Pages: [1]   Go Down

Author Topic: Colour Space Question.  (Read 5176 times)

William Walker

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1134
    • William Walker Landscapes
Colour Space Question.
« on: March 23, 2010, 09:26:03 am »

Hi

I use a Pentax K10. There are two "in camera" options for colour space: sRGB or Adobe1998. I shoot only in RAW.

When I work in Lightroom or CS4, everything is set up for ProPhoto.

Are there any problems with this? Is what comes from the camera limited by those colour spaces or can one make adjustments during editing that will "expand" into the Prophoto gamut?  





Logged
"What can be asserted without evidence can also be dismissed without evidence." Christopher Hitchens

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #1 on: March 23, 2010, 09:42:06 am »

Those settings only affect the JPEG. If you shoot raw, you can select (encode) any color space the converter supports. The raw has no defined color space at this raw state so you’re fine.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

William Walker

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1134
    • William Walker Landscapes
Colour Space Question.
« Reply #2 on: March 23, 2010, 09:50:36 am »

Quote from: digitaldog
Those settings only affect the JPEG. If you shoot raw, you can select (encode) any color space the converter supports. The raw has no defined color space at this raw state so you’re fine.

Thanks Andrew - that was the answer I was hoping for!
« Last Edit: March 23, 2010, 09:51:06 am by WillytheWalks »
Logged
"What can be asserted without evidence can also be dismissed without evidence." Christopher Hitchens

PeterAit

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4559
    • Peter Aitken Photographs
Colour Space Question.
« Reply #3 on: March 23, 2010, 03:23:50 pm »

Quote from: WillytheWalks
Hi

I use a Pentax K10. There are two "in camera" options for colour space: sRGB or Adobe1998. I shoot only in RAW.

When I work in Lightroom or CS4, everything is set up for ProPhoto.

Are there any problems with this? Is what comes from the camera limited by those colour spaces or can one make adjustments during editing that will "expand" into the Prophoto gamut?

The camera's color space settings are applied to jpegs it generates and are irrelevant for RAW. Yes, you can "expand" into the prophoto color space when processing your RAW files.
Logged

William Walker

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1134
    • William Walker Landscapes
Colour Space Question.
« Reply #4 on: March 23, 2010, 03:51:01 pm »

Quote from: PeterAit
The camera's color space settings are applied to jpegs it generates and are irrelevant for RAW. Yes, you can "expand" into the prophoto color space when processing your RAW files.

Thanks Peter!
Logged
"What can be asserted without evidence can also be dismissed without evidence." Christopher Hitchens

feppe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2906
  • Oh this shows up in here!
    • Harri Jahkola Photography
Colour Space Question.
« Reply #5 on: March 23, 2010, 03:55:14 pm »

Quote from: WillytheWalks
Are there any problems with this? Is what comes from the camera limited by those colour spaces or can one make adjustments during editing that will "expand" into the Prophoto gamut?

I'll chip in with yet another addition  when you open a RAW in Lightroom it is automatically rendered in ProPhoto, so it's already in ProPhoto even before you make adjustments.

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Colour Space Question.
« Reply #6 on: March 24, 2010, 08:46:13 am »

Quote from: digitaldog
The raw has no defined color space at this raw state so you’re fine.
The DigitalDog ignores the opinions of Eric Walowit, Thomas Knoll and Chris Murphy when asserts that the raw file has no color space. Perhaps he is getting too old to learn new tricks or simply does not want to admit that he was shown to be wrong by an upstart in a previous thread.

A digital camera does not have a gamut in the usual sense, since it will output values for anything that is put in front of it, i.e. there are no out of gamut colors and hence no gamut (see here for an explanation). However, you can put a color target in front of a camera and obtain useful information. For example, drycreekphoto.com has "gamut" plots for some older digital cameras and one can interactively compare the camera response to the gamuts of ProphotoRGB, Adobe RGB and sRGB. Only ProPhotoRGB can represent all the colors captured by the camera, making it the preferred space into which one should render raw files.

Like any other color space, the camera space can be characterized by 3x3 matrix coefficients as shown for the Nikon D3x on the DXO site; as Mr. Knoll stated, one can use these coefficients to obtain a pretty good rendering of the image.

[attachment=21029:D3xColorSpace.png]


Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #7 on: March 24, 2010, 09:43:00 am »

Quote from: bjanes
The DigitalDog ignores the opinions of Eric Walowit, Thomas Knoll and Chris Murphy when asserts that the raw file has no color space.


