NOTES

Note A1- Lab measurement

The version of Lab we will use is CIE L*a*b*, simply label Lab. L stands for Lightness and runs from 0 to 100. 0 is no light (or Black) and 100 is full light (or white). The a and b components measure color in a somewhat non intuitive way with endpoint color similar to those shown here. The b scale runs from a light blue (-127) to a yellow (+127). The a scale goes from a blue/ green (-127) to a red/magenta (+127). At a=0, b=0 there is no color, only gray. We will be using these scales to measure grayscale lightness and to look for and measure color casts (and tonality). We choose to use this scale because its devise independent character. Unlike RGB and k measures the Lab numbers will not change from space to space. Four types of reading are taken in Lab. They are listed below: L(ps) L(x), a(x), b(x) Measures read from grayscale patches defined in Photoshop PS3. This will give us information on how various spaces work. Measures read from an X-rite Spectrophotometer. To obtain these readings we first print grayscale patches in ABW based on various combinations of paper, ink and settings. These will then be read by the Spectrophotometer and converted into Lab numbers. Lightness measures read from a scanner using methodology similar to that discussed above. Scanner readings will not in general be as accurate as the spectrophotometer readings but will be good enough for side by side comparisons where we mostly want to know if two patches are the same or different and do not need to depend on the absolute value of the reading. L(ss) Measures taken from a screen save of a soft proof (under a particular profile). These will be mostly useful to see how things appear on the monitor.

L(sc)

Note A2 Standard Settings

When not explicitly mentioned the following settings are being used to make prints. Media Setting: Enhanced Matte Paper Actual Paper: Enhanced Matte

Color: Advanced B & W Photo Print Quality: SuperFine 1440x720, High Speed - OFF, Finest Detail- OFF Color Toning: Neutral Tone: Dark All other Printer Color Adjustments set at 0 Color Management: No Color Management Profile: Adobe RGB (1998) The primary print is made on the Epson 3800. In most cases a second set of prints were made on the 4800 printer just to very that results generalize to a second printer.

Note B1-Ink Usage

In the main part of the text we look at ink usage in ABW via the use of a microscope. Here we take a second cut at this issue by printing a large number of gray patches and observe the ink usage via the Epson monitor of ink levels. The patches we print are 8.5 x 11.0. These were printed as follows: k 100 54 29 15 Pages printed 10 20 40 80 L(ps) 0 50 75 87

The intensity measure k is here defined in the gamma 2.2 working space. The ink levels reported below are printed in ABW with the settings-No Color Management, dark, neutral and are printed on 3800 on Enhanced Matte paper.

Comments: Mostly full strength black ink is used at the black point, but a small amount of Light Cyan, Light Magenta and Yellow are also used. At the midpoint (L=50) light black is the dominant ink followed by light magenta and light cyan. At lower density levels light light black is used. Below is a summary over the 150 pages printed.

Comments:

There is no evidence that full strength cyan or magenta, A small amount of yellow is used

Note B2 Color cast by color settings

To study color cast by color settings we proceed as follows. A 21 step wedge is printed for each of the three settings-black, color and advanced black and white. These 63 patches are read by the spectrophotometer and the resulting color scores are plotted below.

Enhanced Matte settings, printed on enhanced matte paper.

Comment:

One of the things one would like from a setting is a smooth linear color transition from the papers white point to its black point. ABW appears superior to both Black and Color in this regard.

Note C Conversion between various Spaces

One of the advantages of the Photoshop Color Management setting is that it converts automatically between any work space and any other work space or output profile. Epson recommended the color management setting No Color Management when using ABW. Here there is no automatic conversion with this setting, thus the working profile can affect the final print (and the appearance of the image on the screen). For these reasons we study spaces and profiles and their effects here. Again we note using Photoshop Color Management would avoid these issues but requires a profile to make a print. ABW profiles are discussed in the main portion of the paper.

C1 Gamma 2.2 v sRGB -It is sometime claimed that sRGB is a gamma 2.2 space. I take this claim to mean that for R=G=B and k=R/255 the same lightness values result from the two space. My calculations support this claim for all non small values of L. Although there may be small differences in at the dark end of the scale they are generally too small to see. From a practical point of view, one can treat sRGB as visually equivalent to gamma=2.2.
From the point of view of what appears on the screen, a screen save demonstrates this.

The above is a screen save of two 21 step wedges defined in different work spaces. Lightness values L(ss) are noted on the gray patches.

