of bird to show at all and the composition is quite good. The normhead photo has value because it is an original one. Normal people don't look at the artistic value or the technical merits. They would not recognize a real monet with than something similar made by anothe artist with similar technique.

Does it means that people would not prefer the same photo with better light, if facebook didn't over compress the thing, if the composition was even better? Of course. But this alternate version is not present to compare. And between 2 photographers if the one with the better photos (even just what he took, not at all the technical merits) has less contacts on facebook or don't make effort to promote it, it will not be seen. But even a very average photos with good promotion with get thousand, hundred thousand like or maybe even make it to flickr explore because of the power of the community the photographer managed to make arround it.

As to the camera and lense, it is a part of what help the photo shine. Depending of the subject, the gear and so own it can be more or less important, but it has it's usage. A lensbaby is not sharp and is quite creative and interresting in some case, totally counter productive in others. The good photographer doesn't conclude no lense ever sharper than lens baby is necessary but instead think of when to use it and when not to use it !

And we need some logic too. When you rate a lense, you rate a lense. You show what it is. That the goal. The numbers of DxO are neutral. Many for example think the number of Pentax lenses are bad but when you read DxO reviews of for example DA20-40 or many ltd, DxO is quite positive about it in their explanations. They don't say the lenses are bad at all. When one choose the gear and so on, he should also know something about it.

Despite all that is said you'll see the difference between a DA*55 and a 18-200 in pratice at least for deph of field. And when taking some macro shoot, or when shooting wildlife, more pixels can also be quite handy.

That the same when you buy a car, you should choose the motor power depending of your intended use. If you plan to just go to work with lot of traffic jam you don't need that much. If like a friend you need to carry a horse from time to time, you need a powerfull car with the capacity to handle the heavy load. If you want to feel the car and drive fast in small roads, a sport car might be better. Does it means that everybody need the powerfull motor needed by the sport car or the SUV to carry the horse? Well hell not. Is it bad or good? It is simply neutral. It is your car, your life, your money.

Like for the photos, it is your camera, your lenses, your photos, your money. Choose wisely for the intended use.

[VoiceOfReason]

The best test of a lens by far is this. Do the results please you? Do you get a good image? Then you have a good lens and camera.