The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Capture One or LR6?

dougpeterson

Workshop Member
With the round-tripping part, you lost me. What exactly do you mean? TIA
Previous versions of C1 didn't have the current robust set of tools to round trip through Photoshop and back. In the current version, you can "edit with" in Photoshop or other 3rd party application, return the resulting tiff and see the results within c1. You can also view layered tiffs, and through process recipes control the naming, and location of the resulting round trip'd file.
 

torger

Active member
I stopped using C1 when I found out it did not support Pentax MF files and that the round-tripping from PS or other filters (Silver Efex Pro etc) was not as robust as with LR.

I only print via QImage and as discussed here before, it does a much better job (IMHO) than either LR or PS.

I am not particularly fussy about color accuracy as long as I see what I like on the screen and the print matches it. I think color is a highly subjective aspect of an image. How does one know what the scene looked like, unless you put in a Macbeth color chart in every shot? Skin tones too are so variable, but arguably easier to 'correct' since we intuitively know what they should look like. Landscapes, buildings, objects, flowers, who is to say what the correct RGB value should be or the saturation?

Overall, I am quite happy with LR at present, it could certainly improve in certain areas, but it checks most of the boxes for me.
Just to note, Capture One profiles are not designed for accuracy. It's designed for a pleasing look, and is as such highly subjective.

Color accuracy using established scientific models is only possible for reproduction use case, where you copy say an painted artwork by photographing it under controlled light and print it. No contrast curve is involved in that process.

For a real scene you need to add contrast for the output medium otherwise it will look dull, as the output medium generally has lower dynamic range and brightness. This is a normal psychovisual phenomenon. Adding a single global S-shaped contrast curve is a bit simplistic but very effective, and that's what we've been doing since the film days and what the camera profiles still do today.

When we add contrast many things happens to the colors which needs compensation, and there's no standard to follow, each manufacturer does it differently. Pleasing skintones may be more about how the tones are rolled off gradually into the whitepoint and how saturation is controlled in the shadows than about the hue itself. All is controlled by the camera profile. The camera profile can also to some extent control color separation, it can be increased in some ranges and decreased in others. One example is that some profiles desaturates close-to-neutral colors to unify neutrals but also separate them from others. Various more or less subtle subjective adjustments that add up to a total color behavior and look.

I guess my point is that there's much more to pleasing color behavior than what the hues are, and the bundled Lightroom profiles are generally less well designed than bundled Capture One profile, at least if you ask me. It's all subjective though. It think one can objectively say though (not with 100% certainty though as I don't have all facts) that Adobe's profile design methods are much simpler than Phase One's, but as color is about taste that does not guarantee that you will like their color better.

Personally I just make my own profiles, and I love the freedom I get from that, I don't feel locked in to a particular manufacturer's look, and while accuracy is not really possible (as discussed above), it is possible to make colors look more realistic than most manufacturers provide with their bundled profiles, and to me that doesn't hurt. It's a good reference point. I rather have that than my colors change radically as soon as change camera brand or in some cases even just camera model. Then I can on top of that base look make color adjustments for creating atmosphere/mood or whatever the subject and context requires.
 

ErikKaffehr

Well-known member
Re: Capture One or LR6? Another sample

Hi,

Today I did run some tests for some issue with one of my Sony lenses but I have also tested one of the Contax Zoom lenses I have for the Sony. Day was clean blue sky with very consistent but harsh light. In this shoot I also included the Planar 100/3.5 and my Canon 24-105/4L. The outcome was sort of nice. The issue with the Sony lens was confirmed. The Contax 35-135/3.3-4.5 is remarkably good and the Canon 24-105/4L is no slouch at all. What about the Planar 100/3.5, it really shines at it is totally consistent across the field.

