Color model and more. Article from the reader of Radozhiva

This article, especially for Radozhiva, prepared Roman Kurbatov

Color space

Color space. Illustration taken from gballart.com

A few days ago, Arkady posted the article written by the reader “Color profiles in photography" The article touched upon a number of rather important topics and asked questions of interest to many photographers. But a number of inaccuracies committed caused a fierce discussion in the comments, which only confirmed that in our midst there are a number of misconceptions regarding color.

This note does not claim to be complete and absolute truth, the author is not a color expert. However, a long-standing interest in this topic allows me to talk about this area in more detail. I will try to the best of my ability to give more precise definitions of some concepts, correct inaccuracies in the article and debunk some myths from the comments. I can send all readers who are not indifferent to color theory to the works of Alexei Shadrin and Ilya Borg - the most famous experts in this field throughout the post-Soviet space.

We will not discuss concepts such as color, spectrum, color temperature, we will bypass the physiology of vision - this is a topic for a separate note. We will only touch on those moments that are of interest, first of all, to photographers.

Let's start with the definitions. Important to distinguish color model и color space.

Color model Is an abstract mathematical model that indicates exactly how we will encode color data. We choose a set of characteristics, each of which is described by some numerical value. A set of such values ​​encodes a color. So, for example, the LMS color model is the most physiological. It uses three numbers to characterize the response of the three types of retinal cones in the long, medium, and short wavelength ranges. RGB is an additive color model. It is based on the fact that any color can be represented by mixing three monochromatic sources of different intensities. By the way, it doesn't have to be exactly red, blue and green, any others will do, if only the color of the third one cannot be represented by the color of the other two (for example, red, yellow and green will not work, since yellow is obtained by mixing red and green ). RGB was chosen because the peaks of the human cone sensitivity fall exactly on red, green and blue colors (conditionally). HSL (HSB) is a model that allows you to express color through its hue, saturation and lightness (brightness) - this is close to artists. There can be more than three parameters. So, CMY is a subtractive color model used in printing. Unlike RGB, where color is “added” to black by increasing the intensity of its mixed components, CMY “subtracts” colors from white by applying and mixing paint. CMYK appeared because it is rather difficult to get pure black from the three basic colors - cyan, magenta and yellow, the output is rather gray, and in printing, black is especially important - this is the font color. Therefore, black color (blacK) was also added to the original CMY, as a result the model became four-component. By the way, due to this, in the CMYK model, visually the same colors can be represented by different combinations. Of the several options, preference is given to where the amount of paint used will be minimal. But progress does not stand still, so printers use more different inks to reproduce more colors (color gamut) Modern photo printers are most often six-color; for unsaturated shades of magenta and cyan, the corresponding light version ink is used. Therefore, we can talk about the color model CcMmYK. Moreover, in the most advanced models, the number of basic colors reaches 12 (red, green, yellow, light and dark gray, matte and photo black are added, gloss instead of ordinary cyan). Conventionally, this is an 11-12-component model.

There may be less than three components. The first Technicolor color films worked with a two-component system (red and green), giving only a subset of the visible colors. This scheme is still popular in the form of tinting in the style of Teal / Orange. Grayscale and sepia can also be considered a color model (achromatic) where the color has only one - the luminance component. Well, for example, Pantone is also a color model, tabular (discrete), where the color is set by the number from the catalog.

Color space - a more specific concept. It defines all the colors that can be represented using a specific color model. Take, for example, your monitor. Its matrix reproduces color in the RGB color model. The color and brightness of the LEDs + filter characteristics set the color space (of your specific monitor) with a specific color gamut. The same goes for the camera matrix. The printer-paper system also has its own color space. Here with this particular set of ink on this particular paper you can reproduce this set of colors. Abstract (ideal) color spaces are used as standards and set some conditional parameters to which the device corresponds to one degree or another. The basis for comparison is the CIE XYZ color space (CIE 1931), adopted as a standard in 1931. It is noteworthy in that it contains all the colors visible to humans and, in fact, is tied to the characteristics of human vision (the reaction of color-sensitive cells). Since there are three coordinates in this color space, its graph should be three-dimensional, but a two-dimensional projection of this graph is used to describe the color itself (excluding brightness), all famous locus.

