View Full Version : Patch count upper limit

12-03-2010, 01:42 PM
What is the upper limit for patch count in a given chart? Suppose I use 2000 mm x 500 mm size. Is it 6000? There must be a point where any large chart size only translates into more of the same patch set to be added to the chart?

Also, there is this thing called, the law of diminishing returns. The regular IT8.7 test chart is made up of 1617 patches. The ECI2002 target has 1485 distinct patches. Given that all raw measurements are internally smoothed out in ColorFlow, at what point is more patches not yielding additional benefits, in terms of characterization? Does the internal profile allow for more than the regular 17x17x17 grid points in the CMYK to PCS direction?

There must be a benefit of using ColorFlow's extra device patches. But I'm not sure.

Best / Roger

12-04-2010, 07:57 PM
Hi Roger,
ColorFlow creates color characterization chart with "randomized" patch placement of the IT8/7.4 data set. There are also additional CMYK solid patches that are scattered across the chart to provide density across the sheet reporting (comparison and characterization reports). As such, anything above the 1617 patch count will likely be repeats.

The benefit to using extra patches is to get a better sampling size for color response generation. Granted, we are still improving the QA of measurement data coming into the software, the software does its best to determine that the patch that you measured is truly the expected color that it should be. This is a tricky thing to do, but greatly enhances the performance, because, as you know the better measurement you can do, the better color model we can create.

So, it is better to have more patches so long as the patches are representative of the print condition that you are modelling. If there are print quality issues between sheets measured, then you are introducing anomalies into the model; this is highly undesireable.


12-06-2010, 02:05 PM
Here are a few related questions.

1. Is there a benefit to using the ECI target which I believe has more unique patches than an IT8? If I import an ECI data file will Colorflow use all its patches or just throw away the extra data?

2. Running targets all over the sheet makes it very important to get the density even across the whole sheet. Would there be any advantages to running only one target on the sheet to make the critical area smaller and easier to keep at the correct density?

3. I believe the P2P target has many more near-neutral patches than the IT8 or ECI. Would it make sense to allow the software to accept both an IT8 and a P2P on the same sheet to make use of the extra data.

4. Are the extra solid patches you mentioned used to adjust the values of other nearby patches or just to determine the quality of the sheet?

And a feature request.
Allow colorflow to average multiple imported data files instead of making me average them in ProfileMaker.

12-08-2010, 07:15 PM
Hey Jamie,
Here are some answers:

1. ColorFlow will create a color model from ECI 2002 data as well as IT8/7.4 data. If you have measured an ECI, you can simply import the CGATS formatted data into the measurement dialog. ColorFlow will not throw any patches away inherently because the data is from either ECI2002 or IT8.

2. The advantage is only short-lived and the approach of adding a characterization chart in a single "sweet spot" on the press is a hotly debated topic. On the one hand, everyone agrees that it is easier to characterize a condition when the area that you are controlling and need to maintain uniformity over is small. In my view, I always ask the question "Do you run all your production work on only 8-10 inches of your printable area on press?". Of course, the answer is no. So, then why only characterize this area? Without a doubt, it is easier to do this, but I believe it doesn't set the correct expectation of a press operator. Uniformity is one of the single most important variables to control during a characterization, and it should be also controlled during normal production operation as well. By running a characterization target that spans the printable area of the press, you are able to report on density fluctuations across the sheet and will give you valuable insight down the road when someone asks "Hey why is this image in the upper right hand corner of the press so warm and everything else is in perfect grey balance?". The report functionality in ColorFlow will give you this information so long as you characterize the entire printable area.

3. I do have a feature request PR open for this. The PR number is DHER-1439

4. At this point, the solid patches are used for the purpose of reporting only. Specifically, they function to provide data to report on density across the sheet values that i mentioned above.

5. I have created PR DHER-1469 for this request. A brief note, though. You can do this so long as you measure multiple sheets in a ColorFlow created chart. The only thing we don't support at this time is the creation of a single measurement set by importing externally obtained measurement files.

Thanks a lot for the feedback, Jamie! Always appreciated.


12-08-2010, 07:35 PM
Thanks for the info. I really appreciate your time and attention on this forum.