But, this posting is about colour rendition. Check the screen dump below, P45+ with C1 on the left, Sony A7rII in LR at center and P45+ with LR on the right. Crop has been chosen to get some interesting colour so no idea to compare sharpness…

Some observations:

  • LR uses DCamPro profiles
  • Colour rendition is a bit different
  • The images processed in LR are pretty close, but the yellow wall is more intense on the P45+in LR
  • Yellow wall is more similar between P45+ and Sony
  • Small flowers blue in Capture One and bluish purple in LR

Original screen dump: http://echophoto.dnsalias.net/ekr/Articles/Shoots/DivSamples/LR_VS_C1/Screen5.png


Best regards
Erik
 

Jay Emm

Member
A few examples of C1's latest version here - might be some tools amongst this lot that assist your needs???

http://blog.phaseone.com/top-5-reasons-switch-capture-one-pro-9/

FWIW, i've found tethered mode in C1 much better than LR, but LR is "free" with my PS sub so I have stuck with it. On the other hand, LR is slow and the modular interface isn't appealing after many years on Aperture (1st world problem I know, to be fair LR does it's job just fine).
 

dchew

Well-known member
Re: Capture One or LR6? Another sample

Some observations:

  • Small flowers blue in Capture One and bluish purple in LR

Best regards
Erik
Erik,
I'm very curious about the blue because that is one difference I've seen with C1: blue is blue and purple is purple. I can't get the same distinction in LR (is this what Torger means by "separation"?). Do you also find that to be the case?

Dave
 

GrahamWelland

Subscriber & Workshop Member
The obvious question to Erik is what colour were the blue/purple flowers in real life?

A common problem with digital files and raw rendering is the classic bluebell/purple bell/pink bell flowers. In the UK it was something that would drive me crazy as some cameras and their raw converters produced distinctly different colours for the same shot. Now we know of course that the human eye is sometimes fooled by the fact that what the camera captures and the eye sees/processes can be very different both in absolute terms and in context of other colours and tones.
 

ErikKaffehr

Well-known member
Hi Graham,

A very good question. The best way is to go back and perhaps collect a sample and measure with a spectrometer.

Still, I am pretty sure that they were something like deep bluish purple, because I know that colour renders blue in Capture One and I made mental note when shooting the pictures to check how they will render in C1.

  • These colours on "the purple line" are tricky as there are no corresponding spectral colours.
  • What is interesting here is that it clearly indicates that it is colour profile and not sensor differences are causing the problems.
  • Just to say, Jack would say that it is easily fixed by the colour editor, that is quite true.
  • What is interesting here may be that the Sony and P45+ sensors are quite close, when using profiles generated by the same tool under same conditions.

I have made an experiment a few years ago, under more controlled conditions. In that experiment the colour of the bluish deep purple petals was measured:

http://echophoto.dnsalias.net/ekr/Articles/OLS_OnColor/SimpleCase/

P45+ C1 (daylight)P45+ LR (dual illuminant, using DNG Profile Editor)
Spectral samples, petalsSpectral samples, blades



Best regards
Erik


The obvious question to Erik is what colour were the blue/purple flowers in real life?

A common problem with digital files and raw rendering is the classic bluebell/purple bell/pink bell flowers. In the UK it was something that would drive me crazy as some cameras and their raw converters produced distinctly different colours for the same shot. Now we know of course that the human eye is sometimes fooled by the fact that what the camera captures and the eye sees/processes can be very different both in absolute terms and in context of other colours and tones.
 
Last edited:

jng

Well-known member
Hmmmm... I was messing about in the garden a few weeks ago with my V system 120mm f/4 Makro-Planar mated to a Flexbody and IQ160 back. Below is a screenshot comparing the same file processed in either C1 or ACR. Nothing very artistic, mind you, but I thought the intense colors of the fuchsia would provide an interesting test case for the present thread. The C1 conversion (left) was made using C1's stock daylight profile, whereas the ACR conversion (right) used a custom dual illuminant profile. Color balance was set using a grey card as reference. The C1 file was exported as a tiff and for the sake of this comparison both files were opened up in Photoshop with no other adjustments.

