Need help with correct colors
Hi,
I'd like to ask if someone could help me with default colors for sony nex cameras. I love how C1 handles skin tones and I'm seriously considering migration from LR to C1, but one thing about C1 bothers me a lot.
On the photos below you can see 2 versions of the same photo taken at Ait Benhaddou right after sunrise. The first foto is developed from a raw in LR 5.7 with all default settings and the second photo is developed from the same raw in C1 8.3.2 again with default settings.
Lightroom
Capture One
We can argue which one looks better, but the problem is that the walls/tower under the sunlight are way to brown/red in C1 comparing to the real colors I remember. LR doesn't represent colors with 100% fidelity either, but is much closer to what was the real ones.
And the situation repeats itself on almost every photo taken during golden hours where the subject is a limestone or clay. If fact almost all my photos taken during my trip to Morroco look like they were taken on Mars in C1.
I'd very much wanted to move to C1, but I need to figure out how to avoid that effect I described above.
Any help is very much appreciated!
Best regards,
Vlad.
I'd like to ask if someone could help me with default colors for sony nex cameras. I love how C1 handles skin tones and I'm seriously considering migration from LR to C1, but one thing about C1 bothers me a lot.
On the photos below you can see 2 versions of the same photo taken at Ait Benhaddou right after sunrise. The first foto is developed from a raw in LR 5.7 with all default settings and the second photo is developed from the same raw in C1 8.3.2 again with default settings.
Lightroom
Capture One
We can argue which one looks better, but the problem is that the walls/tower under the sunlight are way to brown/red in C1 comparing to the real colors I remember. LR doesn't represent colors with 100% fidelity either, but is much closer to what was the real ones.
And the situation repeats itself on almost every photo taken during golden hours where the subject is a limestone or clay. If fact almost all my photos taken during my trip to Morroco look like they were taken on Mars in C1.
I'd very much wanted to move to C1, but I need to figure out how to avoid that effect I described above.
Any help is very much appreciated!
Best regards,
Vlad.
0
-
[quote="VladimirCZ" wrote:
but the problem is that the walls/tower under the sunlight are way to brown/red in C1 comparing to the real colors I remember.
Some of us have been complaining for a long time about Capture One's excessive bias towards red/orange/brown - this overly warm shift. "Taken on Mars" sums it up nicely.
And we're still complaining.
You'll be told that you can "fix" the problem (you can't) by ramping down the reds in the Colour Editor, and saving the result as a new default profile; but what you (and we) really need is profiles from Phase One that are much less hot "out of the box".
What you'll get is denial that there's a problem.0 -
Vlad:
Sorry, you can beat me but from looking your examples I'm unable to see or say which one I'd consider more natural - or even further "more correct". Starting at the sky region both do look pretty neutral to me (considering the clouds) but the second seems to have a bit more overall colour saturation. If it's really a bit more saturation applied by the software as default on import or perhaps just a tad lower overall brightness levels is an uncertain factor.
From my POV one thing should be taken into account: "default" of RAW conversion SW 1 is not equal to "default" of No. 2 and not of No 3... Each "cook" makes their own "sauce" as how to interpret the RAW data from a certain camera brand and a certain sensor type. What can be checked and eventually adjusted / colour managed is the colour reproduction of certain defined patterns with measuring the screen reproduction (adjustments if necessary) and printed output reproduction (adjustments if necessary) so that the colours are kept as close as possible through the entire processing line.
However, this isn't what's meant with "default" of each RAW conversion SW. First off, in most cases the maker of the RAW conversion SW doesn't have access to the information of the sensor maker and their "sauce" of what processing they are submitting the data in the process of the RAW data generation. So this data is "interpreted" by the converter SW maker - and LR has a different interpretation from P1 and again different from DXO and so on.
