Colour Profile Problems Part 2

Keith Reeder

January 28, 2015 18:10

Thanks for keeping this going, Peter. Doubtless the thread will be locked again, but until then...

Here's the gist of what I was writing last night in response to Chris' post just as Drew pulled the plug...

The suggestion that we should raise support cases for this is all well and good - and I'm generally an advocate of engaging Phase One by its preferred support arrangements - but here, a support case is patently the wrong approach.

Why?

Well, I know from personal experience (and ownership of affected cameras) that this problem exists for the 7D, the 70D and the 7D Mk II: that means it has been known about (and - God knows - talked about on here often enough) since around early 2010, when, IIRC, 7D support was introduced in Capture One.

It's now early 2015.

I could raise a support case for those cameras - but we've had people on here complain about the 5D Mk III's profile, and you can bet your pension that other cameras produced by Canon since the 7D have the problem too.

I don't own those cameras. How then, can I raise support cases for them? And if there's no support case, does that mean their profiles won't get fixed?

At least one Nikon D7000 user has made similar complaints on here, too - Canon users are unlikely to be alone in this.

This is a systemic issue - it's not a case-by-case, camera-by-camera problem which would suit the support case process.

It's not as if there's any ambiguity or lack of clarity about exactly what the problem is, either - there's absolutely no need to "investigate" or "analyse". Too many current profiles are too red/orange. End of story.

This isn't something that needs support cases, it needs a root-and-branch review of whatever it is Phase One is doing these days to create its profiles. Support cases would just be a band-aid.

So no support case from me: Phase One addresses this systemically (properly) or it's Optics Pro 10, Lightroom and Photo Ninja, all of which produce far better colours from my bodies, for me.

Capture One's erstwhile "[color=#0000FF:2dfqj1mq]USP[/color:2dfqj1mq]" - its colour rendering - is now a good reason not to use it.

0

RichardT

January 28, 2015 19:20

I raised a support case for the 7D MkII with Phase One today.

Regards,
Richard

0

Peter Jones

January 28, 2015 21:42

Well written Keith. I fully support and agree with you.

My intro to C1 was v7.3/4. I am sure the colour tones were quite acceptable then. Now, alas, I too am moving away from C1 for raw conversion. But it is still good for tifs.

As an aside, I find the C1 profile for my Canon 5D2 is actually quite good - when applied to a raw 7D file.

I will dig out my Photo Ninja stuff again!

Peter

P.S. Thanks for introducing me to USP. So that is how I survived when I was self-employed!

0

ChrisM

January 28, 2015 23:14

[quote="Keith Reeder" wrote:
Thanks for keeping this going, Peter. Doubtless the thread will be locked again, but until then...

Here's the gist of what I was writing last night in response to Chris' post just as Drew pulled the plug...

The suggestion that we should raise support cases for this is all well and good - and I'm generally an advocate of engaging Phase One by its preferred support arrangements - but here, a support case is patently the wrong approach.

Why?

Well, I know from personal experience (and ownership of affected cameras) that this problem exists for the 7D, the 70D and the 7D Mk II: that means it has been known about (and - God knows - talked about on here often enough) since around early 2010, when, IIRC, 7D support was introduced in Capture One.

It's now early 2015.

I could raise a support case for those cameras - but we've had people on here complain about the 5D Mk III's profile, and you can bet your pension that other cameras produced by Canon since the 7D have the problem too.

I don't own those cameras. How then, can I raise support cases for them? And if there's no support case, does that mean their profiles won't get fixed?

At least one Nikon D7000 user has made similar complaints on here, too - Canon users are unlikely to be alone in this.

This is a systemic issue - it's not a case-by-case, camera-by-camera problem which would suit the support case process.

It's not as if there's any ambiguity or lack of clarity about exactly what the problem is, either - there's absolutely no need to "investigate" or "analyse". Too many current profiles are too red/orange. End of story.

This isn't something that needs support cases, it needs a root-and-branch review of whatever it is Phase One is doing these days to create its profiles. Support cases would just be a band-aid.

So no support case from me: Phase One addresses this systemically (properly) or it's Optics Pro 10, Lightroom and Photo Ninja, all of which produce far better colours from my bodies, for me.

Capture One's erstwhile "[color=#0000FF:5md5ttht]USP[/color:5md5ttht]" - its colour rendering - is now a good reason not to use it.

