Skip to main content

⚠️ Please note that this topic or post has been archived. The information contained here may no longer be accurate or up-to-date. ⚠️

Colour Profile Problems Part 2

Peter Jones
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
0

Comments

225 comments

Date Votes
  • dee jjjaaaa
    [quote="ChrisM" wrote:
    and the basis for a default .icc profile has to be color accuracy if it is to serve as an all purpose profile.


    that's just one of many opinions... a different one is the the default profile shall ensure that regular users like what they get and "color accuracy" is not the case for most users... but a valid suggestion is that P1 might consider introducing "faithful" profiles too for all cameras by default...
    0
  • Peter Jones
    Hi Grant and Christian

    Done it! I have never used "Save target as ..." before. So, I can now see for myself what effect it has.

    Regards, Peter
    0
  • Peter Jones
    Thanks again Grant. New preset installed, new camera icm created. Still learning! (I referred to Niels Blogs on this.)

    For me a new light has been turned on. Your main point is that it doesn't matter if the "official" Cap1 profiles (mostly generic) are a little off. We can adjust a typical image to suit and save the result as an icm file, which can become our new default camera profile and starting point. So the workflow is simplified and outputs are controlled and consistent.

    I have to change my mindset from thinking that the Cap1 profiles are sacrosanct!

    Previously I have made good use of User Styles. Now I need to rethink. There are definitely implications for the orange "glow" of my 7D and (to a lesser extent) my 5D2.

    I think D's post is valid. Cap1 has many options but it is only when we dig deeper that we find the more subtle nuances. Use them or lose them to an ageing memory!

    Well, I've got my spade out of the garage.

    Peter.
    0
  • FirstName LastName
    Christian,

    Thanks for your reply to my query confirming the new profile is generic. I'll download it and give it a try in the next day or so and let you know the out come from a Nikon perspective.

    John
    0
  • Christian Gruner
    [quote="NN635459933720599202UL" wrote:
    Christian,

    Thanks for your reply to my query confirming the new profile is generic. I'll download it and give it a try in the next day or so and let you know the out come from a Nikon perspective.

    John


    Please notice that what I made was not a color-profile, but simply an adjustment using the Advanced Color Editor, and saved as a preset. That is why it (the preset) will work with and apply to any camera CO supports. But adjustments have been made from the pictures we received from the 7d2. As such, the adjustments values are extremely 7d2 specific and will not translate well to other cameras.
    0
  • FirstName LastName
    Christian,

    Thanks, I can see that looking at a picture from a D810. Its gone to flat so I'll have a go at seeing if I can adjust it further over the next day or so.

    John
    0
  • SFA
    [quote="NN635459933720599202UL" wrote:
    Christian,

    Thanks, I can see that looking at a picture from a D810. Its gone to flat so I'll have a go at seeing if I can adjust it further over the next day or so.

    John


    John,

    Having some tiny experience with something similar using another application I would counsel that you are much more likely to find success more rapidly starting with the colour profiles provided for the body or bodies you have and then applying the sort of colour adjustment approach outlined by Christian and described in the Videos.

    Starting with something intended for you camera (even if you are not entirely or at all happy with it) should at least get you closer to your target than working with a starting point that is further away from where you want to be.

    Or, to go a stage further, accept flat as a starting point by using the Linear curve of the camera's profile (Not the Film curves) and develop a look that meets your own preferences. Or even looks - there is no reason why you should be completely constrained. That's what all the Presets and Styles are about.

    The challenge I find is that once I start refining my own "settings" I may never feel entirely comfortable with them. With a "provided" setting one always has the excuse that one has done the best one can with what is provided. When you are in more complete control that excuse evaporates. 🤓



    Grant
    0
  • Urukhai
    [quote="SFA" wrote:

    Nearly everything I shoot is outdoors under varying light (second by second). I cannot imagine that any profile will deal with that for me. I expect discrepancies, anomalies and outright errors from the camera or conversion software.