This “horseshoe” shows all the colors visible to man. Its border defines pure spectral colors (blue numbers are the wavelength in nanometers). Note that the monitor, even the best, reproduces only part of them. That is, we can determine some color in the CIE XYZ coordinates, we can compare them, the average healthy person can distinguish them, but we can’t always reproduce them. Since the capabilities of all devices are different, and the color gamut and characteristics of CIE XYZ in practice (so far) are not achievable, we needed “simpler” standards that correspond to techniques from different years.

The narrowest color gamut in the sRGB space is the oldest standard. Cheap matrixes of laptops and TVs have a color gamut of 60 to 90% of the color gamut of sRGB. Dear ones are approaching 100%. I emphasize once again - sRGB, like other standard color spaces, are ideal models. Real devices have different characteristics. This standard was proposed by Microsoft and HP.

The Adobe RGB space is slightly wider in the field of green and cyan colors. Adobe proposed it and began to actively use it in its products. Why was this needed? Photo printers and offset printing systems are generally inferior to monitors in color gamut, even the best of them reproduce only part of the colors from sRGB, but in the field of pure colors - yellow and cyan, as well as their mixture - green - are superior. At the time of adoption of the standard, almost no monitor could reproduce these colors, but photos when printing received more realistic color reproduction.

To date best samples monitors cover up to 99% of the space of Adobe RGB.

Apple, as usual, has gone its own way and is actively promoting another standard - DCI-P3used in digital cinemas. This color space is inferior to AdobeRGB in green, but superior in coverage in red. A number of Samsung mobile device displays, iPhone and MacBook screens are 100% DCI-P3.

There is an even wider standard, ProPhoto RGB, developed by Kodak. Its base colors have been chosen to further overlap the CIE XYZ coverage. A certain number of combinations will fall on imaginary colors that a person is not able to distinguish (it cannot be said that they do not exist in nature, because color is a characteristic of human perception, the concept of “color” does not exist in nature). However, any photographic footage scanned with ProPhoto RGB as a profile can be reproduced more and more accurately as display technology advances.

Why is all this needed and how does it work? Digital images are defined by a set of numbers. If the dot is encoded in the RGB standard, we know that 0,0,0 is black (lack of color, minimum brightness), 255,0,0 is red, and 255, 255, 255 is white (maximum brightness). But which one is white? And which one is red? The brightness of all devices is different, the color temperature is also. The standard color space (CPU) sets the white point and the boundary points of digital coverage. For any CPU model RGB, a triangle is obtained with vertices representing pure colors for each of the components - red, green and blue. Working in a specific standard, we set interpretation. CMS (color management system) - the color management system of our OS or graphic editor knows how to interpret the colors that we process and how to display them based on the capabilities of our monitor. This pixel is 0,255,0, provided that the sRGB space has these CIE XYZ color coordinates. If the space is Adobe RGB - the coordinates are different - this is more saturated green. And for ProPhoto RGB there is no such color in CIE XYZ at all, it is an extraordinary color even in this space, but we still display something on our monitor, some kind of maximum saturated green.

Here we need to introduce such a concept aswind profile. They are distributed in the form of ICM files and are used for various purposes, but generally serve to indicate the correspondence of a given color space to the universal CIE XYZ space.

The same set of image bytes will be played differently depending on which profile is assigned to the file. This can be either one of the standard profiles (sRGB / AdobeRGB / ProPhoto), or some custom one. For example, you can create a profile file that performs conversions that swap the green and blue channels of an image. If you actually swap these channels and attach the corresponding profile, then any program that works with the CMS of the operating system will play such a file normally. If the profile is not attached to the image, the default profile is assigned. And here the nuances begin.

Most often, sRGB is assigned as the default profile. For example, VKontakte, in order to save space, always cuts off the built-in profile. Previously, the profile was simply ignored, so, for example, if you uploaded an Adobe RGB or ProPhoto RGB image (or with any other profile) to the VK network, it was reproduced with distortion, the colors seemed more faded. Now the system automatically transcodes such images into sRGB.

