[Rondec]

Why do you all think that people are actually buying K-1 IIs and upgrading their K-1s if the K-1 II is such a step back? It can't be that they aren't seeing these sorts of threads, because I see a lot more of them than any that say the K-1 II is great.

Maybe it is that people actually are more bothered by noise than by loss of detail and are willing to deal with one to get the other? Clearly there must be a reason, because (a) it costs a fair amount of money to convert your K-1 and (b) for folks buying new, the K-1 is still available and is quite a bit cheaper than the K-1 II. Regardless, I don't see it as a step back and am willing to have a little less detail to have cleaner high iso with minimal work. Then again, I am not taking photos of red swatches of material, so if that was the main thrust of my photography I would probably be pretty dissatisfied.

[normhead]

I have downloaded the files so I can see for myself the K-1MKII files have better color fidelity right from ISO 100. There is much better spatial presence due to the better continuous tone which creates better turn and shadow reflected light. The bottles have better volume due to this. There is a much better sense of space between the bottles and the background. The K-1 images look flat and linear in comparison. As ISO increases the K-1MKII holds better light and dark value. Better tonal uniformity. The K-1 images as ISO increases just become even more flat and linear due to the color noise. The K-1 files loose both light and dark values. The whites get darker and the blacks get lighter due to the color noise.

The K-1MKII files compared against the K-1 are not a step back or inferior in anyway to the K-1 files. Maybe people can use post processing and third party plugins to mitigate out the color noise to get K-1 images to look more like the K-1MKII files I am just glad I don't have to deal with that anymore. Download the IR files for the D850 and A7RIII. They are not even close to being better than the K-1 let alone the K-1MKII.

What we have been told is that the next generation sensors will be BSI which have better high ISO performance like the Sony sensor in the D850 and A7RIII but these BSI sensors are not even better than what the K-1 produces. The only thing they have is more resolution but the high ISO performance pretty much mirrors the K-1. As ISO increases a blanket of color noise covers the image.

The Accelerator Unit was already proven in the KP to produce better high ISO files. It does the same thing for the K-1MKII.

What happens when Ricoh puts the Accelerator Unit in the GRIII. I am looking forward to that.

At this point I'm more interested in why you say that than what you're saying. We need some evidence.

[MJKoski]

IR images are showing the same effect as lenstip LPM measurement drop. Like I said in April there is no free lunch. Multiple EV gain in noise levels is not possible compared to other Sony 36MP cameras without "cheating" a bit.

[Mikesul]

IR images are showing the same effect as lenstip LPM measurement drop. Like I said in April there is no free lunch. Multiple EV gain in noise levels is not possible compared to other Sony 36MP cameras without "cheating" a bit.

Good point.

[normhead]

IR images are showing the same effect as lenstip LPM measurement drop. Like I said in April there is no free lunch. Multiple EV gain in noise levels is not possible compared to other Sony 36MP cameras without "cheating" a bit.

I often remove noise by blurring the background in post. Kengoh often does it by shooting a scene before a bird arrives at low ISO, and then pasting a high ISO bird onto a low ISO background. There are other ways to deal with noise than noise reduction, and as a rule, I try to avoid using NR because we have noted in the past, it makes images mushy. Ioten on noisey images push NR as high as I can before the images loses all semblance of detail, and often end up with a compromise, or the blur tool.

For this type of image the K-1 is better for me, based on IR samples. ISO 1250 on a K-3 (preferred highest noise ISO 640 which lets me post without any NR,) The bird is as photographed, the back ground is blurred. It's my opinion after studying the sample photos the K-1 would be better for this type of image. Resolution is definitely preferred over NR for this image. The noise on the bird isn't even an issue. The feathers are so uneven in their colouration the noise is masked and pretty much invisible. On the consistent tones in the back ground it was oppressive.



For this type of shot the K-1 mkII would be just fine. Shape and richness of colour are more important than detail. But I'm not carrying two of those beasties.


Anyway, that's my bottom line. I want the K-1 images more often than I want the K-1 mkII images, and the K-1 images are more than adequate for my use.

[RGlasel]

Also look at the shirt texture behind the bouquet. This appears to be a focus difference. the K-1 shirt is more in focus, where the K-1ii bouquet is in focus.Trying to get everything the same for comparisons is crazy hard.

It's vastly more difficult in real life.

