Calibrating the Epson 3800 printer for use in printing digital negatives for alternative photographic processes such as platinum/palladium and gum printing.
Calibrating the Epson 3800 for QTR negativesThis is simply a recap of the steps I took to calibrate my Epson 3800 to print negatives for palladium printing that would eventually find their way to gum prints. I need to thank Clay Harmon who got me started and then stayed with me in this process with coaching. Others also chimed in via the HybridPhoto forum (http://www.hybridphoto.com/) and were very helpful. And of course this could not be done without the wonderful QTR software developed by Roy Harrington worth every penny of $50 at http://www.quadtonerip.com/ So, with further adieu here is what I did to get to my first 3800 print:
Step 1: Set the Black Point _______________________________________________ 2 Step 2: Set the White Point _______________________________________________ 2RESULT OF FIRST CURVE _________________________________________________ 5
Step 3: Linearize the Curve _______________________________________________ 8Dialing in UV densities______________________________________________________ 12
Special Tips in BLUE
Step 1: Set the Black PointAs in any calibration technique you need to set the black point for your specific printer and paper combination. I am using Fabriano Artistica Hot Press paper. For chemistry I am using pure PD and Ferric Oxalate only no Na2. Using a 31 Step Wedge I came to a base printing time of 7 minutes 30 seconds.
Step 2: Set the White PointIn this step you need to make an assumption for the ink level as a starting point. There is where the advice of others (like this doc for example) or your own previous experience is your place to start. It is at this stage we begin to use QTR to print a negative. We did not need QTR to find the Black Point as that is totally set by exposure time. Within this first negative here are some key points (thanks to Clay for these): We need to set a default ink level that will lay down enough ink to block the UV light and establish as close to paper white as possible We will have to Linearize our curve in the next step. We need to enter the default end points of the curve in THIS step other wise we will be chasing our tails when we enter the curve later Start using QTR with default endpoints from the beginning
What do I mean by this? Lets look at the first curve I tried
# QuadToneRIP curve descriptor file # for ultrachrome k3 inks - these profiles will be using Matte as the black ink PRINTER=Quad3800 CALIBRATION=NO GRAPH_CURVE=YES (turn this on so you can see the curve as it is generated. It is a big help to learn how to look at this curve to anticipate response before you make a print) N_OF_INKS=8 DEFAULT_INK_LIMIT=35 (this and the ink levels below were the suggested starting points from Clay) Here is the dialog from the forum on how I got this started: LIMIT_K= Clay says - I'm hoping Michael chimes in here with some thoughts too. But my first thought BOOST_K=40 is that you probably want to add yellow only on the high end of the curve instead of the LIMIT_C=4 shadows where you have it now. Your copy curve now for yellow is LLK. I generally have it LIMIT_M=4 track the K curve, which gooses the density in the high values where you need it. I also LIMIT_Y=4 might add small amounts of the other colors (probably ink limits around 3-4 % just for the heck of it, like so: LIMIT_LC=4 COPY_CURVE_C=K LIMIT_LM=4 COPY_CURVE_M=K LIMIT_LK= COPY_CURVE_Y=K LIMIT_LLK= COPY_CURVE_LC=LK # # Describe Usage of each Ink: K,C,M,Y,LC,LM,LK # All Inks of Printer must be listed # # Gray Partitioning Information (this section is all default settings) # N_OF_GRAY_PARTS=3 GRAY_INK_1=K GRAY_VAL_1=100 GRAY_INK_2=LK GRAY_VAL_2=33 GRAY_INK_3=LLK GRAY_VAL_3=12 GRAY_INK_4= GRAY_VAL_4= GRAY_HIGHLIGHT=0 GRAY_SHADOW=1 GRAY_GAMMA=1 GRAY_CURVE="0;0 100;100" (This is the important point I learned from Clay. Set these endpoints in your first attempt so that all ink levels are being laid down are in the context of a curve with these end points. Later we will create the actual linear points)COPY_CURVE_LM=LK
(These were also at Clays suggestion to enable proper curving of the other inks. Note that all lite inks are curved the same and all non-lite inks curve the same) COPY_CURVE_C=K COPY_CURVE_M=K COPY_CURVE_Y=K COPY_CURVE_LC=LK COPY_CURVE_LM-LK
RESULT OF FIRST CURVE
Black Point - good (as expected since it is exposure dependent). This is a scan (obviously) and you will need to take my word for it that there was density difference in the shadows all the way through. White Point I thought that too much ink was being laid down no density difference past about 95.
