2014-03-12 by <luciano@...>
After the discussion on UV density and the profile posted by Sandy King and the paper and files of Michael Rosenberg, I have been making some progress with my curve but now I am stuck as it seems that no matter what I enter in GRAY_CURVE has any effect. I tried both inserting an acv file (created in CS6) or values in the format "xx;yy xx;yy etc" but the step wedge printed does not show any difference from the one printed without a curve.
Also when I load the profile with the curve (created in CS6) I get this error (this is why I tried to enter the values manually):
As a further test I generated quad files from the same profile without curve, with the curve as acv, and as text directly in the profile. I then compared the resulting quad files (renamed) and they were all exactly the same as the quad with NO curve. I am not sure about the validity of this test since I don't know the inner workings of QTR, but I think I am doing something is wrong.
The profile I am using is:
PRINTER=QuadR2880
CURVE_NAME=r2880-gum-FaGF-WN-PM-v4-curve
GRAPH_CURVE=YES
N_OF_INKS=8
DEFAULT_INK_LIMIT=55
BOOST_K=0
LIMIT_K=5
LIMIT_C=5
LIMIT_M=1
LIMIT_Y=5
LIMIT_LC=10
LIMIT_LM=1
LIMIT_LK=15
LIMIT_LLK=1
N_OF_GRAY_PARTS=0
GRAY_HIGHLIGHT=0
GRAY_SHADOW=0
GRAY_OVERLAP=100
GRAY_GAMMA=1
GRAY_CURVE=test1.acv
N_OF_TONER_PARTS=0
TONER_HIGHLIGHT=10
TONER_SHADOW=10
TONER_GAMMA=1
TONER_CURVE=
N_OF_TONER_2_PARTS=0
TONER_2_HIGHLIGHT=10
TONER_2_SHADOW=10
TONER_2_GAMMA=1
TONER_2_CURVE=
N_OF_UNUSED=0
CURVE_K="0;0 100;100"""
COPY_CURVE_C=K
COPY_CURVE_M=K
COPY_CURVE_Y=K
COPY_CURVE_LC=K
COPY_CURVE_LM=K
COPY_CURVE_LK=K
COPY_CURVE_LLK=K
I have used several versions of this profile to print on regular photo paper instead of OHP to be able to scan the page and measure the changes. I have one step wedge printed with one of the older profiles so I can guesstimate what it's going to happen.
After scanning the step wedge, I open it in CS6, convert it to LAB and keep only the L channel, convert back to RGB, adjust Levels so the 0% and 100% step are indeed at that value, and read all the other steps.
I am basically following the procedure described by Ron Reeder.
My readings are:
IN OUT
0 0
5 4
10 14
15 25
20 35
30 51
40 65
50 75
60 81
70 88
80 92
85 94
90 96
95 98
100 100
The data to create the curve are then:
IN OUT
100 100
96 95
86 90
75 85
65 80
49 70
35 60
25 50
19 40
12 30
8 20
6 15
4 10
2 5
0 0
Any ideas? I am sure I am doing something wrong!!
Thanks,
Luciano
2014-03-13 by <mprosenberg@...>
Luciano,
When you enter the curve in Gray_Curve= the change in the curve is small. Try placing the curve in Curve_K=. This combines the Kadj and correction curves, resulting in a better fit. Then leave Gray_Curve= blank.
Mike Rosenberg
2014-03-13 by Roy Harrington
GRAY_CURVE makes no difference because there are no "GRAY" inks defined in the profile.
All the GRAY_'s go together, TONER_'s go together and, TONER_2_'s go together.
Only the LINEARIZE affects the whole set.
Roy
Show quoted textHide quoted text
On Wed, Mar 12, 2014 at 4:49 PM,
<luciano@...> wrote:
After the discussion on UV density and the profile posted by Sandy King and the paper and files of Michael Rosenberg, I have been making some progress with my curve but now I am stuck as it seems that no matter what I enter in GRAY_CURVE has any effect. I tried both inserting an acv file (created in CS6) or values in the format "xx;yy xx;yy etc" but the step wedge printed does not show any difference from the one printed without a curve.
Also when I load the profile with the curve (created in CS6) I get this error (this is why I tried to enter the values manually):
As a further test I generated quad files from the same profile without curve, with the curve as acv, and as text directly in the profile. I then compared the resulting quad files (renamed) and they were all exactly the same as the quad with NO curve. I am not sure about the validity of this test since I don't know the inner workings of QTR, but I think I am doing something is wrong.