    The response of a camera to light depends on the illuminant (light source). As long as the light spectrum is the same you don't need another profile.
    There are demo on the web showing that you don't need different profiles for sunny day or cloudy day. But when the light source changes, say from daylight to LED, you probably need to profile again even though the color temperature is the same. In other words, you need different profile for daylight 5000K and artificial 5000K

    In Adobe dcp, you just need a dual-illuminant profile and it works reasonably well across different light sources. If icc profile fails the next second after profiling, then it was an absolute wrong design in CO to use icc rather than dcp profile then.
    0
  • Urukhai
    [quote="Christian Gruner" wrote:

    Take a look at these (made with CO7, but the same goes for CO8):
    - http://blog.phaseone.com/tweak-the-defa ... ur-camera/
    - http://blog.phaseone.com/nail-down-exac ... orrection/
    - http://blog.phaseone.com/move-your-cust ... computers/


    In all these demo, one common observation is that the picture contained only a few colors. So one can tweak the blue without affecting too much on green. In real life, when I change the hue of orange, will it affect also red and yellow?

    I believe colors are interlinked. Just like the fix for 7D2 here. We change blue and green to fix the color of orange.

    As I said in earlier post, I have no doubt that Color Editor is effective to tweak colors to your liking. But I failed to see how changing the sliders up or down without any measurement is considered as color management.
    0
  • ChrisM
    [quote="SFA" wrote:

    Nearly everything I shoot is outdoors under varying light (second by second). I cannot imagine that any profile will deal with that for me. I expect discrepancies, anomalies and outright errors from the camera or conversion software. I can shoot 10fps and get 3 different sets of results from a second of finger on the shutter button. For me Accurate Colour is a myth - but way off colour is a problem. With my cameras (not incredibly new) I hardly ever get important colour management issues, though when they do appear I fully understand how frustrating they can be.

    Grant

    Grant,

    This is one of the most persistent misunderstandings i.m.o. A good profile cán and wíll give you accurate color across a varying range of lighting circumstance. Perhaps a color picker will give you different values for the same color in different lighting, or when changing the white balance, but that is not the culprit. What matters is that the colors are faithful in their context, and that is what a good profile will give you. When changing white balance or lighting source all the colors shift but the overall context remains intact. And a simple tweak of the white balance will restore the colors.

    The problem with an inaccurate profile is that it upsets the balance of colors through shifting individual colors in hue and lightness, and when you then try to correct a color by targeting it, you inadvertently change neighbouring colors. That's just fine for creative purposes, but it doesn't cut it when trying to get a faithful default profile.

    Chris
    0
  • Peter Jones
    Hello Chris

    I think you've got it right. So what, in practice, have we learned?

    Question: What have Cap1 and quantum mechanics got in common?

    Answer1: Just when you thought you understood something you realise you haven't asked the right question - and probably never will.

    Answer2: Your comfortable chain of thought leads you back to answer1.

    So let's try again. As regards Cap1:

    Generic color profiles with the same number of bytes are identical. Correct?
    The same generic profile is often used for a few cameras. Correct?
    The greater the number of such cameras, the more likely the profile will be "off"?
    If a "generic" profile size is unique to a camera it is specific for that camera?
    V2 profiles are specific for a specific camera. Correct?

    The famed Cap1 "look" is not coded in Generic profiles. It is in the algorithms of Cap1. Correct?

    Users can tweak colour settings to change the "look" but they should be aware that this degrades the intention of the generic profile that Phase recommends. Particularly if the starting point profile is for a different camera. Correct?

    In "Base Characteristics" the Film Standard option is the doorway to the Cap1 "look". The linear curve option removes the "look". The "No Colour Correction" option is a "faithful" rendition. All correct?

    So, when tweaking the colours in an image there are two basic starting points. One is to boost the colours in a linear/no colour image. The other is to reduce the colour tones in a standard image. Which route does Phase recommend?

    But the simplest way of all is for Phase to produce a specific, "natural look" profile for each camera. I note that ACR "standard" profile sizes are unique to each camera. I can't find the DPP4 profiles.

    Answers on a postcard please. Thank you.