I suggest you read this too: http://dougkerr.net/Pumpkin/articles/Sensor_Colorimetry.pdf

Quote
If we are going to state a color, perhaps as a description of a pixel of
an image, with the expectation that the receiving system will be able
to direct its display to reproduce that color, we have to get very
specific about the details. Once we have done that, we have defined a
specific color space. In the case of color spaces based on the RGB
model, this includes:
• Precisely defining, in scientific notation, the chromaticity of the
three primaries upon which the description of color will be
predicated.
• Defining exactly how the “amounts” of each primary are to be
stated: how much is “one unit” of a primary; are the numerical
values the actual amounts or some nonlinear representation of
them, and, if so, following exactly what nonlinear function; and so
forth.

Further (the meat of the piece and which syncs up IMHO with what Eric wrote (A camera has colors it really can capture and encode as unique values compared to others, that are imaginary to us. They don't exist).
Quote
In fact, a design following the Luther-Ives conditions may not be
attractive either. Thus we may well compromise even further and use
a design that does not follow those conditions.
If the Luther-Ives conditions are not met, then colors having
metameric spectrums (that is, different spectrums but nevertheless
having the same color) will in general give different sets of outputs
from the three sensor channels. In other words, such a sensor will not
give a reliable indication of the color of the light. It would be said to
have metameric error. As a consequence, the behavior of such a
sensor cannot be said to be consistent with any color space (not even
a “private” one only applying to that sensor design).

Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Colour Space Question.
« Reply #8 on: March 24, 2010, 10:58:20 am »

Quote from: digitaldog
I suggest you read this too: http://dougkerr.net/Pumpkin/articles/Sensor_Colorimetry.pdf

Further (the meat of the piece and which syncs up IMHO with what Eric wrote (A camera has colors it really can capture and encode as unique values compared to others, that are imaginary to us. They don't exist).
Andrew,

Thanks for the link to Doug Kerr's article. My own bookmark to his site merely directed me to the AT&T personal web page and I had thought that Doug had gone offline. His article  is the best that I have seen for the intelligent layman. I have two comments.

1) Doug agrees with Thomas Knoll that the mosaic form in a Bayer array is a red herring as to whether a mosaic image is really color. As Mr. Knoll explained, the demosaicing process early in the rendering process does assign RGB values to each pixel in the array. Mr. Kerr assumes this in his evaluation.

2) Perhaps we should call the camera color space a "quasi space" as Mr. Kerr suggests, since the Luther-Ives conditions are not met by any current sensor and meeting these conditions would require negative numbers and would have other problems. Imaginary primaries are nothing new, and are incorporated into ProPhotoRGB and CIE XYZ. Nonetheless, the "quasi space" of current cameras is treated like a real color space when one uses matrix math to render the image into a true color space such as ProPhotoRGB. Eric Chan has said that ACR uses matrix math supplemented by lookup tables for its rendering.

That cameras can capture colors that are "imaginary" to us has little bearing on the topic IMHO. There are many polymorphisms in the makeup of the human retinal pigments and some persons are tetrachromats and can perceive "colors" differently from the rest of us. These colors are real to those tetrachromats.

I agree that one can make valid arguments that the camera does not have a real color space, but in practice it is convenient and necessary to assume that they do have a color space so that we can use matrix conversions. Perhaps we should refer to it as a "quasi space" and end the argument.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #9 on: March 24, 2010, 11:48:57 am »

Quote from: bjanes
1) Doug agrees with Thomas Knoll that the mosaic form in a Bayer array is a red herring as to whether a mosaic image is really color. As Mr. Knoll explained, the demosaicing process early in the rendering process does assign RGB values to each pixel in the array. Mr. Kerr assumes this in his evaluation.

Hence, my language over the years (which you seem to find “offensive“ based on your kind words above) “a raw file is essentially a Grayscale file.” Not A Grayscale file but something like it. There is no question it has the potential to be, and will be a document with a defined and unambiguous color space.

Quote
2) Perhaps we should call the camera color space a "quasi space" as Mr. Kerr suggests, since the Luther-Ives conditions are not met by any current sensor and meeting these conditions would require negative numbers and would have other problems. Imaginary primaries are nothing new, and are incorporated into ProPhotoRGB and CIE XYZ. Nonetheless, the "quasi space" of current cameras is treated like a real color space when one uses matrix math to render the image into a true color space such as ProPhotoRGB. Eric Chan has said that ACR uses matrix math supplemented by lookup tables for its rendering.

Again, if you look at what I wrote above, I clearly said “the raw has no defined colorspace”. There’s nothing I know of in a raw file that tells any number of raw converters what the color space is (or is supposed to be). The converter creator has to make an assumption here which is what I got so clearly written from Doug’s fine piece.