The profile I am using is:
PRINTER=QuadR2880
CURVE_NAME=r2880-gum-FaGF-WN-PM-v4-curve
GRAPH_CURVE=YES
N_OF_INKS=8
DEFAULT_INK_LIMIT=55
BOOST_K=0
LIMIT_K=5
LIMIT_C=5
LIMIT_M=1
LIMIT_Y=5
LIMIT_LC=10
LIMIT_LM=1
LIMIT_LK=15
LIMIT_LLK=1
N_OF_GRAY_PARTS=0
GRAY_HIGHLIGHT=0
GRAY_SHADOW=0
GRAY_OVERLAP=100
GRAY_GAMMA=1
GRAY_CURVE=test1.acv
N_OF_TONER_PARTS=0
TONER_HIGHLIGHT=10
TONER_SHADOW=10
TONER_GAMMA=1
TONER_CURVE=
N_OF_TONER_2_PARTS=0
TONER_2_HIGHLIGHT=10
TONER_2_SHADOW=10
TONER_2_GAMMA=1
TONER_2_CURVE=
N_OF_UNUSED=0
CURVE_K=";0;0 100;100"""
COPY_CURVE_C=K
COPY_CURVE_M=K
COPY_CURVE_Y=K
COPY_CURVE_LC=K
COPY_CURVE_LM=K
COPY_CURVE_LK=K
COPY_CURVE_LLK=K
I have used several versions of this profile to print on regular photo paper instead of OHP to be able to scan the page and measure the changes. I have one step wedge printed with one of the older profiles so I can guesstimate what it's going to happen.
After scanning the step wedge, I open it in CS6, convert it to LAB and keep only the L channel, convert back to RGB, adjust Levels so the 0% and 100% step are indeed at that value, and read all the other steps.
I am basically following the procedure described by Ron Reeder.
My readings are:
IN
OUT
0 0
5 ; 4
10 14
15 25
20 35
30 51
40 65
50 75
60 81
70 ; 88
80 92
85 94
90 96
95 98
100 100
The data to create the curve are then:
IN OUT
100 100
96 95
86 90
75 85
65 80
49 70
35 60
25 50
19 40
12 30
8 20
6 15
4 10
2 5
0 0
Any ideas? I am sure I am doing something wrong!!
Thanks,
Luciano
--
Roy Harrington
roy@harrington.comwww.harrington.com
2014-03-13 by <sanking@...>
In the curve I copied there are no toner parts, just one system where all the inks follow the curve of the K ink. So instead of using LINEARIZE, as I did in the original profile, it is possible to equalize input/output values for digital negatives by putting a curve in the K. After reading Roy's previous comment I realize that since there are no gray parts in my profile the lines GRAY_HIGHLIGHT=4, GRAY_SHADOW=4, GRAY_GAMMA= 1 are not operative, so I took them out for the following example.
To calibrate, you would install the profile with CURVE_K= "0;0 100;100", then build a curve to equalize input/output, and insert the numbers in the CURVE_K = You should definitely see a difference between the corrected and uncorrected files
#PRINTER=QuadR3000
#CURVE_NAME= Carbon100 #Based on Carbon Tissue #222, 16 grams of Black Cat Ink with tissue coated to a wet height of 0.9mm.
#Profile when printing a step wedge gives a DR (density range) of approximately log 2.50.
#With Tissue #222 sensitize with a 8% potassium dichromate solution, diluted 1+1 with acetone, approximately 10 ml per 8X10 sheet, applied in two coatings with a roller. 350 ULF units.
GRAPH_CURVE=YES
N_OF_INKS=8
DEFAULT_INK_LIMIT=55
LIMIT_K=25
LIMIT_C=25
LIMIT_M=1
LIMIT_Y=25
LIMIT_LC=1
LIMIT_LM=1
LIMIT_LK=25
LIMIT_LLK=1
CURVE_K= "0;0 6;10 10;15 16;21 24;29 30;35 38;41 43;45 51;51 55;55 67;65 73;70 82;78 88;85 93;90 100;100 100;100"
COPY_CURVE_C=K
COPY_CURVE_M=K
COPY_CURVE_Y=K
COPY_CURVE_LC=K
COPY_CURVE_LM=K
COPY_CURVE_LK=K
COPY_CURVE_LLK=K
Sandy
2014-03-13 by <luciano@...>
Thanks for all the replies.
Roy, what can I say, last night my brain was out on vacation!!
Mike and Sandy, thanks for the suggestions, I will try them this afternoon and let you know.