The following literally happens. Each pixel of the original image is converted from RGB to XYZ thanks to the embedded profile. It is then converted from XYZ to sRGB using the standard sRGB profile. (You can do the same in Photoshop using the Convert Profile command.) Then the profile is removed from the recoded image file by the VK server. If such a file is opened, for example, in a browser, then it will be assigned a default system profile - in most cases it is sRGB. Photoshop, with appropriate settings, will offer to assign a profile to the image without a profile. If you see that the colors of your image are distorted, you can try to assign (the Assign Profile command) some other of the standard profiles. It is worse if you are using a wide gamut monitor in any of the advanced modes - Adobe RGB or DCI-P3. In this case, your file will be interpreted accordingly, and all colors will automatically add saturation, becoming more poisonous. That is why during normal work, surfing, it is better to switch the monitor to normal mode - sRGB.

That's not all :) Besides standard (abstract) profiles, there are also device profiles. They describe the characteristics of your specific devices so that CMS can take them into account when processing images - color reproduction. These are profiles of all input devices (scanners, cameras) and output (printers, displays).

Input devices are easier. They have approximately the same characteristics, rarely degrade over time, so a factory profile is enough. The scanner is calibrated at the factory, when installing the drivers, its profile is installed, which performs the necessary corrections, producing an image that is reducible to the desired standard at the input.

Now the most painful (and important) topic is cameras. From the matrix of the camera we get an image in RAW format. It DOES NOT HAVE AN APPOINTED PROFILE! The profile is assigned to the final JPEG file if it is created by the camera (Adobe RGB or sRGB, selected in the camera settings) or the image received at the output of the RAW converter. (You should also talk about RAW in a separate article, for now, take it for granted). Preference is given to sRGB, since the final output device is still predominantly sRGB devices. In addition, it is highly unlikely that your image REALLY contains all the colors that are included in the extended coverage of Adobe RGB. But even if there is this particularly bright especially sunny sunset or especially juicy grass, you must be sure that the end user will see these colors. That is, he has a wide gamut monitor, the profile will not disappear on the road when transferring a file, or you immediately print after processing your image on an expensive offset printing machine.
If you are processing a RAW file, everything is a bit more complicated. RAW is a black and white image obtained by transmitting light through filters of three colors. As a result of complex mathematical transformations and a process called debyerization (demosization), we get a color image located in the color space of your specific (model) camera. Profiling does not need a camera instance; the spread among them should not be large. When opening the RAW file of the corresponding model, the converter applies the corresponding (standard) camera profile... It allows converting the image to the CIE XYZ space in which the converter works. The camera does it automatically when creating a JPEG file. That's why when new camera models are released, they need support from the converter. This is where the notorious skinton problem lies. If you are not satisfied with the color of your camera - play with the profiles - converter settings. You may find a suitable one. After converting RAW, you can assign already image profile. We talked about him above. If AdobeRGB suits you and you know what you are doing, work in it. If not, use sRGB.

8 or 16 bit problem... Profiles and color depths are independent of each other. Profiles operate with matrices, the values ​​of which are normalized for your bit depth. Internal coefficients are real numbers from 0 to 1. If 8 bits are required, we will normalize to base 255 (2 to the 8th power - 1), if 16 bits - to base 65535 (2 to 16 bits - 1). By and large, if there is no particular need to work in 16-bit color (usually sky posterization) - don't work. Anyway, the final image is converted to 8 bits. All the advantages of a multi-bit (12-14 bit) RAW file should be used in the converter.

Monitors... Like any other output device, monitors are the most problematic. Color depends on viewing conditions and degrades over time. Basically, good and expensive monitors work fine with their factory settings. However, even profiling won't hurt them (not calibration, printers and scanners are calibrated). For cheaper devices, profiling is essential. Profiling is the procedure for building a monitor profile for specific conditions. You set a certain color temperature and brightness on the monitor, connect a special device (colorimeter or spectrophotometer) to it and start the program that comes with the device. It displays certain colors on the screen, the device reads them and builds a profile that takes into account the characteristics of your monitor. Due to the monitor profile, certain adjustments are made to the displayed CMS image so that the images on different output devices look plus or minus the same at least in some subset of colors. Of course, the capabilities of different devices differ, but profiling will eliminate or minimize problems such as ghosting and incorrect white balance.

