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
-
Super, thanks. 0 -
Hi Richard.
Your most recent images: The DPP version is the brightest overall. The tree trunk is very similar in all three images (a quasi grey card?). The sunlit fence and the edge of the bin: compared with DPP the green and blue channels for 5D2 and 7D2 are low, with 7D2 the lowest. When I 'normalise' the data so that all the red readings are the same as for DPP, the 5D2 and 7D2 green and blue are still low (DPP bin 85 47, 5D2 bin 78 52, 7D2 bin 69 39) (DPP fence 130 112, 5D2 fence 125 108, 7D2 fence 118 96). These results suggest to me that Phase has the grey card readings about 'right' and the main issue is with brown colours where the green and blue channels are weaker than DPP and weaker than what is required for a 'pleasing' outcome.
Did you know that you can change the colour engine in 8.1 to a v7 version? (Prefs, image, default engine) Could you try this please with one of your 7D2 images?
Thanks for your efforts!
Peter0 -
Hi Christian
Thanks for your patience. It occurs to me that there may be very few 7D2 owners who also use Cap1 as their primary raw converter. So why not take up my suggestion to post a few 7D2 raw images so more people can contribute? This is the third time I have suggested this!
Thank you.
Peter.0 -
[quote="Peter" wrote:
Hi Christian
Thanks for your patience. It occurs to me that there may be very few 7D2 owners who also use Cap1 as their primary raw converter. So why not take up my suggestion to post a few 7D2 raw images so more people can contribute? This is the third time I have suggested this!
Thank you.
Peter.
I'm not sure how this would help? Can you explain some more?0 -
[quote="Peter" wrote:
Hi Richard.
Your most recent images: The DPP version is the brightest overall. The tree trunk is very similar in all three images (a quasi grey card?). The sunlit fence and the edge of the bin: compared with DPP the green and blue channels for 5D2 and 7D2 are low, with 7D2 the lowest. When I 'normalise' the data so that all the red readings are the same as for DPP, the 5D2 and 7D2 green and blue are still low (DPP bin 85 47, 5D2 bin 78 52, 7D2 bin 69 39) (DPP fence 130 112, 5D2 fence 125 108, 7D2 fence 118 96). These results suggest to me that Phase has the grey card readings about 'right' and the main issue is with brown colours where the green and blue channels are weaker than DPP and weaker than what is required for a 'pleasing' outcome.
Did you know that you can change the colour engine in 8.1 to a v7 version? (Prefs, image, default engine) Could you try this please with one of your 7D2 images?
Thanks for your efforts!
Peter
Hi Peter,
I changed the default colour engine to v7 and imported the garden image into a new catalog. The colours in the generated JPEG look the same to me:
7D2 Using CO v8 engine + 7D2 profile
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1682_CO_7D2_zpslbugimrt.jpg
7D2 Using CO v7 engine + 7D2 profile
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1682_CO7_7D2_zpszpebvny9.jpg0 -
Hi Christian
I have spent many years getting the most out of technical instruments. In my experience the way to get robust results is to control as many variables as possible. At the moment you are asking people to submit images of all kinds of processed 7D2 files - with all sorts of computer hardware and software settings. As I understand it Phase wants to 'eyeball' them - a highly subjective process, I suggest! And subject to visual colour shifts with adjacent blocks of different colours.
I am suggesting a way forward that gives Phase some quality control, and which enables non-owners of 7D2 to contribute. In essence, Phase selects the raws, posts them in this thread and asks us to process them with a set recipe/style and send the jpegs to you for controlled technical analysis.
I have posted my analysis of Richard's images in this thread. I use PS CC with the filter-blur-average tool and read off the RGB settings. Very simple. And very robust. I am sure Phase must have a similar tool.
I hope this helps.
Peter.0 -
Hello Richard and Christian
Thanks Richard. A very interesting result. It confirms my own tests. The v7 and v8 colour engines give the same result with 7D2 files. Please can you pass on this information Christian. It was Ulf who suggested using v7. It does not do what I was expecting.
So, we seem to be left with the Phase 7D2 icc profile as the culprit?