At full magnification, there definitely is a difference in how the blouse looks, but with a 70mm focal length and f8, I highly doubt the difference is because of different focal points. It looks to me like one camera happened to put more contrast in the whites and the other camera happened to put more contrast in greens and oranges. Was that a result of very slight variances in lighting or subtle differences in how each camera processes images or is it random? It also appears to me that differences in one set of comparison photos don't always appear in other sets of comparison photos. Also, everyone has different colour perceptions and every display device renders colours slightly differently and colour perception and colour rendering with the same eyes and display device changes over time. How much coffee you drink and your hormone levels affect your colour perception. At a small enough scale, even careful colour calibration doesn't give exactly the same results every time and a camera's processor could produce slightly different image files even if the signal read by the sensor is very similar.

There is a simple answer when conflicting results come from carefully controlled tests and that answer is that there is no difference in the factor being tested for. Even if someone selects a K-1 or a K-1 II for how likely certain types of details in certain colour ranges will be visible in the image file, they will inevitably make mistakes either because of elements they missed or didn't perceive correctly and when they go back to re-shoot the scene, some of the elements of the scene will have changed. It's madness.

[Rico]

At this point I'm more interested in why you say that than what you're saying. We need some evidence.

The evidence is in the files. You can see how the K-1MKII outperforms the K-1, D850 and A7RIII using the DPR Studio scene comparison tool as ISO increases. As I keep saying when ISO increases the K-1MKII holds better color. Better tonal uniformity. Better value both light and dark.

The K-1MKII files have better 3D appearance. Better spatial depth. The K-1, D850 and A7RIII all have a curtain of noise that flatten the images making them look more linear compared to the K-1MKII files. You can see that effect looking at the PF test images comparing the K-1MKII and K-1.

[C_Jones]

I would not agree that the K-1 has any superiority or advantage over any aspect of the the K-1 II. When using properly contrasting images, I myself would conclude that the K-1 II is the advanced model due to its image quality and AF. The K-1 is a high performance camera, but I would see the K-1 II as the more advanced performer due to the newer technology it utilizes. No offense to anyone, that is per my experience using those Pentax models.

[Rico]

I often remove noise by blurring the background in post. Kengoh often does it by shooting a scene before a bird arrives at low ISO, and then pasting a high ISO bird onto a low ISO background. There are other ways to deal with noise than noise reduction, and as a rule, I try to avoid using NR because we have noted in the past, it makes images mushy. Ioten on noisey images push NR as high as I can before the images loses all semblance of detail, and often end up with a compromise, or the blur tool.

For this type of image the K-1 is better for me, based on IR samples. ISO 1250 on a K-3 (preferred highest noise ISO 640 which lets me post without any NR,) The bird is as photographed, the back ground is blurred. It's my opinion after studying the sample photos the K-1 would be better for this type of image. Resolution is definitely preferred over NR for this image. The noise on the bird isn't even an issue. The feathers are so uneven in their colouration the noise is masked and pretty much invisible. On the consistent tones in the back ground it was oppressive.



For this type of shot the K-1 mkII would be just fine. Shape and richness of colour are more important than detail. But I'm not carrying two of those beasties.


Anyway, that's my bottom line. I want the K-1 images more often than I want the K-1 mkII images, and the K-1 images are more than adequate for my use.

normhead I can understand the K-1 images being adequate but that doesn't mean the K-1MKII has inferior image quality.

Your hummingbird image is an excellent capture. The difference the K-1MKII would have on such an image is first the out of focus background would have better tonal uniformity. A more buttery bokeh that would not need to be blurred. The brown areas under the wings and tail of the hummingbird would be handled with a more 3D appearance than the splotchy brown that is there. The black beak and area leading up to the eye would be handled with better volume and sense of space.

You can see these types of differences from the IR, DPR and PF test images. In the IR images the bottle caps are a good example to how the K-1MKII handles black in a superior way by holding better black point. The caps have a better sense of volume and separation of space between the background. The K-1, D850 and A7RIII all have a flat linear look compared to the K-1MKII. This starts right from ISO 100.

---------- Post added 06-27-18 at 01:49 PM ----------

Rico, I'm interested in how you're doing your processing. The one area I'm having trouble with on the the K-1II files is the bright red swatch with leaf patterns on it (two above the beer bottle). Would you be able to share your results on ISO 1600 RAW files from both cameras (100% crop on each) and explain how you've got equal or better results from the K-1II? I've tried various different approaches, and whilst I can get reasonably close, I can't get the K-1II to be as good or better than the K-1 in that specific area of the image

Incidentally, I see the same effect in the KP files vs K-3... better noise handling, but at the cost of some fine detail in specific circumstances. However, like the K-1II, I'd say it's a worthy compromise for some photographers.

To keep things equal I take the DNG files then convert them to JPG's using Digital Camera Utility 5.

The reds have more to do with color profiles this is easily seen switching between them. Generally for every manufacture the red channel needs to be adjusted especially dealing with skin tones. Channel mixer in Photoshop is effective in dialing back the un-natural redness to skin tones. I find a little goes a long way. The more post processing applied to an image the more plastic and lifeless they become.

