Pages: [1]   Go Down

Author Topic: Comparing ABW and BW Using standard RGB profiles on a 9800  (Read 4224 times)

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Comparing ABW and BW Using standard RGB profiles on a 9800
« on: December 02, 2017, 08:14:57 pm »

This shows the actual L*, a*, and b* values when printing a b&W image using ABS and standard RGB color managed from a large 5.5k patch set with extra neutrals.  The principal differences are that the transitions are much smoother between adjacent neutrals with the ABW and vary a bit more printing using a color profile.

The charts show rendering from L* 4 to 95. The blue lines are ABW mode.

The first chart is the L* out v in. This was done for ABW using a custom LUT. The RGB managed (orange) lines are from using a standard ICC profile. They are very close so the second chart shows the difference between requested L* and printer L*.

The third chart shows a*, and the fourth chart b*.

Also included is a photograph, processed scene referred, of printed images. The first image was printed with ABW, the second with standard color management. The third and fourth are the same but with saturation greatly boosted to bring out the color shifts. The printed gradient image is from using Photoshop's gradient tool and is not linear but the approximate L* on the image (Photoshop Info Tab) is reasonably accurate.

The most striking difference is the relative smoothness with which the a* and b* change across the tone curve for ABW. The 9800 raw device RGB color response is rather lumpy and ICC profiles, with only 36 LUT spacings, can only compensate for the average when they rapidly change.


« Last Edit: December 02, 2017, 08:39:50 pm by Doug Gray »
Logged

unesco

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 254
Re: Comparing ABW and BW Using standard RGB profiles on a 9800
« Reply #1 on: December 03, 2017, 04:27:49 am »

Hi Doug, Very interesting, thank you!
Have you tried to compare it to QTR? As well as to different resolution of the printer within ABW itself?
Was dithering on/off?
Logged

Alan Goldhammer

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4344
    • A Goldhammer Photography
Re: Comparing ABW and BW Using standard RGB profiles on a 9800
« Reply #2 on: December 03, 2017, 08:47:28 am »

Doug,

Thanks for this.  It pretty much looks like things are pretty close in terms of evenness of the B/W response.  When you prepared your RGB profile did you add extra B/W patches to the set for profiling?  that's what I do when I use ArgyllCMS.

Alan
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing ABW and BW Using standard RGB profiles on a 9800
« Reply #3 on: December 03, 2017, 03:21:21 pm »

Doug,

Thanks for this.  It pretty much looks like things are pretty close in terms of evenness of the B/W response.  When you prepared your RGB profile did you add extra B/W patches to the set for profiling?  that's what I do when I use ArgyllCMS.

Alan
Yes, it has extra near neutral patches as well as a several hundred extra patches in portions of the gamut that have high intrinsic variation. Those are areas of high saturation with lot's of Y. These colors are the highly saturated ones ranging from green->yellow->orange.
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing ABW and BW Using standard RGB profiles on a 9800
« Reply #4 on: December 03, 2017, 03:34:08 pm »

Hi Doug, Very interesting, thank you!
Have you tried to compare it to QTR? As well as to different resolution of the printer within ABW itself?
Was dithering on/off?
I haven't ever used QTR. I don't do much B&W except for making specialized charts with superimposed sine waves for calibrating things like imager plane registration. I had previously just used RGB mode but recently found out ABW produces much smoother response. As a consequence, when making B&Ws I've made a Matlab function that maps directly from a tiff image and converts to the raw device values. It works for Relative Colorimetric as well as Absolute with an average error under .2 dE in ABW mode.

I use and print 16 bit tiff images from Photoshop. Since I run Windows, and printer drivers are 8 bits, I enable dither for 8 bit conversions in the global color settings otherwise the dE values are somewhat higher.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Comparing ABW and BW Using standard RGB profiles on a 9800
« Reply #5 on: December 03, 2017, 04:22:12 pm »

Since I run Windows, and printer drivers are 8 bits, I enable dither for 8 bit conversions in the global color settings otherwise the dE values are somewhat higher.
Got to love what 'noise' (not image data) can do for such reports!
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing ABW and BW Using standard RGB profiles on a 9800
« Reply #6 on: December 03, 2017, 04:43:07 pm »

Got to love what 'noise' (not image data) can do for such reports!

Injected LSB dither at 600DPI or 720DPI isn't close to visible but becomes more problematic at low DPI.  OTOH, without dithering one can get banding on synthetic images with 8 bit drivers though rare. I posted a few gradients where single bit (8 bit driver) changes produced dE76 of just over 1. It varies with printers. Not easy to find. Wrote a Matlab script to search the entire driver space to find the few that were over 1.  16 bit / 8 bit is largely marketing hype.

At higher DPIs the printer driver (8 or 16 bit) will introduce a form of it because even at 2880/1440 dots per inch colors can't be accurately printed without significant adjacent pixel adjustments.
Logged
Pages: [1]   Go Up