    Peter.
    0
  • RichardT
    [quote="Peter" wrote:
    Generic color profiles with the same number of bytes are identical. Correct?
    The same generic profile is often used for a few cameras. Correct?
    The greater the number of such cameras, the more likely the profile will be "off"?
    If a "generic" profile size is unique to a camera it is specific for that camera?
    V2 profiles are specific for a specific camera. Correct?


    Not always. If you compare the files using a program like WinMerge:

    CanonEOS5DMk3-Generic.icm and CanonEOS6D-Generic.icm are identical.

    But CanonEOS7DMk2-Generic.icm has very different contents, even though it is the same size (217,584 bytes).
    0
  • marcoi
    Just a little OT: reading this thread is very interesting for anyone, a lot of things to learn from everybody! really really thanks!
    I don't have a Canon but a Fuji X-E2 and I found that the Generic V2 version (the default one) has the same orange casting as you found in your Canon images. Most visible on skin tones (almost useless portraits). Instead the Generic V1 profile works very well (especially with standard or extra shadows film types). Just to add my experience
    Sorry if too OT 😊
    0
  • SFA
    Chris,

    I think we might be revolving on a terminology issue here - but we might need Christian to comment in order to be sure.

    The 'Look' we have been discussing is the result, so far as I can tell, of a base response "ICC Profile" for the camera output modified by the Film Curve selected. That's not entirely the same as the Curve tool in the application in that it pre-defines a set of values that the Levels and Curve tools then work with.

    The results of the Base Characteristics ICC, without any further notable contrast, etc., tweaks, would be the Linear "Film" option.

    The rest of the "Film" Curves apply further adjustments to create the Look as a starting point. The Look that results is the thing we have been discussing here.

    If the core ICC profile has a problem this should be discernible using the Linear "curve" - which I would assume is basically no curve at all other than ensuring linearity. (The Help information describes it as Reduced Contrast). In this I am thinking more about having a good relationship between colour value interpretations rather than assuming the a colour value matches a specific target number. If the relationships are good then one should be able to shift the numbers coherently up or down as required for creative output.

    Once we have the selected starting point, using or not using a provided film curve, the rest of the editing activity is about tweaking things to our liking. The challenges come when, for what ever reason, we can't do that easily or reliably to our satisfaction irrespective of whether we seek colour accuracy.

    Certain colours can be more challenging than others as we all know. I recall a dayglo orange (years ago before Capture One days) that on screen would flip from deep yellow to bright crimson on the touch of a colour adjustment slider for the main Curve. Very frustrating but nothing I could do to get what I wanted with that one even accepting compromises on colours elsewhere in the image. Working with the same problem today I might seek a different approach and would certainly have additional tools available to try out.

    If we don't like the Film Looks provided there is the option of starting with the Linear option and creating our own preferred Look.

    Providing I am right about the balance in the ICC file things should work out well with minimal effort.

    The challenge for the "acceptable" outcome will always be partly a question of how the scene is lit compared to how we want it to look.

    In an outdoor shot under a clear blue sky in an open area you might assume some form of consistency for colour temperature (except in the shadows). That provides a starting point for all colour adjustments - toning the entire image for example. But in mixed light - part sun part clouds or mixed lighting sources where artificial light is involved, you don't have a single coherent base point. If you want to avoid a lot of local editing and manipulation you accept a compromise so long as no individual part of the image stands out as being a long way beyond what you are comfortable with. A good starting balance should make this easier than a bad starting balance - though I dare say there unexpected lucky exceptions to that rule. Obviously I am mainly thinking in terms of White Balance here.

    White balance settings can do remarkable things where light output is reasonably consistent over a the spectrum to which the sensor is responsive. However get an odd source - Dayglo orange in sunlight, certain types of Mercury based flood lighting, etc., and the consistence goes right out of the window (so to speak). Very specific and sharp spikes in values can be almost uncontrollable by normal means.

    Ultimately if one is not comfortable with the proffered ICC profile and generic Linear curve offering the option would be to create one's own ICC curve based on whatever starting point one chooses. It's a facility that seems to be expected by many in the industry so I suppose that to them the Defaults offered may not be of great significance so long as the profile is somewhere close and gives them something to work with initially.

    My thoughts on what is so often a rather complex and emotive subject influenced by many different conflicting priorities.


