Ink Optimization RoundUP-Procedure

UPDATED: March 5, 2010 - Frozen - No further changes will be made

The IPA Ink Optimization RoundUP tests are described below.

• The project seeks to produce a final White Paper printed book that will include supplier press sheet samples. This paper will not just contain the results of the testing but it will also seek to answer some common questions about ink optimization such as; If I were a printer or print buyer, why would I be interested in Ink Optimization? Why should I care about this? Is it real or a gimmick? Does it have a future if I invest what little money I have? Would my customers care? Will this technology enable my print operation to have a competitive advantage? Are there any savings I can gain in my print procurement contract? In using this technology, will my in-line color conflicts be reduced? Are there really ink savings and how can I capture this data? How do I measure the value this technology brings to my operation? Will I have to invest more in my prepress to make this technology work? Is it “user friendly”? Plug & Play? Are there any common benefits all of these solutions offer? Are their unique aspects to the solutions that I should know about?
• All proofing and offset printing will be in accordance with GRACoL Coated 1.
• The profile and data set used will be the official GRACoL profile and data set from IDEAlliance's website
• Form A will be proofed on an Epson 7900 inkjet device.
• Form B and Form C will be printed offset at manroland as two separate 28x40" 4/0 press sheets.
  • The press used will be 8-color ROLAND 700 DirectDrive 4/4 perfector
  • The paper used is Sappi, McCoy, Grade#1, gloss coated, text, 100lb
  • The prepress system used for the generation of plates will be Kodak Prinergy
• IPA will conduct a parallel test using Adobe Photoshop to provide a baseline measure, where appropriate.
• The tests will be conducted remotely on supplier systems using "web-facing" hot folders or other remote desktop sharing.
• The final image set will not be seen by suppliers before the event, however there will be demo example images available for pre-event testing.
• We will hope to get non flattened PDFs back from suppliers. There are instances, for example Agfa Prepress workflow, where it is necessary to create an 8-bit tiff. If such a situation applies to you, please create a non-screened, 8-bit flattened, composite TIFF targeted for 2400 dpi/175 line screen CTP.

On this page we describe the procedure in terms of Supplier 1, Supplier 2, etc. but in the final document we will use actual supplier names, e.g. Agfa :Apogee InkSave, Alwan CMYK Optimizer, etc.

FORMS FOR TESTS
Form A Download Trial Form Printed Inkjet - 22" x 34"	Form B Download Trial Form Printed Offset - 8" x 10"
	Form C Download Trial Form Printed Offset - 3.25" x 15"

These forms may change slightly due to the number of registered participants in this RoundUP due to size limitations on press. Once registration has closed the forms will be updated to show the final sizes.

TEST 1: ANALYSIS OF CMYK SEPARATION

Description: This test will measure the reduction in CMY ink amount and the corresponding increase of K ink, using pdf file, Form A. This test will be performed by analysis of the processed image, based on the pixel values in each color separation. This test will be conducted by measuring the CMYK values of the reference test form and comparing those values to the form after it has been processed by the different ink optimization solutions. To see if your mathematical calculation of ink coverage is the same as ours - download a test image - see calibrate your system.

Procedure: Suppliers are to process Form A aiming for the GRACoL color space and ink saving. The page is to be processed as a single form, to mimic a real-life workflow situation. IPA will measure the “before” test form and the "after" or processed form, from each supplier. The percent coverage for each channel and total percentages for all channels will be measured and compared to create a graph as shown below. Data will be produced using software program, PressPerCent.

Results: Each supplier’s data will be plotted as shown in the graph. The reader will be guided in their interpretation of the graphs, such as “…in this graph a lower number is better as it represents less use of expensive CMY inks which have been replaced by black ink. Black ink is cheaper in cost and easier to control on press…”. Data will be presented in the following format.

 

TEST 2: ANALYSIS OF DIGITAL DEFECTS

Description: Whenever a file is optimized there is the chance for undesired changes to occur to the page. We will evaluate the forms, as optimized by each supplier, and look for changes that could affect the final printed piece.

