[BigMackCam]

So true mtkeller. DPR could have underexposed the ISO 100 and 200 images at a fast enough shutter speed then just adjusted in ACR. That is what they are doing to the files anyway. In ACR they adjust both exposure and white balance both the temperature and tint. What is the difference if they adjust the exposure by 2/3rd stop or 1.5. My years of reading DPR I don't think anyone there is professional enough to be adjusting the white balance temperature and tint.

I appreciate DPR posting RAW files for download for all the makers. It becomes a crux for them when they make statements that can be easily refuted by just downloading the files for yourself to take a look. When I first saw the DPR baked versions of the K-3III files I said these images can not be this bad. Downloading the RAW files I could see for myself they were very excellent. Better than the K-3III peers using the same sensor.

DPR is up front in the sense they do indicate they imbed a custom baked Pentax ACR profile. They only indicate this in the Info Tab for Pentax cameras and only the Pentax camera DNG files DPR uploads are saved with these Image Settings. All the other brands the files are uploaded set to how the file came out of the camera in ACR they open in Camera Raw default. The argument that DPR must do this as in the case of the K-3III for ACR because there is no profile yet from Adobe for the K-3III is really a mute lie.

The purpose to shooting DNG is that Adobe ACR will always be able to properly handle the file regardless if there is a "camera" profile or not from Adobe as those profiles are really so ACR can read the manufactures RAW fill prefix NEF RAF CR2 CR3 etc. DPR isn't shooting in Pentax PEF so there is no need to have a baked DPR imbedded profile for the Pentax DNG files.

DPR is doing a lot to the files by adjusting the exposure and white balance. They lift a whole layer of micro detail by doing this which makes the JPG files they posted look so terrible. Only one adjustment is really needed in ACR. There is color blooming happening so just adjusting the Saturation back to -10 or so controls this nicely without lifting out any detail. DPR's exposure and white balance adjustments only make the color blooming worse while also lifting out micro detail.

In DCU5 the info shown there for the DPR K-3III DNG files indicates DPR had both Lateral Chromatic Aberration Correction and Diffraction Correction set to On. Plus High-ISO Noise Reduction set to Auto. All three of these settings are doing something to the files whether these effects show up in JPG files only or if the DNG files have these settings imbedded as well I have no clue. Personally I never have those settings turned on.

DPR could have just posted the ISO 100 and 200 files extracted from the Pixel-Shift images instead of posting the blurry images. The blurry images are pointless for any serious evaluation which reflects on their standards. The point of DPR's Comparison scene is to try to keep things equal. DPR doesn't do this by then adjusting the files with special baked in profile settings that make the files look worse.

The K-3III files DPR posted are also very askew tilting down to the left. They are way off compared to the other cameras. Which also effects they way they look especially when zoomed in 200% compared to images properly squared.

B&H Photo in Manhattan has the K-3III in store now both Black and Silver for handling. They had no lens on them so I did not get to see the viewfinder. Excellent ergonomics. Compact size.

To be clear, nothing has been "baked in" to the DNG files. Any adjustments DPR has made - specifically, white balance (probably using the eye-dropper tool - no skill required), exposure (just over half a stop - pretty minor), sharpening, etc. - are stored as values for use by the processing software, so they can easily be de-selected when analysing the image quality.

Darktable is great for viewing DNG files at their "most raw", since you can easily switch off all default adjustments. See below portions of the ISO 100 DNG file, imported into Darktable 3.4.1.1. I've switched off all adjustments including exposure compensation, and reset white balance to the "as shot" (i.e. camera calculated) values...

Attached Images

   

---

Last edited by BigMackCam; 06-17-2021 at 02:20 PM.

[BigMackCam]

Nothing is "baked in" to the DNG files. Any adjustments DPR has made - specifically, white balance (probably using the eye-dropper tool), exposure (just over half a stop - pretty minor), sharpening, etc. - are stored as values for use by the processing software, so they can easily be de-selected when analysing the image quality.

Darktable is great for viewing DNG files at their most raw, since you can easily switch off all default adjustments. See below portions of the ISO 100 DNG file, imported into Darktable 3.4.1.1. I've switched off all adjustments including exposure compensation, and reset white balance to the "as shot" (i.e. camera calculated) values...

... and further to the above, here's what the file looks like in RawTherapee 5.8, again with all adjustments disabled (including RT's excellent but potentially misleading "capture sharpen" feature):

[Rico]

Sorry for being contrary, but processing the DNG to linear TIFF in dcraw shows softness similar to to what DPR posted. See full resolution center crop below:

Attachment 539002

Similar to a data dump from RawDigger, this is about as close as one can get to seeing least-processed output from RAW. For more information on linear TIFF, Google "dcraw linear tiff".

