[ElJamoquio]

...the mere fact that some of the Nikons are tested with the Sigma 70mm Macro and some are tested with the Nikon 60mm Macro is troubling. Presumably for autofocus? Why would they need a lens that autofocuses?

[rawr]

I appreciate the IR numbers, but when a APS-C (D3200) and FF lens (D3X) have the same resolution capability, there's been a bad test. MAYBE if there's no AA filter on the APS-C... but the D3x scores only 80% of the D600? No way.

Well, perhaps an even more disproportionate number is the comparison between the 24MP RX1 and the 36MP D800... why should their resolution numbers be identical?

[badsykes]

The stock reply to that is "The difference in price between a 645 and a D800 is such that it's not economical." That is said by everyone who doesn't own a 645D. The fact that FF users are arguing absolutes in order to justify the D800 (better IQ whether they'd actually use it or not) but then change their argument to argue price and refuse to argue absolutes in terms of the 645D, and the contradiction that implies, always seems to escape them. My question is "do you need maximum resolution or don't you?" Once you start talking price and value then the K-5 comes back into the equation. It's a discussion that goes back and forth between absolute IQ and value, that in the end comes down to, if you can afford a D800 and the high end lenses that go with it, go for it. If you can afford the 645 DSLR go for it. If you can't afford either don't sweat it, you'll probably never notice you don't own them, except in "my sensor is bigger than your sensor " bragging contests.

"Do you need maximum resolution?"
NO

"Do i need to get more from the same sensor and get maximized output?"
YES ... Here comes the K5 iis vs K5 ...Imo i am not very pretentious really ... I want same 16mp sensor without the glass in front.. Is that simple...Same sheep in different coat and optimized...


PS: the 645D is only 7000$ + smc DA 645 25mm f/4 AL (IF) SDM AW Lens for just 5000$ ....
If the 645D will ever costs 3000$ it will it will kill most FF's in image quality ... Of course that camera is not for everyone anyway..Even at that price the design may keep people away..

[normhead]

Well, perhaps an even more disproportionate number is the comparison between the 24MP RX1 and the 36MP D800... why should their resolution numbers be identical?

There are test results for the D800e showing it exceeds 4000 lw/ph (the maximum of their test chart). they admit they had to test 6 or 7 times to achieve those results. So in part what you're paying for is the possibility of achieving an incredible resolution. Not the probability that you will, but the possibility that you will. If it takes them 7 attempts to achieve the max resolution in the lab, what are the odds you're going to achieve it in the field? The thing is, maybe the stars will align and one of your images will hit an exposure (and focus) bang on. Then you have an image most cameras are physically incapable of even approaching. Will every image be that way? Of course not, but it gives you a chance.

I can guarantee you, the RX1 will not give you 4000 lw/ph no matter what you do to it. It's physically impossible to achieve more than 4000 lw/ph on a sensor that is only 4000 line width high. So there's no worries about what can be done in the top end, whether or not you can actually do it is another question. Couple this with the fact that a DP2 Merrill produces images that are in every way comparable to a D800 image, and you quickly figure out that what you have to do to achieve maximum resolution is just as important as the actual resolution numbers.

( The highest test I've seen for a k-5 was 2900 lw/ph with I believe a Sigma 70 macro... so , individual numbers for tests can really vary.. to the point i tend to look at them as ball park figures, worth considering, but not taking them as gospel fact.)

---

Last edited by normhead; 05-21-2013 at 10:48 AM.

[Big Dave]

I always wanted a full frame DSLR to make my collection of legacy prime lenses work as designed and got tired of waiting for Pentax to deliver. The answer was to buy a used Canon 5D mark 1. It has fantastic IQ and will take most of my primes without lens modification. It even gives me focus confirmation with a chip equipped adapter. It was very cheap at $600 and came like new condition with a Canon grip and all accessories in the box. My Pentax M42 lenses fit, but the K-mounts lever is too long as 7mm body intrusion is all the it can handle. I didn't realize what an improvement FF made in low light. My 17mm f4 fish eye is awesome on full frame. On the down side it lacks live view,video, higher ISOs and a hi-res 3 inch LCD. My pentax DSLR has all of that stuff anyway. The Pentax is the go to camera for telephoto,micro and general family use. The 5D is for high ISO ,landscapes and architecture. I only own one Canon EF lens, a EF-28-135 IS USM which I got direct from Canon at a fantastic price. So I don’t think that I’ll buy a Pentax full frame, even if I could afford it.

[carlwagee]

I am not sure that a full frame would have a reason for existence by the time that Pentax would have one out. Rapid advancement of technology might invalidate a full frame. As well as software. And, no, I would not buy one. Not unless I had money laying around that had no purpose. Say, lottery.