Printers calibrate harder. Their color gamut is very different from monitors for the worse - the white dot is determined by the color of the paper, black - the saturation of the dried ink. In addition, the change of ink or paper leads to a change in the color of the print. Therefore, calibration requires each set of materials used. Only a more expensive spectrophotometer is suitable for calibration. After you make a special printout, it is analyzed by the device, a profile is built that you can select when printing on this type of paper. This will avoid constant game settings and give reproducible results. In addition, the printer profile can be used in Photoshop in the mode of screen proofing. In it, Photoshop simulates the characteristics of paper and ink on a monitor screen and you can further adjust the image, adjusted for changes made by the printer, to get a better result. In good printers, they can provide you with a printer profile.

By the way, most modern printing devices receive RGB data at the input, which the printer automatically converts to CMYK, CcMmYKk or other, more complex models, which we talked about above. Therefore, there is practically no point in working in CMYK. It is unlikely that you will be dealing with a machine that makes four prints with different colors that are fully consistent with the contents of each channel of the CMYK file.

You will find more articles from readers of Radozhiva here.

Add a comment:

 

 

Comments: 41, on the topic: Color model and more. Article from the reader of Radozhiva

  • anonym

    Everything is perfectly chewed. Including, for those who like to "hyip" about the skin tone. Respect, etc. :)

  • Andrei

    Sensible.

  • Andrei Other

    many terms, many letters, nerd

    • Valentine

      Delete such comments, littering is not the case.

      • Andrei Other

        why delete, for the truth? article sucks, and maybe even copy-paste

        • Alexey S.

          For disrespect for another's work.
          This article was not for you.
          If you don't like it, please pass by. Do not write ... on the fence, even if your conscience allows you, but you lack reason or desire to master a lot of letters or provide an article that you could share with the city on this site.

          Y - respect. It is the heart of a healthy society.

          • Andrei Other

            This is not work

        • NE

          And the blog is not yours. The owner will decide.

        • Onotole

          Your comment sucks, maybe even copy-paste.
          I - for it to be removed.

    • Georgy

      And here is the nerd? The question is about digital images and their output to devices.
      People interested in photography are interested and important.

  • Valentine

    The author, I would like to ask why it is claimed that the displays do not talk about calibration, but only about profiling? The X-Rite and DataColor sites write for displays as calibration. Shadrin Alexey, to whom you refer, also talks about calibrating the monitor. Why did the term calibration cease to be applicable to displays?

    • Novel

      Thanks for the question, I was driven, this is a significant clarification.

      Monitors are calibrated and profiled (characterized).
      Printers (scanners, possibly cameras) are only profiled /

      Profiling - taking characteristics, building the dependence on what you entered at the input with what you got at the output, in order to take into account the features of the device in the future, a description, roughly speaking, of the CPU in which the monitor works.
      Calibration - changing the characteristics of the device to match certain parameters. During calibration, the user adjusts the brightness and contrast, sets the white point, and then the calibrator creates a LUT table for adjustments.

      • Onotole

        By the way, TVs are just being calibrated. Sometimes it can take up to a whole working day. For the most part, TVs are lousy in terms of color, but some individual models can surpass in coverage, uniformity, smoothness of tones, contrast, etc. even the most expensive monitors. But, only after calibration.

  • Peter Sh.

    Very informative and affordable, thank you!

    In general, this is a very important topic and it is extremely important to understand at least the basic points with profile settings. I agree with the author of the previous article - if you don't know exactly why you need all these Adobe RGB, install sRGB everywhere.

    White balance is also a pain. Here you took pictures on your cool calibrated photo monitor, everything is perfect. And the people will look at this whole thing at their usual consumer goods, without any calibration. On their assorted smartphones. And there your masterpieces will not look entirely masterpiece. Something like this…

    • Valentine

      Why with white balance, there is a question with color in general. But I would not worry about consumers. It is important that when editing there is a calibrated monitor and you set the colors correctly. The viewer can cool or warm the picture, but he will see all the images including wallpapers and system pictures, which are usually prepared very high quality. Here in commerce, if the order is responsible, it is better not to blindly trust the customer’s displays: either show on your tested device, or print again on the tested device.

    • Novel

      I agree with Valentine. Yes, and smartphones, mainly IPS-matrix, and there with the factory settings everything is more or less normal. Sometimes it is even recommended, if it is not possible to calibrate the monitor as expected, to bring the image to the image on the smartphone (you can take a few).

    • NE

      Well, if initially everything is not entirely good, then what to say about what will happen next ...

  • Q12

    More articles - more mistakes. Error on error.