FWIW...the corners actually performed a bit better.


Steve

Steve I don't know what file you are showing. Is this the blurry ISO 100 or 200 file? Those files are useless for any serious scrutiny. I don't have any explanation why the ISO 100 and 200 files are so blurry.

For all the DNG files DPR is turning off the ACR default Detail and Color Noise settings which makes their K-3III JPG images soft. It will make a TIFF file look just as soft. It's not a mystery why the DPR JPG files are soft it is why they do it in the first place. They do this for the other brands which make their images look soft as well but DPR does not upload the other brands RAW files with these applied Image Settings baked in they only do this with the Pentax DNG files they upload.

[slartibartfast01]

Steve I don't know what file you are showing. Is this the blurry ISO 100 or 200 file? Those files are useless for any serious scrutiny. I don't have any explanation why the ISO 100 and 200 files are so blurry.



For all the DNG files DPR is turning off the ACR default Detail and Color Noise settings which makes their K-3III JPG images soft. It will make a TIFF file look just as soft. It's not a mystery why the DPR JPG files are soft it is why they do it in the first place. They do this for the other brands which make their images look soft as well but DPR does not upload the other brands RAW files with these applied Image Settings baked in they only do this with the Pentax DNG files they upload.

How can you seriously say that a file that shows an issue is useless for any scrutiny? The fact that it is blurry is enough reason to scrutinise it.

[stevebrot]

Steve I don't know what file you are showing. Is this the blurry ISO 100 or 200 file? Those files are useless for any serious scrutiny. I don't have any explanation why the ISO 100 and 200 files are so blurry.

IMGP0436.DNG, the ISO 100 capture. Sorry for the lack of clarity. I have updated my post. As for being useless, perhaps you misjudge my methods. The data is SOC DNG. Any Pentax-specific DPR funny business would have to be with the camera attachment or maybe with the focus method.


Steve

[Rico]

IMGP0436.DNG, the ISO 100 capture. Sorry for the lack of clarity. I have updated my post. As for being useless, perhaps you misjudge my methods. The data is SOC DNG. Any Pentax-specific DPR funny business would have to be with the camera attachment or maybe with the focus method.


Steve

stevebrot I wasn't questioning or slighting your methods so there isn't any thing to misjudge. It's DPR's methods that are being questioned. I don't see how the ISO 100 & 200 DNG files DPR posted are worth any analysis they are so blurry. What's the point when it is so unclear why? Especially when DPR acknowledges they do have ISO 100 & 200 files without the blurry problem extracted from a Pixel-Shift image. They could have shot at a speed slow or fast enough to get a sharp image then adjusted the exposure using ACR. They are changing the exposure for the files anyway.

DPR has comparison images of the sharp and blurry ISO 100 files with links to the files which are only crops. It is clear one is sharp the other is blurry. Strangely those two comparison images have the same file number IMG0481-01 the blurry image and IMG0481-02 the sharp image. This actually means the files they posted for comparison are the same file. One can not be mechanical shutter and the other Electronic Shutter with the same file number. I just noticed when taking a second look at what DPR posted. That type of numbering the _01, _02 at the end of the file number happens when saving multiple versions of the same file in ACR.

DPR has a habit for this kind of stuff. When DPR did the K-1MKII comparison against the D850 and A7III the lighting was so poorly done the purple fringing it caused made any real comparison useless. All three cameras it is evident. All the high contrast white and black transitions have either purple or green fringing. The print of the people next to the Fab Four the purple fringing is so bad across it you can see Newton's Rings. In that comparison DPR claimed the K-1MKII Accelerator Unit was removing micro detail by focusing in on the small black white and grey lettering which is high contrast where purple fringing happens as evidence denoising was causing the loss of micro detail.

The lighting used for the K-3III there still is the same overall purple fringing happening across the entire picture plan it's just less evident but is easily seen along the high contrast white and black areas of the background and targets. It really causes a visual hum for the entire image which is part of the color blooming. Pushing back the Saturation in ACR cleans this up in the files I have downloaded but can not completely make it go away.

[stevebrot]

stevebrot I wasn't questioning or slighting your methods so there isn't any thing to misjudge. It's DPR's methods that are being questioned. I don't see how the ISO 100 & 200 DNG files DPR posted are worth any analysis they are so blurry. What's the point when it is so unclear why?

I basically said that you did not understand what I did.

