[Mistral75]

(...)
Do you understand the SIMPLE FACT - ANY LENS of ANY brand will have HIGHER resolution at the camera with higher MP.
It's pure mathematics tests of lenses. Klaus uses IMATEST.

The resolution of system camera+lens+soft are always higher with higher MP.
(...)

Not necessarily. The whole system: sensor + AA filter (+ IR filter + protective glass but they have a lesser impact) will play a role and is to be taken into account.

The same phenomenon appears in the tests of the same lenses made by Photozone on NEX-5 (14 Mpixels) then NEX-7 (24 Mpixels): some lenses take full advantage of the increase in pixel count in the centre but not so much in the borders and corners. Sometimes, the absolute figures are even lower with NEX-7 than with NEX-5. See for instance the tests of SEL 18-55mm f/3.5-5.6 and SEL 16mm f/2.8 or have a look at the comparison graphs I posted at Chassimages:

Essai des objectifs en monture Sony E (NEX) sur Photozone and the next two posts. It's in French but the graphs are easy to understand ("bords" = borders and "coins" = corners).

---

Last edited by Mistral75; 05-20-2012 at 11:57 AM.

[Mistral75]

Anyway, the test of the DFA 43 mm f/1.9 limited made by Photozone shows two things:

- the lens takes full advantage of the high pixel count of K-5 in the centre but not in the borders and corners
- the resolution in the corners of an APS-C sensor is weak (and gets only slowly better when stopping down) and should be even worse in the borders and corners of a 24x36 sensor of the same pixel density.

It's not insulting this superb lens to recall that it appeared on the market in 1997, fifteen years ago, and has been optimized (compactness, resolution, bokeh and so on) for film cameras, the resolving power of which is said to be close to the one of 8 Mpixels Bayer digital sensors, and not for the 36 Mpixels sensors we see today.

If it were designed today, it might enjoy a more homogeneous resolution to the detriment of size, weight, price and, may be, bokeh quality.

Nobody's perfect.

[thibs]

Still; a lens not getting better resolution with higher res sensor is OK. Lens is limiting factor. That's it.
Lens getting worse res with higher res sensor though is... well let's say something's fishy.

As for D800 users, if they pixel peep or expected better res, than I don't see anything strange. Disappointing maybe but not inconceivable.

[SuperK5]

The thing concerning me is that the K-30 will have 16 MP- (not that that is too little, I'd be fine with a 6MP ist) but other cameras in its class are coming out with 24 and are in the same price range.

Try telling someone who is new to photography and looking at an entry level DSLR that MP doesn't matter... of course most are going to want to go to the more-known camera brand that has more MP. I know it took me awhile to learn that it is the photographer- not the camera that matters.

Plus they don't have Aston Kutcher saying "Come on ladies touch my Megapixels"
Digital Cameras, Ashton COOLPIX, DSLR Cameras | NikonUSA.com

[clover]

Maybe a confusing point is the density

For a APS-C DSLR,
16MP means approximatively 43000 pixels per millimeter
24MP means approximatively 64000 pixels per millimeter

For a Full frame DSLR,
24MP means approximatively 28000 pixels per millimeter
36MP means approximatively 41000 pixels per millimeter

Which means 36MP Full frame sensor has a lower density than 16MP APS-C Sensor.

EDIT :
There is a chart here.

[sjwaldron]

Plus they don't have Aston Kutcher saying "Come on ladies touch my Megapixels"
Digital Cameras, Ashton COOLPIX, DSLR Cameras | NikonUSA.com

Pentax USA (or elsewhere) should have a contest in cooperation with Pentax Forums to select one of us as the face of Pentax based on Pentax in action. I think that's a lot better idea than Nikon had with Ashton.

[GDRoth]

Pentax USA (or elsewhere) should have a contest in cooperation with Pentax Forums to select one of us as the face of Pentax based on Pentax in action. I think that's a lot better idea than Nikon had with Ashton.

We may not be the most photogenic bunch.........

[sjwaldron]

We may not be the most photogenic bunch.........

