RE: Printing of photography
RE: Printing of photography
- Subject: RE: Printing of photography
- From: David Remington <email@hidden>
- Date: Fri, 24 Dec 2010 21:14:17 -0500
Hello,
There are two distinct uses of targets under discussion here.
A physical target included in the frame during capture as a means of optimizing and characterizing the exposure (input reference), and a digital image of a target included in the frame of an image that is delivered to a client or a vendor (output reference).
The original post questions the practice of including a target in the capture, leaving it in place through image processing/color tonal matching, and sending it on to the client in the image. Ken and Robin point out shortcomings in that workflow and I am in agreement. In addition to the metamerism issues discussed, exposure and tonal response may need to be optimized for the subject leading to an "incorrectly" rendered target. I have discussed this topic before on other boards and venues so if you have read some of this before forgive me for including a bit of cut and paste ;)
In our studio we use a range of targets for set-up, calibration, and system benchmarking. We have used most of the usual suspects at some point or another. We do not photograph a target in-frame with the intent of retaining it as a capture record. Color and tonal corrections needed to accurately reproduce artwork and those need to accurately reproduce a color checker or other reference are almost always unique, and can be significantly different. Exposure and, in some cases, the tonal response curve need to be adjusted based on the material and the lighting condition. e.g. parchment can be very reflective and it is desirable to maintain texture detail in the capture. Correct exposure for the white patch on a color checker or a Q13 will result in overexposure for a parchment manuscript (and other light or reflective materials). We use diffused lighting and an overhead bounce to photograph illuminated/gold leaf manuscripts. We are deliberately introducing flare and compensate using the capture tonal response curve. Other lighting conditions and classes of originals call for other methods. Our practice has long been to (separately) correct the original to the original and the target to the target. The target image is then associated with or pasted into our master files and can be included with images sent out to a patron or for reproduction. For a long time we used modified Kodak Q-13 targets shot at various resolutions, with our various cameras, individually proceed for accuracy. The logical, and greatly simplified, extension of that practice was to make the target image from scratch recently we switched over to using a synthetic target. I consider the target images we supply with files to be an aid for output reference.
Here is an example. (note, jpegs delivered through our IDS are untagged sRGB)
http://ids.lib.harvard.edu/ids/view/21337876?buttons=y
Here is the rundown on targets we use or have used at some point.
I want to emphasize that while targets are very useful for system evaluation and help assure consistent results, they do not by themselves assure accurate results.
24 patch Color Checker/ Mini Color Checker: used to set camera gray balance and exposure and for monitoring exposure consistency. Correct exposure for the piece is often different from correct exposure for the target so we use a weighted target exposure. The color checker is also useful for initial evaluation of camera profiles.
Color Checker SG: Used for making camera profiles.
Color Checker Passport: Tried it out with the Canons.
Robin Myers digital gray card: We sometimes use these for camera gray balance as well.
Kodak Q-60: Used for film scanner profiles.
Slant edge target/ USAF-1951 target/ Kodak TL-5003: We have used all of these to evaluate detail and resolution visually or using software.
Image Associates Golden Thread targets: We have uses these to benchmark system performance.
I have though about this a lot but am always open to getting feedback from and hearing the opinions of others wrestling with the same issues.
--David
On Dec 20, 2010, at 3:02 PM, email@hidden wrote:
> Send Colorsync-users mailing list submissions to
> email@hidden
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://lists.apple.com/mailman/listinfo/colorsync-users
> or, via email, send a message with subject or body 'help' to
> email@hidden
>
> You can reach the person managing the list at
> email@hidden
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Colorsync-users digest..."
>
>
> Today's Topics:
>
> 1. RE: Printing of photography (Robert Rock)
> 2. RE: Printing of photography (Robert Rock)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 20 Dec 2010 14:28:11 -0500
> From: "Robert Rock" <email@hidden>
> Subject: RE: Printing of photography
> To: "'David Ramsey'" <email@hidden>, "'ColorSync
> Users Mailing List'" <email@hidden>
> Message-ID: <001d01cba07c$0c58b4c0$250a1e40$@pchaninc.com>
> Content-Type: text/plain; charset="us-ascii"
>
> Personally, I wouldn't make any tweaks or corrections until AFTER the first
> proof. The pre-press or printer's job is to print a proof showing accurate
> reproduction of the chart/strip. If this is done and the color of the
> artwork is still not right, even though the test strips are right on, then I
> would tweak from there. I have found that in MOST situations, if the test
> strip/target appearing in the image is reproduced accurately, the color of
> the artwork or subject will also be very close indeed. Of course if the
> artwork contains special pigments or substrates which make it difficult or
> impossible to reproduce, then certainly another method of color
> control/correction must be investigated. But again, for MOST end use (museum
> catalogs, artbooks, etc...), matching the test strip is usually close enough
> in our experience. We have been printing calendars, posters, artbooks,
> etc...for a number of years for many publishers, and only in rare instances
> does this procedure not work (for example, some of the Klimt paintings
> which have lots of metallic gold, or in some contemporary works containing
> vibrant pigments, etc...).
