[superpowerpinger]

Pentax K-1 Mark II Camera Review

[biz-engineer]

From the article: "Bottom Lin: If you’re the owner of loads of Pentax glass and desperately need better performance from Pixel Shift Resolution, the K-1 Mark II makes sense as an upgrade. Otherwise, it’s a fairly modest improvement over its already capable predecessor."

For me, the bottom line is a different one, and I believe a lot of customers are facing the same dilemma:
- poor choice of native mount lenses for new Canon and Nikon mirrorless model
- superior performance of Canon and Nikon new mirrorless models, not proven enough to justify the spending, I would get more for the money if I buy a medium format system instead of spending the money to not have a mirror
- only the DFA150450 and DA560 long lenses from Pentax, but if I even want good long lenses from C & N, they'll be in EF and F mount.
- given the investment in mirrorless, EF and F mounts are compromised, why should I spend money into a N or C DSLR now (*) ?
- I rarely make use of HD video, whenever I record I don't use HD because I don't like enormous files sizes, so, I have no interest in 4K whatsoever.
- the Pentax K1 system isn't yet outdated, they are many camera models that deliver image quality inferior to what the K1 and K1 II deliver, and any camera system delivering equal or superior IQ is significantly more expensive than a K1/K1II.

So, at the moment, if I ever spend any money on camera gear, it'll be on maintaining my K mount system that I'm already invested in.
In two to three years from now, the selection of mirrorless lenses will be clarified(**), there will be other choices, but we aren't there yet.

(*) Buying a Nikon D750 or Canon 6DII or even 5DIV is a mistake, we all know that EF and F mount will die whenever Z mount and RF mounts become popular because it won't be possible to adapt new Z and RF lenses onto DSLRs.

(**) I don't believe in lens adapters, lens adapters should be used only for fitting a special lens but not fitting a full set of lenses.

[joergens.mi]

Seems to be reasonable, can be corrected by color checker passport and generating a profile. with it.
The same is possible with all cameras.

The difference method is good, it would be interesting to get the same by the three channels with sign. Then you can identify the color error of the makers firmware.

maybe there is a possibilty to minimize the delta E's, when you accept that white (D4 by the K1) isn't white anymore or the tahe the mean value of all greys per channel to define a "new " grey because all have different delta E's

[beholder3]

The Best Cameras for Color Reproduction, Ranked

Sadly this is yet another poorly (really, really poorly) documented "test".

Obviously "color" can only be judged in JPG / TIFF or similar, not raw files. So some "recipes" have to be applied. In camera the recipes are the color modes plus white balancing. In raw converters you have the camera color profiles (e.g. .dcp files) which apply base curves, contrast, saturation etc.

Completely undocumented which settings they used.

Even the pdf on test methodology leaves it completely open.


From a color science perspective we need to know the exact illuminant used and the color recipes / color modes used in camera.

As it stands they could have used 3000K WB and "bleach bypass" color mode on any random camera.

And how do you pick the right JPG color mode in the camera? They all have a dozen. Just use the default one that comes up?

Any user who really cares about color would never use JPG straight ooC and even worse: the default color mode and even worse: the default color mode without any adjustments.
That is behaviour like you get it from mobile phone users.

[TonyW]

^^ This.

Image Engineering sets the camera color space to sRGB using the factory default for a color profile.

This was enough to convince me that the testing and conclusion very poor - unless of course a user only ever shot JPEG AND happened to prefer what the camera manufacturers committee had decided was a pleasing rendering at a default or standard preset (what about all the other JPEG presets how do they reproduce colour?)

[RonHendriks1966]

K-1 is as good as the Lumix LX100 compact camera

[biz-engineer]

Color is a very subjective thing. Personally, I use the Pentax neutral style most of the time , and I usually don't like other of the Pentax jpeg styles except in some rare cases. But, for default jpeg colors I prefer : Fuji and Canon. I find Nikon and Sony way too saturated. And in this test, the Canons are some of the last ranked models, which is weird.

---------- Post added 07-01-19 at 10:48 ----------

This was enough to convince me that the testing and conclusion very poor

Colors can't be tested in a lab, except full coverage of sRGB colors space, which all brands are able to cover. Color is a matter of people preference. Here this "test" is one more of the content marketing articles where the content is poor.

[MatKus]

This was already discussed some time ago, just can't find link right now.
But 2 main problems with this article IMHO:
1) problem with numbering. There are 2 #1, 3x#2, 2x #4 and 2x #5. This way K1 is numbered as #6, but in fact it is on 11th place.

2) Problem with color reproduction is, that number are not exactly, how it looks.
For me more importent is if picture looks correct for human eye, not for computer. Sometimes on this color chart i have problem to see any difference, yet values have big difference, somethimes for my eye difference is big, yet values are very close. So this comparison is IMHO totally worthless.

[beholder3]

Instead of taking home the message that the Nikon Z7 can't compete with even with a compact camera in this area (tabloids might write "second worst color reproduction camera of the year" ) there is one little interesting bit in the otherwise worthless result presentation:

They ALL fail miserably with the reds.

[biz-engineer]

For me more importent is if picture looks correct for human eye, not for computer. Sometimes on this color chart i have problem to see any difference, yet values have big difference, somethimes for my eye difference is big, yet values are very close. So this comparison is IMHO totally worthless.

I tried to develop a color calibration method, I failed, but I learned a lot:

1) Eyes are very sensitive, small variation of color mix can be appreciated on the overall image but much less on individual color patches
2) RGB colors numbers must be taken noise free, and also the color temperature and intensity of light source plays a role...meaning patches of colors have to be large enough to have the noise (even at ISO100) averaged out.
3) Pentax AWB color temperature is NOT the value that national standard labs report, BUT my eyes say that the Pentax color temp from AWB is more correct than national standard!

---------- Post added 07-01-19 at 11:04 ----------

They ALL fail miserably with the reds.

Clear, all camera struggle with the reds, that's clearly where larger sensors suffer less, regardless of what the articles may say.

[beholder3]

Clear, all camera struggle with the reds, that's clearly where larger sensors suffer less, regardless of what the articles may say.

As a matter of fact (at least for Pentax) the cameras do not struggle from hardware side. The developers creating the camera profiles struggle.

I checked a number of my images where in Lightroom the red colors appeared oversaturated and burnt out.

The raw showed two things:

1. The reds were fine in raw data
2. Actually the two green channels were 0.5 to 1 stop closer to burning out than red.

So poor .dcp color profiles were the main issue.

[Rondec]

Personally, I like the natural color profile the best. Not that I shoot jpegs frequently, but that's the preview my camera gives me when I am shooting raw and that's what I go with. Bright, which I think is the default, seems to oversaturate stuff. But this sort of thing is very personal. I see people all the time who seem to like their white balance really warm or their saturation bumped up to 11. It isn't my favorite thing, but if they like it then that's probably fine.

Regardless, I think the K-1 and K-1 II are the best cameras I have used with regard to white balance and preserving reds.

[biz-engineer]

The reds were fine in raw data

I had the same experience. Interestingly, when I use some auto-correction in software, the converted images look good to the eyes, although some of tone gradation of the red channel is degraded.
Depending on the red content of an image, regardless of the camera, it is not possible to have both correct exposure and fully correct colors in sRGB, simply because the color space won't allow it.
In sRGB, if you want to get a loyal tone gradation of the red, you have to cheat on other colors, or vice versa, regardless of the camera used.