Procedure: This test will utilize Forms A and B. The optimized Forms A and B will be opened in Adobe Acrobat and evaluated using the “Output Preview” for some common defects. The defects that will be evaluated are vector and raster objects optimized to same CMYK values, visible artifacts on bitmap type, any objects that are missing from the forms, it text still text, have primary / secondary colors been contaminated, and simple overprints.

Results: The forms will be analyzed in Adobe Acrobat and the table below will be constructed. The table will be filled out to indicate if a software has defects or not.

Digital Defects in Ink Optimized Forms
	Raster/Vector CMYK Values	Bitmap Type	Omitted Objects	Text still Text	Color Contamination	Simple Overprints
Adobe Photoshop				X	X	X
Supplier 1		X
Supplier 2					X
Supplier 3	X	X
Supplier 4				X
Supplier 5		X	X			X
Supplier 6			X
Supplier 7
Supplier 8	X
X indicates that there were defects found

TEST 3: COLOR CHANGE IN DIGITAL FILE

Description: A basic assumption in ink saving is that the colors in the image are unchanged from the original file. In this test we estimate the colorimetric change that occurs from the application of ink optimization - using only the digital file.

Procedure: In this test we use the IT8.7/4 target on Form A. We will process the CMYK values in the original target using absolute colorimetry and the GRACoL ICC profile, to obtain L*a*b* values.  Next we will repeat the process with a target that has been optimized by each supplier. We will compute Delta E between the L*a*b* values predicted by the original target and compare that to the supplier optimized IT8.7/4 values. Note that we will not use the official reference file, but we will show a more realistic result by using "profile calculated" values.

Results: For each supplier we will measure the IT8.7/4 from Form A using a software capable of calculating Lab from CMYK using the GRACoL ICC profile. The primary plot will be in terms of average and maximum Delta E (2000) based on 1617 patches.

 

TEST 4: COLOR CHANGE ON PROOF

Description: A basic assumption in ink saving is that, on proof, the colors in the image appear unchanged from the original file. We will use this test to determine the colorimetric change that occurs from the application of ink optimization - at proofing.

Procedure: This test will use Form A. Each supplier will download and optimize Form A. This file will be output on an Epson 7900 located at Fujifilm Canada that has been color matched to GRACoL within an average DE76 of 0.5 with a maximum deviation of less that 5 DE76. Each proof will have an ISO12647-7 control strip added in the proofing RIP and the strip will be measured to ensure that it falls within that tolerance.

Results: For each supplier we will measure the IT8.7/4 from Form A using an i1iO. We will use UV included measurement, white backing, D50/2⁰. A single proofer sheet will be measured, but 2 readings will be taken from the sheet and compared to ensure a "correct" reading. The data will be used to compute Delta E between the target on the supplier image and the reference data set. The primary plot will be in terms of average and maximum Delta E (2000) based on 1617 patches. In addition an "error bar" will be published based on the "non-ripped" control strip.

Upon request a duplicate proof of the form will be made available for each supplier.

 

TEST 5: COLOR CHANGE ON PRESS

Description: Just as we assume the color should be unchanged on proofs after ink optimization, we expect the same on press. To verify this statement we print on an offset press, a matrix of “before” and “after” images. In the central row of the press sheet we repeat multiple copies of the "before" file. Above and below each reference image is a file that has been separated by each supplier. We seek to measure colorimetry from the IDEAlliance ISO 12647-7_Control Strip 2009 embedded in each image. Delta E, relative to each reference image will be computed. Note that each set of images is aligned below the ink keys, and that each image will be compared to its “reference” image and also separately to GRACoL. Control bars at the top, middle and bottom of the sheet will be used to verify uniform printing. 