The above considers the appearance on a monitor. To verify that these two spaces are visually equivalent after printing, an 11 step wedges is developed in each of the two work spaces and printed, the output is scanned and the resulting Lightness values L(sc) are plotted.

We again note that the spaces are very similar in terms of their lightness values. They are also quite linear in L. Other spaces that are the same or very similar to Gamma 2.2 are Adobe98 and 30% Dot Gain.

C2 Gamma 1.8-Gamma 1.8 is noticeably different from gamma 2.2. This can be seen from the step wedges below. These were produced by constructing equal interval wedges in the two spaces. These were then converted to Lab for comparison and display. Numbers shown on the wedge are values of L(ps).

Below is a graph of lightness values from prints (L(x)), printed on the same paper and with the same setting as above.

Comment: Both the screen and print show noticeable difference between the two spaces. Gamma 1.8 produces darker prints (lower values of L(x)). Gamma 2.2 is more linear than 1.8.

C3 Comparisons of gamma 1.8 to 20% Dot Gain Here is a wedge comparison done in the same manner as above.

Comment: Gamma 1.8 and 20% Dot Gain are visually the same. The importance of this is that to the extent that gamma 1.8 causes a soft proofing problem in ABW, the 20% dot gain space will also. Unfortunately many scanners use this space for grayscale images.

C4 Epson Pro38 v Pro48 -Here we look at the gray component of two Epson generated color profiles. The Pro38 series is intended for use with the 3800 printer and the Pro48 is intended for use with the 4800 printer. To the extent that the profiles give similar instructions to their respective printers it gives us information about the interchangeability of profiles. This is of interest because gray profiles are publicly available for many papers for the 3800 and not for the 4800.

First we look at screen saves of softproofs (for Enhanced Matte and Luster paper)

Comments: Visually the 38 and 48 profiles are very similar. This is true both for Enhanced Matte and for Lustre.

Below are RGB and Lightness information from an 11 step wedge converted from sRGB to Pro38 and Pro48. The RGB values constitute the instructions of profile to the printer.

Lightness (L) values that correspond to these sRGB values are: Pro38 Pro48
Comments: RGB patterns are very similar across profiles. The differences that exist are at low L where

1 1

17 17

46 46

66 65

80 80

93 93

they tend not to be visible. The differences tend to be related to color cast and less in terms of Lightness. We conclude that the gray component of these two Epsons profile, are very similar (at least for Enhanced Matte). Other work, not reported here, indicates that similar results exist for the other matte settings with Matte Black ink, but gloss type settings with Photo Black ink, show somewhat larger differences between the pro38 and pro48 series profiles

Note D Color Management Settings

D1

Rendering Intent -This should not matter in a grayscale image. Without color everything is in gamut and there be no rendering issue. To verify this we construct an 11 step wedge (in gamma 2.2) and print it under two rendering intent setting. The resulting prints are read by the spectrophotometer and the lightness values L(x) are plotted.

We see no important difference between these curves.

Printer Color Management v No Color Management -The relationship between printer color management and No color adjustment is studied by the procedure discussed above. Results are, in fact, plotted on the same chart above. Again we see no significant difference. It may be an overstatement to conclude that these settings always produce the same results in ABW but it at least it seems to be true for gamma 2.2.

Note E-Media Type settings

At first one might think media type settings are quite straight forward. Set this for the actual paper used - for Enhanced Matte use the Enhanced Matte setting, etc. This works for Epson papers but not for others. Even for Epson papers, sometime there are better settings than using this method.
Media type settings could in principle effect the following; Ink limits Lightness of the black point Linearity of the response curve Color cast at various points on the response curve

To study the effect of Media Type settings we proceed as follows: A 21 step wedge is printed on four different papers for each of four media type settings. Each of these patches is then read by the spectrophotometer and Lab scores are recorded. In total this involves 4 x 4 x 21 samples each resulting in a score for Lightness L(x) and color a(x) and b(x). E1 Black Point and density -Here we review the L(x) values for black points for the four settings on each for four papers.