A simple question arises here. Does Phase tweak camera icc profiles from one version to another? So when I process a 5D2 image in v7.1.3 it uses a different profile from that in v8.1? If not, if the icc profiles are meant to be the same, can I exchange them in v7 and v8 to see what happens? Has Phase tried this?
Still there are many questions and unknowns for we users to understand what may be going on. I am still hopeful that Phase will enlighten us in due course.
Peter.0 -
Peter,
V7 would not have had a 7D2 profile or profile related film curve.
That suggests that the V7 processing engine is able to make use of the V8 profile and curves when installed.
Given some internal processing differences between the engines and a few additional features (extended Clarity for example) presumably the base conversion is possible but some other tweaks may be missed if doing a comparison on edited files.
However if we are only looking at default conversion options the specific editing tool changes may not be an influence.
On the other hand if there is more to it than that it would be good to get an understanding of what that might be and how to work with it.
Grant0 -
Regarding the repeated claim that colors have changed between v7 and v8: that is not my experience. However, the preview rendition has changed: the preview has stronger colors in CO1 v8 than it did in v7. In v7 the preview was an exact match with the ever reliable photoshop viewer. In v8, I more or less have to ignore the notion yhat colors are a bit too strong, knowing that it won't be the case when I open the exported image in Photoshop. This does not take away from the fact that the 7dII profile seems overly reddish/orange going by the posted samples, but it may feed the notion that something changed for the worse with v8. If it did, than it is about the preview not being as accurate and neutral anymore but suffering from a touch of over saturation.
What is behind this, is a mystery to me, color management is set correctly and nothing has changed in hardware or otherwise. It seems inherent to v8. I greatly prefer a Photoshop like 100% accurate preview, and v7 was very close to that.
Chris0 -
For a sample image, should one take that path, maybe Imaging Resource could be persuaded to allow this one to be used to get an impression of how people see things. (I would suggest sharing off line if they would agree.)
http://www.imaging-resource.com/PRODS/c ... 4A0967.HTM
It has a lot of factors including high ISO, what looks like incandescent or maybe Sodium lighting to make the orange highlights. Skin tone preferences can be covered (albeit in less than helpful light) and it stretches dynamic range. It's very interesting to experiment with it although it may polarise people's opinions - but given the topic under discussion I would see that as more helpful to understand what is generally "acceptable" rather than being disruptive.
I think a much more typical image would be helpful too though preferably one with known colour reference points.
Grant0 -
Hi Grant
Thanks for your thoughts. I think you may well be right. In 8.1 the option to use a v7 colour engine is probably overruled for the 7D2. On the other hand when I tried a 5D2 image in the 8.1/v7 engine there was no difference as I recall. I have taken 8.1 off now and gone back to good-old 7.1.3. Because, even for my 7D and 5D2 the spectre of 8.1 brown colour shifts hangs over me! (The best software to correct this is ACDsee in my experience, but that is another story.)
By the way, do you know if a paid-up 8.1 user can load 7.1.3 (for example) without paying for another license? I ask not for me but for another user.
As you say, "On the other hand if there is more to it than that it would be good to get an understanding of what that might be and how to work with it". Quite! Information is power!
Hopefully there will be some light at the end of this tunnel for 7D2 users.
Peter.0 -
Hi,
while I've been sharing at start the evaluation that the relevance of the discussed colour profile problems is low given the low number of complaints, I'm changing this opinion because of Richard's set of samples taken in the backyard. For me the brown trashbin (?) shows very clearly the shift in colour rendition and saturation which is happening in the 7DII/7DII version in comparison to the other edits. Perhaps a differnt approach towards a solution would be to work on a tweaking on the conversion settings of this specific file (via individual RGB levels separately in the shadows, middle and highlight levels) to match the look of the other variants. With the found numeric adjustments it would be easier to qualify/quantify the amount of present offset of the profile.0 -
[quote="Peter" wrote:
Hi Christian
I have spent many years getting the most out of technical instruments. In my experience the way to get robust results is to control as many variables as possible. At the moment you are asking people to submit images of all kinds of processed 7D2 files - with all sorts of computer hardware and software settings. As I understand it Phase wants to 'eyeball' them - a highly subjective process, I suggest! And subject to visual colour shifts with adjacent blocks of different colours.
