Colour Profile Problems Part 2
Hello all
Hopefully Drew will allow this thread to develop.
The 7D2 colour cast problem is blindingly obvious to me. I have the same problem with Cap1 and 7D.
In order to see it, just process and compare on the same screen. It is differences we are looking at, not absolutes.
Cap1 is not alone in this cast problem. I suspect many raw converters have been tweaked in the last year or so to make outputs look 'good'. But not realistic, I suggest.
I am fed up with fiddling around with colours in the yellow-orange-red-magnolia hue range (i.e. autumn leaf colours). I have now gone back to the camera manufacturers' own software for basic raw conversion. DPP and IDC do a good job to my eyes and on my screen. Then I process the basic tif outputs in 'proper' software, including Cap1.
In the past my main criterion for a good raw converter was noise control. Now, with modern cameras and refined software algorithms, noise is less of a problem and I am focused on colour tones. The least offensive ones (IMO) are now DxO-10 and ACDSee-8. Perhaps by next year it will be Cap1 v9?
Always ready to learn!
Peter
Hopefully Drew will allow this thread to develop.
The 7D2 colour cast problem is blindingly obvious to me. I have the same problem with Cap1 and 7D.
In order to see it, just process and compare on the same screen. It is differences we are looking at, not absolutes.
Cap1 is not alone in this cast problem. I suspect many raw converters have been tweaked in the last year or so to make outputs look 'good'. But not realistic, I suggest.
I am fed up with fiddling around with colours in the yellow-orange-red-magnolia hue range (i.e. autumn leaf colours). I have now gone back to the camera manufacturers' own software for basic raw conversion. DPP and IDC do a good job to my eyes and on my screen. Then I process the basic tif outputs in 'proper' software, including Cap1.
In the past my main criterion for a good raw converter was noise control. Now, with modern cameras and refined software algorithms, noise is less of a problem and I am focused on colour tones. The least offensive ones (IMO) are now DxO-10 and ACDSee-8. Perhaps by next year it will be Cap1 v9?
Always ready to learn!
Peter
0
-
Sorry Grant, I've moved on. This Phase of my photography has come to an end.
Who knows, if Phase One does get a grip I can give Cap1 another try.
Peter.0 -
[quote="Peter" wrote:
Sorry Grant, I've moved on. This Phase of my photography has come to an end.
Who knows, if Phase One does get a grip I can give Cap1 another try.
Peter.
Well, I'm somewhat disappointed that you chose to bail out so soon after regenerating a thread and apparently without much of an attempt to engage with the Phase staffers whose ear you had, presumably, gained. Hardly likely to advance the cause and make it worth your effort to follow up later.
I'm not really sure why you felt obliged to be involved here at all looking back.
But hey, good luck to you wherever you have headed. .
Grant0 -
The answer, Grant, is that I am a facilitator. I have ensured that Keith Reeder and you (and others) have had a platform from which you can air valid concerns about the performance of Cap1. Hopefully Phase will now take this colour issue seriously, but I am not holding my breath. I suspect you have direct contacts in Phase. Compared with this influence a one-off support case from me or Keith is neither here nor there. I wish you well.
Peter.0 -
[quote="Peter" wrote:
The answer, Grant, is that I am a facilitator. I have ensured that Keith Reeder and you (and others) have had a platform from which you can air valid concerns about the performance of Cap1. Hopefully Phase will now take this colour issue seriously, but I am not holding my breath. I suspect you have direct contacts in Phase. Compared with this influence a one-off support case from me or Keith is neither here nor there. I wish you well.
Peter.
I can't see your motivation for facilitation on such a specific matter without following through for your own benefit, nor why you feel your involvement has facilitated anything up to now.
If you have direct contacts within Phase that you have influenced that might have qualified as facilitation. Presumably you don't. Nor do I, other than through a Support Case.