I guess my intent was to say that the softness is real and was real when the shot was taken. Whether it goes away with an electronic shutter capture is not the question. Whether there is strangeness in the way they process or present Pentax shots is not the question. The question is why were those two examples soft from the camera. What they posted is what I found by doing some simple forensics.

To be honest, I don't really care where the photo was posted. If the problem was reported as the chance discovery of an established member here, I suspect the reaction would have been more measured with an eye to helping characterize what may be a very real fault with the goal to potentially assist Ricoh in crafting a remedy.

In case it is not obvious, my opinion is that the persistent, predictable, and frankly pitiful whinging about slights against Pentax by DPReview is not characteristic of our community here at PF nor does it represent any particular animosity of our membership against that site...save perhaps a sort of self-righteous attitude regarding comparison of here with their rather free-wheeling forums.

End of my participation in this off-topic line of discussion.


Steve

---

Last edited by stevebrot; 06-17-2021 at 10:53 PM. Reason: forgot to add sig

[LenH55]

I just finished a year long project digitizing my dad's slides, BW 35mm negative rolls, 5x7 negatives etc. The extended family after bugging me for years is very happy with the results. I used my k5iis, copy stand, and my best 50 mm lens with extender rings to get to 1:1. Now I want to do the same for my personal stuff - 35mm and 6x6 negatives and 35 mm slides. Of course this calls for an extra level if pixel peeping, hence the interest in the k3iii. Actually there are three choices to step up the resolution that I may be interested in, the k3iii, the KP, and the k1ii. The KP is being phased out but I could probably still find one. I don't know if I want to go back to full frame with the Ki as I am getting sort of comfortable with APC. Although way overpriced the k3iii does sound like a great camera but I am not going to buy a camera that I have to do workarounds for fundamental functions due to hardware or software issues. I really don't think they exist but I hope you can understand my hesitation. I'll just watch for a while to see how it all shakes out.
Len

I just finished a year long project digitizing my dad's slides, BW 35mm negative rolls, 5x7 negatives etc. The extended family after bugging me for years is very happy with the results. I used my k5iis, copy stand, and my best 50 mm lens with extender rings to get to 1:1. Now I want to do the same for my personal stuff - 35mm and 6x6 negatives and 35 mm slides. Of course this calls for an extra level if pixel peeping, hence the interest in the k3iii. Actually there are three choices to step up the resolution that I may be interested in, the k3iii, the KP, and the k1ii. The KP is being phased out but I could probably still find one. I don't know if I want to go back to full frame with the Ki as I am getting sort of comfortable with APC. Although way overpriced the k3iii does sound like a great camera but I am not going to buy a camera that I have to do workarounds for fundamental functions due to hardware or software issues. I really don't think they exist but I hope you can understand my hesitation. I'll just watch for a while to see how it all shakes out.
Len

I would buy the K-1 and a dedicated makro lens. Use pixelshift. This for the 135-film. For 120-film I would use a scanner, but the K-1 would still make a good job but miss out on the resolution that 120 gives you.

[stevebrot]

What happens if the camera is intentionally not level/true for this case vs. intentionally perfectly level/true? Those with geared heads having bubble levels would have an advantage here.

[Angulon]

@LenH55,

..."but I hope you can understand my hesitation"

of course, same applies to just about any new acquisition in those circumstances.
.

[LenH55]

For 120-film I would use a scanner.

I do plan on buying the fa 50mm macro lens which I can use with APC, full frame or film. Can you recommend a scanner which would be up to the job?
Len

I do plan on buying the fa 50mm macro lens which I can use with APC, full frame or film. Can you recommend a scanner which would be up to the job?
Len

I am using the Epson V600 but that is a buget version. I do get very good scans from it. What I do is overscan and downsize. This makes it easier to sharpen.

All of the images in this album is scanned with said scanner: (Most not overscanned since it takes time and I am lazy...)

Pentax 6x7 | Flickr

/T

[LenH55]

What I do is overscan and downsize. This makes it easier to sharpen.

Your scanner serves you well. Very nice photos. I especially like the even black and white tonal scale.
I do not have a scanner so what is the basic concept of overscanning then downsizing. How does that help in sharpening.
Thanks, Len

Your scanner serves you well. Very nice photos. I especially like the even black and white tonal scale.
I do not have a scanner so what is the basic concept of overscanning then downsizing. How does that help in sharpening.
Thanks, Len

It is the same principle with all images. Scan for, just an example, 36 MP and downsize to 24 MP. This makes grain/noise lessen in the process. This works well for digital as well. With less noise you can sharpen more since only real detail gets sharpened.