The right photographer can make anyone look nice. Especially if they have a Pentax in their hands.

Maybe they could have "from every walk of life" Pentaxian group.

[bossa]

Jay.

Apologies.

[Blue]

. . .

If Pentax makes a FF, remaking 28-70mm f2.8 and 80-200mm f2.8 cannot be a qualitative answer.
Just look at the 35mm DA, there is something called ghostless treatement... Is it so unusefull ?

The FA 43 LTD was the first Pentax slr lens to receive Ghostless Coating. All the FA LTD and DA LTD series have it. The DA 40 ltd is only ~ 4 years newer in design than the FA 77 ltd and 3 years newer than the FA 31 LTD. Certain other lenses have it as well. Nikon and Canon have some ff lenses still in production that date back to the early 90s just like the FA 50/1.4. Just because a lens was developed during the film era doesn't automatically mean that it is not useful on digital. The laws of physics didn't magically change with the advent of digital sensors. I am not going to write any lens off until it is evaluated on a full frame lens.

Hirikawa Jun designed the 43 and 77 to have a unique look and it could very well be these aspects that is baffling the guys at Photozone. Its not like they do scientific testing there anyway. To do so they would need 3 or 4 lenses from 3 or 4 production runs. As it stands, they doe 1 replication and psuedo-replication. On top of that, they have never been all that warm to Pentax glass. When the retested the DA 70, they still didn't torture test it to the degree they did the FA 77 in the original test. To their credit, the did state that they forced it but that often gets overlooked. I use there site as just one of many things to make initial assessments. The FA 31 ltd did very well in the retests.

[Class A]

Lens getting worse res with higher res sensor though is... well let's say something's fishy.

Exactly.

One reason could be that measurement results are not really comparable between K10D and K-5, just as they are not between cameras of different brands. PZ measures a system response, not just the lens characteristics. PZ makes it very clear that lens performances cannot be compared when different brands of cameras were used to measure them. To be clear such comparisons are not possible not only because sensors with higher pixel pitch can produce higher resolution results. I believe it also has to do with how the calibration regarding how to translate measurements into resolution figures is performed.

This still wouldn't explain shifts of performance peaks (31/1.8) and it is unlikely that it explains changes in performance characteristics (43/1.9) but it may be part of the picture as to why some performance appears to be worse than it was with the K10D.

Klaus speculates himself that the arrangement of his test patterns may have created some measurement artefacts in combination with the K10D's asymmetrical AA filter.

[bossa]

...why so much nikon users of D800 complain about their 24-70mm f/:2.8 AFS...?

Because there's now an extra level of resolution offered by the sensor and whilst the center of the lens is able to keep up, the corners can't. If you down sampled an image from the D800 to 12mp or 21mp the results would be similar to those of the previous cameras. The sensor allows us to see the boundaries of the lens but does NOT make it a worse lens.

There's no way that lens is going to be worse on that 36mp sensor, just worse at the edges relative to the center and at full resolution only. Of course if diffraction is kicking in at f/8 and you like to use f/11 or f/16 for DOF then the D800 is going to show you that more than a 12mp camera at full resolution. This is not the lens being worse on that sensor but the camera showing you things that were always there - diffraction was not so obvious on lesser resolution sensors until present in larger amounts at smaller apertures.

[clover]

The FA 43 LTD was the first Pentax slr lens to receive Ghostless Coating. All the FA LTD and DA LTD series have it. The DA 40 ltd is only ~ 4 years newer in design than the FA 77 ltd and 3 years newer than the FA 31 LTD. Certain other lenses have it as well. Nikon and Canon have some ff lenses still in production that date back to the early 90s just like the FA 50/1.4. Just because a lens was developed during the film era doesn't automatically mean that it is not useful on digital. The laws of physics didn't magically change with the advent of digital sensors. I am not going to write any lens off until it is evaluated on a full frame lens.