Perhaps another point to consider is that "true" or "correct" colours may matter for reproduction of fabrics, artwork, colour catalogs but for most of the real life photography it's essentially a question about "what looks best" - in the eyes of the photographer. It could be perfectly an "envisioned" look. On the blog of the "Image Quality Professor" (http://blog.phaseone.com/) I've seen already several examples of how photos were almost "transformed" by their editing from an original version, pretty dull and not exactly appealing, into a stunning photo as the photographer had it "envisioned" when he/she pressed the release button, but weather, light and so on didn't allow to get close to the "envisioned" version at the moment of image taking. Not to say that this is the only way, but reducing the entire RAW-shooting-approach to an import with some default settings and expection "correct" colours isn't either. I think we may compare it to former film brands where some were considered "too warm", others "too cold / bluesih" but in the end all a matter of taste. In RAW we are able to have "all the film brands" as we are able to "process to taste" and in your case it may mean that you just need to reduce a bit the WB setting and/or increase just a tad the exposure correction (brightess; clarity) to get to the desired look. And if you have a series, all it needs is to correct the first to your liking and copy the setting to the other images.
Hope this helps.0 -
[quote="VladimirCZ" wrote:
Hi,
I'd like to ask if someone could help me with default colors for sony nex cameras. I love how C1 handles skin tones and I'm seriously considering migration from LR to C1, but one thing about C1 bothers me a lot.
On the photos below you can see 2 versions of the same photo taken at Ait Benhaddou right after sunrise. The first foto is developed from a raw in LR 5.7 with all default settings and the second photo is developed from the same raw in C1 8.3.2 again with default settings.
Lightroom
Capture One
We can argue which one looks better, but the problem is that the walls/tower under the sunlight are way to brown/red in C1 comparing to the real colors I remember. LR doesn't represent colors with 100% fidelity either, but is much closer to what was the real ones.
And the situation repeats itself on almost every photo taken during golden hours where the subject is a limestone or clay. If fact almost all my photos taken during my trip to Morroco look like they were taken on Mars in C1.
I'd very much wanted to move to C1, but I need to figure out how to avoid that effect I described above.
Any help is very much appreciated!
Best regards,
Vlad.
Is it possible for you to also share the raw-file with us ?0 -
I am by no means an expert, but I find all 3 of his samples to be underexposed and slightly over-saturated. And when I look at them in C1, the exposure evaluation does show that. I would think that would make it difficult to judge the reds. And if you do an auto exposure adjustment, it bumps the exposure up quite a bit.
Feel free to correct me if I am making some wrong assessments.
Sam0 -
What you'll get is denial that there's a problem.
Keith had well said it.
The other reply you will possibly get is that if you are serious about color accuracy you should have gone your own way to have a spectrometer, ba ba ba and created your own camera profile.
It was my wonder that since CO supports multiple profiles for the same camera, P1 does not need to give up the existing profile that P1 thougth it was the best. Instead they can supply another color-accurate profile for users to choose. Color accuracy is an objective thing that can be quantatively measured.0 -
@Christian Gruner: Sure, I will provide the RAW as soon as I'm back home.
@esl1885: No any adjustments were applied to the first 2 photos. For the third one only those 2 that were already described.0 -
Urukhai:
From my POV you're right - and wrong at the same time. Colour certainly can be measured and such been reduced to numbers. But they have also a very subjective nature, stimulate emotions, are pleasing - or not. And in photography we are hardly working with one only colour / hue but with a composition and their distribution over the image.
Another point is that we would have pretty similar discussion and possibly critics if P1 would assure 100% correct colours according to "measurements".
I'd add that P1, in order to have 100% lucky customers, most probably would need to provide a number of profiles as a result of the multiplication of the number of customers and the number of different cameras they own. 😊0 -
Vlad:
None of us here has seen the real scene other than you, but again you can beat me as I don't find anything substancially wrong with either version. One could argue forever whether the LR version is too cold for evening light or the C1 version is too warm. I opened the RAW file in C1Pro and just for fun clicked on the WB tool on the Auto button and at least on my screen then the colours get pretty similar to the rendering of LR, at least less warm with 4617K/2,7 tint. Perhaps C1 does really interpret the WB metering of a Sony cam a bit too warm as its own AutoWB would suggest a bit colder. Yet WB continues for me a very personal choice / adjustment.0 -
[quote="Michael11" wrote:
Vlad:
None of us here has seen the real scene other than you, but again you can beat me as I don't find anything substancially wrong with either version. One could argue forever whether the LR version is too cold for evening light or the C1 version is too warm. I opened the RAW file in C1Pro and just for fun clicked on the WB tool on the Auto button and at least on my screen then the colours get pretty similar to the rendering of LR, at least less warm with 4617K/2,7 tint. Perhaps C1 does really interpret the WB metering of a Sony cam a bit too warm as its own AutoWB would suggest a bit colder. Yet WB continues for me a very personal choice / adjustment.