Procedure: Suppliers will be required to download and process Form B. Suppliers are required to separate this form and return it to us. The images will be assembled onto an offset press form and put on press. The press will be run to GRACoL reference printing conditions. The press run will be done at manroland in Chicago. The press sheet size is 28 x 40". The paper to be used is Sappi, McCoy, Grade#1, gloss coated, text, 100lb that has published L*a*b* values of (95.8, 2.1, -6.5). Contact David Niles for questions on the paper. Contact Hal Stratton for questions on the press run.

Results: We will measure and compute Delta E between the control bars in the reference image and compare that with the suppliers control bar in their processed image. We will use i1iO, UV included measurement, white backing, D50/2⁰. 5 press sheets will be measured and averaged. The primary plot will be in terms of average and maximum Delta E (2000) based on 54 control patches. The primary plot clearly shows whether a supplier can create ink savings without changing the printed color. This information will be collected and analyzed, however due to concerns raised by suppliers, it may not be shown in the final report if the results appear to be misleading. 

In addition, all of the reference forms and supplier optimized forms will also be compared to the published values for GRACoL. The secondary graph below then allows us to evalute supplier claims that they enable printing that is closer to industry specifications.

TEST 6: METAMERIC SHIFT IN INK OPTIMIZED PRINTS

Description: Metamerism is caused by interaction between the pigments of printing inks and the viewing light source. If we change the amount of colored (CMY) inks in the make up of a color, does this also affect the metamerism of the print? Will the image change color under different illuminants, as shown in the above i1Share graphic? Is the optimized image, more or less metameric compared to the un-optimized print? In creating ink optimization algorithms, do suppliers consider the spectral properties of the inks or simply colorimetric accuracy? We thank ppi Media for suggesting this interesting experiment. 

Procedure: This test will use spectral measurements from the control strip printed on the offset press sheet. Control strip patches will be measured. The spectral data will be converted to L*a*b* (D50) and L*a*b* (F2) and the Delta E between these will be calculated. The Delta E between L*a*b*(D50) and L*a*b*(F2) for the unoptimized image will be computed, and then the Delta E for the optimized printed images will be computed.

Results: For each supplier we will report the average and maximum Delta E calculated between the D50 measurements and F2 illuminant for the 54 patches of the control strip. This data will be plotted along with the Delta E information for the unoptimized form - this acts as a baseline.

TEST 7: VISUAL ANALYSIS ON PRESS

Description: In order to be successful, ink optimized images need to preserve the look of a unoptimized image. If the quality of the optimized images is reduced the results will be unacceptable to many print buyers.

Procedure: A group of experts will visually evaluate Form B on the offset press sheet. The optimized images will be compared to unoptimized images in terms of highlight detail, shadow detail, flesh tones, color change, sharpness, artifacts/abnormalities, and overall similarity to original images. Measured values are great, Delta E is very useful, but at the end of the day printers want to see real images on real press sheets.

Results: Each expert will score each form individually on the criteria on a 1-10 scale and the results for all of the experts will be averaged and compiled into a table, shown below.

Visual Analysis on Press
	Highlight Detail	Shadow Detail	Flesh Tones	Color Shift	Sharpness	Artifacts	Overall
AdobePhotoshop	9	7	8	5	4	5	7
Supplier 1	8	8	10	7	5	10	8
Supplier 2	9	8	9	9	7	7	7
Supplier 3	8	10	7	8	9	9	9
Supplier 4	7	9	10	5	9	6	6
Supplier 5	8	8	10	7	5	10	8
Supplier 6	9	8	9	9	7	7	7
Supplier 7	8	10	7	8	9	9	9
Supplier 8	7	9	10	5	9	6	6

TEST 8: INK KEY SETTINGS FROM PressPerCent

Description: Ink optimization solutions may predict good savings in terms of pixel values, but does that translate to savings on press as shown by predictive ink key settings? In this test we use an independent software program to predict the ink key settings for each optimized form, and use that information to verify that the press would indeed need less ink to run the form.

Procedure: Form A and Form B will be opened in PressPerCent v2.8. The predicted ink key settings will be exported as a text file and plotted.