Hirikawa Jun designed the 43 and 77 to have a unique look and it could very well be these aspects that is baffling the guys at Photozone. Its not like they do scientific testing there anyway. To do so they would need 3 or 4 lenses from 3 or 4 production runs. As it stands, they doe 1 replication and psuedo-replication. On top of that, they have never been all that warm to Pentax glass. When the retested the DA 70, they still didn't torture test it to the degree they did the FA 77 in the original test. To their credit, the did state that they forced it but that often gets overlooked. I use there site as just one of many things to make initial assessments. The FA 31 ltd did very well in the retests.

The Aero Bright Coating is only introduced with the 55m DA* f/:1.4, isn't it ?

I do not mean that Photozone is god, I only suggest that it is important for Pentax to update some lenses. Even they are very prestigious at the film era. You can give a A if your test chosse to valorate some aspects more than other aspects. That is why it is important what one can want or not.

If you can get a little less attached to your Pentax Stuff, you can hear what I try to explain.

Do not misunderstand me. I have a lot of FA lenses, I hope to buy the FF Pentax DSLR, so I can use them as before.

But I am also ready to buy news lenses if needed, if the quality drops so much that I cannt bear it.

About the test methods, average buyer do not try 3 or 4 lenses before buy it. That is why you can only put a variable "mesure" but I do not think you will try a dozen of 18-55mm until you get the best one.
It is just like someone who have just bought a lens, how can he rely on the quality of his lens compare to other lens...?

[clover]

Because there's now an extra level of resolution offered by the sensor and whilst the center of the lens is able to keep up, the corners can't. If you down sampled an image from the D800 to 12mp or 21mp the results would be similar to those of the previous cameras. The sensor allows us to see the boundaries of the lens but does NOT make it a worse lens.

There's no way that lens is going to be worse on that 36mp sensor, just worse at the edges relative to the center and at full resolution only. Of course if diffraction is kicking in at f/8 and you like to use f/11 or f/16 for DOF then the D800 is going to show you that more than a 12mp camera at full resolution. This is not the lens being worse on that sensor but the camera showing you things that were always there - diffraction was not so obvious on lesser resolution sensors until present in larger amounts at smaller apertures.

Again, it depends on what you ask to your lens.

43mm is more homogen tahn 50mm FA on film, from center to border.

You cannot say that losing this kind of quality is nothing. It is important or not in your use.

[Blue]

The Aero Bright Coating is only introduced with the 55m DA* f/:1.4, isn't it ?

Actually, you didn't mention Aero Bright, you stated Ghostless Coating which are 2 different things. However, the DA* 60-250/4 has it. So do at least one of the 645D lenses. However, it is unknown exactly what Aero Bright is since Pentax has been so close to the vest about it.

I do not mean that Photozone is god, I only suggest that it is important for Pentax to update some lenses. Even they are very prestigious at the film era. You can give a A if your test chosse to valorate some aspects more than other aspects. That is why it is important what one can want or not.

I know. I was just pointing that out because it is easy to forget that that is a 1 sample test. It isn't a statistically valid test. It would require 3 or 4 lenses from 3 or 4 production runs as well as more than one body.

If you can get a little less attached to your Pentax Stuff, you can hear what I try to explain.

I am not specifically attached. I am more than capable of using my FA and other top glass and use it on film and digital as needed.

Do not misunderstand me. I have a lot of FA lenses, I hope to buy the FF Pentax DSLR, so I can use them as before.

But I am also ready to buy news lenses if needed, if the quality drops so much that I cannt bear it.

About the test methods, average buyer do not try 3 or 4 lenses before buy it. That is why you can only put a variable "mesure" but I do not think you will try a dozen of 18-55mm until you get the best one.
It is just like someone who have just bought a lens, how can he rely on the quality of his lens compare to other lens...?

I am not talking about the average buyer. Photozone and Lenstip aren't the average buyer. I am talking about an entity that implies scientific testing of the lenses. However, there are people that do try lenses until they get a suitable copy. In fact, they are often the people that are concerned about this sort of measurbation.

However, I am not all that concerned about their test on the K-5. I will cross the full frame bridge if I ever come to it.