I feel pretty much the same.
The C1 default Blue sky looks pretty good - the LR seems to have some magenta .... but that could be correct I guess.
The adjusted C1 sky looks very nice for the blue.
There are parts of the image that change dramatically with small adjustments. It seems extremely sensitive to tweaking as Vlad's very small adjustments to get an acceptable result have shown.
C1 pulls out some interesting colour details for some of the features mostly in the shadows on the left.
I really doubt I would have remembered half of the lighting effect had it been my shot. For this sort of image (not product with required colour accuracy and no skin tones) the end result seems to me to be more important than attempting accuracy that would be difficult to verify. It's the sort of "dominant colour" image that has many possibilities.
I took some pictures at dawn at Ayres Rock in Australia a few years ago. It is a very similar colour to the walls here, had a clear blue sky but also some green vegetation. As the sun appeared and moved the colours were constantly changing. I have no idea which are correct. Perhaps all of them. Perhaps none of them.
Grant0 -
Michael, Grant:
Thanks for the input!:)
C1 pictures look superior to LR, skies, shadow areas, except wall/tower parts. The problem is that both the wall and the tower under sunlight look like it was raining right before the pictures were taken, they are a bit too much brown/dark/saturated.
For the photo from Morocco I can accept that it was 2 years ago and I cannot check the colors on site, but the tower shot was taken in my home city Prague on my usual walk from work to home (usually right before sunset) and I can say for sure that the tower is rather yellow under that lightning conditions.
I like output from C1 much better for most parts except for sunlit clay/lime stone photos taken during golden hours, which is a substantial part of my shots.
Very slight changes helped with the shot from Morocco, but this trick didn't work for the shot from Prague.
I do hope that I will find a solution to this problem.
Vlad.0 -
Vlad:
Isn't "getting it as desired" the main purpose of all that "messing around with RAW"? 😊
Especially when it comes to the WB?
Perhaps I wouldn't go that far to compare a simple import to C1 (and expecting just everything fine) to normal in-camera JPG shooting. But besides, of course, keeping all the captured data which came from the sensor, it's the work on the RAW data which (at least as a potential) sets it apart from the in-camera "ready-to-go" processing.
Far from entering an argument about if I'm doing better than "all the expertise of the camera maker's engineering team" when developping their JPG engine, it's a step of the processing which I like to do by myself. Let alone the question the amount of valueable data which is discarded (prematurely from my POV) in the in-camera JPG processing. Valueable data for snaps? Possibly. Perhaps I'm a bit on the side of a control freak and like to be in command. But for each their own.0 -
Vlad,
Firstly let me say that I understand what you mean by about the Prague shot and I get the same results as you when I open it. That's a good sign.
The LR version has the slightly odd looking sky colour - although it may in fact be reasonably correct of course - and the roof tiles on the tower look a little flat or muddy to me but that could also be correct! The C1 colours are a little stronger.
I had a look at changing the Base Characteristics curve to Extra Shadow and thought that was a better setting for the image which, according to the Histograms, seems a little underexposed by C1's assessment. No surprise, there is a lot of dynamic Range to try to deal with.
Using Auto Exposure with White Balance, Exposure and High Dynamic Range active also seemed to make a useful adjustment.
Similarly using the Levels tool (before any other adjustments in order to see its basic effect) with the Preferences set to work on the red, green and blue channels individually, also looked good to me.
So there are some internal adjustments that C1 seems happy to apply that go beyond the initial formula based interpretation that by nature will be somewhat generic and presumably will have been created based on "typical" and "Averaged" content.
I think on the basis that the small tweaks applied for the Morocco shot did not seem to have the same good effect on the Prague shot goes some way to illustrating how tricky it can be to get a generic balance that will work in all cases as a reasonable compromise.
Personally I always expect to make some notable changes to an image to get the best out of it (in my opinion of course!) but then my cameras seem to produce files that are never strongly affected by anything in the initial interpretation that I would consider to be unwelcome.
