Please explain who can make sense of this
So, CO1 v 7.2.1 is out. I downloaded and installed. For 5 months I have been waiting for Pentax K3 support, 5 whole months that is.
So now I process an image from the Pentax K3 and I embed the camera profile and choose to open in Photoshop.
Photoshop warns me that the image I open does not contain the workspace image (as it should), but the following profile: "Pentax K5 genericv2"
Now understand, I have been waiting for five months for my favorite RAW developer to support the Pentax K3, that has a BRANDNEW 24 mp Sony sensor, and is by no means a K5 rehash.
Five months it took Phase one to recycle the K5 profile into a very questionable support for the Pentax K3
Before I let my feelings run away with me and shout equally questionable terms directed at the Phase one developers on this public forum, what should I make of this?
Really Phase one, could you not even be bothered to create a custom profile for the stellar Pentax K3 in 5 , FIVE!!!, months?
What to say.....
Chris
So now I process an image from the Pentax K3 and I embed the camera profile and choose to open in Photoshop.
Photoshop warns me that the image I open does not contain the workspace image (as it should), but the following profile: "Pentax K5 genericv2"
Now understand, I have been waiting for five months for my favorite RAW developer to support the Pentax K3, that has a BRANDNEW 24 mp Sony sensor, and is by no means a K5 rehash.
Five months it took Phase one to recycle the K5 profile into a very questionable support for the Pentax K3
Before I let my feelings run away with me and shout equally questionable terms directed at the Phase one developers on this public forum, what should I make of this?
Really Phase one, could you not even be bothered to create a custom profile for the stellar Pentax K3 in 5 , FIVE!!!, months?
What to say.....
Chris
0
-
Have you actually tried the K3 profile? It exists.
http://kazemisu.me.uk/images/pentax_k3.jpg
If it's not automatically being called for K3 files, it might be a bug.0 -
[quote="Keith Reeder" wrote:
Have you actually tried the K3 profile? It exists.
http://kazemisu.me.uk/images/pentax_k3.jpg
If it's not automatically being called for K3 files, it might be a bug.
Keith,
Colors are terrible, that is: TERRIBLE in capitals. No true blues at all, very, véry flat and dull. That is impossible with a true profile. This is a rehash of the K5 profile, or I am losing my eyesight. I know the look of Raw images with improper color profiles, and this is one if I ever saw one. Only the reds have been boosted somehow, but this is a BAD profile.
It's simply incredible that Phase one is up to this, there's no other way to put it....
b.t.w. it's easy to rename a profile into any name you would want, that does not mean anything. I believe Photoshop more than than the outer naming in CO1, especially since images are looking as bad as they do. I just cannot believe it, after five months.
Chris0 -
Checking the long awaited K3-generic profile in Colorthink profile inspector gives me this:
Look at the internal name and the creation date........
http://www.vogelsenhunleefwereld.nl/K3_Phaseone.tiff
Chris0 -
What a bummer!
I feel for you.
Hope C1 comes up with a good solution.0 -
[quote="ChrisM" wrote:
... and I embed the camera profile ...
Chris
Why embed a camera profile in the output file?
I'd say you should instead record here the name of the appropriate colourspace, the like of Adobe98, ProPhoto or sRGB.
I believe it will alter the appearance of your output file to your liking. Best of luck.
Cheers,
Mogens0 -
[quote="ChrisM" wrote:
I embed the camera profile and choose to open in Photoshop.
You should never (or at least in very rare cases) embed the camera profile to the output. Here you should use srgb, adobergb or equivalent.0 -
[quote="mli20" wrote:
[quote="ChrisM" wrote:
... and I embed the camera profile ...
Chris
Why embed a camera profile in the output file?
I'd say you should instead record here the name of the appropriate colourspace, the like of Adobe98, ProPhoto or sRGB.
I believe it will alter the appearance of your output file to your liking. Best of luck.
Cheers,
Mogens
Of course I always output to a working space profile, usually a ProPhoto like color space. (I use a color space created by J. Holmes, DCAM 4, that is very similar to ProPhoto but has a nicer tone curve)
My reason for embedding the camera profile was that I did not like the rendering of my K3 images because colors were very flat and lifeless, especially compared to the normally excellent color output of CO1. To my amazement the profile read, as I explained with a screenshot "Pentax K5v2 generic".
I took a look at the color gamut of this profile, and it is exactly the same as the 2011 Pentax K5v2 generic profile indeed, with the exception of the red primary being shifted outwards.
The Pentax K3 has a brand-new generation 24 mp sensor that is capable of excellent color. I simply don't believe that this modified old 2011 profile gets anywhere near the best out of this sensor color wise. I compared colors of CO1 extensively with Raw converters the likes of Lightroom, Raw developer, DxO optics. CO1 was always clearly the better. Not so with the K3, it trails behind color wise to DxO, but even to Adobe standard.