Comments: Lightness values tend to run around 15 except for Entrada which is not quite as black. For all papers studied Lightness scores are highest for the Watercolor setting. Less ink must be used at this setting
Looking first at prints on Enhanced matte paper, we observe that the watercolor setting has a higher L than any other setting. That is, it is less dark. But note the paper is being held constant, all 4 readings are on enhanced matte. The only variable is the media type setting. The setting must be laying down less ink. This is equivalent to a lower color density setting. My interpretation is that Epson is using less ink when it thinks it is dealing with a porous paper like watercolor, in order to maintain detail. This pattern is repeated for each paper indicating further that the result is due to the setting, not the paper. The Velvet setting uses the most ink. The ranking from least to most ink is the same for the four papers. The patterns being very similar for Enhanced Matte, Ultra Smooth and Velvet and less pronounced for Entrada. This is in all likelihood due to a common coating on Epson papers and different coating used by Moab.

E2

Linearity -Lightness values L(x) are plotted for all 11 measured gray values for the four media type settings studied. The top chart is from printing on Enhanced Matte (gamma 2.2, tone dark and color toning neutral. In the other graphs this is repeated for other papers.

Comment:

The four curves are similar. Each starts at the Lightness value (L) for the papers black point and runs in a rather linear fashion to the L value for the papers white point. I do not believe the settings can be differentiated based on their linearity.

E3 Color Cast -Here we study curves with a similar structure to those discussed above but now we are
dealing with the color components of the Lab score ( namely a and b).

All reading on Enhanced Matte paper with the settings as above

Comments: The biggest difference we see here is in the b scores. The Velvet and Enhanced Matte settings produce similar results. The Watercolor and Ultra Smooth produce results similar to each other but different from the other two settings. The difference is that they are warmer (more yellow, less blue). In terms of the a scale, Enhanced Matte and Velvet are similar but Watercolor and Ultra Smooth are not as similar as they were in the case of b. Summary: Watercolor and Ultra Smooth settings produce warmer tones. I assume Epson set these tones this way to match the warmer white point of these papers. The Watercolor setting has a lower ink limit. This, I presume, was set this way to adjust for the tendency of ink to spread on their watercolor paper.

Color caste in photo paper settings

Scans of prints from three media type settings (all printed on Premium Gloss Photo Paper) are shown below:

Comment:

The three settings produce visually equivalent prints

Note F- Advanced Settings

F1 Tone - The default is Darker but I cannot say why as it is too dark for most combinations of ink, paper and images. Below is a 21 step wedge printed at each Tone setting (on Enhanced Matte paper) then scanned to obtain a digital image that could be displayed here.

The above wedges are next compared to prints made by varying gamma in Photoshop and then printing all at the Dark setting in ABW. Gamma was manipulated in Photoshop via a levels slider, namely the middle levels slider. This slider, which we label s, has a default value of 1.0 at which gamma is not changed. Other values of s are related to gamma by the formula g=2.2/s.

Comment: The Tone control in ABW is a gamma type adjustment. At the standard settings listed in Note A2, Dark is visually equivalent to gamma 2.2.

F2

Other color adjustments - A similar procedure was used to study the other ABW settings. First a print was made and scanned based on a high value setting for each of the controls-brightness, contrast, shadow tonality and highlight tonality. Next a wedge was printed and scanned for various Photoshop Image/Adjust settings.

Comments: Brightness increases in ABW can also be obtained via the Photoshop Brightness command Contrast increases in ABW can also be obtained via the Photoshop Contrast command Shadows Tonality is visually equivalent to a particular Photoshop curve. Highlight Tonality is visually equivalent to a particular Photoshop curve.

Note G- Toning
To verify the meaning of Tone settings we print patches from the Neutral, Cool and Warm settings and read resulting L(x), a(x) and b(x) with the spectrophotometer.

FIND Comments
As was show in the Neutral setting produces a slightly warm print on Enhanced Matte paper.

Note G- Gray Profiles

Below are screen save image for a gamma 2.2 images and that image converted to three different gray profiles. The conversion is done with and without the simulation of paper color. The images with the paper color setting are also presented in the main paper. They are shown here for completeness although the ones with the paper color simulator show bigger differences. Below the images is a graph which summarizes Lightness values for the 8 settings. We note a big

simulation affect occurs at the dark end of the scale. Hear the simulation is not primarily a color issue but one of the lightness values that result from a particular combination of paper and ink.

Need for Profiling with Gloss, Luster and Semi Gloss Papers

Here we consider if profiling makes an important difference when printing on gloss (Photo) papers. We print two 11 step wedges in ABW. One with the No Color Management setting and one with Color Management and the EEF (Epson Exhibition Fiber) profile. Resulting prints are scanned and displayed with their l (scan) values super imposed.

Settings

Comments:

Results are quite similar Using Gray profiles make less difference on gloss papers (with photo ink) than with Matte paper (with matte ink).