    Grant
    0
  • Christian Gruner
    Peter, I've written my answers in bold.

    As for a status on a new profile for 7d2 we are still discussing details internally. In the mean time, please use the presets I provided earlier.

    [quote="Peter" wrote:
    Hello Chris

    I think you've got it right. So what, in practice, have we learned?

    Question: What have Cap1 and quantum mechanics got in common?

    Answer1: Just when you thought you understood something you realise you haven't asked the right question - and probably never will.

    Answer2: Your comfortable chain of thought leads you back to answer1.

    So let's try again. As regards Cap1:

    Generic color profiles with the same number of bytes are identical. Correct?
    The same generic profile is often used for a few cameras. Correct?
    The greater the number of such cameras, the more likely the profile will be "off"?
    If a "generic" profile size is unique to a camera it is specific for that camera?
    V2 profiles are specific for a specific camera. Correct?
    There can be certain cameras that react very much like each other, and some even use the same circuitry. In those rare cases, it makes sense to use the same profile, maybe with minor tweaks if required, to ensure the same look. V2 profiles are usually based on user-input.

    The famed Cap1 "look" is not coded in Generic profiles. It is in the algorithms of Cap1. Correct?
    They are in the Color Profiles and the Filmcurves.

    Users can tweak colour settings to change the "look" but they should be aware that this degrades the intention of the generic profile that Phase recommends. Particularly if the starting point profile is for a different camera. Correct?
    We have the generic profile defaulted, to have a good starting-point for the average user. You as a photographer and artist have the creative freedom with the tools provided.

    In "Base Characteristics" the Film Standard option is the doorway to the Cap1 "look". The linear curve option removes the "look". The "No Colour Correction" option is a "faithful" rendition. All correct?
    The Linear presents the raw in a linear state with hard fall-offs in the highlight especially visible. In the other curves this has been tweaked, so the fall-off is a a lot smoother.
    The No Color Correction choice is just what it say, no input color correction what so ever.

    So, when tweaking the colours in an image there are two basic starting points. One is to boost the colours in a linear/no colour image. The other is to reduce the colour tones in a standard image. Which route does Phase recommend?
    The one you as a user feel you can get the best and fastest results from. They are different by design, so the user have options.

    But the simplest way of all is for Phase to produce a specific, "natural look" profile for each camera. I note that ACR "standard" profile sizes are unique to each camera. I can't find the DPP4 profiles.
    The best profile is the one that get you the quickest from input to output. And that is what we hear from the vast majority of our customers. Yes, if skilled enough, they could get somewhat the same colors, tone and mood in other software. They can just get there a lot faster when using CO's starting-point and adjustment-tools.
    Please notice that I said "the best" and not "the most precise". A mathematically correct profile would provide some unflattering imagery, which is not what the vast majority of our users want.
    You are most welcome to write our Support and file a Feature Request for the mathematically correct profiles. They will forward them to the Product Manager. This way your voice is heard, registered and channeled to right people.

    As for a status on a new profile for 7d2 we are still discussing details internally. In the mean time, please use the presets I provided earlier.


    Answers on a postcard please. Thank you.

    Peter.
    0
  • Christian Gruner
    [quote="Urukhai" wrote:
    [quote="Christian Gruner" wrote:

    Take a look at these (made with CO7, but the same goes for CO8):
    - http://blog.phaseone.com/tweak-the-defa ... ur-camera/
    - http://blog.phaseone.com/nail-down-exac ... orrection/
    - http://blog.phaseone.com/move-your-cust ... computers/


    In all these demo, one common observation is that the picture contained only a few colors. So one can tweak the blue without affecting too much on green. In real life, when I change the hue of orange, will it affect also red and yellow?

    I believe colors are interlinked. Just like the fix for 7D2 here. We change blue and green to fix the color of orange.

    As I said in earlier post, I have no doubt that Color Editor is effective to tweak colors to your liking. But I failed to see how changing the sliders up or down without any measurement is considered as color management.