Over all I'd say that the color rendition is pretty similar. Per Graham's comment the purple part of the flower is actually purple in both files and matches pretty closely to the real thing. It may be difficult to see in this screen grab, but as with other images I've processed I feel that there's more texture (for lack of a better word) in the C1 conversion's midtones. The difference is admittedly quite subtle, in any case.

- John
 

Attachments

torger

Active member
Good example showing that differences are generally quite subtle, because that is what's generally the case when comparing profiles. This means that many just say "why bother?" and use whatever at hand, and I think that's pretty valid. There are differences though.

What you should do when comparing is to put both images on top of each other in separate layers, and then toggle on/off, differences become much more clear then. It's much more difficult to see differences side by side.

When doing so one can see that the LR rendering is a bit brighter and more saturated, and thus lose a bit of texture. One can also see that on the purple part of the flower C1 works more with lightness to enhance texture, while LR keeps more saturation, which loses texture but may be more "accurate". This is related to LR's RGB-HSV curve, which I don't think has that good properties in the rolloff to whitepoint. In DCamProf I have chosen an approach more similar to C1 in this regard. The hues differs a fair bit too, although they are in the same ballpark. C1 has more yellow as usual. Without knowing I would guess that LR is a bit more realistic concerning hue.
 

ChrisLivsey

New member
The obvious question to Erik is what colour were the blue/purple flowers in real life?

A common problem with digital files and raw rendering is the classic bluebell/purple bell/pink bell flowers. In the UK it was something that would drive me crazy as some cameras and their raw converters produced distinctly different colours for the same shot.
Blubells in the UK - tell me about it !!!

D500 Capture NX-D



P20 C1 v7



Fuji Pro400H 35mm Plustek OpticFilm 8100
 

ErikKaffehr

Well-known member
Hi,

Sometimes difference are sublime, sometimes not.

Lets consider some cases:

  • You are a studio photographer, adjust your settings for taste, save as a profile/setting and are happy forever.
  • You are a landscape shooter, you get what you get and are happy for life. (that's me)
  • You are a botanist shooting flowers for a book…
  • You are a landscape painter, taking care to match paints to scenery. You train your vision to interpret colours...

Different scenarios, all feasible…

Best regards
Erik



Good example showing that differences are generally quite subtle, because that is what's generally the case when comparing profiles. This means that many just say "why bother?" and use whatever at hand, and I think that's pretty valid. There are differences though.

What you should do when comparing is to put both images on top of each other in separate layers, and then toggle on/off, differences become much more clear then. It's much more difficult to see differences side by side.

When doing so one can see that the LR rendering is a bit brighter and more saturated, and thus lose a bit of texture. One can also see that on the purple part of the flower C1 works more with lightness to enhance texture, while LR keeps more saturation, which loses texture but may be more "accurate". This is related to LR's RGB-HSV curve, which I don't think has that good properties in the rolloff to whitepoint. In DCamProf I have chosen an approach more similar to C1 in this regard. The hues differs a fair bit too, although they are in the same ballpark. C1 has more yellow as usual. Without knowing I would guess that LR is a bit more realistic concerning hue.
 

torger

Active member
Animated GIFs only support 256 colors so the image below is a dithered mess, but I just wanted to show how much easier it is to spot differences between two different renderings when you layer them on top and toggle between them. Here JNG / John's images from a few post backs comparing LR and C1 renderings:

C1-ACR-comparison.jpg

To make a comparison with less difference between the two profiles one should try to make the difference in overall brightness as small as possible, but it's usually not possible to get exactly the same as it varies between profiles how light various hues are rendered, plus that the contrast curves can differ a fair bit.

It can be interesting to see how brightness varies despite the same settings too, as it can vary a fair bit that too between raw converters. A good idea when making a camera profile is try to match the in-camera JPEG brightness, as the camera's auto-exposure is tuned for that.
 

torger

Active member
Just a few example of profile differences seen in John's image (C1 shown):

C1-ACR-comparison-1.jpg

A) Color separation is not necessarily achieved by separating hues more, you can also separate them with lightness. Here in the C1 profile we see that the browns in the bud are rendered much darker than in the LR version, and thus there's more separation. Another trick I've seen in some profiles is to add in say a little magenta into the oranges/browns between greens and reds. This way separation is increased between greens and browns which can be beneficial in landscape photos.