[jsherman999]

There are test results for the D800e showing it exceeds 4000 lw/ph (the maximum of their test chart). they admit they had to test 6 or 7 times to achieve those results. So in part what you're paying for is the possibility of achieving an incredible resolution. Not the probability that you will, but the possibility that you will. If it takes them 7 attempts to achieve the max resolution in the lab, what are the odds you're going to achieve it in the field? The thing is, maybe the stars will align and one of your images will hit an exposure (and focus) bang on. Then you have an image most cameras are physically incapable of even approaching. Will every image be that way? Of course not, but it gives you a chance.

You don't need to achieve the utmost resolution every time in order to see the resolution benefits. Using your same technique on say 16MP aps-c and 36MP FF, the 36MP is going to show more resolution, more detail, unless your shooting conditions are so compromised that you've lost all detail anyway (really low shutter speeds and lots of camera shake or a really bad lens on the 36MP)

Keep in mind that to get the utmost resolution from that K-5, you need to have the stars align too. Same technique, both cameras, and the 36MP shows you more.

(In fact because the aps-c image is magnified more, you need better technique and glass on the aps-c side.)

.

[stevebrot]

(In fact because the aps-c image is magnified more, you need better technique and glass on the aps-c side.)

The sad truth...

[normhead]

You don't need to achieve the utmost resolution every time in order to see the resolution benefits. Using your same technique on say 16MP aps-c and 36MP FF, the 36MP is going to show more resolution, more detail, unless your shooting conditions are so compromised that you've lost all detail anyway (really low shutter speeds and lots of camera shake or a really bad lens on the 36MP)

Keep in mind that to get the utmost resolution from that K-5, you need to have the stars align too. Same technique, both cameras, and the 36MP shows you more.

(In fact because the aps-c image is magnified more, you need better technique and glass on the aps-c side.)

.

The sad truth...

No arguments there gentlemen, with the very close to the same pixel size , you'd need to use exactly the same technique on a K-5 to max out it's resolution as you would on a D800. This brings up another burning question.. does the larger pixel size of say D700 make it less prone to camera shake. It would seem just thinking about it that with a tiny bit of shake a larger sensor would have less chance of moving that stream of photons off the pixel, but I have no idea if this would be true or not in a practical sense.

That would be in a way confirmed by the fact that the same guys who squeezed 4000 lw/ph out of a D800 got the same resolution out of a 645D without going through the same kind of effort. I makes you wonder what they'd have done with a 645D if they used the same micro adjustments they did on the D800.

Keep in mind that to get the utmost resolution from that K-5, you need to have the stars align too. Same technique, both cameras, and the 36MP shows you more.

On one set of tests with the K-5 rated at 2100, the K-5II rated at 2200 and the K-5 IIs rated at 21300 lw/ph, the D600 was rated at 2600 and the D800 was only rated at 2700 lw/ph. So yes you get more with the D800 you get more but not what you might think. If you set 100 line per inch as your desired level for a print you can print 23 inches with a K-5II and 27 inches with a D800. Honestly, if you look at the images sizes, the magnification is not really much of a factor. And the improvement in print size isn't either. Not only that when you reduce your images size you lose lw/ph.. you obviously can't show 2000 lw/ph on an image that is displayed at 1500 pixels deep. (Alternating black and white lines artificially created would give you 1500 distinct line and therefore 1500 lw/ph as a theoretical max, and everything I've seen is that the best a camera can represent is about 80% of that theoretical max).

So as far as I can tell, in practical terms, you're looking at only about a 25% increase in magnification going from APS-c to a D800 based on available sensors and current testing, and what they resolve, and reduced to one dimension sizes Of course if that 25% is of use to you, then it's a bargain being able to get a D800. But lets not get carried away with concepts, we haven't tested. Looking at all the numbers the first thing that stands out is there is simply no way to explain them using simple concepts and simple math. The numbers simply don't turn out to mean in the tests what some imply they do.

Bottom line, doubling your file size and sensor size gets you a file that by linear dimensions is only 25% bigger. You have to double your magnification to get that, but in terms of lw/ph, you don't get anywhere near double the linear size. It's one of those facts where it's so simple, no one knows what it means in the real world.

Doubling your sensor size gets you 2700 lw/ph compared to 2300 lw/ph using the same test with no teasing. Thats one tester's reality. But then, who knows what the truth is. Simple fact is, if you read through different tests from different sites, they all can't be right. None the less certain trends emerge.

