Pages: [1]   Go Down

Author Topic: Exact color match (5D, Nec 2690, Solux, B9180)  (Read 6183 times)

ihelfant

  • Newbie
  • *
  • Offline Offline
  • Posts: 4
Exact color match (5D, Nec 2690, Solux, B9180)
« on: June 22, 2008, 04:25:24 pm »

Hi all -- this is a bit long, so please bear with me:).  

I've been working on setting up a digital darkroom.  I have a neutral gray 4x 4 foot table painted in Munsel 8 with "wings" to provide a viewing area, overhead Solux 4700 K track lighting about 6 feet above the table, a NEC 2690 monitor with spectraview software and XRite 94 colorimeter (currently calibrated to 120, D65 and gamma 2.2), and a HP B9180 printer.  I blacken out the single window with a good blind.  I use a 5D and a 40D and shoot raw with the cameras set to RGB (although I think that the in-camera RGB only affects the jpeg files when I shoot both raw and jpgs concurrently).  I import into lightroom Beta 2 and also work with Photoshop CS3.  I've been printing color 13x19 images, mostly to HP Professional Satin paper and also to Hahnemuhle Fine Art Smooth Paper using soft-proofing in CS3.  I'm getting very good correspondence between the overall colors on the screen and print with the monitor calibrated to a luminance of 120 (also up to 140).  I've also tried D50 but find the colors correspond better at D65.  I use Prophoto as a workspace in CS3 and I understand that Lightroom uses Melissa RGB, which is very close to Prophoto.

Nevertheless, I find that I get a more exact correspondence between print and screen when I set the Color Settings in CS3 to "desaturate monitor colors" by 15 or 20 percent (and this is in the softproofing mode).  I know that you're not supposed to do this when you soft proof, but I'm wondering if anyone else has had the same experience or can lend any insight?  Could this just reflect the difference btw reflective light and the light from the monitor?  

Another question.  I bought a GretagMacbeth color chart and photographed it right on the table with the 4700 K Solux lighting.  I then imported the raw file into CS3 with the white balance set to 4700 K.  Both when I just leave it like that and when I click on one of the neutral gray squares to render the color balance exactly neutral (ie, equal RBG values for the gray square, which affects all the other colors too), I find that some of the colored squares correspond quite exactly (the monitor colors look exactly like the color chart colors) while others are substantially off.  The same thing happens when I print out the image; some squares match very closely btw print and original and others are noticeably different.

Can anyone offer any insight or suggestions as to whether I should pursue all of this any further?  I'm getting great prints, so perhaps worrying about this level of exactness is taking it too far?  I'm especially interested in your opinion about the desaturation setting in Photoshop and whether any of you use this in conjunction with softproofing.

One other thing -- a technically oriented friend suggested that one of the reasons that Lightroom may not be incorporating soft proofing in version 2 is that the designers have basically finessed the issue by incorporating a sort of "under the radar" soft proofing within the very color system utilized by Lightroom.  Does anyone here print straight from Lightroom even for their fine arts or important prints without using Photoshop's soft proofing?
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #1 on: June 22, 2008, 06:00:12 pm »

Where did the paper profiles come from?

That you get nearly ideal soft proof but only with the desatruate on, gives me the impression the soft proof table in the output profile is just a bit off. You can use this destaruate (although I'd be very careful about having it on and off only when necessary) but its not ideal.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

papa v2.0

  • Full Member
  • ***
  • Offline Offline
  • Posts: 206
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #2 on: June 22, 2008, 07:21:27 pm »

soft proof ing is still a major challenge in the graphic arts.
unless you have the same adaptive white point and luminance levels you will never get a screen to print match.

Even the gamut differences between the two systems make it obvious that there will never be a one 2 one match.

if your main industry is output image, gauge that not that screen

Its not easy.
Trial and error best way
keep a note of all your settings and then you can make a subjective decision about your output compared to screen appearance.
Logged

Allen Phillips

  • Newbie
  • *
  • Offline Offline
  • Posts: 1
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #3 on: June 22, 2008, 11:52:09 pm »

The color chart does not match exactly because you need to profile your camera .... just a gray balance is not enough. To get an accurate match of the Macbeth chart you need to profile your camera then apply that profile to the image.... in most cases you'll also need to edit the camera profile for the most accurate result.
« Last Edit: June 22, 2008, 11:59:15 pm by Allen Phillips »
Logged

BobDavid

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3307
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #4 on: June 23, 2008, 10:17:13 am »

Quote
The color chart does not match exactly because you need to profile your camera .... just a gray balance is not enough. To get an accurate match of the Macbeth chart you need to profile your camera then apply that profile to the image.... in most cases you'll also need to edit the camera profile for the most accurate result.
[a href=\"index.php?act=findpost&pid=202954\"][{POST_SNAPBACK}][/a]

Camera profiling will get you very close. However, the problem with camera profiling is that it is highly dependent on lighting. If you profile your camera in the studio under highly controlled lighting, you will only get accurate color when you replicate the lighting scheme used to make the profile. Camera profiling is essential if you photograph artwork on a copystand.
Logged

ihelfant

  • Newbie
  • *
  • Offline Offline
  • Posts: 4
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #5 on: June 23, 2008, 01:53:11 pm »