    I selected the green and yellows (as well as the reds) in my preset as they also seemed to be a bit too saturated. They were not selected to affect Orange, and the green does not affect the orange. All this was using my eyes comparing. In this case, there were no reason complicate things by starting to measure values.
    Blue was not at all touched.
    I would suggest you try the tool in practice. You can make some extremely color-local modifications.
    0
  • Peter Jones
    Christian

    Thank you for your fulsome response. Much appreciated food for thought!

    Have a good day.

    Peter.
    0
  • ChrisM
    In reply to the overall responses that my plea for more color accurate default profiles raises:
    I specifically brought up the Sony A7r "standard" profile in the latest v8, to illustrate my conviction that a good profile is as much pleasing as it is accurate. The human eye responds very well to accurate color reproduction because it is in line with what we see around us in the real world day by day. In many cases, photograpehers like to (heavily) post-process their images, because they want to add a unique and specific look to their images, or want their photographs to look more colorful or otherwise pleasing than the real-world scene. Many times a creative freedom is more important in this regard than real-world accuracy.

    I find that this approach to photography heavily taints the whole discussion about what the default .icc profiles in CO1 should offer. But color accuracy is not an expression of a drive to achieve mathemetical precision. Let's not forget that there are forms of photography that make use of the "pleasing" factor already present in the real world scene: nature or landscape photography especially. An accurate rendering of images can make them look immensely pleasing if the scene captured was pleasing in itself.
    However: adding a pleasing look in the belief that accurate color reproduction will most likely not be pleasing, can have very much the adverse effect, and this whole discussion is very much the proof of that. People are asking back for their images rendered without the pleasing look.

    My take on it is all in the Sony A7r profile: faithful color reproduction across a large variety of images. Immensely pleasing at the same time.

    Chris
    0
  • SFA
    Chris,

    As part of my investigations around this thread and related "how to deal with the internet masses" reading I stumbled across a post on what seems to be an official Adobe support forum that pointed out (it has been "promoted" to a top of the thread view so I assume it is correct) that all RAW convertor vendors have to buy cameras and reverse engineer what they find in order to get their results.

    Now, it's a 4 year old post and things may have changed and I guess Silky Pix have some inside knowledge as part of their bundling. The only bundled software I can recall (other than DPP from Canon) was getting PS Elements 2 with a Pentax Optio about a decade ago. Now RAW of course, only jpgs. So no big deal.

    My guess - and it is absolutely a guess - is that having a deal with camera manufacturer allows access to inside information (indeed it may be a requirement to make use of it) that bypasses the "reverse engineering with permission" process and in effect delivers exactly what the manufacturer wants to be delivered as THEIR USP for image standards.

    Which, if the manufacturer has done a good job with respect to their analysis of the marketplace and its users buying their products, should most likely match >90% of their customer's reported preferences. Hence you like the results.

    As a Canon user I appreciate the results that DPP offers. I don't use it, however, because I found it slow and clunky years ago and there is no way I would feel comfortable working with it trying to process a few thousand shots after a weekend event. But as a tool for producing excellent colour accuracy for Canon files it would be immensely disappointing if found wanting.

    If (or maybe when) I am not shooting large numbers of shots, I may reconsider what DPP offers .... or maybe not since C1 seems perfectly fine anyway!

    Perhaps I should switch to Sony?


    Grant
    0
  • ChrisM
    [quote="SFA" wrote:
    Chris,

    As part of my investigations around this thread and related "how to deal with the internet masses" reading I stumbled across a post on what seems to be an official Adobe support forum that pointed out (it has been "promoted" to a top of the thread view so I assume it is correct) that all RAW convertor vendors have to buy cameras and reverse engineer what they find in order to get their results.

    Now, it's a 4 year old post and things may have changed and I guess Silky Pix have some inside knowledge as part of their bundling. The only bundled software I can recall (other than DPP from Canon) was getting PS Elements 2 with a Pentax Optio about a decade ago. Now RAW of course, only jpgs. So no big deal.

    My guess - and it is absolutely a guess - is that having a deal with camera manufacturer allows access to inside information (indeed it may be a requirement to make use of it) that bypasses the "reverse engineering with permission" process and in effect delivers exactly what the manufacturer wants to be delivered as THEIR USP for image standards.