Fair enough Keith, if you like DxO or Lightroom colors better, that's the way to go.
But I have been through all that when Pentax K3 support was late. Adobe's profiles are flat and lifeless if you ask me, and no way that a support case will ever make them give you a better one. DxO forces the restrictions of the Adobe RGB color space on you with their camera profiles, to which they are calibrated and beyond which they cannot display colors, really incomprehensible with today's large gamut printers. I asked them, and no way they will give you camera profiles in a larger color space: their answer: we give you the option to provide your own .icc profile if you want more than sRGB or Adobe RGB. DxO is fine if you can live with that, but don't expect your nature images printed on high quality papers to reflect your cameras capabilities. You dismiss the benefits of Phase one tailor made support through their support cases a bit too easily. I think it is a rare form of support that is worth the effort, simple as that.

Chris

0

AiDon

January 29, 2015 04:21

I really don't have any problems with the color profiles for either the 7D or the 5D Mark II/III and when C1v8 came out I moved from using Lightroom to Capture One.

I prefer the color and tone rendition in Cap 1.

Don.

0

Peter Jones

January 29, 2015 10:07

A thought.

The grass is always greener in the next field - until we spot an even better-looking field.

Peter

0

Urukhai

January 29, 2015 14:16

Can some one from CO shed some light on how the icc profiles for CO8 Cultural Heritage Edition were made?

One bullet point there said the CH Edition has "Specialized ICC Profiles for accurate color reproduction"? This CH Edition will have a price tag of EUR 4999.

Why the same icc profile making process cannot be applied to other cameras in CO8 normal version?

0

Keith Reeder

January 30, 2015 19:55

[quote="AiDon" wrote:
I prefer the color and tone rendition in Cap 1.

And that's fine: but too many people are posting on here explicitly and specifically to complain about this exact issue.

I'm also quite heavily involved in the Lightroom and Optics Pro ecosystems too, and you simply do not see an analogue of this issue on their forums.

This ongoing, longstanding, demonstrable and repeated complaint is uniquely Capture One's, and it goes beyond simple "personal preference" of one default colour rendition over another.

I've extolled (and sometimes defended) Capture One's colours for years, and they've clearly changed - that's the issue here.

They've changed, and in direct comparison to what went before, the change is for the worse.

I'm dealing with it well enough, though: I'm not using Capture One any more.

But that's surely not what I paid for...

0

NNN635077039496462564

January 31, 2015 03:24

Yes, agreed wholly with Keith.
It's absolutely pointless wasting time and energy trying to render
adequate, suitable colors with Capture One Pro. No hope.
I tried all the possible ICC profiles available in C1P with RAWs shot on Nikon D4, D800E
and D810 bodies and every and each result was worse that that of competitors.
Yes, there is a lot of improvements in the late CO8P versions, but not enough yet.
For example, new Nikon D810 Generic Profile are much, much better than everything before,
but, anyway, the colors outputted using it, are noticeably worse than those of competing
applications, especially skin tones, especially in shadows.
And not to admit this means not to believe own eyes.
I stopped using C1P as a working converter several months ago. Now I've activated
only one of my two full CO8P licenses. And only for testing and comparing purposes.
And every time I'm comparing the results to those of competing apps I'm getting
evidence that C1P is losing, losing, losing...
By ethical reasons I don't call those competitors here, on Phase One forum, but,
I'm sure you all know which ones I mean.
So, farewell, Phase One Team, no deal with you in the foreseeable future.
And no more cent respectively.
And, please, don't advise me to create a support case and send you some samples -
I already did it in June 2014 and you know very well what came out from it.
No repeating again.

Best regards,
Mike.

0

Keith Reeder

January 31, 2015 10:37

[quote="NNN635077039496462564" wrote:
Yes, agreed wholly with Keith.

Thanks, Mike.

And, please, don't advise me to create a support case and send you some samples -
I already did it in June 2014 and you know very well what came out from it.

Hmmm... That's rather telling, isn't it?

If the problem persists even after support cases have been raised, doesn't that suggest - pretty unequivocally - that the issue here isn't a camera-by-camera issue; and that by logical implication, fundamental flaws exist in Phase One's profile creation process?

That's what it says to me. My suspicion that this would be the case is one very good reason why I haven't raised any myself - a piecemeal approach to addressing systemic faults is pointless.

0

NNN635077039496462564

January 31, 2015 13:19

Hi Keith,

I can only repeat again and again: I absolutely agreed that "...fundamental flaws exist
in Phase One's profile creation process." And, in my opinion, without any "?". As well as I absolutely
agreed with your conclusion that "... a piecemeal approach to addressing systemic faults is pointless."
The problem is too deep, too fundamental and, I would say, it's too neglected case to cure it with aspirin.