But when you have an expert look the UV density readings for the NEGATIVE and you see the following Min Density 0.09 (this is good as it can get because that is simply substrate and no ink) Max Density 2.38
(These were read using and Xrite 810 densitometer. By the way I would recommend getting a new bulb and calibration card from Xrite if you get one of these machines off eBay)
So, what does this mean grasshopper? It means that while it looks like too much ink is being laid down actually NOT ENOUGH is being laid down. Your negative needs a density of around 2.7 to 2.9 here is Clays response to this print and UVs I finally had time to look at the numbers you posted on the spreadsheet. What I think may be some of the problem is that you are not getting enough maximum density on your negative. On all of my negatives, I am shooting for a max UV logD transmission density 2.8-2.9. Your maximum density is 2.38 So while I did not have separation in the highlights of the print it was NOT because of too much ink it was they were not separated by that endpoint that I printed to which was NOT white enough. So
Get at proper density negative first then begin to worry about separation So I did a few more tries at different ink levels until I landed on the following curve # QuadToneRIP curve descriptor file # for ultrachrome k3 inks - these profiles will be using Matte as the black ink PRINTER=Quad3800 CALIBRATION=NO GRAPH_CURVE=YES N_OF_INKS=8 DEFAULT_INK_LIMIT=40 (it took 4 tries to get the ink level needed and to get more density in the highlight end of the curve I goosed the yellow to 12) LIMIT_K= BOOST_K=43 LIMIT_C=4 LIMIT_M=4 LIMIT_Y=12 LIMIT_LC=4 LIMIT_LM=4 LIMIT_LK= LIMIT_LLK= # # Describe Usage of each Ink: K,C,M,Y,LC,LM,LK # All Inks of Printer must be listed # # # Gray Partitioning Information # N_OF_GRAY_PARTS=3 GRAY_INK_1=K GRAY_VAL_1=100 GRAY_INK_2=LK GRAY_VAL_2=33 GRAY_INK_3=LLK GRAY_VAL_3=12 GRAY_INK_4= GRAY_VAL_4= GRAY_HIGHLIGHT=0 GRAY_SHADOW=2 6
GRAY_GAMMA=1 GRAY_CURVE="0;0 100;100" COPY_CURVE_C=K COPY_CURVE_M=K COPY_CURVE_Y=K COPY_CURVE_LC=LK COPY_CURVE_LM-LK At this point my UV density was 2.89!! This was done by having the K Boost at 43 and Yellow at 12 while it tracked the K ink. These two things together boosted the ink in the highlight controlling part of the negative. This became my base curve to start to linearize.
Step 3: Linearize the CurveOk, at this point we have a good black Point and a good White Point so lets work on the middle. My issue here was how to start what do I move and how far do I move it? Well guess what you already have the answer in front of you if you have done UV Density measures! Clay provided me a table that contained Ideal UV Density for Palladium Printing. It looks like this:
So I took this and laid it on a spreadsheet next to my ACTUAL densities that I measured on my working negative and all of that looked like this: Explanation of what you are looking at Step Step Number from my palette PS Reading the K Value in Photoshop of that wedge of the step Ideal this is the same as % On Negative from Clays table UV of Base this is my actual UV density reading using Xrite 810
The table tells me that for a Negative Density of 30 my UV reading should be 0.99. But my negative has a density of 0.59 The 0.99 occurs at Step 48 or 49 This is how I generate Input:Output curve points. For this example my Input if 30 and my Output is 48 So in the curve this looks like 30:48
Here is what my first linearization curve attempt looks like: # QuadToneRIP curve descriptor file # for ultrachrome k3 inks - these profiles will be using Matte as the black ink PRINTER=Quad3800 CALIBRATION=NO GRAPH_CURVE=YES N_OF_INKS=8 DEFAULT_INK_LIMIT=40 LIMIT_K= BOOST_K=43 LIMIT_C=4 LIMIT_M=4 LIMIT_Y=12 LIMIT_LC=4 LIMIT_LM=4 LIMIT_LK= LIMIT_LLK= # # Describe Usage of each Ink: K,C,M,Y,LC,LM,LK # All Inks of Printer must be listed # # Gray Partitioning Information # N_OF_GRAY_PARTS=3 GRAY_INK_1=K GRAY_VAL_1=100 GRAY_INK_2=LK GRAY_VAL_2=33 GRAY_INK_3=LLK GRAY_VAL_3=12 GRAY_INK_4= GRAY_VAL_4= GRAY_HIGHLIGHT=0 GRAY_SHADOW=2 GRAY_GAMMA=1 # this is base curve 5 and then adding the curve values just from UV no print was made of 5
GRAY_CURVE="0;0 4;15 10;25 20;39 30;48 40;56 50;60 60;66 70;70 80;78 90;83 96;92 100;100" - this is from the spreadsheet above COPY_CURVE_C=K COPY_CURVE_M=K COPY_CURVE_Y=K COPY_CURVE_LC=LK COPY_CURVE_LM-LK Result of this was not good enough here is what a reflection measurement of the print looked like Note that it is TOO LIGHTReflection1.6
0 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101 Step
What did I learn here Ideals are not always Ideal as they do not take into account your paper, conditions, etc. But it is a GREAT place to start and gives you the map point to point for correction
Dialing in UV densities
Now that I have a reflection density measurement I was able to get an idea of where on the curve I was too light and then just tried a couple of more curves till I