    Which, if the manufacturer has done a good job with respect to their analysis of the marketplace and its users buying their products, should most likely match >90% of their customer's reported preferences. Hence you like the results.

    As a Canon user I appreciate the results that DPP offers. I don't use it, however, because I found it slow and clunky years ago and there is no way I would feel comfortable working with it trying to process a few thousand shots after a weekend event. But as a tool for producing excellent colour accuracy for Canon files it would be immensely disappointing if found wanting.

    If (or maybe when) I am not shooting large numbers of shots, I may reconsider what DPP offers .... or maybe not since C1 seems perfectly fine anyway!

    Perhaps I should switch to Sony?


    Grant

    Hello Grant,

    I'm sure that the exchange of inside technical knowledge is a great benefit regardless. But this does not necessarily lead to great results. Pentax cameras e.g. are bundled with a version of Silkypix, but in fact I find the results in CO1 with particularly the Pentax K5IIs (v2 generic profile) superior to the results that come out of the bundled software color-wise.
    I am absolutely convinced that Phase one can produce a very good profile for any camera. Looking at profiles like the Pentax K5IIs v2 generic, there is little left to be desired, even though the Sony A7r standard is to my eyes the best I've tried, and a step up from even the already very good K5IIs profile.
    I don't believe that it is the quality of CO1 that is the decisive factor, or Phase one's lack of resources or inside technical knowledge. After all they sell some very high end MF cameras.
    I.m.o. it is the profiling protocol that somehow neglects or underestimates the necessity of incorporating reliable color reproduction in their generic (default) camera profiles. The conviction that a pleasing look will gain more approval amongst the majority of CO1 users has the upper hand. The fact that some of us (totally) disagree and even find that the bias towards a perceived "pleasing" look can degrade the image quality coming out of CO1 with certain cameras, is simply not enough in itself to change this situation. The change will only come from well managed support cases, that is the lifeline to better IQ for your particular camera that Phase One offers, and it is a viable one if you're willing to put in the effort and have the patience to wait for future v2 profiles.

    Chris
    0
  • Peter Jones
    Hello Urukhai

    In your posts you referred to the QPCard camera calibration process and their 203 "book".

    I thought damn it, let's find my 203 and give it a go! The software has been updated and (in my hands) a little temperamental, but it still works with my 7D.

    I now have a few .icc profiles that I have copied to the Cap1 "generic profiles" folder, and a few .dcp profiles for ACR. For Cap1 the QP-processed images are much lighter and less saturated than Cap1 "Film Standard". This is easily rectified in situ and the orange/red issue has gone. In ACR the difference is less marked but again the ex-QP orange/red is more real-looking to my eyes.

    Dare we think that the QP-generated profiles are as good as it gets for a starting point? (I don't have the X-rite system to compare.)

    Also, I stumbled across an article by Thom Hogan where he contrasts the concept of a "pleasing look" profile (a moveable feast) and an accurate profile.



    Worth a read-through perhaps.

    Peter
    0
  • Urukhai
    [quote="Peter" wrote:
    I thought damn it, let's find my 203 and give it a go! The software has been updated and (in my hands) a little temperamental, but it still works with my 7D.


    Hi Peter, which version of software you were using? The current version on their website, and stated clearly by them, does not work well with C1 yet. According to some old info, one need to generate a 16 bit TIFF from C1 without any color correction before profiling The currrent QPCalibration does not take TIFF. A new version of QPCal is supposed to come at any time that can take TIFF generated from C1.
    0
  • Peter Jones
    Hi Urukhai

    Thanks for your response. I am using QPCard v1.99a. I ignore the official QP comment on CaptureOne. The QP .icc profiles do work with Cap1. And the effect is miraculous!

    Whereas Cap1, ACR and all the others seem to delight in a Disney-style Technicolour palette the QPCard profile processed in Cap1 produces landscapes with colour tones reminiscent of en plein air paintings. In other words more true to life as we see it (not how we think we see it).

    To be sure of this I have re-checked my eyes for any colour problems (I had a stroke 3 years ago), and all is well.