Results:  The original reference file will be plotted, followed by data from the file processed in Adobe Photoshop. Each supplier file will also be analyzed and plotted as shown below. This test on its own does not indicate a good optimization, but it does contribute to our understanding of the relative performance of each system, especially when considered in conjunction with other tests in this series.

 

TEST 9: STABILITY ON PRESS

Description: A claim made about ink optimization is that the lower amount of CMY inks will enable better control on press. To verify this statement we change the target density on the offset press sheet from the optimum setting. The press will initially be run at the normal GRACoL target density and L*a*b* values. After the press has stabilized we will pull sheets for the above tests, then the density will be changed in separate channels so that we have the following press runs – normal density, (cyan density +/- 0.1), (magenta density +/- 0.1), etc. Actual changes will depend on what we see when running on press. All settings and results will be documented. Suppliers are invited to attend the press run.

Procedure: Suppliers will be required to download and process Form C, which has the following explanation.

• C-1 - Photoshop default separation
• C-2 - Supplier separation with typical ink optimization setting
• C-2 - Supplier separation with aggressive ink optimization setting

A Photoshop default separation is called Form C-1. Suppliers do nothing with this form. The suppliers will be asked to optimize form C-2 with a typical ink optimization setting. In Form C-3, suppliers will be asked to separate the form using a more aggressive ink reduction setting. Forms C-1, C-2 and C-3 will be combined and assembled onto an offset press form and put on press. The press will be run to the GRACoL reference printing condition, then density changes will be made as above. The details for the press run are provided within this document.

Results: We will visually analyze the press sheets and complete the following table. We will also measure the control bar using an i1iO for the normal density from each area. We will use UV included measurement, white backing, D50/20. 5 press sheets will be measured and averaged. The data will be used to compute Delta E between the non standard density forms and the normal density. The resulting numbers will be in terms of average Delta E (2000) based on 54 control patches.

Stability on Press
	Standard – 0.1 Cyan			Standard		Standard + 0.1 Cyan
	Color Change Mean (Max)	Rosettes, moire?	Trapping issues?	Rosettes, moire?	Trapping issues?	Color Change Mean (Max)	Rosettes, moire?	Trapping issues?
Supplier 1	1.5 ∆E (6.8 ∆E)	No	Yes	No	No	1.5 ∆E (6.8 ∆E)	Yes	No
Supplier 2	2.5 ∆E (7.7 ∆E)	No	No	No	No	1.9 ∆E (5.1 ∆E)	Yes	No
Supplier 3	1.9 ∆E (5.1 ∆E)	No	No	No	No	1.5 ∆E (7.7 ∆E)	Yes	No
Supplier 4	1.5 ∆E (6.8 ∆E)	Yes	No	No	No	1.5 ∆E (6.8 ∆E)	No	No
Supplier 5	1.9 ∆E (5.8 ∆E)	No	No	No	No	2.5 ∆E (7.7 ∆E)	No	No
Supplier 6	1.9 ∆E (5.1 ∆E)	Yes	No	No	No	1.5 ∆E (6.8 ∆E)	No	Yes
Supplier 7	1.5 ∆E (6.8 ∆E)	No	No	No	No	1.9 ∆E (5.1 ∆E)	No	Yes
Supplier 8	2.5 ∆E (7.7 ∆E)	No	Yes	No	Yes	2.5 ∆E (7.7 ∆E)	No	Yes

 

TEST 10: FEATURE COMPARISON

Description: To assist the reader, the report will include a section that describes various feature comparisons. This grid will include questions such as “Can the user make their own profiles”, or “What options do they have when making profiles”. It will also cover features that are essential to integration into the workflow; such as does the system use hot folders or can it create device link profiles.

What is the MSRP of the ink optimization feature over and above a typical installation of the product. What do you have to pay to licence this feature? Or if standalone product – the price of the product. The ROI on these products can pay for the system in months. We seek to document the cost of the software used to produce the images used in these tests. The price does not need to include the cost of web-based features or a web portal version of the product.