    It will not be right to give recommendations, but I will express comments:
    sRGB / AdobeRGB ... and so on and so forth and so on, needed! all are good! And used by knowledgeable people in every suitable situation. And they were created for different situations and solutions.
    MOST IMPORTANT: understanding that “Color management” is a very complex and multifaceted subject, where theory is not always friendly with practice, where the “human factor” has a huge influence. (Example: p. 396 Dan Margulis. Photoshop for Professionals. The Classic Guide to Color Correction. 5th Edition) and environmental conditions (lighting and / or climatic factors, quality of consumables, measurement equipment). ICC profiles (from the pre-installed ones) were created in ideal conditions, and they are not a panacea !!! Rather, the model to which you need to strive. And if you take, for example, your favorite picture and print it in a dozen minilabs, as I did in my time, then you will be greatly surprised by the result.

    Recommended reading (basic):
    Dan Margulis. Photoshop for professionals. Classic color grading guide.
    A.E. Shadrin. Parenting Dr. Margulis: work on the bugs.
    R.V.G. Hunt. Color reproduction. Sixth Edition. translation Shadrin A.E.
    Adobe Photohop Help - Color settings, Custom CMYK, Custom RGB, Lab.
    Hi-Fi Color / UCR / GCR (GCR1-GCR2-GCR3-GCR4) / NoK / MaxK / Minimize ink consumption / Metamerism / Gamut mapping / Rendering Intent / TVI.
    ProfileMakerHelp.
    Digital proofing.

    http://www.cie.co.at/
    https://forum.rudtp.ru/
    https://alexey-shadrin.com/
    https://photolab.com.ua/articles/10023.html
    http://colorlab.com.ua/efi/
    bonus))
    https://fstoppers.com/pictures/adobergb-vs-srgb-3167

    Each article has great "pearls".
    the funniest of the last:

    However, even profiling (not calibration, printers and scanners are calibrated) will not hurt them (we read the X-Rite help).

    After you make a special printout, it is analyzed by the device, a profile is built that you can select when printing on this type of paper. This will avoid constant game settings and give reproducible results. (we read the help of the X-Rite - linearization, profiling, optimization).

    In good printing houses they can provide you with a profile of the printing device (did you work in printing houses? Have you seen what will actually happen in fact? And what kind of profile will it be? And what if the label or packaging, especially flexography?)

    Therefore, there is practically no point in working in CMYK (Fiction! Dan Margulis. Photoshop for professionals. A classic guide to color correction).

    • Valentine

      And where is error-upon-error? Of the points that you indicated, there are several inaccuracies, but basically - just different interpretations of the questions. I re-read the commentary several times; unfortunately, I cannot say that he somehow substantially supplemented the material presented in the article.

    • Novel

      Wide color spaces are needed and important. To someone. For some purpose. How spaceships and racing cars are needed. You can talk a lot about them for a long time, but this is a site for photographers. Amateur photographers predominantly. I would like to hear about the advisability of using these spaces for amateur purposes with real examples without a list of links.

      Margulis is a good author and an outstanding specialist. OF YOUR TIME. This is a color grading dinosaur that I respect, but which, sorry, is somewhat outdated. Some of his methods have become the standard arsenal of modern programs (Vibrance instead of Martian Method, advanced sharpness by mask), some methods such as color correction by numbers are not applicable in creative photography, where aesthetics are important for the sake of accuracy. Work in CMYK has practically disappeared at the moment, yes. The last time "Professional Photoshop" was published in 2006. Probably, this is the last edition and in subsequent books Dan works in RGB (well, in Lab, of which he is an absolute popularizer and expert).

      About minilabs and a favorite picture ... Well, it is logical, if you have different ink and paper, the color gamut and contrast of the printed images will be different, it’s strange to even discuss it. But if all printing devices are supplied with correct profiles, all printouts will be conditionally the same - without parasitic shades and with adequate contrast.

      In general, the essence of the claims is not particularly clear to me.