> We can go on nit-picking about this, but for most commercial end uses, this
> serves as an acceptable starting point.
>
> Bob
>
> -----Original Message-----
> From: David Ramsey [mailto:email@hidden]
> Sent: Monday, December 20, 2010 12:05 PM
> To: email@hidden; ColorSync Users Mailing List
> Subject: Re: Printing of photography
>
> So, If I do include a color target and I, for example, move the blues
> towards purple is this not going to throw whomever is looking at he image
> off. It seems that if the printer is trying to "match" a target then he/she
> will be incorrectly "correcting" the artwork. This leads me to this
> question, If I am in a color managed workflow (my responsibility to be
> accurate) and I send you or whomever an RGB
> ADOBE98 file tagged as such and you open the file on your color managed
> system (your responsibility to be accurate) would you not see the exact same
> thing as I do? And likewise, should the printer see the same?
>
> I am not comfortable sending special instructions downstream with a file.
> Who knows where or with who it is going to land.
>
> David Ramsey
>
> On Dec 20, 2010, at 11:40 AM, Robert Rock wrote:
>
>> I could not disagree more with Ken's suggestion that NO color target
>> should be used when photographing artwork. In my opinion, there
>> absolutely SHOULD be some known reference included in the photograph,
>> perhaps a Macbeth ColorChecker, or even an old Kodak Q-13 chart.
>> Assuming that the image will be reproduced from this digital capture
>> later on, what would pre- press or the printer use for a reference?
>> You have to assume that they have never seen the image and have no
>> idea what it should look like. Instead of wasting everyone's time
>> sending proofs back and forth, it at least gives them a visual
>> reference that they are in the ballpark when proofing. Ken is right
>> that there are some issues that may affect color accuracy (metamerism,
>> etc...), but there still needs to be a base from which to start. The
>> printer or pre-press person needs to have some reference as to what
>> the image should look like. If there are no color bars or reference
>> data in the image, the first thing the printer will tell you before
>> proofing is that 'he has no idea what the image should look like and
>> that all he did was output to proof...'. After proofing, if further
>> color tweaking and adjustments are desired to more accurately match
>> the original, then have at it.
>> There most
>> certainly WILL be the need for tweaking of the color and/or tonal
>> contrast, if not for reasons that Ken mentioned, perhaps because of
>> the inks used, paper, type of dot pattern, etc...But again, I believe
>> that there should be a starting point or reference for the printer,
>> and that should be dictated by the use of a known color reference
>> included in the capture. One can agree or disagree with this, but
>> bottom line is that 1) Many printers WILL require it, and 2) What harm
>> does it do to include it in the capture? You can always give the
>> instructions to ignore the patch if you so desire.
>>
>> Bob Rock
>>
>> -----Original Message-----
>> From: colorsync-users-bounces+rock=email@hidden
>> [mailto:colorsync-users-bounces+rock=email@hidden]
>> On Behalf
>> Of Ken Fleisher
>> Sent: Monday, December 20, 2010 10:10 AM
>> To: David Ramsey; ColorSync Users Mailing List
>> Subject: Re: Printing of photography
>>
>> On Mon, Dec 20, 2010 at 9:18 AM, David Ramsey
>> <email@hidden> wrote:
>>>
>>> 2 questions;
>>>
>>> 1) I am getting ready to photograph some works on paper for a museum
>>> and I was asked about including "color bars" in the photos. Because
>>> of the nature of the materials used in these works I have to adjust
>>> specific colors as we shoot (Color managed workflow and calibrated
>>> display). Colors are adjusted individually, for example, purples may
>>> go
>> towards blue reds towards magenta.
>>> So if I place color bars in the frame and will this not create havoc
>>> down the line since I am moving individual colors around? Also which
>>> color bars should I use if they are to be included?
>>
>> I do not recommend including a color target when photographing artwork
>> for reproduction for the following reasons:
>>
>> 1. The camera's spectral sensitivities are different from a human's.
>> Because of this, the differences in the material on a color target and
>> the artwork can result in "observer" metamerism. (i.e. getting the
>> color of the target correct does not imply that the color of the
>> artwork is
>> correct.)
>>
>> 2. Because of observer metamerism, the color target is not a reliable
>> indicator of an accurate color capture of the artwork.
>>
>> 3. As you describe, if you make color edits to the image to make the
>> artwork correct, the colors on the target can no longer be used as a
>> reference of any sort, since those colors will necessarily be moved as
>> well.