If they were I think I would be making a lot of use of customized Styles!
I would suggest you raise a Support Case with a suggestion that the base processing might be re-considered and an alternative offered since this might be useful feedback to Phase once "officially" reported.
Grant0 -
Vlad:
The last paragraph in Grant's post is very good advice because from my own experience with the Sony RX10 in its first time of C1 RAW support I know that things may need some tweaking. In the case of the RX10 it was a magenta cast in deep shadows which P1's engineers were able to eliminate in a following program update.
BTW did you take some of the shots parallel in RAW+JPG? It would be interesting to see how the in-camera processing would interpret the scene / lighting situation. Perhaps some shift is happening as the lighting gets warmer (evening). I'm saying this because I remembered the weird numeric AWB values from my RX100M1 which is always around 3,500K for daylight / sunny (as opposed to perhaps expected 5,500K). The look of these images in C1 is just fine but certainly some adjustment was needed to get a match. Now what if there would be some type of similar "derivation" on the A6000 from camera's numeric AWB values to values or a curve where it "should be" as the lighting gets warmer (evening)?0 -
Grant,
I think you nailed it. It seems that LR doesn't bother about shadows during import and leaves it to a user, while C1 tries to recover shadows to make an image, that contains both deep shadows and bright sunlit areas, more balanced right from the start. As a side effect of this kind of processing is oversaturation of the red channel in midtones.
So far, as it was advised by the forum members, changing the curve to Film Extra Shadow is the most universal way to make oversaturation less dramatic, but its not panacea and additional steps are required.
Vlad.0 -
Michael:
Sorry, but I have no OOC jpegs for those 2 raws.
Vlad.0 -
Vlad:
At least I'd be intrigued to have a look at some OOC JPGs just for a comparison of what the AWB and Sony's processing does. Perhaps something for a coming outing. Thinking better, there's another option: running your RAW files through Sony's IDC since it should provide a very similar look as the in-camera processing based on all the info stored in the RAW file (w/o further adjustments in IDC obviously).0 -
Michael:
I've tried Sony IDC, but the results are weird. The images are washed out and the corolrs are way off. I'd probably need to spend more time with IDC settings, but the interface is putting me off.
I'm seriously considering buying X-Rite ColorChecker Passport, but I'm not sure if I can create ICC profiles for C1 with it.
Vlad.0 -
Vlad:
Well, your experience with IDC's UI reflects 100% my own. 😊
But from time to time I see popping some pretty positive comments on other forums about how easy good results can be achieved in IDC. The given explanation is that "Sony knows what's going on with the data and the files..." Since I didn't have otherwise own solid experience I normally prefer not to rise doubts on that - but I neither didn't consider to give IDC another try after some attempts back in 2012 when I got my A99 and it still wasn't supported by C1.0 -
CO does not support the DCP files produced by x-rite. It has come up many times and as far as I can tell is falling on deaf ears. So many people complain about the colors that CO produces, you would think they would support it just to try to shut us up.
I understand the difference between an ICC profile and a DCP one, and it might work with their engine flow, but as they seem to be getting a larger user base it might be good for them to talk to x-rite and see what they could do to help the situation.
Robert0 -
[quote="VladimirCZ" wrote:
Michael:
I've tried Sony IDC, but the results are weird. The images are washed out and the corolrs are way off. I'd probably need to spend more time with IDC settings, but the interface is putting me off.
I'm seriously considering buying X-Rite ColorChecker Passport, but I'm not sure if I can create ICC profiles for C1 with it.
Vlad.
Vlad, you absolutely do not need an obsolete software from X-Rite... and in your specific case with Sony NEX 5N you even do not need a target...
