What is Recipe Color?
What is Recipe to Recipe Conversion?
What is Recipe to CIE_Lab Calculation?
What is CIE_Lab to Recipe Conversion?
What does Gamut mean?
What does Within Page Gamut Mean?
Which recipe is the best one?
Which recipe should I use?
If all the solutions are out of Within Page Gamut then is it possible to obtain the color?
What is Accuracy Estimate?
How close will the new Recipe be?
What paper does the database use?
What if no answer is returned?
What if all the answers are out of gamut?
Q: What is Recipe Color?
A: Recipe Color on Kodak APPROVAL is the ability to specify a color as a combination of any of the Donors available for the machine.
Back to top of page
Q: How do you calculation CIE_Lab Values from a Recipe?
A: CIE_Lab Value of a Recipe is calculated by modelling CIE_Lab using Recipes near the Recipe requested. For each donor set a database is created consisting of pages of Recipe Colors. Each page is a combination of three donors. The set of pages used is {CMK, CYK, MYK, CMY, MOK, YOK, YGK, CGK, CBK, MBK, and CMB}. If a Recipe with a different combination of donors is input then we will return CIE_Lab = 0,0,0. This includes all combinations of four donors.
Back to top of page
Q: How do you calculate Recipes from a CIE_Lab Value?
A:Recipes are calculated by fitting a model to Recipes in the database that have CIE_Lab values near the CIE_Lab request. For each donor set a database is created consisting of pages of Recipe Colors. Each page is a combination of three donors. The set of pages used is {CMK, CYK, MYK, CMY, MOK, YOK, YGK, CGK, CBK, MBK, and CMB}. The results from each page are tabulated and the three best solutions are presented. The "Best" solutions are determined by whether the CIE_Lab Request is within the page and the Accuracy Estimate of the Model.
Back to top of page
Q: What is Within Page Gamut?
A:Within Page Gamut is the distance between the CIE_Lab Request and the three color page. If the Within Page Gamut is negative then the CIE_Lab Request is on the page, the Recipe is valid, and we indicate this with a Yes. If the Within Page Gamut is positive then the Recipe is off the page, and we indicate this with a No. Within Page Gamut is calculated as the difference between (the distance of the CIE_Lab Request to the Center of the Page), and (the distance of the Closest Surface Point on the Page to the Center of the Page). If this difference is positive then the CIE_Lab Request is further away from the Center than the Closest Surface Point, indicating that the CIE_Lab Request is out of Gamut on this page. If the difference is negative then the CIE_Lab Request is within Gamut on this page. Note that the Recipes on each page are sparsely sampled, so that a small Gamut number (<= 1.5), may still be within Gamut on the page. If the Within Page Gamut is positive then the Recipe Returned is out of gamut and the accuracy will be worse than the accuracy estimate by the amount we are off the page. This is indicated by showing the Accuracy Estimate as {Within Page Gamut ± Model Accuracy} when the Within Page Gamut is Positive, and {0 ± Model Accuracy} when the Within Page Gamut is Negative.
Back to top of page
Q: What does Gamut mean?
A: Gamut is the color space that is achievable on the machine. The Gamut will change slightly for each Donor Set that is available. Special colors such as Orange (DO01), Green (DG01), Blue (DB01), White (DW01), and Metallic (DS01), will extend the Gamut. Each three color page looks like an orange slice in terms of machine Gamut. Each of the donors forms a vertices along the edge of the orange slice. The center and inside edges of the pages, touch adjacent pages. Therefore it is possible for each Within Page Gamut to be positive, indicating that the CIE_Lab Request is off each page, and the CIE_Lab Request is within the Machine Gamut.
Back to top of page
Q: What is Recipe to Recipe Conversion?
A: Recipe to Recipe Conversion is used to compute a Recipe in a new Donor Set that will match a Recipe in an existing Donor Set. The Recipe returned will be a good starting point for running a Recipe Grid and obtaining a visual match. Note that Recipe to Recipe Conversion first predicts the CIE_Lab value for the input Recipe. Then the CIE_Lab Value is used with the new Donor Set to calculate the new Recipe. The accuracy estimate of the new Recipe does not include the error in the CIE_Lab Estimate.
Back to top of page
Q: What is the best Recipe?
A: In general the best Recipe is the first one listed. However there are conditions when you might want to choose a different Recipe. For instance the best Recipe may include a Special Donor that you do not have. Or the best Recipe might include a Special Donor while the second best Recipe is composed of CMYK. The Accuracy Estimate and the Within Page Gamut are used to automatically order the three best solutions. The order is equal to the minimum of Accuracy Estimate if Within Page Gamut is less than zero, or Accuracy Estimate plus Within Page Gamut if Within Page Gamut is greater than zero. Accuracy estimates within 1 are probably not statistically different from each other.
Back to top of page
Q: What is Accuracy Estimate?
A: Accuracy Estimate is an estimate of the error from modeling the data, plus an error due to rounding the Recipe to the nearest click, plus the error due to the System Accuracy.
Back to top of page
Q: How good will the Recipe be?
A: If the CIE_Lab Request is Within the Page, (Within Page Gamut < 0), then the Accuracy Estimate will indicate one standard deviation of error for the Recipe. This means that there is a 66% chance of obtaining the CIE_Lab Request to within ±Accuracy Estimate. There is a 90% chance of being within ±2 x Accuracy Estimate. The Accuracy Estimate is similar to a Delta E metric. If the Within Page Gamut is > 0 then the CIE_Request is outside of the page by the Within Page Gamut. Therefore the Recipe Out will be off by the Within Page Gamut ± Accuracy Estimate. Recipe to Recipe Conversion will be slightly worse as the Accuracy Estimate does not include the error in Estimating the CIE_Lab Request. This additional error is shown as CIE_Lab Error. In general a Recipe Grid will need to be run with the Recipe returned in order to make a better visual match. The Recipe returned should be a good starting point. The Within Page Gamut number will tell you how close we are to being able to obtain the CIE_Lab Request.
Back to top of page
Q: What paper does the database use?
A: The database is built using Centura Gloss 100 lb Cover paper. The measured CIE_Lab value of the Paper with prelaminate and intermediate is L* = 95.2, a* = 1.0, b* = -3.9.
Back to top of page
Q: What if no answer is returned?
A: If no answer is returned then a database page for the requested recipe does not exist. At the present time only three color combinations are modeled. For instance a CMYK recipe will not return a solution. Try eliminating the lightest color to obtain an estimate. Valid three color combinations include, {CMK, MYK, YCK, CMY, MOK, YOK, YGK, CGK, CBK, MBK, and CMB}.
Back to top of page
Q: What if all the answers are out of gamut?
A: If the accuracy estimate is a positive number ± a tolerance, then that answer is out of gamut. If all of the answers are out of gamut then the solution is not within the gamut of the donor set chosen. We have seen poor results the further out of gamut the answer is as we sort by gamut and accuracy and do not distinguish between an error in L* verses a* and b*. For instance an error on the blue page, CMK, may look as good as an error on the Green page, CYK, if the L* error is equivalent to the error in a* and b* for an out of gamut color. When all of the answers are out of gamut, use common sense in selecting the best three color combination. For instance use the CMK solution for blue, and the CYK solution for a green recipe request.
Back to top of page
Last Revised:
February 25, 2005