>>
>> 4. Trying to match the colorimetry of target colors all the way to the
>> printed page is an incorrect goal anyhow. If you do this, your
>> reproductions might be colorimetrically correct (if you are lucky),
>> but will likely appear dark and flat. Instead, you must try to match
>> the color appearance, which usually involves opening up the image and
>> increasing contrast (many perceptual rendering intents achieve this
>> for us when printing, but not so for image capture).
>>
>> I recently spoke on this very topic at "Currentl Practices in Fine Art
>> Reproduction" which was a symposium at the Rochester Institute of
>> Technology. You can view a video of the talk at
>> <http://tinyurl.com/3x5rv6v>. This was a panel discussion on target-
>> based workflows and my part begins at 29 minutes into the video
>> (though you may find the entire panel interesting.).
>>
>> Ken Fleisher
>> Color Scientist
>> National Gallery of Art, Washington, D.C.
>> _______________________________________________
>> Do not post admin requests to the list. They will be ignored.
>> Colorsync-users mailing list (email@hidden)
>> Help/Unsubscribe/Update your Subscription:
>> .com
>>
>> This email sent to email@hidden
>>
>>
>
>
>
> ------------------------------
>
> Message: 2
> Date: Mon, 20 Dec 2010 14:44:27 -0500
> From: "Robert Rock" <email@hidden>
> Subject: RE: Printing of photography
> To: "'David Ramsey'" <email@hidden>, "'ColorSync
> Users Mailing List'" <email@hidden>
> Message-ID: <002801cba07e$53667c60$fa337520$@pchaninc.com>
> Content-Type: text/plain; charset="us-ascii"
>
> Of course, I might add, as some have pointed out, if the creator/originator
> of the image capture KNOWS what the images should look like, and is adept
> at color correction and can supply a digital file that has already been
> accurately corrected, then by all means that is the file that should be sent
> to the printer, along with color reference if available. Maybe I should have
> been more clear...I'm talking about situations where images are captured,
> sometimes by an outside photographer, and then archived away. Maybe years
> later, a publisher requests an image for reproduction, maybe there's been a
> turnover of museum employee's, or any number of other scenarios. I would
> think it would be beneficial to have at least some reference in the image.
>
> Bob
>
> -----Original Message-----
> From: Robert Rock [mailto:email@hidden]
> Sent: Monday, December 20, 2010 2:28 PM
> To: 'David Ramsey'; 'ColorSync Users Mailing List'
> Subject: RE: Printing of photography
>
> Personally, I wouldn't make any tweaks or corrections until AFTER the first
> proof. The pre-press or printer's job is to print a proof showing accurate
> reproduction of the chart/strip. If this is done and the color of the
> artwork is still not right, even though the test strips are right on, then I
> would tweak from there. I have found that in MOST situations, if the test
> strip/target appearing in the image is reproduced accurately, the color of
> the artwork or subject will also be very close indeed. Of course if the
> artwork contains special pigments or substrates which make it difficult or
> impossible to reproduce, then certainly another method of color
> control/correction must be investigated. But again, for MOST end use (museum
> catalogs, artbooks, etc...), matching the test strip is usually close enough
> in our experience. We have been printing calendars, posters, artbooks,
> etc...for a number of years for many publishers, and only in rare instances
> does this procedure not work (for example, some of the Klimt paintings
> which have lots of metallic gold, or in some contemporary works containing
> vibrant pigments, etc...).
> We can go on nit-picking about this, but for most commercial end uses, this
> serves as an acceptable starting point.
>
> Bob
>
> -----Original Message-----
> From: David Ramsey [mailto:email@hidden]
> Sent: Monday, December 20, 2010 12:05 PM
> To: email@hidden; ColorSync Users Mailing List
> Subject: Re: Printing of photography
>
> So, If I do include a color target and I, for example, move the blues
> towards purple is this not going to throw whomever is looking at he image
> off. It seems that if the printer is trying to "match" a target then he/she
> will be incorrectly "correcting" the artwork. This leads me to this
> question, If I am in a color managed workflow (my responsibility to be
> accurate) and I send you or whomever an RGB
> ADOBE98 file tagged as such and you open the file on your color managed
> system (your responsibility to be accurate) would you not see the exact same
> thing as I do? And likewise, should the printer see the same?
>
> I am not comfortable sending special instructions downstream with a file.
> Who knows where or with who it is going to land.
>
> David Ramsey
>
> On Dec 20, 2010, at 11:40 AM, Robert Rock wrote:
>
>> I could not disagree more with Ken's suggestion that NO color target
>> should be used when photographing artwork. In my opinion, there
>> absolutely SHOULD be some known reference included in the photograph,
>> perhaps a Macbeth ColorChecker, or even an old Kodak Q-13 chart.