1) http://www.ludd.ltu.se/~torger/dcamprof.html
2) http://forum.luminous-landscape.com/ind ... c=100015.0
3) spectral data for your camera from http://www.cis.rit.edu/~dxl5849/projects/camspec/
SONY NEX-5N = RED, GREEN, BLUE = from 400 to 720nm in 10nm steps :
0.0073643 0.055237 0.051835 0.043321 0.033681 0.028963 0.025 0.031318 0.034458 0.033059 0.036819 0.04791 0.075209 0.083833 0.040534 0.025882 0.029435 0.096315 0.40981 0.59492 0.48306 0.4419 0.32839 0.26576 0.18549 0.14261 0.10679 0.073901 0.045969 0.011987 0.0019017 0.00074491 0.00024202
0.0087465 0.075143 0.10312 0.12927 0.14942 0.19179 0.24858 0.37672 0.45492 0.50906 0.67036 0.85874 0.93855 1 0.87676 0.85895 0.66097 0.54573 0.3987 0.29141 0.14386 0.086686 0.046569 0.032661 0.021849 0.016681 0.013952 0.012859 0.011204 0.0038035 0.00078707 0.00032949 0.00013348
0.033016 0.35254 0.49797 0.60149 0.65736 0.78336 0.73926 0.776 0.72276 0.61365 0.4557 0.30535 0.17946 0.12249 0.073361 0.04487 0.020265 0.012304 0.0085751 0.0064746 0.0032885 0.002511 0.0020362 0.0023356 0.0027482 0.0036074 0.0036988 0.0031634 0.0022407 0.00069044 0.00012112 6.3737e-05 4.2452e-05
that's it... you are free to create any ICC or DCP profiles now
if you are on Windows platform and can't build binaries yourself then you can use what I supply in the topic @ LL : 0.9.3 build for Windows (mingw = dcamprof.exe + libgomp_64-1.dll + HTML manual, no PDF this time) : https://app.box.com/s/74jijebztt5nupgypmvqzqq2i5p4i4yo
if you are on Mac and can't build yourself, ask here - I think I know people who do build it for mac, so that I can get that for you.
if you are on some linux and can't build yourself then you can ask A.Torger, the author, himself for binaries.0 -
Hi folks
I have seen a few posts by dja on colour issues. Am I the only reader for whom all this technical stuff is way over my head?
I assume dja knows what he is writing about, but it is all gobbledegook to me. No offence, and no criticism implied. But what does it all mean, and where are the simple outputs such as colour profiles created by the dja procedures?
The fact remains that judging by the amount of interest in the forums one of THE most problematic areas for Phase is raw colour conversions and colour rendition in output files. Perhaps it will ever be thus?
Peter.0 -
[quote="Peter" wrote:
Hi folks
I have seen a few posts by dja on colour issues. Am I the only reader for whom all this technical stuff is way over my head?
I assume dja knows what he is writing about, but it is all gobbledegook to me. No offence, and no criticism implied. But what does it all mean, and where are the simple outputs such as colour profiles created by the dja procedures?
The fact remains that judging by the amount of interest in the forums one of THE most problematic areas for Phase is raw colour conversions and colour rendition in output files. Perhaps it will ever be thus?
Peter.
you need to invest of your some time into actually reading the manual that A.Torger provided and the discussion @ LuLa forums... a lot of other people did contribute things (software, methods, even CMF/SSF data, etc) that you at least can try to simply pickup - nobody is asking you to discover america on a small boat across the ocean, just bother to research a price and buy a ticket, so to say...
just do it slowly, if you have questions - join the discussion @ LuLa0 -
deejjjaaaa:
Thanks a lot for the useful information. I will have a look at the solution you recommended.
Vlad0 -
Michael:
As you recommended I took few snaps in a RAW+JPG mode. Additionally, I tried to play with the DXO tiral as well.
You can see the results of my experiment:
Original RAW:
Original OOC jpeg:
C1 default output:
DXO default output:
On my monitor DXO is closet to reality and OOC jpeg is second. C1 output is way too yellow/brown. This was also confirmed by my wife and she's very sensitive to colors.
After playing for few weeks with C1 I came to a conclusion that the program generates most pleasant skin tones. People look sun-tanned which is nice even if not always true, but for this we have a heavy price to pay. Almost in every landscape photo mid-tones are heavily shifted toward red/brown hues and it seems there is no easy way to fix it.
I do hope that P1 somehow will fix it in the new versions of color profiles for A6000.
Vlad.0 -
[quote="VladimirCZ" wrote:
Michael:
As you recommended I took few snaps in a RAW+JPG mode. Additionally, I tried to play with the DXO tiral as well.
You can see the results of my experiment:
Original RAW:
Original OOC jpeg:
C1 default output:
DXO default output:
On my monitor DXO is closet to reality and OOC jpeg is second. C1 output is way too yellow/brown. This was also confirmed by my wife and she's very sensitive to colors.