As I currently have no evident colour interpretation problems affecting any of my existing kit I am not in a position to help development directly. From your previous posts it seems you could be in a position to help with examples and assessment but despite spending time making that position public you choose not to participate in influencing possible changes. That's not facilitation as I understand it.
I have, in past times, been on the developer side of complex software products supporting users and working with them to create and develop products that would help the success of their employer's businesses. (For clarity - nothing at all to do with photo editing.)
In the days before Internet forums we ran hosted User Groups open for all licenced users and prospective users to attend. Now and again someone would turn up, make a lot of noise, create tension for a while and then disappear having achieved nothing except wasting a lot of time for a lot of people. Such an approach did little for their employer's credibility from anyone's point of view, especially that of the other customers.
Sometimes when reading product related forums I get this extremely strong sense of deja vu.
Good luck wherever you are headed.
Grant0 -
Grant,
Sorry if I am a disappointment to you. But the bottom line is that this is a Phase problem, not yours and not mine.
If pushed further I would have to confess that I no longer have confidence in Phase One. This thread would not exist if someone from Phase had been prepared to admit that there is a problem and that they are actively looking into it, with feedback to us. But this is not the Phase way. I am trying not to judge. It is their business. So perhaps we should just let them get on with it.
Peter0 -
[quote="Peter" wrote:
Grant,
Sorry if I am a disappointment to you. But the bottom line is that this is a Phase problem, not yours and not mine.
If pushed further I would have to confess that I no longer have confidence in Phase One. This thread would not exist if someone from Phase had been prepared to admit that there is a problem and that they are actively looking into it, with feedback to us. But this is not the Phase way. I am trying not to judge. It is their business. So perhaps we should just let them get on with it.
Peter
Peter,
Presumably just a week ago when you started this thread you had a different opinion?
Grant0 -
Grant, you are very tenacious!
Sadly I have had my doubts about Cap1 for some time now. This colour business has brought things to a head. But I wanted to give Phase one more chance. Hence my contributions to this thread. I feel I have done sufficient to ensure that Phase will look again at a systemic problem. Job done. Now it is down to them.
Remember, we are the paying customers, not a testing panel.
Peter.0 -
I would like to repeat that we still would like to see more examples on Your issues. So far only 1 canon user has sent us images regarding this thread.
So, if you feel you have color-issues on your Canon 7d2, please come forward with your problematic files.0 -
Hello Christian
Thank you for your reminder post. You may get something useful in return but you may wish to consider another approach.
Why not post a couple of raw files with potential orange-red issues and invite Canon users to process them to a basic recipe of your choosing (perhaps posted by you), and submit their Cap1 outputs of your image for a controlled comparison? Each user would have to specify what version of Cap1 they have used - hopefully some 7.x as well as 8.x.
I would be happy to contribute to such an exercise.
Peter.0 -
Well, some good news at last.
I removed 8.1 in despair of ever getting decent colours again. Then, just to see what happened, I loaded up v7.1.3 (recommended to me in an old support case). And guess what. My ex-5D2 CR2 files processed just fine, with yellow-orange-red-magenta tones matching those from DPP, and the whole image looking more balanced, colourwise.
It's an easy test. You may wish to try it.
Peter.0 -
[quote="Peter" wrote:
Well, some good news at last.
I removed 8.1 in despair of ever getting decent colours again. Then, just to see what happened, I loaded up v7.1.3 (recommended to me in an old support case). And guess what. My ex-5D2 CR2 files processed just fine, with yellow-orange-red-magenta tones matching those from DPP, and the whole image looking more balanced, colourwise.
It's an easy test. You may wish to try it.
Peter.
+ 1.
I see the same difference as Peter.
Regards,
Willy0 -
I just checked with QPCard yesterday and they said they will have a tool in 2 weeks time to create icc profile for CO8. First available in Mac and then in Windows.
If that is true and the tool works well, I hope this will put an end to the color profile problems.
Just keep my fingers crossed.0 -
Thank you Urukhai
I too have QPCard. They have taken some time to extend it to Cap1. I will be very interested to try it out.