Quote
That cameras can capture colors that are "imaginary" to us has little bearing on the topic IMHO. There are many polymorphisms in the makeup of the human retinal pigments and some persons are tetrachromats and can perceive "colors" differently from the rest of us. These colors are real to those tetrachromats.

I would submit that if we can’t see it (an imagery specification of a color), its not really a color. Again, it might be semantics but in terms of human vision, color is a perceptual properly. If we can’t see it, its not a color (it can be a particular wavelength of light, but a color coordinate outside the spectrum locus isn’t a color in the sense we perceive it).

Quote
I agree that one can make valid arguments that the camera does not have a real color space, but in practice it is convenient and necessary to assume that they do have a color space so that we can use matrix conversions. Perhaps we should refer to it as a "quasi space" and end the argument.

Yes, no question its convenient and necessary to make these assumptions or I can’t fathom how we’d end up with something from the raw data.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Colour Space Question.
« Reply #10 on: March 24, 2010, 08:13:28 pm »

Quote from: digitaldog
Hence, my language over the years (which you seem to find “offensive“ based on your kind words above) “a raw file is essentially a Grayscale file.” Not A Grayscale file but something like it. There is no question it has the potential to be, and will be a document with a defined and unambiguous color space.

Again, if you look at what I wrote above, I clearly said “the raw has no defined colorspace”. There’s nothing I know of in a raw file that tells any number of raw converters what the color space is (or is supposed to be). The converter creator has to make an assumption here which is what I got so clearly written from Doug’s fine piece.
Colorspace information is given by reference. The raw file does contain the identity of the camera, and the characteristics of the Bayer filters are the same within manufacturing tolerances for a given camera model and can be determined and applied separately. Apparently a method for determining color matrix is defined in ISO standard 17321 as reported by DXO for various cameras as per my previous post. The DNG spec does contain spaces for two matrices for two illuminants. They are not supplied by most cameras, but are filled in by the DNG converter. For example with my D3 they are:

33432 Copyright ASCII     Public Domain
50708 UniqueCameraModel  ASCII     Nikon D3
50721 ColorMatrix1
SRATIONAL
 9336/10000 -3405/10000 14/10000
 -7320/10000 14779/10000   2763/10000
 -914/10000  1171/10000 8248/10000

Since no sensor satisfies the Luther-Ives criteria, these coefficients are usually determined by a least squares fit and are not exact. Some raw converters might use a different set of coefficients to better optimize other colors. I agree that the camera space is not as well defined as in such spaces as ProPhotoRGB, but that does not meas that a camera color space does not exist. See Chapter 6 of the Adobe DNG Specification: "Mapping Camera Color Space to CIE XYZ Space". Some colors would be rendered more accurately than others according to how these coefficients are determined. The situation is analogous to what might take place when I send some files for printing at a service bureau. I might not embed the profile, but could tell them the profile is for a specific printer and they could then assign the profile prior to printing.
 


Quote from: digitaldog
I would submit that if we can’t see it (an imagery specification of a color), its not really a color. Again, it might be semantics but in terms of human vision, color is a perceptual properly. If we can’t see it, its not a color (it can be a particular wavelength of light, but a color coordinate outside the spectrum locus isn’t a color in the sense we perceive it).
As I mentioned previously, different people perceive color differently. Some have red-green color blindness and others are tetrachromats. What if I can see a color and you can't? Is it still a color?
Logged

Wayne Fox

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4237
    • waynefox.com
Colour Space Question.
« Reply #11 on: March 24, 2010, 08:23:09 pm »

Quote from: bjanes
The DigitalDog ignores the opinions of Eric Walowit, Thomas Knoll and Chris Murphy when asserts that the raw file has no color space. Perhaps he is getting too old to learn new tricks or simply does not want to admit that he was shown to be wrong by an upstart in a previous thread.

A digital camera does not have a gamut in the usual sense, since it will output values for anything that is put in front of it, i.e. there are no out of gamut colors and hence no gamut (see here for an explanation). However, you can put a color target in front of a camera and obtain useful information. For example, drycreekphoto.com has "gamut" plots for some older digital cameras and one can interactively compare the camera response to the gamuts of ProphotoRGB, Adobe RGB and sRGB. Only ProPhotoRGB can represent all the colors captured by the camera, making it the preferred space into which one should render raw files.

Like any other color space, the camera space can be characterized by 3x3 matrix coefficients as shown for the Nikon D3x on the DXO site; as Mr. Knoll stated, one can use these coefficients to obtain a pretty good rendering of the image.

