Pages: [1]   Go Down

Author Topic: Randomize patches or not in I1Profiler  (Read 915 times)

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Randomize patches or not in I1Profiler
« on: October 27, 2017, 08:08:21 pm »

I have two printers. A Canon 9500 II and an Epson 9800. If I make a single set of patches as opposed to two sets with the same patches but at different locations for averaging, I've found a significant difference.

This is data from profiling using the i1iSiS 2 xl. Some info is specific to that. Some applies equally to the I1Pro.

The 9500 patch color depends slightly on the color of the patches adjacent to it. The 9800 does not. However, the 9800 patch color varies slightly (about +/- .25 dE RMS based on the horizontal paper position. The 9500, OTOH, depends on the colors of adjacent patches. The variation is slightly less, about +/- .2 dE. This is true on all papers.

Additionally, when profiling matte paper. Specifically the Canson Rag Photo, there is a bit of color smear but only on the 9500. This showed up on the white patch (RGB 255,255,255) which is unprinted. On a target sheet the patches to the left of the white patches were all orange/yellow shades and some of the pigment transferred, possibly through friction, to the white patch resulting in the *b shifting to +2.9 instead of the actual white point of the paper which is at b*=.9. Verified with the i1iSiS and 3 different I1Pros. This shift causes Absolute Colorimetric printing to be way off as it shifts the entire tint of the brighter areas about 2 dE into the blues (opposite the yellow bias of b*). No visible effect on Perceptual and Relative of course.

I addressed this effectively by randomizing the patches and adding additional white patches for 9500 targets. While the additional white patches varied more than other patches of the same color, randomizing their location did eliminate the extreme shift seen on the b* and they all remained nearer neutral.

As for the 9800, I made 3 sets of target files, each randomized differently. This reduces somewhat the variations that occur from patch location sensitivity. Also, it's a subtle effect that mostly occurs in saturated colors, especially the more saturated yellows. The neutrals have only about half the variation from position probably because there are 3 neutral inks. pk, lpk, and llpk.

So, all in all I'd recommend randomizing even a single patch set. It can help for printers like the 9500. But it doesn't hurt with both of my printers.
« Last Edit: October 27, 2017, 08:24:06 pm by Doug Gray »
Logged
Pages: [1]   Go Up