B) Here we see on along the edge of the flower that the purple color becomes light and desaturated, while LR keeps more saturation. Bright saturated colors easily clips the gamut, and then you lose texture/tonality, it becomes more flat. A subjective adjustment used by C1 here is then to desaturate and lighten those colors rather than just clipping them. Less accurate perhaps, but usually a more pleasing result. Overall we can see that the C1 profile is less saturated, almost certainly a bit under-saturated, but it makes rendering flowers much easier/robust.

C) This area is a variation of the same theme, less saturation, more modulation with lightness => texture more visible.

Not seen in this particular image, but another very common subjective adjustment is to render deep blues much lighter than realistic, for the reason that our eyes have poor sensitivity in the deep blue range and thus texture/tonality gets flat unless we lighten it.
 

jng

Well-known member
Animated GIFs only support 256 colors so the image below is a dithered mess, but I just wanted to show how much easier it is to spot differences between two different renderings when you layer them on top and toggle between them. Here JNG / John's images from a few post backs comparing LR and C1 renderings:
Layering the two versions to facilitate comparisons is a nice tip. I will try this the next time I'm deciding between two different versions of the same image. Thanks!

In this particular comparison I was frankly surprised at how similar the two conversions are. I don't usually keep all possible parameters the same as I'm generally more interested in the final look of an image than in making A-B comparisons between the different software. I rarely use ACR anymore for converting my P1 back files as I've grown to prefer my results with C1. But this is entirely subjective.

John

EDIT: I missed your more detailed comparison before posting my reply. Yes, the differences you highlight - which can be more dramatic in other images I've worked on in the past - are in large part why I prefer the C1 conversion although I suppose futzing more in ACR and PS could offset some of the differences.
 
Last edited:

ErikKaffehr

Well-known member
Re: Capture One or LR6? Another sample

Hi Dave,

I have seen this before, not least when scanning film, that the purples are sort of tricky (*). It is easily done that a profile turns them into red or blue. It may be argued that a good profile should keep purples purple.

What I have seen is Capture One is pushing saturation on blues (and greens) quite aggressively. Check the colour patch in the enclosed image. It is a comparison of a shot of an IT8 chart with it's reference values. the blue field here is about Delta E 16.8 off it's target value, but the patches are very close. How come? C1 pushed the colour widely outside the gamut of Adobe RGB, which is about what my screen can show. So there is a large difference in colour, just that my screen cannot show it.

Screen Shot 2016-07-23 at 17.35.29.jpg
Best regards
Erik

(*) I am in a bit of hurry, but I can try to explain it a bit later…

Erik,
I'm very curious about the blue because that is one difference I've seen with C1: blue is blue and purple is purple. I can't get the same distinction in LR (is this what Torger means by "separation"?). Do you also find that to be the case?

Dave
 

torger

Active member
High saturation colors is a mess to handle for cameras and camera profiles. The native raw colors of a camera is generally quite desaturated, so the profile needs to separate the channels more, thus increasing noise. With high saturation colors only one or two raw channels have much signal, so the process becomes more unstable. Saturated colors also exaggerate the difference between the human observer and the actual camera filters.

Simply put, they are more difficult to get right.

And then after hue has been determined there's the problem of gamut clipping. If raw converters had been designed in a modern and color scientific way the camera+profile would just mimic a "colorimetric measuring device" and the raw converter would take care of all contrast handling, gamut compression, clipping etc, but that's not how it works -- the profile needs to handle it all (although inelegant from an engineering perspective it's not all bad though, as it means that a custom profile can control the fundamentals of color in any raw converter supporting ICCs or DCPs and thus "fix" the "poor" rendering of LR for example). Anyway, compressing the gamut and clip while keeping texture and maintaining the feeling of a bright saturated color is not an easy task.