Thanks for all those really helpful comments!  I have the impression that profiling a camera is a) quite expensive and  necessary mostly for studio or other extremely precise and exacting work that involves color-matching.  Therefore, I imagine that would be taking it too far for my purposes (primarily landscape, nature, portrait, and cityscape photography).  I was intrigued by Andrew's comment that the soft-proofing issue may reflect my use of generic paper profiles (yes: I'm just using the one's available from HP for the B9180).  I suppose it may be worth spending the 50 or more bucks to get exact profiles of the one or two papers I mentioned, which are the ones I'm using most; that would eliminate another variable, at least.  I'm still curious if any of you print fine art or "important" enlargements directly from Lightroom w/out feeling the need for softproofing (I assume not).  I'm also very curious about whether anyone else uses CS3's desaturation function to get a more exact softproofing match.  Thanks again for your insights! -- Ian H.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #6 on: June 23, 2008, 02:20:33 pm »

Quote
Thanks for all those really helpful comments!  I have the impression that profiling a camera is a) quite expensive and  necessary mostly for studio or other extremely precise and exacting work that involves color-matching.  Therefore, I imagine that would be taking it too far for my purposes (primarily landscape, nature, portrait, and cityscape photography).

Yup, skip it. Especially if you're using an Adobe Raw converter or one by Apple.

Aside from that, once you do get data in a working space inside of Photoshop, the matching should be working no matter what profile for the capture device you may or may not be using.


Quote
I was intrigued by Andrew's comment that the soft-proofing issue may reflect my use of generic paper profiles (yes: I'm just using the one's available from HP for the B9180).  I suppose it may be worth spending the 50 or more bucks to get exact profiles of the one or two papers I mentioned, which are the ones I'm using most; that would eliminate another variable, at least.

Printer profiles have two tables. One controls what goes to the printer, the other the soft proof. Its possible they are not in sync.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

rdonson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3263
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #7 on: June 23, 2008, 03:22:13 pm »

Quote
I'm still curious if any of you print fine art or "important" enlargements directly from Lightroom w/out feeling the need for softproofing (I assume not). 

I'm also very curious about whether anyone else uses CS3's desaturation function to get a more exact softproofing match.  Thanks again for your insights! -- Ian H.
[a href=\"index.php?act=findpost&pid=203131\"][{POST_SNAPBACK}][/a]

1) Nope - the only thing I print from Lightroom is sports posters on a high gloss product where color fidelity isn't critical.  It just has to pop.  Virtually all my fine art work uses Qimage rather than Lightroom for printing anyway - softproofed in CS3.

2) Nope - how exact a softproof match are you trying to achieve?  One of the skills of printing in my mind is learning the translation from screen (emissive) to paper (reflective).  If softproofing gets me 90% of the way there I'm a happy camper.  If its less than thatn I consider tweaking the softproofing in the profile.
Logged
Regards,
Ron

jackbingham

  • Full Member
  • ***
  • Offline Offline
  • Posts: 205
    • http://www.integrated-color.com
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #8 on: June 24, 2008, 07:35:25 am »

Quote from: BobDavid,Jun 23 2008, 02:17 PM
Camera profiling will get you very close. However, the problem with camera profiling is that it is highly dependent on lighting. If you profile your camera in the studio under highly controlled lighting, you will only get accurate color when you replicate the lighting scheme used to make the profile. Camera profiling is essential if you photograph artwork on a copystand.

Camera profiling is not highly dependent on lighting. You do indeed need to profile your camera in highly controlled conditions so it will work under varied conditions. The better the data the more universal the profile. Kodak chips in general require a tungsten and daylight profile. Other's most often only one. That being said since, you can't use custom profiles in Camera Raw or Lightroom it's moot.
Logged
Jack Bingham
Integrated Color Corp Maker

Tim Lookingbill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2436
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #9 on: June 24, 2008, 04:35:19 pm »

..."I bought a GretagMacbeth color chart and photographed it right on the table with the 4700 K Solux lighting."...

I wonder how the spectral qualities of the Solux lamps used in this manner affects accuracy shooting Raw. Is it possible to get accurate results just clicking one of the gray squares to get R=G=B in the Raw capture of this target shooting under these lights?

What was the uncorrected color cast in the Raw capture using these lights and how off looking was it?
« Last Edit: June 24, 2008, 04:36:48 pm by tlooknbill »
Logged

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2115
    • Web
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #10 on: June 24, 2008, 08:40:00 pm »

Quote
Kodak chips in general require a tungsten and daylight profile. Other's most often only one.

Interesting to hear you say that. With several of the non-Kodak chips I've studied the sensor response appears to vary quite a bit between tungsten and daylight. That said, it is certainly possible to have a single profile that generates visually pleasing results for both conditions.
Logged
Eric Chan

jackbingham

  • Full Member
  • ***
  • Offline Offline
  • Posts: 205
    • http://www.integrated-color.com
Exact color match (5D, Nec 2690, Solux, B9180)
« Reply #11 on: June 25, 2008, 08:03:31 am »

Clicking on the gray patch to get a gray balance will be far more accurate than measuring the color temp of the bulbs, (color meters are not great) or taking the manufacturer's word for it, and then setting the camera to a specific color temp. This will vary from camera to camera and raw converter to raw converter by a wide margin. Set the camera to daylight, click on a gray patch.
Logged
Jack Bingham
Integrated Color Corp Maker
Pages: [1]   Go Up