And what about the support case mentioned in my previous post you can, if you wish, of course,
have a look at this my recent talking with Paul Steunebrink
http://forum.phaseone.com/En/viewtopic.php?f=57&t=17822&sid=e9bbc1b30be4d352f73d4650c6251050
Initially we have said not too pleasant things, but then have overcome the misunderstandings
and the talk has become quite constructive and frendly.
There I've told what I think of the current situation with CO8P.
I just suppose that repeating here everything I've told Paul in those posts is, well, not too good
and it would be better you just have a look at the posts.
Again, if you wish...
Thank you for you posted thoughts, which are so similar to mine.
But, to say honestly, I'd like so much that there would be no reason for such our thoughts and concerns
and CO8P would be working just brilliantly... May be in the near future?

With best regards,

Mike.

0

ChrisM

January 31, 2015 13:58

We all draw our own conclusions about CO1. Mine are favorable, as I cannot see any other raw converter giving me the same true to life results for my nature and wildlife photography. But I also believe we share the same wish, that Phase one would simply shift their profile making in the direction of neutrality and accuracy, instead of something as wildly subjective and unreliable as "pleasing". The moment colors start to fundamentally shift away to different hues, there is something inherently wrong with the profile. I took some images of wild swans, they feature a yellow patch across the beak that is very close to the area of orange hues in the spectrum. In DxO the patch was yellow, in Pentax own software (Silkypix) the patch was yellow. In Adobe ACR the patch was yellow. In CO1 v8 the patch was heavily orange tainted. It demands a few ours at the color editor to create a better profile, and it worked. But its due to the shift towards red that this happened in the first place. Create a "pleasing" profile, and you open up a can of worms for images that contain colors with hues that are very close to adjacent (but distinctly different) hues. Phase one can never check their profiles with thousands of real world images, and that is where it goes wrong. Stick more to "neutral" and "accurate" and let the user add the "pleasing" sauce, and you avoid a lot of color issues for a lot of images. Alternatively, add two profiles by default: one home cooked (pleasing) CO1 profile, and one accurate neutral profile.
Still: I would not want the flat lifeless neutrality of Adobe ACR, nor do I care for the severely restrictive (crippled color space) DXO. In other words: always put things in perspective. If it were the case that Phase one is giving everyone a headache and all is rosey with Adobe or DxO, then I would join your laments, but that is far from the case.

Chris

0

RichardT

January 31, 2015 15:50

[quote="ChrisM" wrote:
We all draw our own conclusions about CO1. Mine are favorable, as I cannot see any other raw converter giving me the same true to life results for my nature and wildlife photography. But I also believe we share the same wish, that Phase one would simply shift their profile making in the direction of neutrality and accuracy, instead of something as wildly subjective and unreliable as "pleasing". The moment colors start to fundamentally shift away to different hues, there is something inherently wrong with the profile. I took some images of wild swans, they feature a yellow patch across the beak that is very close to the area of orange hues in the spectrum. In DxO the patch was yellow, in Pentax own software (Silkypix) the patch was yellow. In Adobe ACR the patch was yellow. In CO1 v8 the patch was heavily orange tainted. It demands a few ours at the color editor to create a better profile, and it worked. But its due to the shift towards red that this happened in the first place. Create a "pleasing" profile, and you open up a can of worms for images that contain colors with hues that are very close to adjacent (but distinctly different) hues. Phase one can never check their profiles with thousands of real world images, and that is where it goes wrong. Stick more to "neutral" and "accurate" and let the user add the "pleasing" sauce, and you avoid a lot of color issues for a lot of images. Alternatively, add two profiles by default: one home cooked (pleasing) CO1 profile, and one accurate neutral profile.
Still: I would not want the flat lifeless neutrality of Adobe ACR, nor do I care for the severely restrictive (crippled color space) DXO. In other words: always put things in perspective. If it were the case that Phase one is giving everyone a headache and all is rosey with Adobe or DxO, then I would join your laments, but that is far from the case.

Chris

I fully support the above points. The thing I don't understand is why anyone would consider the exaggerated orange bias to be "pleasing"?

Richard

0

SFA

January 31, 2015 16:10

I don't think I have seen any claims for the recent Canon (where the forerunner of this thread started) profiles to have been created as "pleasing".

Moreover it's worth remembering that there are 3 modifying curves and Linear to check out within each profile.

It strikes me that it is entirely possible that the test results, on whatever hardware was available, were based on files that do not necessarily correspond to files that are to be found in the wild ... but who is to know that unless there is a serious volume of feedback that provided acceptable and unacceptable results for comparison - whether or not the concept is based on being 'pleasing'.

I would mention that another product I have used extensively (though less more recently) allows people to develop their own curves should they wish to take on the technical challenge. I have come across images where it was almost impossible to get anything even nearly right - especially with a very specific and narrow section in the orange/red spectrum. It's also an area where different opinions about preferences will always abound.