I am suggesting a way forward that gives Phase some quality control, and which enables non-owners of 7D2 to contribute. In essence, Phase selects the raws, posts them in this thread and asks us to process them with a set recipe/style and send the jpegs to you for controlled technical analysis.
I have posted my analysis of Richard's images in this thread. I use PS CC with the filter-blur-average tool and read off the RGB settings. Very simple. And very robust. I am sure Phase must have a similar tool.
I hope this helps.
Peter.
I get your point, however, this is not a hardware issue, but an issue with that specific camera using the defaults in CO. So no recipes or styles in play either, thankfully.0 -
So looking into the files sent to us, I have made a Color Balance preset for You guys to try out (this way you can see what is being modified).
It is basically a selection of the brown trashcan shown earlier, hue selection expanded to cover some more green and yellow, saturation range decreased and moved a bit towards more saturated, transition smoothed, and finally the selection was desaturated a bit.
The preset was then used on all the images we have received to check for consistency.
The result of this preset seem similar to the output of the 40D/5D2 files we have received as well.
You can save this preset as a new ICC profile, and then set the new ICC as default for the 7d2. Then the profile will automatically be used for new images.
Here's the download-link: ftp://ftp.phaseone.dk/Support/Download/ ... d.copreset
The file has to be placed here:
Win: C:\Users\[USER]\AppData\Local\CaptureOne\Presets60\Color Editor
Mac: ~/library/application support/capture one/presets60/color editor
If this will make into a service release in one form or another, has not been decided yet, so I cannot comment on that.0 -
Hi Christian,
Thank you very much for the test colour-corrected preset.
I have been checking a lot of my 7D2 images comparing it with the generic 7D2 profile and am very pleased with the results - this is a definite improvement, the colours look much more realistic 😊
Even photos where I hadn't actually noticed the orange effect before, are looking better.
I really hope that a profile like this can be included in a future release of Capture One.
Regards,
Richard0 -
Hello
For what it's worth, this is a summary of my latest findings. Please note that I do not have a 7D2 but I have analysed images processed with 8.1.
Today I loaded 8.1 onto a laptop that has never seen Cap1. So a virginal baseline. I have 7.1.3 on my PC, which has had 8.1 on it, but not now. I only use sRGB (not Adobe RGB).
The following conclusions are based on the colour (RGB) profile for parts of processed images, selected and analysed in PS CC with the filter-blur-average tools
For my 5D2 there is no difference between 8.1 and 7.1.3. Compared with these outputs, when I use DPP4 there is a small difference at the red end of the temperature spectrum in that the green and blue channels from Cap1 are less prominent. Differences at the blue end are trivial.
For my 7D there is no difference between 8.1 and 7.1.3. Compared with these outputs, when I use DPP4 there is a significant difference at the red end of the temperature spectrum in that the green and blue channels from Cap1 are definitely less prominent. Differences at the blue end are minor.
For the 7D2 images (thank you) the Cap1/DPP4 image colour data discrepancies are much more obvious at the red end and would be a concern to me if I had a 7D2.
In all cases there is minimal cast in mid-tone greys.
I find I have gone full circle in all this. It does seem that Phase has progressively introduced a colour shift at the red end of Canon profiles over 5+ years, and they have reached a tipping point for user acceptability with the 7D2.
I have little or no problem with my 5D2. The 7D can be a problem with the strong colours of autumnal leaves. All I want is a good starting point! My workaround is to do basic raw conversions in Cap1 at two temperatures and then blend them to suit, while at the same time changing the screen brightness. Yes, this takes an extra minute or so, but the result is (for me) a robust starting point that saves time in the long run.
I have posted an article on screen brightness and image colour temperature in the "Workflow" forum. This may seem counter-intuitive on first reading but it works for me.
Hopefully we can put this saga to bed now. Many thanks for all your contributions. It has been a very productive learning experience.
Peter Jones.0 -
Hi Peter,
your points are very interesting; I totally agree about wanting a good starting point.
Here is an example to show the effects of the experimental color-corrected 7D2 profile.
First using the original 7D2 Generic profile in all its orangeness...