[attachment=21029:D3xColorSpace.png]
We have been down this road before ... an exhaustive thread arguing whether or not a raw has a color space.  I"m not sure why we need to dredge it up again ...

From a purely technical perspective whether or not a raw file has color space seems to have little bearing on the average user .. from a practical point of view a raw file is just that and during the conversion it is rendered into the color space of choice ... the only color space that really matters.  Seems anything beyond that is just debate for the sake of debate and not very helpful ...
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #12 on: March 24, 2010, 08:25:14 pm »

Quote from: bjanes
Colorspace information is given by reference.

Is it the correct reference or something assigned so to speak?

Quote
The raw file does contain the identity of the camera, and the characteristics of the Bayer filters are the same within manufacturing tolerances for a given camera model and can be determined and applied separately. Apparently a method for determining color matrix is defined in ISO standard 17321 as reported by DXO for various cameras as per my previous post.

Again, that appears to mean someone has assigned some specifics, but I don’t know that means a lot (one can for example define one color space onto another, an untagged document and that assignment may or may not be correct). I’m playing devils advocate to some degree. My point is, someone can say the data represents this or that, it may or may not. After reading Doug’s piece, sure seems that this is hardly empirical.

Quote
The DNG spec does contain spaces for two matrices for two illuminants. They are not supplied by most cameras, but are filled in by the DNG converter.

Again, what provides doesn’t seem to indicate anything about a specific color space (real, imagery or otherwise) does it?

Quote
Some raw converters might use a different set of coefficients to better optimize other colors.

That sounds like what I understand, begging the question, if two different converters both use differing coefficients, which is right and which is “wrong” and how do we (or anyone else) decide?

Quote
I agree that the camera space is not as well defined as in such spaces as ProPhotoRGB, but that does not meas that a camera color space does not exist.

It appears to exist when someone at some point says it does.

Quote
As I mentioned previously, different people perceive color differently. Some have red-green color blindness and others are tetrachromats. What if I can see a color and you can't? Is it still a color?

Agreed. But I don’t think its correct to say that colors that fall outside the CIE chromaticity diagram, colors no human can see, could be called a color considering the perceptual aspect of color.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #13 on: March 24, 2010, 08:26:45 pm »

Quote from: Wayne Fox
We have been down this road before ... an exhaustive thread arguing whether or not a raw has a color space.  I"m not sure why we need to dredge it up again ...

You’d have to ask the fellow who dredged it up again <g>.

Quote
From a purely technical perspective whether or not a raw file has color space seems to have little bearing on the average user .. from a practical point of view a raw file is just that and during the conversion it is rendered into the color space of choice ... the only color space that really matters.  Seems anything beyond that is just debate for the sake of debate and not very helpful ...

Damn straight! Or we could argue how many ICC profiles can dance on the head of a pin.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Colour Space Question.
« Reply #14 on: March 24, 2010, 09:45:58 pm »

Quote from: Wayne Fox
We have been down this road before ... an exhaustive thread arguing whether or not a raw has a color space.  I"m not sure why we need to dredge it up again.

From a purely technical perspective whether or not a raw file has color space seems to have little bearing on the average user .. from a practical point of view a raw file is just that and during the conversion it is rendered into the color space of choice ... the only color space that really matters.  Seems anything beyond that is just debate for the sake of debate and not very helpful ...
Hardly! How do you think that the camera color space (if it doesn't exist or is irrelevant) gets transformed into XYZ or ProPhotoRGB? I would suggest you also read Chapter 6 of the DNG spec on how this transformation is done.

These points need to be reiterated since many people continue to think that their camera's color space is sRGB.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #15 on: March 25, 2010, 09:16:11 am »

Quote from: bjanes
Hardly! How do you think that the camera color space (if it doesn't exist or is irrelevant) gets transformed into XYZ or ProPhotoRGB? I would suggest you also read Chapter 6 of the DNG spec on how this transformation is done.

These points need to be reiterated since many people continue to think that their camera's color space is sRGB.

I’d fully agree with you that the misconception that some have that their cameras color space is sRGB (or whatever setting on the camera body) needs to be clarified. That’s a bit different than the debate about the raw data and its color space.

If the raw data, prior to being introduced into a raw converter that makes some assumption about its color space to render it, what is the color space? Do we agree that until the stage that a converter makes this assumption, what’s the color space of this data, sitting on my CF card in the camera?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Colour Space Question.
« Reply #16 on: March 25, 2010, 01:04:43 pm »

Quote from: digitaldog
I’d fully agree with you that the misconception that some have that their cameras color space is sRGB (or whatever setting on the camera body) needs to be clarified. That’s a bit different than the debate about the raw data and its color space.