Sadly I don't have any of the more recent bodies with which to experience the problem first hand and that limits my options for experimentation.

Grant

0

Keith Reeder

January 31, 2015 16:12

[quote="ChrisM" wrote:
Still: I would not want the flat lifeless neutrality of Adobe ACR, nor do I care for the severely restrictive (crippled color space) DXO.

Sigh...

Chris, comments like that just demonstrate a lack of knowledge on your part of how to use these converters properly, not of any failings in them; and a personal bias which is irrelevant to the discussion.

ACR/Lightroom is as "flat and lifeless" as you make it; and Optics Pro uses Adobe RGB as its internal colour space, which is more than enough for any 16 bit processing any of us are likely to do; as and when 32 bit/wide gamut processing becomes the commonplace - ie when it becomes necessary - we'll see a colour space change in Optics Pro.

Again, this issue is not about whether we "care for" one converter's default colour rendering over another. It's about a much-demonstrated excessive bias towards orange/red, way beyond what many of us consider to be "pleasing", in direct comparison to Capture One's older profiles, which still had enough warm bias to give the "Capture One Look", but which were/are massively less skewed towards the (frankly) fluorescent.

0

Keith Reeder

January 31, 2015 16:22

[quote="SFA" wrote:
It strikes me that it is entirely possible that the test results, on whatever hardware was available, were based on files that do not necessarily correspond to files that are to be found in the wild

If this was a one-off complaint, that'd be a fair comment, Grant - but it has been ongoing since the 7D/Capture One 6 at least; and affects cameras from other manufacturers too.

Which begs the question: what kind of files have Phase One been profiling with since 2010?

Rather, I'm pretty sure that around that time, Phase One actually announced something about adopting a new profiling regimen...

0

SFA

January 31, 2015 16:50

[quote="Keith Reeder" wrote:
[quote="SFA" wrote:
It strikes me that it is entirely possible that the test results, on whatever hardware was available, were based on files that do not necessarily correspond to files that are to be found in the wild

If this was a one-off complaint, that'd be a fair comment, Grant - but it has been ongoing since the 7D/Capture One 6 at least; and affects cameras from other manufacturers too.

Which begs the question: what kind of files have Phase One been profiling with since 2010?

Rather, I'm pretty sure that around that time, Phase One actually announced something about adopting a new profiling regimen...

Well that would all fit Keith but there do seem to be some odd things happening in some areas that don't appear to be problems in others. The same appears to have been true in recent times for some applications that use dcraw as the basis for conversion. What has made such results less acceptable one wonders? Or perhaps I should say "more difficult to achieve to wider satisfaction".

I don't disagree with the criticism made. For example the originally offered sample file probably would not appeal to everyone based on the Film Standard curve. However apply the alternative Film Extra Shadow curve and things got very much closer to the desired output indicated by the jpg sample provided. So in that particular skin tone case on might argue that V8's "Auto" feature might be better adjusted to pick an alternative to the Film Standard for predominantly portrait images of some skin tones. No doubt similar observations might be made for other subject matter.

This all seems like an opportunity for mutually beneficial engagement of all parties to improve "the breed" for a wider audience. It would be nice to see that happen.

Grant

0

Peter Jones

January 31, 2015 18:39

This thread makes very interesting reading for me. It is good to read the contributions from people who have a broad experience and interest in getting the best out of raw files.

I still periodically see what Cap1 does with an image. But every time I turn elsewhere to find a starting point that is easier to work with. As I have written before, my simple approach is to do a basic raw conversion with DPP or IDC (with no colour adjustments except WB) and go from there with tif-based adjustments. Far less stressful!

Please, Phase One, allow this thread to continue. It is important for us all.

Peter

0

NNN635397352773288520

February 01, 2015 10:44

I know that they want us to make a support case, but is it possible to make it in danish? Because im not the best to explain myself in english. Drew was talking that color is a matter of taste or how to say.. But like some said in this thread say orange skin will never be pleasing. I wish that the colors wasnt dominated by orange and red. Some times so much that it looks burned out. I wasted many hours fighting only with the colors..Just to make it a bit realistic in the rawconversion. It you have a realistic start point its easy from there to make more pleasing colors it the photo needs it and that doesnt mean orange..But when the start point looks like something not from this planet. It can only give you headache. And naturally colors for me is what the eye see.

0

Christian Gruner

February 01, 2015 12:17

Yep, you can send it in Danish, as we have Danish speaking supporters as well. However, you might get a quicker response writing in English, as everyone in Phase One speaks English.