      • Q12

        MINILAB does not operate with ink! You need to read and study carefully! Think wider, not being limited to the word “PHOTOGRAPHER”! A.E. Shadrin. Education according to Dr. Margulis: work on mistakes. THESIS 20 ..., there is the WHOLE answer, do not be lazy to study. Only 3 pages of text.
        What is reality? A snapshot in your home archive? So this is one thing, and if “PRINTED PICTURE”, it’s a different result every time. And I will note from practice “Calibrated Monitor” plus the “i1” case-set absolutely does not guarantee success, but rather gives an understanding of how it works and where there may be problems, and how to solve them. In general, “… work on errors is 41 pages, and everything is very well laid out there, everything is on the shelves”

        “Work in CMYK has practically disappeared at the moment”, I have been doing this for 18 years now in CMYK, and so far I have not seen the sunset ...

        • Q12

          All “minilabs” have “PROCESS” and standards that describe “exposure of photographic paper”…. and I EVERYWHERE FOR THAT MOMENT ASKED FOR THE SAME MATERIALS !!!

      • Q12

        AS AN ADDITION: Output ICC (printing device-printer-plotter-printing machine) profile, which describes the “inc limit”, “TVI”, “color separation” scheme and “gamut mapping”, after which you correct the places and problems that you did not like it. And the most important thing is to remember that the result of the monitor and the paper will be different, “even after a hundred thousand calibrations”

  • NE

    Thanks a lot, Roman. There is a rhetorical question to the topic of the “notorious skin tone”: I wonder why (IMHO) profiles in ACR for D800 suck, but for d5100 everything is quite normal ....

    • Novel

      Perhaps a bad profile. I do not know the features of the internal kitchen converters. It is interesting, by the way, to compare the results between different converters, in theory, each of them uses its own set of profiles.

      I have never done this personally, but you can create a custom profile if you have a ColorChecker. I ordered non-original ones from Lushnikov, they are much cheaper and with accurate data measured by the spectrometer for each specimen. Make a few with different lighting conditions and use the one you like.

      You can do the same with 3D Lut Creator every time you shoot, photographing the target and building a profile on it, but this is already more intense.

      It's just that it turns out like with a film or with a tube sound. More accurate results will seem worse because you are used to certain distortions. For example, there is such a profile from Kodachrome 64, pulled from the RPP. It slightly raises the contrast of the whole image, but it is especially active on the skin, making it a little lighter, more even and yellow. Retouching for the poor, one might say. But in fact - an unreliable picture.

      • NE

        I made my profiles with the help of Lushnikov's target. Including for RPP. In ACR, the result seemed to me worse than with the "family" and did not use it. In RPP I use my profiles. In general, in my opinion, profiles should be done for each ISO used and for each light source ...

        • Valentine

          ACR profiles with a name similar to the built-in one give a slightly different picture compared to the on-camera jpg or native converter. But, sometimes, you can pick up profiles from another camera (which give the best result) and feed it ACR as from your camera.
          Often, the Adobe Camera Standart profile is not liked by users, but in general it is usually quite high-quality and universal. On its basis (as by the way and on the basis of other available profiles for ACR), you can build your own, modifying them in one direction or another (tone curve, color, etc.). These profiles will be saved, they can even be assigned by default (replacing the standard).
          Finally, with the colorchecker, you can apply the results to specific images without building and loading a profile, but this is done in Lightroom, and not in adobe camera raw. Here, if ordering is in place with the checker and the ability to use it, good accuracy is ensured.

          • NE

            Thank you

  • scif

    uhhhhhh, as much as typography smell)))))))))))

  • Sergei

    Very informative, thanks!

  • Andrii

    Privit all.
    The topic is given, but I can't go back to that photo of the hobby. May I have a PROBLEM: having set the printer profile with photo-handles in your lightroom and the result of color when converting from RAV into a jeep, "jump" a lot. (especially often greens are transformed into a bowl of acidic greens). Chi warta zamorochuvatis with this profile is enough for a simple SRG?
    Dyakoy, yakshko htos help)

    • Dmitry

      The order of work is as follows (I tell you how for FS, in Lightroom we look for similar items)

      1. In Color Setup in Photoshop settings, we set the working space as: Prophoto RGB, AdobeRGB, or, if you have a screen with a normal color gamut, then just SRGB. (but not a monitor profile!)
      2. Next, go to Soft Proofing Setup (color proofs) and set the ICC profile of the printer, taken from the photo lab.

      3. The main idea is that the processed photo on the screen (and we are editing it in one of the RGB spaces selected in point 1) will need to be further modified to make it look approximately the same as when we print it on the screen . That is, for printing, you will need to make a separate copy of the photo and additionally bring it with contrast, color \ tone, etc., to compensate for the distortions introduced by the printer. Such a copy with edits will look awful on the screen, but when printing, distortions are compensated and the printouts will be very similar in color and tone to the original processed photo displayed on the screen.

