bs kite Offline Upload & Sell: On
|
Here is the ColorPerfect response to me. I believe it is a candid and thorough response. I appreciate that. The graphs he sent do not appear. Hope you make sense of it.
Christoph Oldendorf | ColorPerfect.com <[email protected]>
11:38 AM (7 hours ago)
to me
Robert,
My question: IS KODACHROME INCLUDED WITHIN THE FILM CHOICES IN YOUR COLORPERFECT PLUGIN.
There are no built in slide films. In contrast to color negative film it is often not a goal to eliminate the film's characteristic curves in printing.
I will ask the same question in a different way. Please give me CANDID answers.
There are no CANDID answers. Many folks use CP for slides but it'll mean understanding many details and figuring out what to do.
I'll put some very non candid items in below and you can take the trial version of ColorPerfect and find out wether there is something in it for you...
Best regards
Christoph Oldendorf
-----------------------------------
With slide film there is one obvious distinction to negative film and that is the fact that in principle the slide is the intended final image. Many users create linear scans of slides and process them in ColorPerfect's ColorPos mode in some limited way they like and are very happy with the results.
There are no characterizations for slide film built into ColorPerfect. With the negative film and digital camera characterizations the same goal applies to somewhat different extents. The goal is to get a recording of a scene that is not intended as a final image but was designed to best store the scene (in order to be made into a final image later) back into a form that has color integrity. That means a state in which basic physical relations exist that allow gray clicks into the preview image and other adjustments in ColorPerfect to work beautifully. In case of digital cameras we are also able to ascertain color fidelity. In case of negative film color fidelity depends largely on the film’s specific dyes and sensitivities.
Anyway, for slide film there is a complication. If we assume that the slide is not already in the state it should be as a final image and try to approach it with the same logic we use on color negatives we’ll find a significant difference. You mentioned Kodachrome so let me show that difference using the characteristic curves of Kodachrome KR64 compared to today’s Kodak Ektar 100. In case of the latter – the negative film – we can see that while the curves for R G B are significantly apart each has a clear mostly straight line section between shoulder and tail that will record most of our scene’s relevant detail:
[]
A characterization defined by three film gammas (slope of the straight line portion of the curve) works reasonably well here when combined with ColorPerfect’s other tools.
Now the Kodachrome KR64:
[]
The three colors are much closer together – a requirement for slide film to produce a natural looking positive image – but the curve is not as good a fit for three plain Gammas. What we do see here is a plot of density versus log exposure. For log Exp it suffices to know that 0.3 is equal to 1 f-stop and for density it’s enough to know that more is darker and for the slide the film’s DMAX is dark enough to seem opaque to the eye.
Now for the slide we have different slopes in the brightest 3 f-stops than in the darker f-stops of an imagined capture and since not all slides must contain as bright areas that might as well mean that one image requires one characterization while a second image requires another. The steeper Gamma in the darker regions also somewhat serves as a substitute solution to Black Point which as such doesn’t exist in the analog realm.
A system dedicated to slide film might be possible to create but we’re less than certain if it could really work well and we did not attempt it yet.
What you should do is experiment and see what works well on a series of images. The FilmType / SubType / FilmGamma calibration system shown to the end of this video: http://www.colorperfect.com/ColorPerfect-video-tutorial/ColorNeg-and-PerfectRAW-feature-overview/ will also work on slide film in ColorPos mode but because of the situation I tried to outline above a given set of Gammas will probably not work equally well on all images from even the same film. It depends on what you are trying to do.
-----------------------------------
When processing linear scans of slides as produced by this work flow:
http://www.colorperfect.com/scanning-slides-and-negatives/scans/Hamrick-Software/VueScan/
note that ColorNeg is set to expect linear scans as input by default. The latter is set to expect Gamma encoded input. There is a little button displaying either "G" or "L" on the start panel it should show "G" in ColorPos mode. With your Velvia Scan press it so that it shows "L" for linear input before you start. Among the options (which are per mode, so ColorPos has its own) is a setting "G/L Defaults G" untick that and you'll always start with "L" as is the case for ColorNeg mode. The reason ColorPos mode expects gamma encoded images originally was so that it wouldn't give folks trouble when editing pre-existing images.
-----------------------------------
I'll also paste in an item below that might be helpful along with this resource:
http://www.colorperfect.com/working_spaces.html?lang=en
Let’s discuss a bit of analog color photography first. I hope you have seen actual quality color photographic prints somewhere - fine art ones that is, printed by someone who knew what he was doing and not your average quick lab ones. Maybe even your own or maybe you’ve been to museums during the reign of analog photography and absence of digitally aided printing techniques. Long story short: That did work out just fine. There were wonderful color photographs made in the second half of the 20th century.
Now let’s take a look at color negative film and color photographic paper. What are they, what do they do and how do they do it (simplified in places)? Let’s envision a natural scene of some kind. Countless surfaces reflecting light. Each reflecting different wavelengths to differing amounts etc. Depending on the light source - e. g. midday or late sun there will also be varying metameric effects etc.
A pretty complex system to model but film does not model it precisely (neither do digital cameras or any three primary color system). There is tremendous abstraction and that’s a good thing. Color negative film (ignoring the 4th color layer technology marketing gimmick) has three light sensitive layers. Each is sensitive to a certain range of wavelengths - or rather a combination of that and the fact that color negative film is a layered design including filter layers that make sure that certain wavelengths do not reach certain light sensitive layers. These filters are not sharp cut offs but are rather steep gradients so what we get are three spectral averages, that peak at some wavelength and more or less quickly degrade from there. They are roughly sensitive to Red, Green and Blue portions of the spectrum and are found in most film data sheets as plots so let’s see one:
[]
This is from Kodak’s Ektar 100. So, from our multifold spectral responses in the scene we get to three averages at every position the film can record.