Regarding the color-issue, please make sure to send the problematic raw-files to us in a support-case. Preferably with a description of the light-source(s) used (daylight, sun, flash, so on). They will be forwarded to R&D and evaluated using the description provided.
Why do we need a description? There can be a lot of difference in the lighting and it's temperature/tint. For non-Team Phase One brands, the default ICC-profile is done for daylight/sunny situations at around 5600 Kelvin.

I will also encourage people in this thread to post examples of their specific color issue, side by side or single examples.

0

Peter Jones

February 02, 2015 22:14

Hello Christian

Thank you for your Phase input to this thread. Please allow me to make the following comments.

You refer to "problematic raw files". But this is putting the cart before the horse. Raw files have their own integrity. They are the camera's record of the light that the lens passes on to the sensor. In this sense the raw file is a perfect record. It is what we do to the raw file afterwards by processing in camera, and in photo software such as Cap1, that is inherently problematic.

Photo software is designed to cater for a range of subjects and scenic types. Each image is mapped from the raw file to its in-process equivalent according to what each image "needs". Most software will produce some kind of default output that is the starting point for personal processing. Default settings can be available as a range of presets.

If the starting point is not a close match with what the photographer saw (or thinks he/she saw) an experienced photographer can use the software to tweak it here and there. But if the mismatch is too great it can require a lot of work to get the colours and tonal structure looking "right" - like trying to correct a complex lens distortion.

I have seen enough Cap1 posts in recent months to know that I am not alone in having problems with Cap1 in the orange-red portion of the spectrum. With all due respect it does seem that Phase One has changed Cap1 output colour algorithms in a way that is creating problems for some of your most experienced users.

Your post refers to 5600K as your benchmark/starting point. This seems a little too red to me, and in any case you will know that there is no such thing as "the" colour temperature for a photo. Highlights tend to red and shadows tend to blue, in all photos. I assume Cap1 is programmed to allow for this, and also tweaks the saturation differently at different parts of the spectrum. But it sounds as though the final outcome of Cap1 settings/adjustments are judged by eye? In which case, how true is this judgement, and how consistently is it replicated from one version of Cap1 to the next?

I agree with Keith Reeder. Recent versions of Cap1 have acquired a systemic colour problem that needs sorting out at your end.

Thank you once again for your interest.

Peter.

0

ChrisM

February 02, 2015 23:13

[quote="Keith Reeder" wrote:
[quote="ChrisM" wrote:
Still: I would not want the flat lifeless neutrality of Adobe ACR, nor do I care for the severely restrictive (crippled color space) DXO.

Sigh...

Chris, comments like that just demonstrate a lack of knowledge on your part of how to use these converters properly, not of any failings in them; and a personal bias which is irrelevant to the discussion.

ACR/Lightroom is as "flat and lifeless" as you make it; and Optics Pro uses Adobe RGB as its internal colour space, which is more than enough for any 16 bit processing any of us are likely to do; as and when 32 bit/wide gamut processing becomes the commonplace - ie when it becomes necessary - we'll see a colour space change in Optics Pro.

Again, this issue is not about whether we "care for" one converter's default colour rendering over another. It's about a much-demonstrated excessive bias towards orange/red, way beyond what many of us consider to be "pleasing", in direct comparison to Capture One's older profiles, which still had enough warm bias to give the "Capture One Look", but which were/are massively less skewed towards the (frankly) fluorescent.

Keith,
the discussion would gain a lot of transparency if the pounding at the address of Phase one were filtered out and DxO optics and Adobe were no longer served up as alternatives that get it right and that can serve as an example to Phase one of how it should be done. You and others that are frustrated with the imperfections of Phase one's generic .icc profiles may not see it this way, but this indictment at the address of Phase one comes across as rather depraved. There is simply no basis for stating that Adobe or DxO offer a better starting point as raw converters, other than your own subjective like/dislike or your own typical workflow.
To get the lifelessness out of Adobe's standard profiles, I need a lot more work myself than I need to create my own general customized .icc profile in the color editor in CO1 that can act as a starting point for all images of a particular camera, and the result to my eyes are superior. ALL raw converters only act as a starting point, even Jpeg is ONLY an interpretation of an image rendered by a camera maker's Jpeg engine. DxO is severely hampered by its crippling implementation of a rather odd color space (read up on how Adobe rgb was created, e.g. in the articles by Joseph Holmes) that shares the red and blue primary with a color space (sRGB) that was created for CRT monitors!. Sure, you can claim that Adobe RGB is all that professionals will ever use/need, but given todays quality printers, that is highly debatable to say the least. If you were truly interested in color spaces and their function in digital imagery, you would not so readily allow for a forced limitation Ã la DxO with its rigorous clipping of highly vivid colors. Capture One does its color management in the only way that is right: start with the camera's color space, let the user select a working space, and allow for outputting into any suitable color space for further editing/printing/viewing on screen whatever. Give the user all controls and the tools to get the right output. DxO is just there to make it easy for you by limiting your choice and flooding you with presets and "auto's".

