I was thinking about buying one of the scanner-based printer profilers. But after reading all the tales of endless tweaking and frustration, I decided to hold off. Perhaps the technology behind these products (or the cheap scanners that they rely upon) needs to mature. If the scanner is the weak link in these systems, I'm wondering if it is possible to build a printer profile without a scanner or spectro, using only feedback from human visual judgments. I'm thinking of something like Epson's print head alignment procedure, where the user prints out a series of samples, makes a choice, inputs that choice to the software, and interates if necessary. Suppose I have a target page with a set of calibrated color patches (IT-8 or Macbeth Colorchecker, or something optimized for this particular task). I run the profiling software, and in its first phase it prints out a test page carrying several variants of the color patches on the target page. So for example, there might be a calibrated blue patch on the target. My printed test page has a series of blue patches, which vary around the target color, from the cyan-ish side to the magenta-ish side. I compare the test and target pages under my preferred lighting source, and pick out the test patch which best matches the target blue. Then I enter the code for that test patch into the software. I do the same for several other color series. The software then generates a new test page with refined settings, I enter more data, and repeat until the test page provides a good match. When I confirm the match, the program generates a profile. I have not used any of the existing profiling products, but I know that some of them contain visual-based editing facilities. What I am wondering now, is whether it is possible to create a printer profiling system that uses ONLY visual judgements for feedback. Perhaps an eyeball-based system would not match the precision of an expensive hardware spectro. But maybe the results would not be much worse than the results people are getting from scanner-based profilers. I also wonder if an eyeball-based product could be marketed for a more attractive price than the current options. Since I don't own a flatbed, a scanner-based profiling system is going to cost me something like $300 to 500. Why wouldn't it be possible to sell an eyeball-based program, bundled with calibrated target, for $99 or less? -- Julian Vrieslander <mailto:julianv@mindspring.com> - Turn off HTML mail features. Keep quoted material short. Use accurate subject lines. http://www.leben.com/lists for list instructions.