[BigMackCam]

To keep things equal I take the DNG files then convert them to JPG's using Digital Camera Utility 5.

The reds have more to do with color profiles this is easily seen switching between them. Generally for every manufacture the red channel needs to be adjusted especially dealing with skin tones. Channel mixer in Photoshop is effective in dialing back the un-natural redness to skin tones. I find a little goes a long way. The more post processing applied to an image the more plastic and lifeless they become.

Thanks - that's interesting, and makes some sense to me.

Could I ask a favour? Would you be able to post 100% crops of the red swatch area from your DCU5-processed Imaging Resources K-1 and K-1II ISO 1600 images (similar to those I posted previously)? I'm entirely prepared to accept that colour profiles and red channel adjustment can make a big difference, and I'd be delighted if in fact the K-1II's images can show even equal detail in that red swatch - as, so far, achieving that has evaded me

[Rico]

Thanks - that's interesting, and makes some sense to me.

Could I ask a favour? Would you be able to post 100% crops of the red swatch area from your DCU5-processed Imaging Resources K-1 and K-1II ISO 1600 images (similar to those I posted previously)? I'm entirely prepared to accept that colour profiles and red channel adjustment can make a big difference, and I'd be delighted if in fact the K-1II's images can show even equal detail in that red swatch - as, so far, achieving that has evaded me

BigMackCam which IR ISO1600 files in particular? NRO PSR etc? The settings between cameras are much different depending on the file and even differ for each camera file to file. The K-1 1600 PSR files has High ISO NR set to Auto with Image Tone set to Natural. The K-1 used had FW Version 1.00. The corresponding K-1MKII ISO1600 PSR file also has High ISO NR set to Auto but the Image Tone is set to Bright. The Contrast and Sharpness are both set to 1. The Color Temperature has been Fine Tuned M-G to Green 1.

So for starters both cameras are having in camera High ISO NR applied where it could be turned off. Having different Image Tone settings and shifting the Color Temperature is a likely source to why you are having difficulty comparing the Red swatches.

I downloaded the PSR and the NRO thinking it means NR Off. Checking the K-1 file it does show that all NR was turned off but they used Bright Image Tone instead of Natural as they did with the PSR image. The K-1MKII NRO file also has Image Tone bright and NR Off the Color Temperature is also adjusted Green 1. Both Contrast and Sharpness set to 1. Why these settings are being changed all over the place makes no sense to not only comparing the files to other cameras but comparing the settings for each camera. If they are using Natural for PSR image and Bright for NRO the files can not really be compared properly as the images have a different appearance both in color and tonal value which is clearly visible.

The DPR test images at least have similar settings for both cameras but they adjust the Color Temperature to Amber 1. High ISO NR is set to Auto so both cameras are applying in camera NR. DPR also has lens correction functions turned on for both Lateral Chromatic Aberration Correction and Diffraction Correction. The use of the FA77 introduces purple fringing all over the images. The 85 used for the D850 does the same thing. These are horrible lenses to be used for such comparison scenes.

[reh321]

IR images are showing the same effect as lenstip LPM measurement drop. Like I said in April there is no free lunch. Multiple EV gain in noise levels is not possible compared to other Sony 36MP cameras without "cheating" a bit.

This all depends on what your main goals are. Not everyone looks at images under magnification. When I showed slides, I sat everyone behind the projector because I wanted them to see the trees, not the cell structure in the leaves.The k-1ii is selling well in Japan - don't assume that these people are purchasing as a result of ignorance .... maybe they see what they like.

[BigMackCam]

BigMackCam which IR ISO1600 files in particular? NRO PSR etc? The settings between cameras are much different depending on the file and even differ for each camera file to file. The K-1 1600 PSR files has High ISO NR set to Auto with Image Tone set to Natural. The K-1 used had FW Version 1.00. The corresponding K-1MKII ISO1600 PSR file also has High ISO NR set to Auto but the Image Tone is set to Bright. The Contrast and Sharpness are both set to 1. The Color Temperature has been Fine Tuned M-G to Green 1.

So for starters both cameras are having in camera High ISO NR applied where it could be turned off. Having different Image Tone settings and shifting the Color Temperature is a likely source to why you are having difficulty comparing the Red swatches.

I downloaded the PSR and the NRO thinking it means NR Off. Checking the K-1 file it does show that all NR was turned off but they used Bright Image Tone instead of Natural as they did with the PSR image. The K-1MKII NRO file also has Image Tone bright and NR Off the Color Temperature is also adjusted Green 1. Both Contrast and Sharpness set to 1. Why these settings are being changed all over the place makes no sense to not only comparing the files to other cameras but comparing the settings for each camera. If they are using Natural for PSR image and Bright for NRO the files can not really be compared properly as the images have a different appearance both in color and tonal value which is clearly visible.