So lets get this discussion back to its core: Phase one, like any Raw converter, gives you a starting point by means of their support for a camera model, no more than just that: a starting point, like a vendor Jpeg engine, like DxO, like Adobe. They have a unique support system by means of which they allow user feedback. If CO1 is your software of choice: use this support, if you like Adobe or DxO better: start using those and run into theÃr limits eventually, and find that they don't offer Ã¡ny sort of customized support, they will basically tell you to accept their software "as is" or otherwise move elsewhere. I know from personal experience.

If you refuse to open a support case and provide material for Phase One to work on, fine, but then one should not expect much in return. The endless bickering about how Phase One has or must have changed its calibrating routines and how that is for the worse, is tiresome and pointless. Just open a support case, provide proof in samples and that way we all have something to gain from this discussion. But from the bashing and calling for Adobe and DxO, no one will get any wiser.

Chris

0

Urukhai

February 03, 2015 05:38

Capture One does its color management in the only way that is right: start with the camera's color space,

I am not quite sure about this. Color Space is just one thing. I see Color Management in CO workflow is broken.
I can have my monitor calibrated to certain delta E. I can also obtain fairly accurate icc output profile from printing house.

But the icc camera profile in CO was just arbitrarily set to someone's liking.

In the LR solution, I can use either an Xrite ColorCheck or QPCard to make a "calibrated" dnp profile for my camera in just a couple of minutes. The results are consistently satisfactory so far.

But for CO, I have tried using Rawdigger & TIFF from CO + Argyll to make an icc profile. But they never look accurate.

Whether the default profile from CO is good is just my secondary concern provided I can make my own.

Unfortunately there weren't any tutorial or video officially from Phase One on how to "objectively" calibrate a camera profile via Color Editor. That's why I see Color Management in CO workflow is broken.

0

SFA

February 03, 2015 12:13

My understanding is that the colour interpretation from the RAW data results in the LINEAR curve with any profile provided.

That's the baseline interpretation from which the other (Film) adjustments are derived.

If LINEAR is not what you can live with then there are grounds for claiming that the base interpretation is wrong. However "wrongness" is still a matter for user judgement since all converters that I have personal experience of using have produced different results, perhaps only marginally different in some cases but different nevertheless, even when using the same core demosaicing "engine" (for example dcraw). That to some extent is the nature of a product having a USP in the marketplace.

In addition to Liner C1 offers 3 "Curves" to go with the profile and will typically default to one of those before presenting the image in an "enhanced" form on screen. Before 8.1 I never saw anything other than "Film Standard" selected. It mostly works pretty well for typical daylight outdoor shots for the cameras I use. As you might expect really. For alternative daylight light conditions one of the other curves might be a better starting point but one can usually make adjustments from "Film Standard" anyway or simply go to Linear and do your own thing.

Since 8.1 (maybe earlier?) the "Auto" selection seems to be the chosen route to C1 picking a curve. So far I have not noticed it selecting anything other than Film Standard but I confess I have not spent a lot of time checking.

One might have criticism of the Auto selection.

Experimenting with some Canon 7D2 files from some of the recognised internet photo "test" sites shot in different conditions I think the "Film Extra Shadow" curve offers a consistently more agreeable starting point than does "Film Standard". The difference seems to be mostly WB and exposure related. The results are close to the Out Of Camera jpg for the same image - though that may not be a step forward in all cases. Actually for certain important areas the results are much better of course but here we are discussing the stronger colours especially orange/red interpretation. On the other who knows how the camera settings were set for internal jpg production?

From that point I find I was a couple of clicks away from a number of different but potentially equally pleasing results form the images even when they involve some inconvenient lighting conditions of the sort that would make discussion of accurate colours rather meaningless. (Mixed light sources at higher ISO values for example).However I think that (the concept of absolute colour accuracy) is a separate discussion from this discussion about Colour Profiling for more general photographic purposes and should be kept separate.

But the curve are just curve offered to assist and perhaps save some time when shooting with some "typical" results and seeking a film like response. There is no reason at all why people should feel a need to use them and so the goodness or otherwise of their results ought not to be part of a final decision making process about whether on not default C1 is worse or better than default Brand X. If that is so important to use we should be creating our own interpretations for defaults and starting with them.

However, coming back to my earlier comment about the LINEAR interpretation .... if that does NOT give a good starting point for you to develop something you like then certainly that would be a good reason to suggest that the core calibration and profiling seems to be awry and making life more difficult for all.