>> Assuming that the image will be reproduced from this digital capture
>> later on, what would pre- press or the printer use for a reference?
>> You have to assume that they have never seen the image and have no
>> idea what it should look like. Instead of wasting everyone's time
>> sending proofs back and forth, it at least gives them a visual
>> reference that they are in the ballpark when proofing. Ken is right
>> that there are some issues that may affect color accuracy (metamerism,
>> etc...), but there still needs to be a base from which to start. The
>> printer or pre-press person needs to have some reference as to what
>> the image should look like. If there are no color bars or reference
>> data in the image, the first thing the printer will tell you before
>> proofing is that 'he has no idea what the image should look like and
>> that all he did was output to proof...'. After proofing, if further
>> color tweaking and adjustments are desired to more accurately match
>> the original, then have at it.
>> There most
>> certainly WILL be the need for tweaking of the color and/or tonal
>> contrast, if not for reasons that Ken mentioned, perhaps because of
>> the inks used, paper, type of dot pattern, etc...But again, I believe
>> that there should be a starting point or reference for the printer,
>> and that should be dictated by the use of a known color reference
>> included in the capture. One can agree or disagree with this, but
>> bottom line is that 1) Many printers WILL require it, and 2) What harm
>> does it do to include it in the capture? You can always give the
>> instructions to ignore the patch if you so desire.
>>
>> Bob Rock
>>
>> -----Original Message-----
>> From: colorsync-users-bounces+rock=email@hidden
>> [mailto:colorsync-users-bounces+rock=email@hidden]
>> On Behalf
>> Of Ken Fleisher
>> Sent: Monday, December 20, 2010 10:10 AM
>> To: David Ramsey; ColorSync Users Mailing List
>> Subject: Re: Printing of photography
>>
>> On Mon, Dec 20, 2010 at 9:18 AM, David Ramsey
>> <email@hidden> wrote:
>>>
>>> 2 questions;
>>>
>>> 1) I am getting ready to photograph some works on paper for a museum
>>> and I was asked about including "color bars" in the photos. Because
>>> of the nature of the materials used in these works I have to adjust
>>> specific colors as we shoot (Color managed workflow and calibrated
>>> display). Colors are adjusted individually, for example, purples may
>>> go
>> towards blue reds towards magenta.
>>> So if I place color bars in the frame and will this not create havoc
>>> down the line since I am moving individual colors around? Also which
>>> color bars should I use if they are to be included?
>>
>> I do not recommend including a color target when photographing artwork
>> for reproduction for the following reasons:
>>
>> 1. The camera's spectral sensitivities are different from a human's.
>> Because of this, the differences in the material on a color target and
>> the artwork can result in "observer" metamerism. (i.e. getting the
>> color of the target correct does not imply that the color of the
>> artwork is
>> correct.)
>>
>> 2. Because of observer metamerism, the color target is not a reliable
>> indicator of an accurate color capture of the artwork.
>>
>> 3. As you describe, if you make color edits to the image to make the
>> artwork correct, the colors on the target can no longer be used as a
>> reference of any sort, since those colors will necessarily be moved as
>> well.
>>
>> 4. Trying to match the colorimetry of target colors all the way to the
>> printed page is an incorrect goal anyhow. If you do this, your
>> reproductions might be colorimetrically correct (if you are lucky),
>> but will likely appear dark and flat. Instead, you must try to match
>> the color appearance, which usually involves opening up the image and
>> increasing contrast (many perceptual rendering intents achieve this
>> for us when printing, but not so for image capture).
>>
>> I recently spoke on this very topic at "Currentl Practices in Fine Art
>> Reproduction" which was a symposium at the Rochester Institute of
>> Technology. You can view a video of the talk at
>> <http://tinyurl.com/3x5rv6v>. This was a panel discussion on target-
>> based workflows and my part begins at 29 minutes into the video
>> (though you may find the entire panel interesting.).
>>
>> Ken Fleisher
>> Color Scientist
>> National Gallery of Art, Washington, D.C.
>> _______________________________________________
>> Do not post admin requests to the list. They will be ignored.
>> Colorsync-users mailing list (email@hidden)
>> Help/Unsubscribe/Update your Subscription:
>> .com
>>
>> This email sent to email@hidden
>>
>>
>
>
>
> ------------------------------
>
> _______________________________________________
> Colorsync-users mailing list
> email@hidden
> http://lists.apple.com/mailman/listinfo/colorsync-users
>
> End of Colorsync-users Digest, Vol 7, Issue 283
> ***********************************************
David Remington
Manager, Digital Imaging and Photography Services
Digital Imaging and Photography Services
D-70 Widener Library
Cambridge, MA 02138
p 617-495-4701
f 617-495-0403
w http://imaging.harvard.edu
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden