Message
Re: [colorvision_group] Idea for profile accuracy
2006-03-09 by CDTobie@aol.com
In a message dated 3/9/06 8:28:32 AM, jpgentry@... writes:
Forget trying to make patch reading easier/faster, it's already a
breeze.
I'll certainly second that one...
But here's an idea for profile accuracy.
This could be done with a modification of the software or outside the
software with a nifty Excel spreadsheet (and I hope this wasn't
already thought of but it may have been.)
You really don't want to average profiles, you just want to cast out
anomalies. So you must take a minimum of three readings per patch.
The program (spreadsheet whatever) will look at the readings on the
three and searh for the reading that is out of bounds keeping an
average of the two closest reading (or just one of them.) This would
be a sure way to catch a false read on a patch.
If I found that I made even one undiscovered patch error per chart (or even per several charts) then this type of multiple reading scheme would interest me more... but I don't see why errors are happening and not being noticed in the first place. Let me be frank: I can go to my hotel room, after a long day on a show floor (talking to all of you ), and a long evening out doing business, and partaking of more refreshments than I ever do at home, do a couple of hours or email, and about two in the morning, sleep deprived, I get around to reading the target prints that other booths gave me to color manage their booth printers. I read a few of them, check the results in the morning, and they are all just fine. So maybe its just me, but reading a chart a hair slower once, instead of sloppily the three times needed for averaging, seems the quicker and easier solution. The "hair slower" speed for me is about five minutes per 225 patch chart... I can do it in four if I'm wide awake, and hurry... even then I might reread two or three patches (within that four minutes), and not miss a single one that I didn't know I blew, when I blew it.
Here's how color geeks used to check this: open the TIFF version of the target that you used to profile your printer/ink/paper in Photoshop in AdobeRGB, and go to Custom Proof Setup, and apply your custom profile to the target. Toggle between softproof, and nonsoftproofed views (Command or Control Y, depending on platform) and see if any of the patches show as being out of line. We built this feature right into PrintFIX PRO, with the Raw/Split/Measured option for any measurement set, and now we're adding quick key commands to let you toggle between them instantly. How can you miss patches, when you get the option to do this check, and to remeasure any suspicious patches on the spot, before even building your profile?
That is what PrintFIX PRO is about: bringing years of color expert's knowledge and techniques to all users. Tuning paper white and ink black previews, selecting an optimal media setting, all of those tricks are expert methods now available to anyone who'll take the time to read the help, and give it a try.
Oh, on the point of how carefully does the nose of the Spectro be placed on the patch:
The nose of the spectro is 10mm wide. The opening in it is 7mm wide. The measured area, within that 7mm illuminated area, is 4mm wide. This allows for "over-illumination" necessary to accurately measure translucent materials. So when you notice that the nose of the spectro has crowded the edge of the patch, and remeasure it, chances are that you still weren't off enough to have effected the reading. In fact, since the gap between patches is typically about a 1.5mm, you would have to be over the line, and have the nosecone onto the next patch before it would effect readings. Try it in the spot measure function, you'll find this to be true. So a lot of concerns about reading accuracy and reading times, are just obsessing about things that simply don't matter. If you read a chart slowly and precisely, and then read it fast and dirty, as long as you scan for errors before building the profile, I don't think you'll be able to tell one from the other... just (as Red Green would say) keep your stick on the ice...
C. David Tobie
Product Technology Manager
ColorVision Business Unit
Datacolor Inc.
CDTobie@...
www.colorvision.com