K-x Vs. K-r Vs. K5

[airconvent]

Seem like the Kx is slightly .... ( jus a bit) better . From What I see from the pic lah

That's what the TS was trying to say. If you are upgrading only to improve on IQ, then moving to K-r does not make sense. In anycase, I can't see much difference except for the first set of images. Now I feel a little but better with my K-x (bought 2 weeks before kr lanuched). But of course if there is opportunity, I will still go for the kr..just to be current

 

[fengwei]

不能见风，但可以听到风吹...听证会还相信。总之，眼睛也可以很容易受骗...如果不知道去哪里找...或怎么看!

Here is some comparison tests shot in RAW (courtesy Devorama) and compiled into a nicely organized ISO ladder in a single image for viewing convenience:

Are these from your testing results? Where are the K5 comparison samples?

I can see the focusing is slightly off w/ the Kr, to me the results aren't valid at all just because of this. Same as those K5 sample photos from DPR if they keep using an ancient FA50 which obviously has some optical problem.

 

[airconvent]

I can see the focusing is slightly off w/ the Kr, to me the results aren't valid at all just because of this.

Hi Fengwei
I noticed this too but technically, if the image is out of focus, wouldn't it appear to have lesser noise as it would also be out of focus?
Also, you mentioned about use of an FA50mm lens for DPR's tests. I thought that for an FF lens designed for film which technically have almost infinite "pixels", it should be able to handle the rigours of a high-megapixel camera. Does this have something to do with the optics instead?
Thanks

 

[fengwei]

Hi Fengwei
I noticed this too but technically, if the image is out of focus, wouldn't it appear to have lesser noise as it would also be out of focus?
Also, you mentioned about use of an FA50mm lens for DPR's tests. I thought that for an FF lens designed for film which technically have almost infinite "pixels", it should be able to handle the rigours of a high-megapixel camera. Does this have something to do with the optics instead?
Thanks

Noise has nothing to do w/ focus. But when you look at those photos, the noise level is pretty much similar. If there is some difference, it's very minimal to even differentiate these two camera. But you look at those photos, Kx ones seem sharper, thus look better to the viewers.

The FA50 sample DPR has been used for Pentax camera reviews in the past a few years, and everytime the photos came out from that lens look softer (comparing w/ other ones) w/ more purple fringes. I'd say they should use our DA35/2.8 macro lens for all the test, so at least their test samples would show the capability of our cameras instead of the weakness of that copy FA50.

 

[sircam]

I noticed this too but technically, if the image is out of focus, wouldn't it appear to have lesser noise as it would also be out of focus?

Yes. You are right. The fact that the K-r is slightly out of focus actually indicates that the K-r's noise should be a bit smoother too i.e. slightly out of focus too... but it isn't. This means another small plus for the K-x's fantastic sensor. Nothing I would rave about however, as the difference is quite slight. I don't like to nitpick on such issues, because, for all practical purposes, the sensors of both cameras are almost identical.

I thought that for an FF lens designed for film which technically have almost infinite "pixels", it should be able to handle the rigours of a high-megapixel camera. Does this have something to do with the optics instead?
Thanks

Yes. DA lenses are optically optimized for incoming light to fall more perpendicular to the sensor plane away from center than FA lenses are. This optical optimization is slight however... nothing ground-breaking.

Light sensitive film grains are several times smaller than even the smallest pixels on modern sensors. They also overlap each other at random angles unlike pixels on a sensor which are all neatly arranged in a linear array. Thus angular incoming light also is captured by film grains much more efficiently than CCD or CMOS sensors.

Sensor manufacturers use micro-lenses on modern sensors to improve their angular light gathering ability... but these still aren't as efficient as film. Tomorrow's sensor designs may incorporate curved sensors to further improve their angular light gathering ability.

Furthermore, there are some additional coating benefits the newer DA lenses have over their older FA cousins also which do provide a perceptible benefit to IQ.