After playing for few weeks with C1 I came to a conclusion that the program generates most pleasant skin tones. People look sun-tanned which is nice even if not always true, but for this we have a heavy price to pay. Almost in every landscape photo mid-tones are heavily shifted toward red/brown hues and it seems there is no easy way to fix it.
I do hope that P1 somehow will fix it in the new versions of color profiles for A6000.
Vlad.
Hello Vlad,
Let's not forget that DXO at "default" settings does a lot of auto correction on the image, including EV correction. Your particular image looks underexposed here, it is a very bright sunny setting, and the CO1 image clearly suffers from the underexposure. To do a realistic comparison between DxO and CO1, you could either disable DxO's auto corrections, or use the "A" (auto) correction with CO1. Notice that Sony cameras (I have the Sony A7r), tend to underexpose. I always need +0,7-+1 EV compensation on the A7r.
Chris0 -
Let me join in here with another example. It is just an illustration and certainly not suitable to derive exact parameters. But I do see the same problems as the OP and there is a definite pattern in all images coming out of C1. It is a pity as it effectly means I can't use C1 to process my pictures. But back to my stupid out of focus example. It is a color chart I printed ages ago on my inkjet printer. On the top there is also a white card and an 18% grey card but I didn't use them to set any parameter. I shot RAW+JPEG and processed with default settings in C1. Taken with NEX-6.
Oops, I just realize I can't attach images to posts on this board, is that true? I don't have online storage. So, hm.
Anyway, I find that the jpeg out of camera is much closer to what the printout I photographed looks to me. This got nothing to do with taste or what looks better or a precise photometric reproduction. It is just what *looks* closer to the real object.
I'd really appreciate if C1 could take care of this isssue. I really want a solution.
If I find time I might look at what deejjjaaaa suggested.
Michael0 -
[quote="deejjjaaaa" wrote:
Vlad, you absolutely do not need an obsolete software from X-Rite... and in your specific case with Sony NEX 5N you even do not need a target...
1) http://www.ludd.ltu.se/~torger/dcamprof.html
2) http://forum.luminous-landscape.com/ind ... c=100015.0
3) spectral data for your camera from http://www.cis.rit.edu/~dxl5849/projects/camspec/
SONY NEX-5N = RED, GREEN, BLUE = from 400 to 720nm in 10nm steps :
0.0073643 0.055237 0.051835 0.043321 0.033681 0.028963 0.025 0.031318 0.034458 0.033059 0.036819 0.04791 0.075209 0.083833 0.040534 0.025882 0.029435 0.096315 0.40981 0.59492 0.48306 0.4419 0.32839 0.26576 0.18549 0.14261 0.10679 0.073901 0.045969 0.011987 0.0019017 0.00074491 0.00024202
0.0087465 0.075143 0.10312 0.12927 0.14942 0.19179 0.24858 0.37672 0.45492 0.50906 0.67036 0.85874 0.93855 1 0.87676 0.85895 0.66097 0.54573 0.3987 0.29141 0.14386 0.086686 0.046569 0.032661 0.021849 0.016681 0.013952 0.012859 0.011204 0.0038035 0.00078707 0.00032949 0.00013348
0.033016 0.35254 0.49797 0.60149 0.65736 0.78336 0.73926 0.776 0.72276 0.61365 0.4557 0.30535 0.17946 0.12249 0.073361 0.04487 0.020265 0.012304 0.0085751 0.0064746 0.0032885 0.002511 0.0020362 0.0023356 0.0027482 0.0036074 0.0036988 0.0031634 0.0022407 0.00069044 0.00012112 6.3737e-05 4.2452e-05
that's it... you are free to create any ICC or DCP profiles now
I tried that but got stuck quite early. Could you lend me hand? This is what I got:
Copied the data above into a file nex5.txt.
Ran./dcamprof.exe txt2ti3 -f 400,720,10 nex5.txt nex5.ti3
and got nex5.ti3.
Now I am lost. The manuals and post (039 pages) were just a bit too much. Any suggestion?
Thanks,
Michael0
投稿コメントは受け付けていません。
コメント
55件のコメント