Purple is a special challenge due to it's mix of red and blue, with many cameras it's like balancing on a knife's edge; you have relatively weak blue and weak red and that should point somewhere along the line of purples, just a small change makes it jump a large distance, that is purples can be even more unstable than other saturated colors.

Differences between camera's hardware responses (in terms of color filters) also become more evident with saturated colors, and it's harder for a camera profile to "fix" things in that range.

I'd love to do some more experimentation with high saturation colors, as I think it's an area I could improve DCamProf, or at least improve my understanding. It's more difficult to experiment with it though as saturated colors of high quality is harder to come by. Flowers are excellent, but they eventually fade :-/
 

tjv

Active member
Anders, the more I read your comments the more I realise that I need to give your profiling tool a try...

Is it easy to print a calibration target and run the profiling process, or does one need a commercially available patch?
 

ErikKaffehr

Well-known member
Hi,

I think it is preferable to go with ColorChecker. The standard one is just fine.

If you print a target of your own you need to measure it. A printed target will just contain different mixes of the inks, while the ColorChecker has a wide mix of spectral responses. Anders suggests the normal ColorChecker is a pretty good starting point.

Shooting the test target takes some care.

DCamProf is now command line driven, no problem for Linux jerk, like me, but may be an obstacle for some others.

As a Lightroom user, I would start off buying a ColorChecker and use it with Adobe DNG Profile Editor or ColorCheckerPassport software. Both are free. After that I may look into using DCamProf.

Which back do you have? It wouldn't be difficult to make a decent single illuminant profile if there is some decent quality test chart shot with that back including a ColorChecker.

Best regards
Erik



Anders, the more I read your comments the more I realise that I need to give your profiling tool a try...

Is it easy to print a calibration target and run the profiling process, or does one need a commercially available patch?
 

jng

Well-known member
Hi,

As a Lightroom user, I would start off buying a ColorChecker and use it with Adobe DNG Profile Editor or ColorCheckerPassport software. Both are free. After that I may look into using DCamProf.

Which back do you have? It wouldn't be difficult to make a decent single illuminant profile if there is some decent quality test chart shot with that back including a ColorChecker.

Best regards
Erik
I made a dual illuminant profile for my IQ160 using a Colorchecker Passport and Adobe DNG Profile Editor. Although I still prefer C1, the custom profile has markedly improved my conversions in ACR. I originally tried generating a profile with the X-Rite ColorCheckerPassport software (both single and dual illuminant) but found that it couldn't handle the file size generated by the 60 Mp back. This was back in November 2014 - X-Rite tech support advised me that, at least at that time, their software could not handle files larger than 30 Mb, which rules out using this software for uncompressed IQ160 files.

tjv: I recall that you're using a Credo 60 so this limitation likely applies to your situation as well. The Adobe DNG Profile Editor works fine, in any case. And given my inability to use command line, Anders' tools are beyond my capabilities unless I get some professional help! :p

John
 
Last edited:

torger

Active member
I can make you a DNG profile in five minutes if I just get a raw file with a CC24 shot in your preferred light, something similar to D50 is a good all-around light. Minimized glare, and even light, minimized perspective distortion of the target, and I need to know if the colorchecker was made before or after november 2014. A colorchecker passport is okay too.

Making a C1 icc profile is a bit messier, but you can make a quick test in LR first to see if you like the result, and if so we can make a C1 ICC later.

Making a basic DCamProf profile is not that difficult, but if it's the first time you use the command line it can look a bit scary. The "easy way" to make a DNG profile is described here: Making a camera profile with DCamProf and there's a corresponding section for Capture One.

Currently I only provide source code, but builds of the DCamProf program for mac and windows can be found in the Lula DCamProf thread, OSX link OSX lula message and Windows link Win lula message
 
Last edited:
Top