 

[airconvent]

Noise has nothing to do w/ focus. But when you look at those photos, the noise level is pretty much similar. If there is some difference, it's very minimal to even differentiate these two camera. But you look at those photos, Kx ones seem sharper, thus look better to the viewers.

The FA50 sample DPR has been used for Pentax camera reviews in the past a few years, and everytime the photos came out from that lens look softer (comparing w/ other ones) w/ more purple fringes. I'd say they should use our DA35/2.8 macro lens for all the test, so at least their test samples would show the capability of our cameras instead of the weakness of that copy FA50.

Yes. You are right. The fact that the K-r is slightly out of focus actually indicates that the K-r's noise should be a bit smoother too i.e. slightly out of focus too... but it isn't. This means another small plus for the K-x's fantastic sensor. Nothing I would rave about however, as the difference is quite slight. I don't like to nitpick on such issues, because, for all practical purposes, the sensors of both cameras are almost identical.

Yes. DA lenses are optically optimized for incoming light to fall more perpendicular to the sensor plane away from center than FA lenses are. This optical optimization is slight however... nothing ground-breaking.

Light sensitive film grains are several times smaller than even the smallest pixels on modern sensors. They also overlap each other at random angles unlike pixels on a sensor which are all neatly arranged in a linear array. Thus angular incoming light also is captured by film grains much more efficiently than CCD or CMOS sensors.

Sensor manufacturers use micro-lenses on modern sensors to improve their angular light gathering ability... but these still aren't as efficient as film. Tomorrow's sensor designs may incorporate curved sensors to further improve their angular light gathering ability.

Furthermore, there are some additional coating benefits the newer DA lenses have over their older FA cousins also which do provide a perceptible benefit to IQ.

Hi Fengwei/Sircam
Thanks for the clarification.

 

[armadillo]

sircam, thanks for sharing your hardwork :thumbsup:

 

[Gengh]

Yes. You are right. The fact that the K-r is slightly out of focus actually indicates that the K-r's noise should be a bit smoother too i.e. slightly out of focus too... but it isn't. This means another small plus for the K-x's fantastic sensor. Nothing I would rave about however, as the difference is quite slight. I don't like to nitpick on such issues, because, for all practical purposes, the sensors of both cameras are almost identical.

I agree that the K-x's noise level seems to be a wee bit lower than the K-r's. But the focus has absolutely nothing to do with the noise. You can take a completely OOF picture, and the noise will look the same. Noise "grains" do not go OOF because they're something that happens at the sensor level. The focus will only affect the perceived sharpness of the pic. But then again, I also agree with you that if we're just talking about noise here, we shouldn't nitpick on a minor focus error.

Yes. DA lenses are optically optimized for incoming light to fall more perpendicular to the sensor plane away from center than FA lenses are. This optical optimization is slight however... nothing ground-breaking.

Light sensitive film grains are several times smaller than even the smallest pixels on modern sensors. They also overlap each other at random angles unlike pixels on a sensor which are all neatly arranged in a linear array. Thus angular incoming light also is captured by film grains much more efficiently than CCD or CMOS sensors.

This is not the usual reason given to explain digital sensors' problem with incoming light at an angle. Can elaborate on this, or provide a source? I'm interested to learn more.

Just wondering also, does anyone have any complaints about excessive vignetting from non-DA lenses? I know I haven't from the few older lenses I have. This issue about digital sensors' lower sensitivity to light at an angle may be exaggerated to help the manufacturers sell new lenses....

Sensor manufacturers use micro-lenses on modern sensors to improve their angular light gathering ability... but these still aren't as efficient as film. Tomorrow's sensor designs may incorporate curved sensors to further improve their angular light gathering ability.

Curved sensors will require a whole new system (AF module, lenses for optimal imaging on curved surfaces, implementation of viewfinder, etc.) to be designed and built from ground up. I think it's a great idea too (the lenses are waaaaay easier to design), but I won't be crossing my fingers that such a system will be available for the mass market anytime soon.