Now, I am not going to imply that I now more about color than Phase one engineers, but they will have a hard time convincing me that the way they made the K3 profile is proper practice.
What I can not understand, is why a best in class 24mp Apsc camera that without exception gets the highest ratings in reviews, is treated with lackluster support by Phase one this way, with a re-used profile from 2011!!, while cameras like the Olympus EM-10 or even the Nikon coolpix are served with multiple lens profiles and properly designed color profiles. And this almost 6 whole months after release. I can try to fill this in, but I don't want to spell out negative assumptions when this leads absolutely nowhere.
I would however appreciate some clarification.
This is what the title of this post refers to.
@Phase one: Ricoh has plans with Pentax, a FF camera is planned for this fall, high quality lenses are on the roadmap. It is worth taking Pentax seriously, even when they now do not make up a large part of your user base.
Chris0 -
ChrisM, as you from our Support Team, we are looking into the issue about the profile and will take the appropriate action after this analysis.
I cannot give any time-estimates.0 -
I located a K-3 raw file on the web and downloaded it to my computer. A Colorchecker is present in the image.
I imported it into Lightroom, Adobe Standard Profile, and I opened it in COP7.2.1, using the Pentax K3 Generic Profile.
Minor colour differences as expected, but both exhibited vibrant colours.
I processed the image from COP7 to PhotoShop, where it look equally good.
Based on this I cannot see anything fundamentally wrong with the PO Pentax K3 Profile.
Cheers,
Mogens0 -
[quote="mli20" wrote:
I located a K-3 raw file on the web and downloaded it to my computer. A Colorchecker is present in the image.
I imported it into Lightroom, Adobe Standard Profile, and I opened it in COP7.2.1, using the Pentax K3 Generic Profile.
Minor colour differences as expected, but both exhibited vibrant colours.
I processed the image from COP7 to PhotoShop, where it look equally good.
Based on this I cannot see anything fundamentally wrong with the PO Pentax K3 Profile.
Cheers,
Mogens
Thanks for the effort Mogens, but for what it is worth: Adobe standard profiles are generally not known as life-like color profiles. They represent extremely neutral and rather dull color renderings, hence the inclusion of camera specific color profiles for a lot of brands, including the Pentax K3. The color out of CO1 for all my cameras prior to the Pentax K3, and especially the Pentax K5, was much better than Adobe standard. Did you ever try to shoot a colorful building in bright daylight, remember its colors and open the image in Adobe standard profile? Not much left of the life-like colors whenever I tried. That was the reason in the first place for getting CO1.
Anyhow, Phase one is taking a look, so I'll wait for what comes out of that. If there's no more in store for the K3 than Adobe standard colors, then let them at least explain why they did not make a quality profile from scratch.
regards
Chris0 -
I just did another conversion of said image in Lightroom, after first having created a custom K3 profile, based on the Colorchecker in the image. Greater difference between Adobe Standard and the custom profile than I'm used to see when using my Nikon cameras.
However, in either case, the most notable difference is with the red patch: Much cleaner and slightly brighter in the C1 conversion!!! I know the COlorchecker from personal experience, and I find the rendition offered by the C1 Pentax K3 Profile very good. One other example: The Sky Colour patch was identical for all conversions I did.
Funny with the reds though. I have seen C1 accused of producing too dull reds, I believe that was for Nikon cameras, but as noted, the Pentax K3 Profile red is bright and punchy.
Chris, if we should take this analysis further, I'd like to propose that you post here a K3 RAW file, taken with your camera which include the Colorchecker in the image. I'd be happy to do C1 and Lighroom conversions of it and post the results here.
Cheers,
Mogens0 -
I put together a small image comparison to illustrate what I have been posting over the last couple of days on this forum.
I carefully selected an image that features the critical blue (mis)rendering caused by an improper K3 profile, but would like to add that it is also the green gamut that causes problems with this profile. A profile that is, as I tried to prove/illustrate, not the result of a real and careful profiling of an actual K3 camera, but the result of a lackluster tweak of an 2011 Pentax profile that was originally made for the K5, but that was earlier also "recycled" with an exact copy/paste, to serve as the generic profile for the Pentax K5IIs, a successor to the K5 that did not feature an AA filter, and that had a different color response from the K5, for which this profile was made. So the last flagship Pentax camera that Phase one actually DID profile, was the K5 in 2011. This explains the obvious irritation that is present in my posts these last days, all the more since support for the Pentax K3 took six months.
I have corresponded with Phase one elaborately in the past on this bad habit of recycling old profiles with regards to the top Pentax Apsc cameras, moreover since lesser cameras like the Nikon coolpix or the Olympus EM-10 are treated with much more care, not only proper profiles, but also including a multitude of lens profiles. In fact, it seems that Pentax is the only camera brand that is treated this way by Phase one, and that leads to the tone in my posting.