Beyond that if the adjustment tools fail to offer the finesse required when fine adjustments are necessary then they too are open to critique.

Both of the above are, in my opinion, far more critical to the process of creating the image we want to see than what choices have been made when generating the "Film" curves.

Maybe it's time for a set of "Digital" curves as well? And perhaps some form of ability to save "User Curves" - something distinct from Styles.

My thoughts for what they are worth.

Grant

0

ChrisM

February 03, 2015 14:47

[quote="SFA" wrote:
My understanding is that the colour interpretation from the RAW data results in the LINEAR curve with any profile provided.

That's the baseline interpretation from which the other (Film) adjustments are derived.

If LINEAR is not what you can live with then there are grounds for claiming that the base interpretation is wrong. However "wrongness" is still a matter for user judgement since all converters that I have personal experience of using have produced different results, perhaps only marginally different in some cases but different nevertheless, even when using the same core demosaicing "engine" (for example dcraw). That to some extent is the nature of a product having a USP in the marketplace.

In addition to Liner C1 offers 3 "Curves" to go with the profile and will typically default to one of those before presenting the image in an "enhanced" form on screen. Before 8.1 I never saw anything other than "Film Standard" selected. It mostly works pretty well for typical daylight outdoor shots for the cameras I use. As you might expect really. For alternative daylight light conditions one of the other curves might be a better starting point but one can usually make adjustments from "Film Standard" anyway or simply go to Linear and do your own thing.

Since 8.1 (maybe earlier?) the "Auto" selection seems to be the chosen route to C1 picking a curve. So far I have not noticed it selecting anything other than Film Standard but I confess I have not spent a lot of time checking.

One might have criticism of the Auto selection.

Experimenting with some Canon 7D2 files from some of the recognised internet photo "test" sites shot in different conditions I think the "Film Extra Shadow" curve offers a consistently more agreeable starting point than does "Film Standard". The difference seems to be mostly WB and exposure related. The results are close to the Out Of Camera jpg for the same image - though that may not be a step forward in all cases. Actually for certain important areas the results are much better of course but here we are discussing the stronger colours especially orange/red interpretation. On the other who knows how the camera settings were set for internal jpg production?

From that point I find I was a couple of clicks away from a number of different but potentially equally pleasing results form the images even when they involve some inconvenient lighting conditions of the sort that would make discussion of accurate colours rather meaningless. (Mixed light sources at higher ISO values for example).However I think that (the concept of absolute colour accuracy) is a separate discussion from this discussion about Colour Profiling for more general photographic purposes and should be kept separate.

But the curve are just curve offered to assist and perhaps save some time when shooting with some "typical" results and seeking a film like response. There is no reason at all why people should feel a need to use them and so the goodness or otherwise of their results ought not to be part of a final decision making process about whether on not default C1 is worse or better than default Brand X. If that is so important to use we should be creating our own interpretations for defaults and starting with them.

However, coming back to my earlier comment about the LINEAR interpretation .... if that does NOT give a good starting point for you to develop something you like then certainly that would be a good reason to suggest that the core calibration and profiling seems to be awry and making life more difficult for all.

Beyond that if the adjustment tools fail to offer the finesse required when fine adjustments are necessary then they too are open to critique.

Both of the above are, in my opinion, far more critical to the process of creating the image we want to see than what choices have been made when generating the "Film" curves.

Maybe it's time for a set of "Digital" curves as well? And perhaps some form of ability to save "User Curves" - something distinct from Styles.

My thoughts for what they are worth.

Grant

My own overall thoughts on this, are that a good camera profile should be accepted as a valid starting point, and the work then has to be done with the (very good) tools that CO1 offers. But when a profile has inherent behavior of shifting colors to different hues by default, then the profile is pretty much hopeless, and any work put into it a mere waste of time. The inherent shifting of colors is not caused by either WB or Curve selection. WB change or Curve change will influence the overall look of an image, and may make some colors look more (or less) saturated, and look a bit different due to a color cast, but you will not see the serious shifting of colors that a bad profile causes.
To use my own experience as a result: the generic v1 profile for the Pentax K3 made purple shift towards dark blue, it made red shift to yellow and oversaturated the red in any WB or curve selection. It made the blue of skies shift towards grey. I spent literally hours to correct this profile, eventually gave up, opened a support case and switched to DxO. The generic v2 profile corrected 90% of the color shifts, and was vibrant as well as pretty accurate on most colors. But it still did feature a shift from yellow to orange due to the general color casting that Phase one seems to do lately on its camera profiles, preferring orange/red. Still, this is a good profile, and far from beyond repair in the color editor. It took me an hour to edit the profile and save it as a new default for the Pentax K3 that beats the pants of brand x and brand x (not going to name them anymore).
Something similar happened with the Sony A7r, but much less severe. Nonetheless, the v2 (or "standard" as Phase one has chosen to call it for some reason, instead of simply "generic v2"), is a far superior profile to the generic v1, in fact: it"s the single best color profile for a camera that I have ever used.
In short: Phase one messes up with some of their profiles badly for some reason, but they dÃ³ almost always come up with a very good v2, and also offer personalized support when you cannot live with a profile and take the effort to show why.
The better (v2) profiles can easily be worked on to great results.

