[Digitalis]

This is not intended to be a formal review, but more of a practical examination on image quality between the Pentax K10D, K7 and the current Pentax K5 IIs.


First thing that struck me with the K5 IIs is that the images seemed to have muted colour saturation and lower contrast than my K10D - I can chalk this up to the fact that the K10D has a CCD sensor - which are known for producing excellent colour rendering at low ISO, though the K5 IIs has a dynamic range that is considerably greater than the K10D which is perhaps why the images seemed to have lower contrast. There are differences in colours between the K7 and K5IIs, using the X-rite color checker target under flash illumination I have observed the K7 is much more sensitive to reds than the K5 is.

Since it is nearly 40 degrees outside at the moment, I thought some NIR image tests were in order, and I was surprised at what I found. I did some tests with the DA15mm f/4 ED ASPH Limited and a Hoya R72 IR filter. The DA15mm f/4 would always have a hotspot on the K7 in the centre of the frame, it also suffered from the most appalling astigmatism I have ever seen from a modern lens in the extreme corners of the image*. What is interesting is that both of these Issues completely disappeared when I used the same lens and IR filter on the K5 IIs. Despite the differences in red sensitivity I find that the low noise and greater detail due to the absence of the Bayer AA filter that the K5 IIs is an excellent choice for Infrared photography.


Unsharpened crops from the extreme corner from the DA15mm f/4 ED ASPH Limited with Hoya R-72 IR filter - f/11 15s @ ISO 400


I'll keep adding to this, as I continue testing the new K5 IIs. Incidentally of all my pentax lenses focus accurately on the K5 IIs - no adjustments needed. Unfortunately the same can not be said for my sigma lenses, so I am going to be busy calibrating them for the camera.

PS. the DA15mm f/4 again shows considerable improvement in image quality in the extreme corners on the K5 IIs


Unsharpened crop from the extreme corners of the DA15mm f/4 ED ASPH Limited - 1/30th f/11 @ ISO 100

*I have reason to suspect the Bayer AA filter is responsible for the degradation of image quality at the corners when using wide angle lenses for visible and IR imaging. MgF2 and CaF2 - which are both used for their Birefringent properties and are good candidates for being used as a Bayer AA filter are transparent across a great spectrum - however the refraction and dispersion properties change depending on the frequencies of light passing through them. I suspect the high angle of incidence from a wide angle lens and the wavelengths involved is causing greater dispersion as the angle of light that passes through the Bayer AA filter increases - but that is just my theory on what is going on. It is difficult to find accurate dispersion and refraction information on these materials that covers IR wavelengths.

---

Last edited by Digitalis; 01-03-2013 at 12:09 AM.

[JinDesu]

Some fantastic thoughts on the camera. The improvement looks fantastic, but then again the normal k-5 sensor is a big improvement over the k-7 already. Love the report on the AF though.

[lightbulb]

Let the dragon do the talking; All sizes | C01 | Flickr - Photo Sharing!

[xhanatos]

not enough moire

thanks for the the comparison shots.

[Digitalis]

Moire is like a rake hidden in tall grass - just waiting for a careless step. I have worked with many cameras that do not have Bayer AA filters in them and I consider the tradeoff to be worth it. IMO moire it isn't as bad a problem as people make it out to be - I have shots that are affected by moire from cameras that have Bayer AA filters in them, there is only so much camera makers can do to prevent it. I just prefer to deal with it my own way.

[fanofcc]

HOW if i try to remove the AA filter on my K5? @_@

[Digitalis]

later this afternoon I was doing some Studio work with my Pentax FA*200mm f/4 ED Macro - I wrapped up early, so I took some macro shots of my watch to compare the output from the three APS-C pentax cameras I keep in my studio.I am really impressed how despite a handicap in resolution and dynamic range the K10D has, how close it comes in terms of resolution to the K5IIs. The Pentax K7 is the obvious loser here, despite the legendary resolution of the FA*200mm ED the Bayer AA filter really gets in the way.


All images shot at f/11 1/180th @ ISO 100 - all images unsharpened 100% crops

[bkpix]

I'm an owner of -- and a big fan of -- the K-5IIs. But to make a useful comparison, you really need to sharpen the images from cameras like the K-7 that have a built-in AA filter.

In these particular images I don't see detail captured in the K-5IIs image that's not visible in the other two; the only apparent difference is sharpness, which can be fixed.

That said, in my own printing, I've found sufficient extra detail in 20x30 prints from the IIs compared to the old K-5 that I'm happy I bought the new camera. (For that, and for its ability to autofocus in next to zero light.)

[Digitalis]

But to make a useful comparison, you really need to sharpen the images from cameras like the K-7 that have a built-in AA filter.

No you don't. The point of the comparison is to show the cameras for what they are, if I sharpen the K7 output, to be fair I would have to sharpen the output from the K5IIs and K10D as well and the K5IIs will again take the lead in ultimate IQ.

[RBellavance]

No you don't. The point of the comparison is to show the cameras for what they are, if I sharpen the K7 output, to be fair I would have to sharpen the output from the K5IIs and K10D as well and the K5IIs will again take the lead in ultimate IQ.

I disagree somewhat. Showing out-of-camera jpegs is not very useful. What should be compared is the optimally-processed output of each camera. That is the only way to show what each truly is capable of.

[Digitalis]

Showing out-of-camera jpegs is not very useful. What should be compared is the optimally-processed output of each camera

I agree In-camera Jpegs are useless for comparisons, but that isn't what I am doing here. What I am showing is the RAW unedited output from each sensor, people can draw their own conclusions from there.

[bkpix]

Well, there's a bit of a philosophical issue here. There is no way to examine unedited RAW files except as a series of 0s and 1s. The "straight out of the camera" file is just a Jpeg that's been converted using some default setting. So I agree with RBellavance that the most useful comparison is between two jpegs that have been processed as well as possible.

[JinDesu]

What's wrong with comparing unsharpened RAW outputs that were submitted through the same processing?

If I, as a user, decide to choose the camera that gives me the sharpest image without post process sharpening, wouldn't seeing unsharpened RAW outputs be useful to me?

When doing the majority of my photo editing, I apply very light sharpening to all my files. I don't really spend the time and effort to go into Lightroom or Photoshop's sharpening tool to maximize sharpness without artifacts. So seeing the base sharpness from unsharpened RAWs should tell me more than enough.

[bkpix]

Well, as I said, it is a slightly philosophical question. There is a hidden assumption in comparing "unprocessed" images that sharpening isn't natural somehow. But shooting a camera with an AA filter assumes you're going to do some sharpening to put the image back together, so it seems odd to me to compare one without sharpening at all.

Practically speaking, though, I'd rather compare the best possible image available with good postprocessing from each camera.

[Rondec]

I would say that at low iso, with the images presented, you would be hard pressed to see a difference. The big differences to me with the K5 sensor (I don't have the K5 II or IIs) are increased dynamic range at low iso (not evident here), better high iso performance (also not evident here) and significantly improved handling over previous generations of cameras.