      4. For the convenience of this work, there is a mode of Proofing (Soft Proofing). When we turn it on to a photo that has already been edited and looks good on the screen, then the color profile of the printer is superimposed on it, which we tied in step 2. And the picture begins to look close to how the printer prints it (for example, a stray hue appears, some colors will become less saturated, etc.). That is, emulation of how it will look on paper. In general, with the color proof turned on, the picture on the screen will be distorted and not at all the way we wanted to see it.

      5. You need to create or a copy of the photo for printing or make a separate adjustment layer for it, turn on Soft Proofing, and without turning it off, move the brightness, contrast, curves, black point sliders, select Rendering Intent, etc. - so that the picture simulated "for the printer" again became similar to the original photo (switch between 2 images, pull the second to the view of the first).

      6. As soon as the copy for printing (with softproofing enabled) becomes close in color to the original image (without soft proofing), the copy for printing can be saved and sent to the printer without further color correction. By the way, if you turn off Soft-Proofing, then the copy for printing on the SRGB screen will look distorted in colors / tones, but this is normal.

    • Dmitry

      Important! Screen profile or printer profile - we don't make it a workspace! Only RGB (optional)!
      Printer or monitor profiles are used only for emulation in Soft Proofing mode!

      • Andrii

        Dmitry, you’ll cripple you.

        • Dmitry

          And, I forgot to say, as soon as we bring the copy for printing to the right condition, so that when the software proof is on, it looks like the original edited photo, it needs to be truly converted to the color space of the printer. (Edit-Convert to profile (we select the printer profile from the photo lab)), the software proof was only temporary emulation. And only then, the copy for printing translated into the color space is saved in jpg or tiff and put a check mark that we save together with the color profile of the printer. And so we give it to the photo lab, telling them that extra. color grading is not required.

          • Andrii

            so i'm robbing now, ale so doesn’t want to copy three photos). Dyakoy, more time!

    • Translator Andriy (Andreyki)

      Hello to all.
      I am interested in this topic, but I can’t figure it out a little and photo is a hobby. I have such a problem: I installed a printer profile with a photo drunk in my lightroom, and as a result of the color, when converting from RAO to a jeep, they “jump” a lot. (Especially often green turns into very acid green). Is it worth bothering with that profile will a simple SRGB be enough?
      Thanks if someone helps)
      WELL FUCK DO YOU WRITE IT? The article is in Russian. Comments too. And this AndriyAndriy.

  • Dim

    Shot on smart and enjoy, but no it's too easy

    • Onotole

      Someone is holding an espresso machine at home, while someone is enjoying 3-in-1 coffee.
      The taste and color.

      • Charles

        Forgot more cezv fans :)

  • Denis Korzun

    In general, my own experience (I have been scanning more than 20, and printing) says one thing - that you need to store in 16-bits and preferably in a wide color space that covers your sensor, or photographic film!

    In the case of film, the following color profiles are more optimal (they are wider than Adobe RGB (1998), but narrower than ProPhoto RGB - and rightly so, the unused color gamut is excluded):
    1 DodnRGB4.icm
    2 Beta RGB
    3 BestRGB
    4 Ekta Space PS 5, J. Holmes

    For a figure, it is definitely difficult to talk about color gamuts of matrices - manufacturers practically do not provide information about this.
    In my opinion, it makes sense to use fairly wide profiles profiles (preferably with a gamma of 2.2):
    MaxRGB.icc (wider than ProPhoto RGB - good for large format photo printers)
    ACES-elle-V4-g22.icc (Widest color gamut - comparable to Lab, even wider)
    ProPhoto RGB (preferably with a gamma of 2.2) - aka LargeRGB-elle-V4-g22.icc
    ACEScg-elle-V4-g22.icc (extreme coordinates within LAB - narrower than ProPhoto RGB)

Add a comment

Copyright © Radojuva.com. Blog author - Photographer in Kiev Arkady Shapoval. 2009-2022

English-version of this article https://radojuva.com/en/2018/07/color-profile-space/

Version en español de este artículo https://radojuva.com/es/2018/07/color-profile-space/