So here are the images, all images are at absolute default Capture one setting, and treated absolutely identical to give an as much as possible true representation of the profiles.
If there is any doubt on this, I will be more than happy to upload the original Raw file to Phase One, so they can verify the results.
First image is the image with the K3generic profile:
http://www.vogelsenhunleefwereld.nl/K3generic.jpg
The sky in this image is NOT rendered as it really was, the blue is drained out, and it has turned almost grey. Yet the eyes of the Oystercatchers are overly flaming red (see my remark at the bottom of this post), the result of (arbitrarily?) shifting the red primary outwards for this profile, something which I noticed when loading the profile into the profile inspector from Colorthink pro. The blue and green primary by the way were not modified, but directly copied over from the K5 profile. But let's stop this tech talking, let me show the same image at exactly the same settings, but rendered with the profiles of two cameras that feature an identical Sony 24mp sensor to the sensor that is used in the Pentax K3, profiles that have no old internal date or naming, and that I expect are the result of truly profiling the actual camera in house.
First the Sony Nex 7:
http://www.vogelsenhunleefwereld.nl/SonyNex7generic.jpg
Ah, the blue that was in the sky, is returning!! The red flaming eyes are more naturally red. Still, this profile is not exact for the K3, greens are not rendered properly. But this is just to illustrate that it is not the K3 camera that is unable to capture subtle blue tones in skies under subdued weather conditions. I don't doubt that the K3 generic profile will show vivid primary blues, but that is not what a good profile is about.
Next is the Nikon D5300, that possibly features the same sensor as the Pentax K3, as it was released after the K3:
http://www.vogelsenhunleefwereld.nl/NikonD5300generic.jpg
Getting closer and closer to te way the sky really was: subdues blue-ish. No tweaks in the color editor can repair the K3 profile to satisfaction.
I hope I sufficiently motivated my call for Phase one to get a Pentax K3 in house and create a true profile for this camera, as is done with all other camera brands. Again: if Phase one doubts the exactness of this comparison, I will be more than happy to upload the original Raw file, so they can duplicate these results on their own calibrated high end monitors.
I would also like to add, that the reason that I am putting so much weight on this issue, is that CO1v7 is simply the best Raw converter around, no comparison, WHEN they supply a proper quality profile that is!
Note: converting to sRGB for the web does not show all the points I make realistically, e.g. the flaming red eyes. But take my word for it. I use a calibrated Eizo monitor.
Chris0 -
Good luck!
The Nikon D800 profile isn't very nice either and there have been plenty of complaints from users but nothing from PhaseOne for months now. Its so frustrating 😭0 -
[quote="sizzlingbadger" wrote:
Good luck!
The Nikon D800 profile isn't very nice either and there have been plenty of complaints from users but nothing from PhaseOne for months now. Its so frustrating 😭
Thanks.
But just to be clear: I would not be complaining if the K3 profile had some issues related to taste, if these issues could be solved with use of the color editor, and the profile resaved as a corrected .icc profile.
The profile for the D800 is actually the result of profiling the D800 camera, that is what I presume to be the case for all cameras. Not so with the Pentax K3.
The issues with the K3 profile are i.m.o. not repairable with use of the color editor, the profile is just not correct overall, and I take it that that is because it is not the result of profiling the K3, but the result of tweaking the old K5 profile. The result is literally that images do not come to life, but are flat and lifeless in especially subtle blues and subtle greens.
Raw images need a good and balanced profile to bring the images to life.
If Phase one profiles the K3 and releases a profile that is not entirely up to my taste, but does bring images to life, than I will readily accept that and use the color editor to tweak to my own taste.
Chris0 -
Chris,
Out of interest have you compared the different profiles you mention in both Linear and Film replication mode?
Your comments yesterday (along with some other things unrelated to C1) prompted me to make a quick comparison using a previously edited file from my older Canon body (1D3).
Cloning a previously edited file and adding some colour read out points I changed the Base Settings and looked at the file with Sony a7 and Pentax K5 (mostly V2) profiles and was surprised that the differences were not so easy to see.
The files I used was strong in greens but light in standalone blue so not directly comparable with your observations BUT it was interesting that the readouts were generally very similar - especially in grey areas where colour and luminosity might be a point or two different but were generally consistently different across all measures.
The image features a large area of a red/orange colour that was not dissimilar to your bird's eyes and beak. Although some difference were evident I had to look carefully to spot them comparing images side by side on screen. The reported colour and luminosity values were usually quite close with exception of one colour - not always the same colour showing the different at each point.