    We are now drifting away from the core topic of this thread. I am thinking of opening a new thread in the Workflow folder.

    Cheers for now.

    Peter.
    0
  • Dave R
    [quote="Christian Gruner" wrote:
    [quote="NN635459933720599202UL" wrote:
    Christian,

    Thanks for your reply to my query confirming the new profile is generic. I'll download it and give it a try in the next day or so and let you know the out come from a Nikon perspective.

    John


    Please notice that what I made was not a color-profile, but simply an adjustment using the Advanced Color Editor, and saved as a preset. That is why it (the preset) will work with and apply to any camera CO supports. But adjustments have been made from the pictures we received from the 7d2. As such, the adjustments values are extremely 7d2 specific and will not translate well to other cameras.


    Actually it works very well for Nikon D700 NEFs, I have a picture with a lot of red brickwork open in front of me now in both Capture One and Lightroom. Without the preset the brick in the Capture One rendering shows all the problems described in this thread, with the pre-set the Capture One rendition and the Lightroom rendition (using Adobe Standard Profile) are very similar.
    A thought though there is a choice of 17 different profiles for my Camera in Lightroom some of which give far less believable brick tones than the basic Capture One Profile. 🤓
    0
  • Keith Reeder
    [quote="Urukhai" wrote:
    In all these demo, one common observation is that the picture contained only a few colors. So one can tweak the blue without affecting too much on green. In real life, when I change the hue of orange, will it affect also red and yellow?

    I believe colors are interlinked. Just like the fix for 7D2 here. We change blue and green to fix the color of orange.

    As I said in earlier post, I have no doubt that Color Editor is effective to tweak colors to your liking. But I failed to see how changing the sliders up or down without any measurement is considered as color management.

    This.

    Personally I accept unreservedly that I can create a preset which tweaks the base profile; but this is the epitome of putting a band-aid over a bullet hole in that it can only ever really address colour problems in images which are spectrally similar to the image that the preset was based on.

    We have camera profiles specifically to provide the most "usable out of the box" image in the majority of cases, and putting the onus on users to correct profiles that fail to do that, is not what I would hope to hear as a "solution" from Phase One.

    Still, to take a "positive" from this, at least Phase One is now actively acknowledging the problem with the 7D Mk II profile.

    Note that it exists with the 70D too, to my personal knowledge.

    [quote="ChrisM" wrote:
    This is one of the most persistent misunderstandings i.m.o. A good profile cán and wíll give you accurate color across a varying range of lighting circumstance. Perhaps a color picker will give you different values for the same color in different lighting, or when changing the white balance, but that is not the culprit. What matters is that the colors are faithful in their context, and that is what a good profile will give you. When changing white balance or lighting source all the colors shift but the overall context remains intact. And a simple tweak of the white balance will restore the colors.

    And very much this. No further comment necessary, this says all that needs to be said about "why profiles?", and (by inference) also explains why a tweaked preset isn't a fix.


    [quote="marcoi" wrote:
    Just a little OT: reading this thread is very interesting for anyone, a lot of things to learn from everybody! really really thanks!
    I don't have a Canon but a Fuji X-E2 and I found that the Generic V2 version (the default one) has the same orange casting as you found in your Canon images. Most visible on skin tones (almost useless portraits). Instead the Generic V1 profile works very well (especially with standard or extra shadows film types). Just to add my experience

    Very helpful - and further evidence, I would suggest, that this is a pervasive (dare I say "systemic" again?) rather than a localised issue; further evidence too (assuming that v1 is older than v2) that the newer the profile, the more likely it is to make things look like the inside of a Jaffa Cake.
    0
  • Peter Jones
    Hi Keith

    Goodness knows how many hours I have spent supporting this thread. But I don't regret a minute of it. Because I now feel competent and knowledgeable enough to challenge what I see as a skewed conventional wisdom of what constitutes a good digital camera colour. It is unfortunate that all this has come to a head with Phase Cap1, because I do not believe it is just a Phase One problem. I acknowledge the way Christian Gruner has helped me to understand some of the dark arts of raw conversion!