The DPR test images at least have similar settings for both cameras but they adjust the Color Temperature to Amber 1. High ISO NR is set to Auto so both cameras are applying in camera NR. DPR also has lens correction functions turned on for both Lateral Chromatic Aberration Correction and Diffraction Correction. The use of the FA77 introduces purple fringing all over the images. The 85 used for the D850 does the same thing. These are horrible lenses to be used for such comparison scenes.

Thanks for the reply

The specific images I'm comparing are:

Digital Cameras, Pentax K1 Digital Camera Test Image

... and:

Digital Cameras, Pentax K-1 II Digital Camera Test Image

Regarding the difference in settings - NR, colour, image tone etc. - these shouldn't affect in my own processing, as I'm using Darktable. They should only affect straight-out-of-camera JPEGs, or RAW files when loaded into DCU specifcally (since DCU reads any camera settings and applies appropriate adjustments to RAW files in the software). In Darktable, I have all processing modules switched off except for the essentials (which are set identically)...

Incidentally, one of the reasons I like Darktable so much is that it's possible to switch everything off. You get to see a RAW file in all its gory detail, without any of the processing that many other RAW development programmes (LR included) carry out by default

[Ian Stuart Forsyth]

Incidentally, one of the reasons I like Darktable so much is that it's possible to switch everything off. You get to a RAW file in all its gory detail, without any of the processing that many other RAW development programmes (LR included) carry out by default

You still have to worry about how much base line exposure is being add to the k1 file, for both the k1mk1 and mk2 they are using a baslineexposure of-0.5 so when it comes to seeing what is found in the raw file you have to also compensate for that multiplication ( lightness added to the raw file) so that the image is mapped to what you would find using the cameras default mapping. This puts what you would see in the raw file much darker than shown in a raw converter.

There is also the problem of white balance that will also not allow you to see the red, green, blue channels and how the fall within the histogram of that raw file . For a person to see how the RGB falls within the raw file you need to know how much those channels are being multiplied, if we take a look at the these values for one of the IR shots you will see that the red channel is being multiplied by over a stop while the green is not and the blue is multiplied by 1/3 stop there about. To override this you need to see the raw file without this and use a UniWB where the channels are not being multiplied. To see how the channels are being multiplied look in the exif data



If you look at the as shot neutral you will see 0.475 , 1 , 0.697
What this tells us that for this WB red is being multiplied by 1.1 stops the green has none and the blue is 0.43 stops
here you can also see the baseline exposure that tells adobe how much the raw file has to be lightened -0.523

If we where to see the raw file it would appear darker (because of the baseline exposure) and would have a very strong green blue color because of the light being used.

If we where to look at the raw file it would look something like this

You can see in the Raw histogram how the 3 channels have captured at different levels with the red being the one furthest from clipping and why the raw image has a green blue cast.

How to override ACR there is 2 very useful reference bellow the first is
Where Are My Mid-tones? - Photography Life

Once you know how to override ARC you can further override the base line exposure that is stored within the raw file so that it will not be include in the ARC. For the k1 at iso 800 it is -0.523

Where Are My Mid-tones? Deriving Hidden Baseline Exposure Compensation - Photography Life

Once you account for this behind the scene processing you will start to see the raw file more as it is

What is UniWB? | byThom | Thom Hogan


Remove the channel multiplying



and you are left with what is found within the raw file. ( the image above is not using the correct UniWB as I did not have time to build one and just eyeballed one in )

If you are interested in the correct UniWB set the- as shot neutral- to 1,1,1 and save the to the raw file.

[drache]

This all depends on what your main goals are. Not everyone looks at images under magnification. When I showed slides, I sat everyone behind the projector because I wanted them to see the trees, not the cell structure in the leaves.The k-1ii is selling well in Japan - don't assume that these people are purchasing as a result of ignorance .... maybe they see what they like.

I think the k-1ii is an improvement on almost all fronts. There are one or two things I'm unsure of, adding the ability to disable the accelerator would fix all of them. This means I can ignore what might not be an issue but geek out if I want to. I think the accelerator is making the image better out of the camera, which is important because I feel many images will go straight to web. The question I have for a Ricoh engineer is, is the accelerator circuit cutoff variable? Can a user defined range be added or a toggle in the menus.
I think your right reh, the images are better looking faster and people aren't internet scientisting these things.
I like to do that though
<flogging dead horse image>
:-D