The most noticeable differences to the 'naked eye' were in the green and green/yellow (grass in this case) areas where values of single colours could be significantly different for the chosen point. This is not a surprise of course - different sensors and different systems in the bodies are very likely to lead to at least some variation of the results presented. I was actually more surprised that in Linear mode, at a glance, they looked so similar given the somewhat different age and size of my Canon sensor! Even the Histograms were quite consistent although one colour or another (mostly green it seemed) could shift along the luminosity scale - which pretty much reflects what happened with the images as far as I could tell. One would expect that.
I am looking at this through 7.1.6 (I don't see anything obvious in 7.2 or 7.2.1 that suggests I should rush to upgrade when 7.1.6 works well for me.) so I have not tried the K3 profile in the same experiment. It occurred to me that you might find something closer to your liking and easier to tweak that final yard using one of the other profiles - likely one that is intended for a camera with the same base sensor but you never know - something totally random may work as well!
As I mentioned earlier part of my interest in this is that I have seen things change 'elsewhere' even for existing and well established profiles when some in application technology has been updated. Not a dramatic update - just an evolutionary change to take in the most recent sensor and RAW file developments. It as if the development of sensors has forced a more complex colour reference data set and so a more challenging interpretation is required that address only specific parts of the data. Oddly in some cases (again 'elsewhere', not within C1) the edited files look fine but the previews look entirely wrong for no obvious reason. I am starting to think that the entire process is becoming even more complicated than we may have thought it was!
Given the way camera sales for the volume market players are going in the hobbyist and DSLR markets I would imagine they would be unwilling to share technical data that covers the complexity without some sort of fee for a Right to Use licence. One might expect this to add to the cost of software development. I assume that at some point we may all need to consider how we would feel about absorbing that cost - and about having accepted in principle an approach that means such costs would be thought of as 'acceptable' for all future developments.
It must be fascinating to be an industry insider.
A few thoughts for what they are worth. I look forward to following how this develops.
Grant.0 -
Hi Grant,
first of all: I am not that worried about the tone curve, or preset, either lineair or film. I always make my own custom tone curves in CO1 anyway and save them as a preset. It is the color gamut description that is essential. Digital cameras are good at "seeing" luminosity, but are quite bad at "seeing" color. Without a proper definition of the way a digital camera sensor sees colors, it becomes a hard task getting either accurate or pleasing images color wise.
The thing is: I would not know how to create a Quality color .icc profile for my camera myself, if I knew I would certainly do that, instead of getting frustrated at the lack of commitment Phase one has shown with regards to proper support for Pentax cameras.
If I lived nearby, I would gladly pay them a visit and drop off my Pentax K3 and my Zeiss Makro planar 50/2.
I played around with the thought of browsing through 24mp Sony sensor cameras and choose a profile to start my own tweaking. But I don't know if you tried: it's hard to get it right for one image, let alone globally.
A good profile makes the difference between great images, and so-so images. My K5 images, and also K5IIs images after tweaking the phase one profile carefully (but that was the same sensor!), developed in CO1 always got compliments about how good they looked colorwise.
If someone on this forum can tell me how to cook up my own quality K3 profile for use in CO1 with the same great results, I'm done, but I doubt that...
Chris0 -
Hi Chris,
I have some small experience seeking to tweak colour response values related to another product that uses different core technology (as far as I know) and is somewhat simpler I suspect. However the process is still instructive about the challenges of making things look vaguely right!
It occurred to me that using the Linear interpretation was about as close as one could get to seeing the interpreted RAW based on what the camera delivered. It might not provide unambiguously accurate colours but the differences visible between instances of the same basic sensor as supplied by different manufacturers might be informative even if it does not point to an immediate solution that suits your needs.
I understand what you are suggesting about how difficult it can be to get a usefully balanced colour interpretation from which to start adapting for different shots. In fact it's anomalies I have seen developing whilst using updated versions of other products that make me wonder if something rather odd may be going on 'under the hood' in the camera systems involved. My simplistic way of looking at it would explain (though not necessarily correctly!) why for some cameras it seems to be difficult to get an over all balance using the other software I mentioned even though addressing individual components seems to be relatively straightforward.
I sometimes wonder if an entirely different approach might be required - though I have no idea what that might be.
Grant0 -
The latest release of CO1 (v7.2.3) includes a new profile (generic v2) for the Pentax K3, which is much much better than the original generic profile (which for the record was a modified K5 generic profile).
Colors are now true to life and accurate again.
The only issue is a noticeable overall color cast towards yellow, which makes many shots in dull light look too "optimistic", as if the light was sunnier than it was, and a bit too colorful. But this is easily corrected with either a preset from the colorbalance tab, or a resaved .icc profile from the advanced color editor, shifting the yellow hue of the colorspectrum towards green a bit.
Such an edit was impossible with the original generic profile: I needed about 20 of them in the advanced color editor, judging by eye with a colorchecker raw file, to get even close to accurate colors, and still the new CO1 profile is superior. Which goes to show how important a good profile is.
Chris0
Post is closed for comments.
Comments
18 comments