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1117_CO_7D2_generic_zpsrikzs9ua.jpg
Next using the 7D2 Color Corrected profile....
http://i1381.photobucket.com/albums/ah220/Python54/phase_one/IMJ_1117_CO_7D2_color_corrected_zpszcj8lbsf.jpg
Much better 😄
Regards,
Richard0 -
Hi Richard
You are a happy bunny now! Interesting that the colour adjustments effectively boost blues and greens, entirely consistent with my findings.
Your chosen images feature a lot of sandstone buildings - always a difficult hue to get just right!
I suspect Phase will continue to work on this, but at least we have a quick fix now.
Nothing about Nikon and Pentax? So is it mainly a Canon problem? I expect Keith Reader will have something to say!
Cheers for now, and thank you for the feedback.
And thanks to Christian for coming up with the goods!
Peter.0 -
[quote="Peter" wrote:
Interesting that the colour adjustments effectively boost blues and greens
why is it interesting as they are somewhat opposite to orange... what other colors you 'd suggest to "effectively boost" ?0 -
Hello D....
The answer is in your question. Spot on!
Peter0 -
Christian,
Thanks for organising this fix for Canon. Could you do something similar for Nikon users please as we have similar problems and I believe Pentax users will also be seeking for a similar fix?
Thanks
John0 -
You can use the preset on any type of image in CO. As you can see from content of the preset, the adjustment made is very simple indeed. 0 -
[quote="Christian Gruner" wrote:
You can use the preset on any type of image in CO. As you can see from content of the preset, the adjustment made is very simple indeed.
I 'd suggest for your "Phase One’s Image Quality Professor" to write a simple posting in his blog specifically addressing this 7DmkII conversions are too "orange"(for some users, with default settings), as it is simple it will take 20-30 min of his time but provides a good place to link in all future discussions here and there...0 -
[quote="deejjjaaaa" wrote:
[quote="Christian Gruner" wrote:
You can use the preset on any type of image in CO. As you can see from content of the preset, the adjustment made is very simple indeed.
I 'd suggest for your "Phase One’s Image Quality Professor" to write a simple posting in his blog specifically addressing this 7DmkII conversions are too "orange"(for some users, with default settings), as it is simple it will take 20-30 min of his time but provides a good place to link in all future discussions here and there...
I don't see a reason to be that specific, however, there are already some post on the Color Editor, that basically covers the same theme, just for everyone to read, and not only 7d2 users.
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/0 -
[quote="Christian Gruner" wrote:
I don't see a reason to be that specific, however, there are already some post on the Color Editor, that basically covers the same theme, just for everyone to read, and not only 7d2 users.
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/
as you can see it is not helping... people need to be repeated again and again and in many ways... take any forum the threads about that camera/converter is too red, this camera/converter is too green are perennials... so it helps to repeat something that is obvious to "Professor" for hoi polloi on a regular basis... 😎0 -
[quote="Christian Gruner" wrote:
[quote="deejjjaaaa" wrote:
[quote="Christian Gruner" wrote:
You can use the preset on any type of image in CO. As you can see from content of the preset, the adjustment made is very simple indeed.
I 'd suggest for your "Phase One’s Image Quality Professor" to write a simple posting in his blog specifically addressing this 7DmkII conversions are too "orange"(for some users, with default settings), as it is simple it will take 20-30 min of his time but provides a good place to link in all future discussions here and there...
I don't see a reason to be that specific, however, there are already some post on the Color Editor, that basically covers the same theme, just for everyone to read, and not only 7d2 users.
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/
The tools in CO1 for editing colors yourself are fantastic. It's liberating to let go of the search for acurate color and just experiment with the color balance tool and the advanced color editor. But eventually you always want to return to believeable colors, and the basis for a default .icc profile has to be color accuracy if it is to serve as an all purpose profile.
Chris0 -
Hello Christian
I am stepping back from all the bits of detail in this thread to muse on what has been achieved.
Firstly, and most importantly perhaps, we have discovered that it is possible for users to express their collective concerns to Phase One and establish an ongoing dialogue which has been constructive and useful. Perhaps a new paradigm for the future? Support Cases are OK for well-defined one-off issues, but wider problems may justify a different approach.