If the raw data, prior to being introduced into a raw converter that makes some assumption about its color space to render it, what is the color space? Do we agree that until the stage that a converter makes this assumption, what’s the color space of this data, sitting on my CF card in the camera?
I think that the camera space (or quasi space if you will) is determined by the colorimetric characteristics of the sensor, which would be reflected in the 3x3 matrix used by the raw converter or the in camera JPEG engine. Such a matrix can be derived by the methods outlined by the ISO 17321-1:2006 or by other means. DXO gives the ISO 17321 matrix coefficients for various cameras and the DNG converter fills in the matrix values for the camera model. I do note that the daylight matrix for the D3 in the DNG and DXO D50 values are different. Perhaps Eric Chan could elaborate if he is reading this thread.

The raw converter does not merely guess what matrix to use, but rather uses the matrix appropriate for the camera model and illumination. Again, these matrices do fully define the space, unlike the matrix values for a fully defined space such as ProPhotoRGB or sRGB. However, this quasi space is sufficiently well defined to enable a useful rendering of the raw file using them. The metamerism index can be calculated as Mr. Kerr explains in a separate article.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #17 on: March 25, 2010, 01:12:21 pm »

Quote from: bjanes
I think that the camera space (or quasi space if you will) is determined by the colorimetric characteristics of the sensor, which would be reflected in the 3x3 matrix used by the raw converter or the in camera JPEG engine. Such a matrix can be derived by the methods outlined by the ISO 17321-1:2006 or by other means.

Do you own that ISO article? I don’t. But it says in the teaser: “Two methods are provided, one using narrow spectral band illumination and the other using a spectrally and colorimetrically calibrated target.“ and I’m wondering again, like making ICC profiles for cameras, the target and illumination would (should) play a role on the results and thus the color space.

Quote
Perhaps Eric Chan could elaborate if he is reading this thread.

That would be real useful.

Quote
The raw converter does not merely guess what matrix to use, but rather uses the matrix appropriate for the camera model and illumination. Again, these matrices do fully define the space, unlike the matrix values for a fully defined space such as ProPhotoRGB or sRGB. However, this quasi space is sufficiently well defined to enable a useful rendering of the raw file using them. The metamerism index can be calculated as Mr. Kerr explains in a separate article.

I’m not suggesting its just a guess. I’m suggesting that lots of folks do this differently making some assumptions and thus, the actual “color space” or, if you want to now call it qusi space, is not well defined. At least at the state that the raw data is sitting on my card.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Colour Space Question.
« Reply #18 on: March 25, 2010, 05:22:32 pm »

Quote from: digitaldog
Do you own that ISO article? I don’t. But it says in the teaser: “Two methods are provided, one using narrow spectral band illumination and the other using a spectrally and colorimetrically calibrated target.“ and I’m wondering again, like making ICC profiles for cameras, the target and illumination would (should) play a role on the results and thus the color space.
Unfortunately, I don't have that reference. It costs too much for my tastes. Imatest Master can calculate the matrix coefficients from various test charts and Norman gives a reference to an article by Stephen Wolf for a method that uses a multicolor chart similar to what might be used for making an ICC profile of a camera. I don't make such profiles myself, but you might want to take a look at the article. Of course, Eric Chan knows all about such methods and I think the DNG profile editor and the Passport editor also make matrices; Eric has previous stated that the DNG editor also uses lookup tables.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Colour Space Question.
« Reply #19 on: March 25, 2010, 07:27:25 pm »

Quote from: bjanes
Unfortunately, I don't have that reference. It costs too much for my tastes. Imatest Master can calculate the matrix coefficients from various test charts and Norman gives a reference to an article by Stephen Wolf for a method that uses a multicolor chart similar to what might be used for making an ICC profile of a camera. I don't make such profiles myself, but you might want to take a look at the article. Of course, Eric Chan knows all about such methods and I think the DNG profile editor and the Passport editor also make matrices; Eric has previous stated that the DNG editor also uses lookup tables.

I’ve built plenty of profiles using plenty of targets and software. Again, that doesn’t mean said profile defines the actual color space. In fact that’s a huge stretch considering you feed the products an output referred, rendered image. So going back to “...can calculate the matrix coefficients from various test charts“ can I assume that like building an ICC profile, the same routine applies? I wonder if DNG profiles are again built upon demosaiced data and/or what assumptions about the “raw colorspace” if any is made and how.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".
Pages: [1]   Go Up