Regards, Peter0 -
Thank you Willy
Your confirmation is reassuring! But there has not been as much interest in this thread as I expected. Are we to assume everyone else is happy with the colour of 8.1 outputs from Canon CR2 files?
I have been 'playing' with my monitor settings - brightness and contrast. The colour tones output using a darker monitor setting look rather 'thin' while those using a lighter monitor setting are more robust and highlights show less clipping. So even these differences in workflow can produce different outcomes. Very interesting!
Regards, Peter.0 -
[quote="Peter" wrote:
Thank you Willy
Your confirmation is reassuring! But there has not been as much interest in this thread as I expected. Are we to assume everyone else is happy with the colour of 8.1 outputs from Canon CR2 files?
I have been 'playing' with my monitor settings - brightness and contrast. The colour tones output using a darker monitor setting look rather 'thin' while those using a lighter monitor setting are more robust and highlights show less clipping. So even these differences in workflow can produce different outcomes. Very interesting!
Regards, Peter.
Hello Peter,
I see a clear difference on my Eizo ColorEdge CG RGB monitor. In files with ICC profile Adobe RGB the yellow is more orange and the greens are more or less green brown. The colour red purple is more red orange. I see it also on my notebook (sRGB) with files in ICC profile sRGB . The deviation is much more in CO 8 than in CO 7. Though CO was always a little yellowish in the greens.
This is not the case in DxO 10 and in Canon DPP 4. These colours are the same. I have all these experiences with the Canon 5D and 5D lll.
Kind regards,
Willy0 -
[quote="NN383" wrote:
I see a clear difference on my Eizo ColorEdge CG RGB monitor. In files with ICC profile Adobe RGB the yellow is more orange...
The shift from pure yellow to orange-like yellow is a very real problem. It also happens with the Pentax K3 profile.
Chris0 -
I find it a bit odd that we have all this talk of colors, but no one is posting real examples, and we only get very few files in the Support?
If You expect to see the the same color-conversion as DPP and ACR or others, you will wait forever.
Capture One have a long standing history for warmer/brown/less red skincolors, from which we have received much praise over the years as well. So, when it comes to skintones, this is our legacy and our default.
That said, we have done v2 profiles as non-default choices on some cameras, when users have been raising concerns, and we might do that here as well. But only if the those users can tell us, in a very concrete, what they want changed.
Buttom-line, if the colors of a camera is a problem, tell us in a support-case with lots of raw examples, side by sides and so on, even your own profiles made in CO.0 -
I submitted a support case on 28th January concerning the profile for the EOS 7D MkII. On 29th January the status of that case was set to "Finished" with a comment: I will make a test case and send it to the image professor for analysis. If we can improve on what we have we will look at including a v2 profile in the future.
Does setting the case to "Finished" mean there will be no further updates to it in the support case journal?
How will I be informed about the outcome of the case? or do I just have to keep waiting to see if a v2 profile appears in the future?
regards,
Richard0 -
[quote="RichardT" wrote:
I submitted a support case on 28th January concerning the profile for the EOS 7D MkII. On 29th January the status of that case was set to "Finished" with a comment:I will make a test case and send it to the image professor for analysis. If we can improve on what we have we will look at including a v2 profile in the future.
Does setting the case to "Finished" mean there will be no further updates to it in the support case journal?
How will I be informed about the outcome of the case? or do I just have to keep waiting to see if a v2 profile appears in the future?
regards,
Richard
It means the Support have finished their part.
If a new profile comes, it will be mentioned in the release notes of that release.0 -
Well if you can't be bothered to inform people of the outcome, it's not surprising that people don't bother to raise support cases is it?
Disappointed,
Richard0 -
It basically comes down to resources, and not about "being bothered"... We would rather help our users up front with their issues, where possible, than spend a lot of resources on keeping up to date with fix-notifications. 0 -
One of the perpetual challenges presented to support systems management is how to communicate successfully with those reporting problems when the corrective activity (if there is one) is medium or long term. Plus, of course, the terminology that is useful for internal status identification is not necessarily helpful for external purposes.