Procedure: Each supplier will fill out a form and answer the questions about the feature set of their software. There will be space for the supplier to add their own comments.

Results: The results will be compiled into a table so that feature sets of various software packages can be compared.

Feature Comparison
	Profiling			Workflow Integration			Other Features
	Industry Standard Profiles	User Creatable Profiles	User Controllable GCR Level	Hot Folders	Create Device Link Profiles	Standalone Application
Supplier 1	X	X		X	X	X	Sharpening Filter
Supplier 2	X	X	X	X	X
Supplier 3		X	X	X	X
Supplier 4	X			X		X	Selective Color Correction
Supplier 5	X			X	X
Supplier 6	X	X	X	X
Supplier 7	X
Supplier 8		X			X	X

 

TEST 11: SUPPLIER TIME FOR PROCESSING

Description: To help ensure that all of the software being used is standard off the shelf software and the images are not modified or tweaked specifically for this test, we seek to monitor the process. Suppliers are required to create a "web-facing" hot folder where we can drop and retrieve files at our convenience, this would allow vendors to set up the process, but not individually manipulate files. In a prepress workflow, the ink optimization stage must not introduce a lengthy delay in the production process. So we seek to record the time for processing of images. If a supplier does not have a viewable hot folder system, they can grant us access via Adobe Connect, WebEx or GoToMeeting or other protocol. The speed of processing will depend on whether the system is being run on a laptop or server farm, but we hope this is not really a critical parameter in this instance.

Procedure: We will identify a given date, and then schedule the test for Agfa - 9.00 am, Alwan - 10.00 am,  GMG - 11.00 am, Kodak - 1.00 pm, etc. During that time we will work with the suppliers to process the test images. We will note the procedure and any issues. We will note from the date stamp on the files the time it takes to process the images.

Results: The results will be noted in a table showing how long it took each supplier to process the image files.

 

 

TEST 12 : TEST PROCESS & DATA ANALYSIS

A trial version of each form is made available as a download. These forms will be used to conduct a "dry run" of the whole process. At a pre-arranged time, IPA will process these forms at the supplier's site, using remote methods. The forms will be analyzed and the data will be used to check and fine-tune the process. The trial run is seen as a rehearsal, the results will not be published. In the final form there will be no new types of elements that were not in the trial form. Both the trial and final forms will be PDF/X1-a. The missing elements are visible on the preview of the final forms, we have simply removed some CMYK images, these will be replaced in the final forms. There will be no change in type of image, there will be no introduction of spot colors, overprints, transparency, or other technical trickery. Simple 4-channel CMYK images are omitted to create the trial forms, these will be re-introduced in the final forms. The scenario is – the agency showed you the form, we got sign off, but the printer did not get all the artwork, the final artwork with all the images are now presented to you for automatic processing and optimization.

At a later date the final forms will be processed at the supplier's site, using remote means. A time slot will be booked with each registered supplier and during this final session the final forms will be provided and processed.

We will collect the forms, analyze the data and provide feedback to suppliers for confirmation. The process for checking of data with suppliers is only to ensure accuracy and technical correctness of the results. If a supplier says - "..there is no way my form could show 200% coverage..." then we know to track and identify any errors in processing and recording.

The final report will be in the form of a White Paper, with full results, graphs and analysis. Print samples from the press run will also be included in the final report. Suppliers will see a copy of the draft White Paper and are invited to critique and comment, however the IPA reserves the right to publish results that we believe are technically accurate.

Suppliers have experience with their products and with user requirements, and can provide helpful and constructive input, with this in mind, suppliers are invited to provide a general comment that will be included in the publication.

A condition of the Ink Optimization RoundUP is that each system entered must be shipping and commercially available — R&D, beta systems or development software and products are not permitted. In the spirit of the RoundUP it is required that you submit entries for all tests. It is a further intention of the RoundUP, that all the submission components will be done on a single system. Suppliers should review all the tests to ensure that the chosen configuration can process all the required components.