In the chemical processing when developing the film the recorded intensity of each average controls the amount of dye deposited at that position. There are three dyes and since we are dealing with a negative they are of the secondary primary colors Cyan, Magenta and Yellow.
In a nut shell the scene gets split into three spectral averages. The three averages then get “ASSIGNED” to three secondary primaries. Of course there are further aspects to this, for instance that brightness is not being recorded linearly but at a negative Gamma significantly smaller than one (x < 1) which explains the film’s latitude and needs to be dealt with in processing by use of an inverse (1/x) positive Gamma. I don’t mean to write a solid description of the technology here I’d just like to loosely look at what’s happening with color.
So, the scene's color was assigned to CMY with some items that will be dealt with and which I’ll ignore in the following. Interesting to note is the fact that color fidelity is mostly controlled by the film. It controls what color of the scene makes it into what average and consequently gets assigned to what dye.
Then what? Printing the negative. In order to do so the three color channels in it need to be balanced so that for neutrals (grays) C=M=Y. That is where CC filters come into the picture but we’ll ignore them. Let’s just say that a suitable filter pack was dialed into our enlarger.
So, an image gets projected onto color photographic paper. The paper again has three color sensitive layers. This time they are sensitive to C M Y and in developing the print the intensities recorded by each will be “ASSIGNED” to yet another set of dyes, being the final RGB this time.
Important here is that the enlarger has little to no influence on color fidelity. Its optical qualities are important, sure but as long as it has a reasonable light source and filters all else is fine. The color photographic paper also has little influence on color fidelity. It needs to be made to suitably separate the CMY intensities that come in from projecting the film, so that they can get “ASSIGNED” to the RGB dyes as intended but the photo is in the averages. The RGB dyes themselves will of course control how the final image looks. Be that more vivid or more pastel but color fidelity is a thing of the film.
O.k. after this wide sweep can you see what I’m getting at? In contrast to many technologies devised during the onset of the digital age that just seem to ignore the most simple physics of photography - for various reasons, early stages of technology, lack of understanding etc. - with ColorPerfect we are building a complex system that uses a much simpler physical model but one that actually works and produces superior results when used right.
So, what is the scanner compared to the analog world as described above? It takes the role of the enlarger but it also represents part of the color photographic paper. The part prior to development, that is the one dealing with the separation of the CMY intensities coming in from the projection of the negative. Does the scanner have an influence on color fidelity (as long as it gives us a plain RAW recording)? No. Does the scanner even have an influence on the final colors (RGB dyes in the color photographic print)? No. Those are modeled by assigning an RGB working space that is suitable for the task. Are different scanners comparable? Yes absolutely. They might use different light sources but the fact that we are dealing with such a limited number of dyes makes the differences between those mostly equal to effects that can be dealt with by CC filters. The scanner is usually not our problem.
Now, for some films assigning Adobe RGB 1998 might work but in my experience it most often does not work all that well. That’s particularly noticeable when you do have a screen capable of displaying the full gamut of that working space. This gets us more deeply into color science and I don’t intend to go there, too now but color lore has it that Adobe RGB 1998 got the green primary by accident from NTSC (which is defined relative to illuminant C while Adobe RGB 1998 is relative to D65) and that the folks designing it “liked” the result of that and kept it. The R and B primaries are identical to those of sRGB…
Anyway, most often assigning sRGB or one of the other normally narrow standard RGB color systems which were all designed to pretty much meet the same goals as color photographic prints work better. So assigning Apple RGB or ColorMatch RGB will work the same. What to use is a matter of preference. The latter two use a Gamma of 1.8 which was not an ideal choice to begin with and the tone reproduction curve of sRGB is equally nonsensical. Still, while having 16-Bit color precision all of that is irrelevant and you can convert the final image to whatever you’d like for 8-Bit use.
One final thing I should probably point out: Can you lose any colors by assigning a standard narrow RGB system? Can any color be out of gamut due to that? No. Simply because by assigning your RGB system you are giving the metric of what fully saturated Red (255) produced by ColorPerfect should mean. The gamut problems you might fear only exist in color system conversions (from wide to narrow). If the image is not as colorful as you’d like most often ColorPerfect’s Saturation command does a better job than fussing around with assigning other primaries.
You can of course convert the almost final image from narrow RGB system to Adobe RGB 1998 or even Pro Photo RGB to be able to use Saturation to create even more colorful tones. Keep in mind though that your screen might not display them and that photographic processes out there might not print them so that should be the exception rather than the standard case.
So what you are supposed to use is a normally narrow RGB system (unless something else works for a given film), Saturation, White, Shadow compression etc. What I use a lot is the luminance mask paired with removing White as illustrated in the general ColorPerfect tutorial video I recorded a while ago.
What I’d like to give you with this write up is confidence. Many users seem to think that the color they get by assigning a standard working space is very arbitrary. It is not. It follows the intended color flow for photographic film and it works very well. The color is somewhat replaceable in this and assigning any other primaries that perceptually work better for a given artistic goal is a valid thing to do as illustrated in the text linked to in my first post here. Usually that is not necessary.
Often people will tell me that they’d like to keep the color properties of the film. Notice that you do. Using the same RGB primaries for all final images (but for differently constructed spectral averages of a given scene as provided by different types of film) is like printing different films on the same color photographic paper. That does not impact the properties of the film.
If you made it all the way down here that is not the end of it but it might be a beginning... What Scanner do you use?
|