It's a tricky one as I know from personal experience over a few decades both as service user and service supplier. However I would strongly suggest that a "Finished" status with no brief extended explanation attached directly to it (Something like "Finished - Passed to development" or "Finished - User confirmed" or "Finished - Added to wish list", etc.) can appear to be somewhat terse and abrupt. (Yes I appreciate that the quoted case offered an more complete explanation but that does not seem to be the situation in all all cases.)
With a growing number of licencees, many of whom will be likely not be entirely used to or comfortable with what they might see as "industrial" style answers that have not been explained to them and in a style alien to their previous experience, it may be worth considering how meanings might be improved and/or the clientele "educated" in advance.
On the user side we users should probably take more time to understand what the support process is about and consider how the supplier needs to supply it and make some allowances for available word choices - language use is almost always going to be imperfect especially in an international multi-language setting.
My apologies for the digression but it seemed to be a pertinent opportunity to make the observation given the flow of the thread.
Grant0 -
Hello all
I have raised this Support Case. It is my attempt at summarising the main points arising from this thread. I await developments!
Dear Ulf
Please see the Win/CO8 thread "Colour Profile Problems Part2"
I am obliged to your Phase One colleagues for allowing this thread to develop. I am particularly grateful to Christian Gunner for his inputs. Slowly I am seeing the bigger picture. Allow me to summarise where I think we are now.
As I understand it the C1 colour engine has two components - individual icc profiles for supported cameras that are (misleadingly) referred to as Generic Profiles, and software algorithms that create the C1 "look". In addition, by special request, Phase can create a V2 camera profile, but this is a side issue and does not apply to this thread.
The C1 software algorithms have been developed over a few years, always focused on making the most of skin tones for portraiture. This "fine tuning" proceeded without serious comment or complaint until v8. Now, suddenly, some users have detected a significant change from v7 to v8 for the colour of Cannon, Pentax and Sony images. They report a significant shift from yellows to oranges and reds to magenta. Phase is aware of these colour shifts. They are intended to support and develop the C1 "look" in skin tones.
An unintended consequence of this colour shift is that the C1 starting point for landscape images has the same yellow-orange-magenta cast. This cast is very difficult to remove. One knowledgeable contributor refers to the v8 linear mapping algorithm as producing a core/baseline colour image that we can tweak in C1 to give a more acceptable "look". But frankly I and others have better things to do with our time! We want a realistic default starting point!
So, what now for Phase/C1? I suggest you and your colleagues should consider the following options, which are the basis for my Support Case.
1. Restore the v7.1.3 colour algorithm in v8, as a user option.
2. Explain what you have done to change the "core" colour profile.
3. Conduct a controlled test to compare outputs from v7 and v8, and let your users know the results.
I appreciate that Christian has asked for users to submit support cases with lots of raw samples, etc. I'm not sure this will tell you anything - too many variables. Why not post a couple of test raw files and a standard recipe/style, from which a few users can produce an output file for examination at your end? Much simpler, and more controlled, I suggest.
Thank you for reading this. I hope we can now move forward.
Peter Jones.0 -
Of course, the perception of a colour is very subjective, but doesn't it lead to a personal taste and so e.g. of the designers of CO?
If I take images from flowers, the nature, landscape etc. than I want the delicate colours of the real nature so as I perceive it. How? Though colours are subjective I can compare the colours of the output with the reality.
I am afraid that yet CO 8 is no longer the just RAW converter to do this job for me. I use CO a very long time and I am unhappy with the results of today.
I will send a Canon .CR file and two jpeg files, one CO and one Lightroom converted as support case to Phase One to judge it there.
Kind regards,
Willy0 -
[url][/url] [quote="Christian Gruner" wrote:
I find it a bit odd that we have all this talk of colors, but no one is posting real examples, and we only get very few files in the Support?