I hope this method of printing the step wedge first on photo paper proves to be quicker and effective, provided that I do not make such blatant conceptual mistakes :-)
Luciano
2014-03-13 by <luciano@...>
OK, I think this is a technical problem and before I go spinning around, I'll check with you.
I use QTR on a Mac, and my CS6 is on the Mac. This is where I generate my curves (acv).
I am using my PC to check the curve with Curve Creator.
I created a curve on the Mac, copied it on the PC and modified the profile (here in my first post) to insert it in CURVE_K (the data are the same as those in the post here with the addition of the Kadj_1.acv from Mike).
When I try to display the curve I get the error: "C:\Program Files\QuadToneRIP\bin\test2.acv did not return valid coordinates at quadprofile.pl line 41"
From a command popmpt I executed the command "qacv test2.acv curvenum". The command displays: "0,0 100,100". Obviously there are a lot more points in the curve.
If I run the same command with one of Mike's curves, I have no problem.
So on CS6 on the Mac I made a small change in the Kadj_CC1a.acv curve from Mike (change one point from 30 to 31, so nothing drastic), saved it a test.acv, copied it to the PC, added to my K curve and got the same error.
So it looks like Curve Creator cannot read curves created on the Mac?
Any advice?
2014-03-13 by David Kachel
OK, I think this is a technical problem and before I go spinning around,
I'll check with you.
I use QTR on a Mac, and my CS6 is on the Mac. This is where I generate my
curves (acv).
I am using my PC to check the curve with Curve Creator.
I created a curve on the Mac, copied it on the PC and modified the profile
(here in my first post) to insert it in CURVE_K (the data are the same as
those in the post here with the addition of the Kadj_1.acv from Mike).
When I try to display the curve I get the error: "C:\Program
Files\QuadToneRIP\bin\test2.acv did not return valid coordinates at
quadprofile.pl line 41"
From a command popmpt I executed the command "qacv test2.acv curvenum". The
command displays: "0,0 100,100". Obviously there are a lot more points in
the curve.
If I run the same command with one of Mike's curves, I have no problem.
So on CS6 on the Mac I made a small change in the Kadj_CC1a.acv curve from
Mike (change one point from 30 to 31, so nothing drastic), saved it a
test.acv, copied it to the PC, added to my K curve and got the same error.
So it looks like Curve Creator cannot read curves created on the Mac?
Any advice?
The last few day I have had problems getting QTR on my Mac to accept
profiles and curves from Mike, made on his PC. I think there is a
translation problem between the two platforms. In my case, QTR on my Mac
refuses to see an .acv curve that is in the exact same location as two other
curves it finds without problem, every time. Photoshop (also CS6) on my Mac
has no trouble loading that same PC-made curve.
A couple of days ago, Photoshop stopped seeing the QTR 3880 printer I was
using, completely. I had to remove the profiles Mike sent me and completely
reinstall QTR from scratch before I could see that QTR printer again. BTW,
the other QTR printers I had installed remained visible. So, something about
moving QTR files from PC to Mac, or vice versa, is causing trouble.
David Kachel
___________________
Artist-Photographer
Fine B&W Photographs
www.davidkachel.com
david@...
Gallery:
www.reddoorfinephotographs.com
director@...
PO Box 1893
Alpine, TX 79831
(432) 386-5787
2014-03-13 by <luciano@...>
I searched the forum a bit more and found the answer! I was in RGB space so I needed to convert to Grayscale to generate the curve. Now I can load my curve without any problem.
2014-03-13 by <luciano@...>
David,
Fortunately mine was human error. So far no issues of translation between Mac and PC, and I hope it stays that way. If you have issues with your curves, check if you are in Grayscale mode, load your curve, make a small change and go back to the original value, save the curve, and it will work.
Luciano
2014-03-13 by Roy Harrington
Yes, the RGB .acv curves can be an issue. Only grayscale ones were accepted. Recently however
I allowed RGB .acv's and just pick the RGB curve (as opposed to the 3 R G B curves). This change
is in the Mac version but not yet in the PC version.
Roy
Show quoted textHide quoted text
On Thu, Mar 13, 2014 at 9:44 AM,
<luciano@luminari.com> wrote:
I searched the forum a bit more and found the answer! I was in RGB space so I needed to convert to Grayscale to generate the curve. Now I can load my curve without any problem.
--
Roy Harrington
roy@...www.harrington.com
2014-03-13 by <luciano@...>
Good to know. So if I switch between platform I better stick to grayscale curves until this change is also on the PC platform.
Thanks for the excellent program and support.
Luciano