Chris

0

SFA

February 03, 2015 14:52

[quote="ChrisM" wrote:

Something similar happened with the Sony A7r, but much less severe. Nonetheless, the v2 (or "standard" as Phase one has chosen to call it for some reason, instead of simply "generic v2"), ........

I would think that this one might have some better/different access to Sony influenced data given the recent relationship and the terminology would fit with the co-branding.

IMO.

Grant

0

Peter Jones

February 03, 2015 20:32

Let's cut to the chase. Phase One must now be fully aware that there is a perceived colour problem with Cap1 outputs. It does seem to be a systemic problem. There is more than a hint in the thread posts that some kind of step-change occurred in the colour rendering algorithms at some time in the upgrade from 7.x to 8.x. Users have speculated on this but only Phase knows for sure. Perhaps Phase will soon be in a position to respond to these concerns.

If I were investigating this problem I would also want to check the monitor calibration of the Phase computers, and compare outputs from different versions of Cap1 with a suitable test image file. Then, perhaps, we can identify how far back we should go to download a version of Cap1 that is as it should be, colour-wise. I would also be interested in possible external factors such as recent camera firmware updates and the extent to which users routinely use polarisers.

Out of interest I just tried Cap1 (and other raw converters) again. The no colour setting is a tempting starting point, and it can form part of a user Style setting, but still I prefer the camera manufacturer's own raw converter for colour rendition. Cap1 outputs just look too "stressed" in comparison. Perhaps Phase can give some thought to providing us with a colour rendition that is intended to emulate that which the camera manufacturer produces. Then I can go back to Cap1 as my preferred raw converter.

Peter.

0

SFA

February 03, 2015 22:22

How is your support Case progressing Peter?

Grant

0

Keith Reeder

February 04, 2015 09:28

[quote="SFA" wrote:
How is your support Case progressing Peter?

Indeed - I'm particularly keen to read Phase One's explanation of how any "local" fix Phase One might deliver to Peter will fix essentially identical issues with all of the cameras - and camera models - out there that he didn't raise support cases for...

Also: this isn't a curve/contrast issue: it's a fundamental - and I'll say it again, systemic - flaw in Phase One's profiling methodology, and this excessive orange/red bias persists to a greater or lessser extent regardless of the selected film curve applied. I've already noted that this bias has always been part of the Capture One "look", and it's in direct comparison to the previous - far more restrained - standard that I can be so resolute about the nature and significance of this issue.

All the "testing" I do - of any converter - is comparison testing; I never draw conclusions in the absence of robust, objective baselines against which to consider them. I'm not expressing an arbitrary, subjective opinion here, just as I didn't back when I raised concerns about the unacceptable "checkerboard" rendition of old Capture One's NR. That's fixed now....

Like Peter, I'm seriously wondering about the accuracy of the monitor calibration of the computers being used to create these profiles.

Does Phase One not use beta testers? This problem should have been caught ages ago in a decent beta testing regime.

Funnily enough, I've recently joined the official user beta testing group of another converter. I was asked to join precisely because I'd picked up some significant rendering/image quality issues that had got by the members of that group (identified their likely causes, too).

I'd have picked this one up, if I'd been a Capture One beta tester.

0

Michael Fritzen

February 04, 2015 12:42

Just out of interest, is the discussed problem a recognized issue over all, or at least a representative number of camera makers?
AFAIK a lot of Phase One's work is the re-engineering of the RAW-sauce each maker cooks. While I understand that for some camera models the problems are longer time already, I'm only wondering which side carries what responsabilities. Also understanding that dicscussing this doesn't get us anyway closer to solutions.

For my part I know only that in the very specific case of the Sony RX10 the output of the processing in its initial support of this camera showed distinct shifts to magenta most noticeable in shadow areas. This has completely gone (not sure if in V7 or in the transition to V8). What says to me that P1 is working on solutions for such issues. Why wouldn't they do so in the case of the discussed issue? May be other professionals shooting with the same equipment tell them to do differently. May be some of the equipment, for unknown reasons, represent a special challenge.

0

Colour Profile Problems Part 2

Comments

Didn't find what you were looking for?