If You expect to see the the same color-conversion as DPP and ACR or others, you will wait forever.
Capture One have a long standing history for warmer/brown/less red skincolors, from which we have received much praise over the years as well. So, when it comes to skintones, this is our legacy and our default.
That said, we have done v2 profiles as non-default choices on some cameras, when users have been raising concerns, and we might do that here as well. But only if the those users can tell us, in a very concrete, what they want changed.
Buttom-line, if the colors of a camera is a problem, tell us in a support-case with lots of raw examples, side by sides and so on, even your own profiles made in CO.
Hello Christian,
It is not difficult to establish that CO1 .icc profiles for cameras can shift colors. Here is a quick example. Regardless of the white balance (which is different), both images at default settings, and difference in exposure, one single thing jumps out: the yellow patch on the beak of the wild swan shifts to orange. No white balance tweaking will change that, and DxO is right here, at least more right than CO1. I could post 100% crops with tweaked white balance, corrected exposure etc.., but the shift is simply so obvious that there is no need.
I raised a support case to get this Pentax K3v2 profile, and I have not raised a second support case, because I am not expecting a v3, and the profile is overall very good. It is simply this shifting of colors that are by themselves already colors on the edge of transitions, that causes the frustration.
I have suggested quite a few times to add a "neutral" profile for each camera, but either this is too much work, or Phase One does not believe in color accurate profiles. It is not a matter of "look" or "taste" that is the issue here. It is simply colors shifting to a color that does not correspond with reality.
regards
Chris
CO1:
http://www.vogelsenhunleefwereld.nl/WildSwan_CO1v8.jpg
DxO:
http://www.vogelsenhunleefwereld.nl/WildSwan_DxOv10.jpg0 -
Hi Chris
To be honest, I'm not sure what to make of your examples. At first sight there is little to choose between them to my eyes. However if pushed I would say the C1 beak is an orangey-yellow, the DxO beak is greeny-yellow when judged against a bluey background.
We can quantify this by using PS CC to isolate the beak and use the blur+average tools. I get the following results:
C1-8 R147 G100 B17 DxO-10 R141 G110 B25
When I do the same for the bluish swan's head the results are:
C1-8 R119 G137 B144 DxO-10 R130 G142 B146
Concentrating on the red channel, the C1 beak is redder than the DxO beak (yes, consistent with my eyes), but the C1 head is less red than the DxO head. And, in the green channel the DxO beak is greener than the C1 beak (yes, that is what I see).
BUT there is likely to be a perceived colour shift with contiguous blocks of different colours side by side, so the same beak in a photo taken with a green water background may look different. This kind of perceived colour shift is exactly why individual support cases will be difficult to analyse by eye. The CC blur+average route is at least objective even if the results are difficult to square with what the eye thinks it is seeing. Could this be relevant to your findings, I wonder?
I'll be interest to read Phase's response.
Peter.0 -
[quote="Peter" wrote:
Hi Chris
To be honest, I'm not sure what to make of your examples. At first sight there is little to choose between them to my eyes. However if pushed I would say the C1 beak is an orangey-yellow, the DxO beak is greeny-yellow when judged against a bluey background.
We can quantify this by using PS CC to isolate the beak and use the blur+average tools. I get the following results:
C1-8 R147 G100 B17 DxO-10 R141 G110 B25
When I do the same for the bluish swan's head the results are:
C1-8 R119 G137 B144 DxO-10 R130 G142 B146
Concentrating on the red channel, the C1 beak is redder than the DxO beak (yes, consistent with my eyes), but the C1 head is less red than the DxO head. And, in the green channel the DxO beak is greener than the C1 beak (yes, that is what I see).
BUT there is likely to be a perceived colour shift with contiguous blocks of different colours side by side, so the same beak in a photo taken with a green water background may look different. This kind of perceived colour shift is exactly why individual support cases will be difficult to analyse by eye. The CC blur+average route is at least objective even if the results are difficult to square with what the eye thinks it is seeing. Could this be relevant to your findings, I wonder?