The beauty of lw/ph is it doesn't care what size your sensor is. It's a sensor independent rating system. It leaves more artificial theories based on magnification etc, which are essentially theories based on two cameras from times gone by when the same film could be used in two different cameras. The obvious flaw in lw/ph is, the test results are so different from different labs. if it were a reliable standard, that wouldn't be true.

The more you get into this the more you realize statements like like

In fact because the aps-c image is magnified more, you need better technique and glass on the aps-c side.

based on my own experience I do think you need better glass. Perhaps not so much better technique.



Taken without qualification as a simple statement of fact, doesn't tell the whole story. Then you look at a print from 16 mp Sigma DP2 Merrill being almost identical to a print made at the same time from by a D800, and you realize, there's a lot moe going on than simple math.

ANd if the largest you ever view your prints is on a screen 1500 pixels deep, you are actually throwing away resolution from a D800 to get to that size. At that point the issue is not who started with more resolution, but with who is throwing away more resolution. Saying you have to magnify your APS-c prints more than your FF prints ignores current practice of using files at a reduced size.

Months ago I suggested that at for an image 3000 pixels across, A D800 image and a K-5 image would be virtually the same ( discarding differences in DoF etc.) and so far, I've yet to see any evidence to the contrary. For the most part our images are not magnified at all they are reduced, and at a certain point (like 3000 pixels in the example above) the original resolution becomes meaningless.

[ElJamoquio]

So as far as I can tell, in practical terms, you're looking at only about a 25% increase in magnification going from APS-c to a D800 based on available sensors and current testing, and what they resolve, and reduced to one dimension sizes

Personally I wouldn't trust any test that requires a lens to have AF.

Take a look at photozone.

[normhead]

Personally I wouldn't trust any test that requires a lens to have AF.

Take a look at photozone.

Personally, I'm not sure I "trust"any of them. When I go to IR and I see the test prints of the same thing and I can see detail with the 645D and D800 I can't see in any other camera's images, that means something to me.... based on sample images, you're going to be hard pressed to convince me any other camera is worth purchasing, although it might be different if I did a lot of action shots, shot in low light at really high ISOs etc.

I'm happy to look at what everyone does, but you also realize, no one site or tester is the whole picture. It's about looking at all the information you have and seeing trends. The evidence is describing something, you have to go out and look at some pictures to understand what?

I look at jsherman's shots, and I think, if I took that type of picture, I'd want that camera. I think that's how it works for most of us. We see something none of our current cameras do, and we think "I want the camera that does that." it's really not an intellectual discussion.

[Fogel70]

No arguments there gentlemen, with the very close to the same pixel size , you'd need to use exactly the same technique on a K-5 to max out it's resolution as you would on a D800. This brings up another burning question.. does the larger pixel size of say D700 make it less prone to camera shake. It would seem just thinking about it that with a tiny bit of shake a larger sensor would have less chance of moving that stream of photons off the pixel, but I have no idea if this would be true or not in a practical sense..

As long as you use the same FOV on all sensors, the sensor size do not affect camera shake. And you will need the same amount of pixels on all cameras to see the same amount of camera shake. More pixels gives higher resolution in the images, so less camera shake is needed to blur pixels.

It's only if using the same lens on D800 and K5 they will show the same amount of pixel blur.

Equivalence is a tool that can be used in this case too.

[normhead]

As long as you use the same FOV on all sensors, the sensor size do not affect camera shake. And you will need the same amount of pixels on all cameras to see the same amount of camera shake. More pixels gives higher resolution in the images, so less camera shake is needed to blur pixels.

It's only if using the same lens on D800 and K5 they will show the same amount of pixel blur.

Equivalence is a tool that can be used in this case too.

Thanks for posting that, another fuzzy area of my fuzzy thinking cleared up.

[rawr]

Personally I wouldn't trust any test that requires a lens to have AF.

Don't diss IR. I guess you have never really looked at how IR do their tests. They focus bracket manually until they get the sharpest result from their test bench.

There is a paper from them somewhere on the IR site along the lines of 'the fallacies of AF' where they outline their experience shooting thousands of test photos over the years, and the problems with AF they have encountered.

EDIT: found the link to their methodology:
http://www.slrgear.com/articles/focus/focus.htm

It's a 2008 article by SLRGear, who are part of IR, so IR certainly are aware of the issues with AF and image testing.

[stevebrot]

Personally I wouldn't trust any test that requires a lens to have AF.

A couple of days ago forum member Nesster posted this article from a 1970 issue of Modern Photography...

How to test your lenses 1970 | Flickr - Photo Sharing!

Click the "..." in the corner to view larger size. The part that is particularly interesting is that their test protocol involved removing the shutter from the equation. That's right...open using "B" in the dark. Provide timed illumination or a modulated flash burst. Close the shutter when the lights go off.


Steve