Datacolor User to User Support Group.

Epson 3880 (photo black) and 4880 (matte black)
Mac OS X 10.6
Photoshop CS5
Datacolor Spectrocolorimeter #1005
Spyder3Print v. 4.1

Kirkland glossy paper for the 3880
Moab Lasal (matte) for the 4880

I created color profiles the the papers and printers using the classic 225 patch target plus the grey target. Both profiles compensate for the optical brighteners in the papers, and both have the "shadow detail" option set all the way to 20.

Printing a full-scale color image to both of the printers and comparing the results:  The Kirkland print shows noticeably more detail in the darkest shadows.  The Lasal print could benefit from "opening" the shadows somewhat more. 

Since the "shadow detail" adjustment has be maxed out, the only thing I can think of to open the shadows more is to work out some kind of curve to do this in Photoshop, and then add the curve to profile-making process.

How should I go about constructing the curve?  I think there *should* be a way of using the colorimeter to measure the output I am getting for some kind of standard image, and then at least estimating the necessary curve corrections.   Are there instructions somewhere online that would help?

Yes, I know that the range of the two papers are different, so the output curves cannot be identical. But I should be able to redistribute brightness values within the existing range so that the darkest values have a bit more separation.

Thanks!

>>Since the "shadow detail" adjustment has be maxed out, the only thing I can think of to open the shadows more is to work out some kind of curve to do this in Photoshop, and then add the curve to profile-making process.

Yes, that will certainly work, though it begs the question: why do you need so much shadow adjustment in your prints? Is your display too bright? Do you just prefer detail to punch in the shadows? Personally, about five points of shadow opening is all I need on a matte Moab paper.

>>How should I go about constructing the curve?  I think there *should* be a way of using the colorimeter to measure the output I am getting for some kind of standard image, and then at least estimating the necessary curve corrections.   Are there instructions somewhere online that would help?

The colorimeter has already produced it's best guess at a smooth density curve from paper white to ink black. What you want seems to be significantly different. So you will need to create that response curve yourself. Open Photoshop's curve creation tool, and in the global curve add a few points to keep the highlights and mid tones nailed down. Then use a point to create an upside down dipper (long straight handle, curve of the dipper going up, not down, to open the shadows). You may need to add one or two more points to adjust the shape of the dipper. Save this curve, then apply it to your image, and print. If it's not what you wanted, open the saved curve, adjust, and try again. When you get just what you want, import the curve into Print SR, and incorporate it into a copy of your profile.    

C. David Tobie
Global Product Technology Manager
Digital Imaging and Home Theater
Datacolor inc. 
cdtobie@datacolor.com
www.datacolor.com

On Nov 7, 2010, at 7:59 AM, "gochatunbdotca" <goch@unb.ca> wrote:

> Since the "shadow detail" adjustment has be maxed out, the only thing I can think of to open the shadows more is to work out some kind of curve to do this in Photoshop, and then add the curve to profile-making process.
> 
> How should I go about constructing the curve?  I think there *should* be a way of using the colorimeter to measure the output I am getting for some kind of standard image, and then at least estimating the necessary curve corrections.   Are there instructions somewhere online that would help?

>> Since the "shadow detail" adjustment has be maxed out (at 20), the only thing I can think of to open the shadows more is to work out some kind of curve to do this in Photoshop, and then add the curve to profile-making process.
> 
> Yes, that will certainly work, though it begs the question: why do you need so much shadow adjustment in your prints? Is your display too bright? Do you just prefer detail to punch in the shadows? Personally, about five points of shadow opening is all I need on a matte Moab paper.

This has puzzled me as well.  Over a wide variety of ordinary lighting conditions the +20 shadow versions of most of my images look more "relaxed" or "natural" (i.e. not over-inked, burned-in, or "heavy").  This is much more obvious with matte paper on the Epson 4880 (using MIS inks).    

Even with my new 3880, which I am using exclusively for photo black glossy prints, +20 shadow prints of natural scenes with wide brightness ranges appear rich and detailed throughout the range.  So far I haven't had any images that looked weak because of the shadows.  The shadows don't *look* dodged or elevated.  Most of my recent work involves more detail than graphic impact...

I just now printed out and measured the standard 21-step Adobe RGB 8-bit greyscale.  I used the +20 shadow profiles I made for the Epson 3880 (Kirkland paper) and 4880 (Lasal matte paper).

The high and low values look pretty normal for these papers and printers (at least for me). 

I'm guessing that with the 3880 the jump from 4.37 (95% black) to 11.47 and then 16.69 are so large because of the +20 shadow settings.

The 4880, with less range in the dark tones, has much less stretch in this region, although it is obvious that something has happened.  It may well be that I cannot expect the visual separation that I get with glossy papers.

Do these values seem appropriate, given my settings?  Would you *expect* real-world images to look "off" if they followed these L-curves?

Here are the L-values:

3880			4880
95.25		96.14
90.89		91.92
85.85		87.27
80.95		82.37
76.34		78.12
72.92		74.77
68.03		68.83
63.63		63.91
58.89		57.66
54.35		53.88
49.33		49.26
44.59		43.24
39.29		38.60
34.09		35.48
29.41		32.66
24.88		29.86
20.63		27.67
16.69		25.34
11.47		21.87
4.37			18.37
3.71			17.32

Myron

On Nov 8, 2010, at 11:34 AM, Myron Gochnauer wrote:

> Do these values seem appropriate, given my settings?

Appropriate is whatever gives you the effect you are looking for...

>  Would you *expect* real-world images to look "off" if they followed these L-curves

Well, the steps at the bottom end are a bit larger than further up the scale, where they would typically be a bit smaller. But thats fine, if thats what works for your images. 

C. David Tobie
Global Product Technology Manager
Digital Imaging & Home Theater
CDTobie@...



Datacolor
www.datacolor.com/Spyder3