I'll be interest to read Phase's response.
Peter.
Hello Peter,
It is not earth shattering, I did state that this Pentax v2 profile is overall quite good. Had this been a serious request for a support case, I would have taken more care to clearly show the color shift. But even with the default quick samples, I believe it is obvious. It is a shift that occurs on certain colors that are in the "danger" zone, and you learn to recognize the shift, even if your eyes adapt, and you don't always know how the colors looked exactly in reality. If it were very obvious, blue turning red ☹️ , the profile would never make it to the release, but it is these subtle color shifts and casts that will be the cause for frustration with some users. The patch on the beak was pre-dominantly yellow, without the noticeable shift to orange in CO1, regardless of the versions of both DxO and CO1. I visit the place often, and when I had noticed the orange shift, I also went back to have a really good look at the wild swans: yellow.
I posted this particular sample, because many times the criticism regarding color in CO1 in the latest versions is regarding the favoring of red/orange/brown. I like the look in CO1, I feel it gives the best final image quality, but a real eye opener was the fabulous profile for the Sony A7r, the second one, called standard. The effect of a profile with that quality, is that you continue to like the look of your images, the more you use CO1. It literally grows on you like all things of high quality. And the basis for that, is that there's literally never an image where you expect a certain color to be off, or wrong. Even if it may eventually not be 100% true, I repeatedly totally believe the colors of the Sony A7r in CO1. What perhaps Phase one should learn from this profile, from how they made it, or whatever lay at the basis of it, is that eventually the best profile is the most reliable profile, not the most pleasing one. A "pleasing" one will become tiresome if it repeatedly mis-reproduces certain colors where it matters: faithful reproduction. A profile that is both reliable as well as good looking (like the Sony A7r standard) never irritates or frustrates.
Chris0 -
Using DPP 4.1.50
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1104_DPP_zpsrrys1lzy.jpg
Using ACR/CS6
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1104_ACR_zpsp7u1te0a.jpg
Using Capture One 8.1.1 with 5D MkII profile
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1104_CO_5D2_zpspbypkpdr.jpg
Using Capture One 8.1.1 with 7D MkII profile
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1104_CO_7D2_zpsfyway41n.jpg
All using default settings and exported to JPEG.0 -
[quote="RichardT" wrote:
All using default settings and exported to JPEG.
Richard,
Could you run a comparison for the different Film Curves for the 7D2 profile please. (I assume you you have left the default set to Auto for the C1 examples and that Auto has elected to work with Film Standard?)
Grant0 -
I've been following this post with interest as unfortunately I've come to a similar conclusion about how C1 displays colours following conversion from RAW files.
I shoot Nikon and with the no more upgrades for their in house NX2 RAW conversion software I had to start looking elsewhere and after trying LR5 I started using C1-7 as a trial last August and I was happy with the results from the limited use it saw before C1-8 was released which I subsequently purchased at the end of November. Since then my frustration levels have quietly increased as I've been spending ever longer trying to get the correct colours with C1 followed by CS6. I had started using various v2 profiles with film or auto but they were getting contrasty so I ended up using the more shadow detail profile which improved things.
I finally realised I needed to review the situation again when white snow wasn't white as it had an orange cast which was very difficult to remove and I had similar results across 3 different top end Nikon bodies including a D810. I reprocessed the same files using the free Nikon NX-D RAW converter which is very slow in comparison but the colours are spot on and they need far less processing !
I tried reloading C1-7 but the trial had expired and I'm certainly not buying it so I cant compare images between the 2 versions.
Whilst I like the C1 interface and features and I'll happily say its the best I've ever used, the colour casts from RAW conversion have pushed me to look elsewhere. As others have already said, can we have a neutral profile ASAP please, like many others all I want is accurate colour !
John0
Post is closed for comments.
Comments
225 comments