You have posted a "quick fix" mini-prog that tweaks the colour output of Cap1 specifically for the 7D2 and which can also be used for the 5D2, Nikon and Pentax (and more?) to assist the user in getting a more realistic starting point. Perhaps this is indeed a generic, systemic fix as originally envisaged by Keith Reeder?
So the "generic" colour profiles for 7D2, etc. in Cap1 remain unchanged. I suspect some of these generic profiles are more generic than we fully realise. For example, the "generic" profiles for the 5D3, 6D and 7D2 appear to be identical, given their same size of 217,584 bytes. Similarly the profiles for the 5D2 and 5D are the same size of 219,320 bytes. The 7D profile is singular, at 219,316 bytes. One more example - the Sony A7, A77 and A99 all appear to have the same profile, 217,572 bytes. I appreciate that basing these observations solely on profile size may be misleading, but the same size down to the last byte? Can you please comment on this. Also, what extra care does Phase take when producing a V2 profile?
I am not " having a go" at Phase over this. I simply wish to be reassured that Phase is prepared to keep camera profiles under review and to refine them as necessary. Compared with just a few years ago, I think hardware developments (cameras, monitors and printers) have moved on as regards colour fidelity, but perhaps software colour algorithms are being left behind - even at Phase? Witness your decision to tweak Cap1 to accommodate the 7D2 orange cast problem. And at the risk of repetition, please bear in mind that all we are seeking is a robust and realistic starting point. The Professor's Blogs, which describe tweaks here and tweaks there are OK for one-offs, but not for a mixed bag of images. (No disrespect to Niels!)
I confess that my confidence in Phase One has been dented by what has emerged in this thread. But not (as yet) sufficient to justify an insurance repair in my local garage!
Peter.
By the way, at the risk of looking stupid I can't get your 7D2 quick fix across to the Cap1 folder. I can copy the lines of code and paste it to Word, but it won't paste into Cap1. What am I doing wrong?0 -
Hello John
I recall that you were thinking of reverting from 8.1 to v7, but the latter trial had time expired. I have some information from a support case that may help you. No guarantees, mind!
In essence you can use your v8 license key to activate an older version, such as v7. There is a nuance here - this definitely works for an older version that has been licensed (I have done this). It may not work if you only trialled it. But worth a try perhaps. Bear in mind that although you can have two different versions of Cap1 loaded up, you should only launch one at a time, and close it before launching the other one. You may also see evidence of "cross-talk" between them when they access support files!
If in doubt, why not submit your own support case.
Peter.0 -
Peter,
What lines of code?
Download the file to somewhere on your system. If your Browser has a Download folder destination that will do.
Copy the file to the folder Christian identified. That will make it available as a User pre-set.
If you want to make is a default to be applied to all files for the camera (which, like me, may only be of academic interest right now as we don't have that body) you can open a 7D2 file, apply the preset and then Save as an ICC profile. The ICC profile can then be made the default profile for the camera.
Job done.
As I suggested a few posts back - there are plenty of user options available to resolve the apparent problem (which I don't deny but for which I don't necessarily believe there is a single all pervasive solution) at a very personal level. Total personalisation if required.
Whether a single new default file (which may possibly be forthcoming) will answer everyone's wishes remains to be seen. I suspect that it may, for many, get them closer to a comfortable starting point (if provided) but ultimately does not provide a new solution - just a different starting point to arrive somewhere that was already possible to reach.
One thing that strikes me looking as what appear to be "default" RAW file conversions from several different software vendors is that nearly all the results are different in some way. There is no standard that is followed other than the assumption that the manufacturer software "gets it right" (even if it is usually impossible to use for some reason!) most of the time.
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.
I suspect that is why Phase have had, so far as we know, few "complaints" across the board for any single camera. When used by pros they expect to deal with such issues anyway (or have one of their people do so ....) and when used by others many may not be shooting stuff that seems TO THEM to be compromised. Once in a while we all come across annoying exceptions ... but that's life.
My thoughts, for what little they are worth.
Grant0 -
Hi Grant
Yes, I knew I would look stupid. For some reason the link that Christian has sent simply opens a few lines of code that I can do nothing with (excuse grammar).
I will digest the bulk of your post idc. Thanks for taking the trouble.
Peter.0
Post ist für Kommentare geschlossen.
Kommentare
225 Kommentare