    The photo industry has become obsessed with "in your face" raw colour conversions. From camera suppliers to software to printers, and to magazines and camera clubs. Witness the HDR fad. Never mind reality, let's make it pop. So with this sweeping background view in mind I feel I can make the following statements based on my experience with a Canon 7D and 5D2 (and, indirectly, the 7D2).

    The "off the peg" generic profiles supplied by Canon are inaccurate/wrong. Too much red. Too little green.

    The Phase drive for a "pleasing look" exacerbates this. Even the "No Color Correction" option is too much (for me).

    The DPP4 outputs are better, but still not true to life.

    The only camera profile I now find acceptable is from the QPCard system. Perhaps the X-Rite system is OK, but I don't have access to it.

    QPCard profiles allow easy image adjustments with the Phase "Exposure" tool, giving very pleasing results. In contrast Phase profiles are as if on a knife-edge - a small tweak here or there completely changes the "look". Phase is more vulnerable to this sensitivity than ACR.

    So, in summary I like the way the Phase Cap1 software handles colour but their raw camera colour presets do not do justice to the original raw file.

    An interesting spin-off from all this colour stuff is that I find the QP profiles provide colour images that produce superior black and white conversions - because the colour palette is tonally more accurate, I suspect.

    Yes Keith, I have no doubt that we have a systemic problem, but it is probably industry-wide (rather than just Phase One) driven by the preferences of (the majority?) of users. Would we be in this position if Technicolor and HD colour television had not been invented, I wonder?

    Regards, Peter.
    0
  • Peter Jones
    Hi there

    In response to a Support Case about my QPCard profile for the 7D I have been advised that while it is clearly functional it does not meet Phase strict quality requirements.

    Here is the Dropbox link to the file:

    https://www.dropbox.com/s/us7kpqks9wx8d ... t.icc?dl=0.

    I have not used Dropbox before (thank you Phase for another computer lesson!) but I assume it does what it says on the tin.

    Download from this link and copy it to:

    Program Files/Phase One/Capture One 8/Color Profiles/DSLR

    Give it a try. Let me know what you think. And as a matter of interest, try it with the 7D2. If Phase can live with generic profiles why not us!

    Thank you for your interest.

    Peter.
    0
  • AiDon
    Hi Peter,

    Thanks for posting the profile but maybe I'm not seeing something here as I tried it on a 7D image and as you can from the posted image with the 7D Generic/High Contrast on the top and the QPCard profile on the bottom there is a vast difference:

    https://plus.google.com/photos/11602543 ... 1741571288
    0
  • Peter Jones
    Hi AiDon

    Yes, for your image and for your settings this looks about right. The top one, with Cap1 generic and high contrast default settings comes from a colour profile that is "tuned" for Cap1 - so the default output is as-designed by Cap1. The bottom image looks just as it does for me for a Cap1-default with QPCard profile. The trick is to add your own processing in Cap1, as described here.

    The QPCard .icc profile is not "tuned" for any specific raw converter (although the QPCard .dcp profile is "tuned" for ACR). So, in Cap1, using the standard contrast setting and the QP profile, I use the Cap1 "Exposure" tool to reduce exposure and brightness, and increase the settings for contrast and saturation. When processed in this way (no colour adjustments note) my photos of forestry, autumn leaves and blue sky look just right (to me) whereas the generic Cap1 settings produce a garish, contrasty scene with no tonal balance. Think Mr Turner!

    Perhaps you can try a different image, perhaps one where the background is in focus so you can see more clearly the QPCard difference with subtle hues?

    Thanks for your interest.

    Peter.
    0
  • AiDon
    Hi Peter,

    Thank you for the explanation. I can see now what you mean about the over saturation of reds etc. in the default profiles as your profile is more exact for blues/greens although it is not suitable for skin tones. You can see the standard profile on the left and the QPCard (with rough adjustments on the right) and it does give more pleasing tones although the brickwork on the path probably has too much magenta in it ... but saturation would fix that.

    https://plus.google.com/photos/11602543 ... 1741571288

    Thank you, I appreciate the input.

    Regards, Don
    0

Post is closed for comments.