Using Atkinson iSis targets in i1profiler?

[narikin]

Can someone advise me which of the Bill Atkinson targets seems a happy compromise for getting a great profile from the iSis XL with i1publish? I see them from 1000-near 6000 patch on the download.

and am I right is assuming that most of these were made for Profilemaker, and if so will they work in the new i1publish at all?

or... if there is really no advantage anymore to using Bill's targets, then please feel free to say so...  though your recommendation of a high end target size is helpful in that case (number of patches not an issue, max quality desired!)

thanks.

[shewhorn]

I can't figure out why the Atkinson profiles get as much attention as they do. There's really nothing special about them aside from the fact that there's a greyscale step wedge included which is something anyone can do using ColorPort (that may have made a difference back in the days of the Epson x600 series and other earlier model printers but not so much anymore, at least from what I've seen). Aside from that, the math used to determine the patches is pretty much identical in Monaco Profiler, and i1Profiler. I've tested the addition of a greyscale test wedge for use in i1Profiler and I came to the conclusion that 1) in most cases it makes little if any difference and 2) sometimes it makes things worse.  What's sometimes a bit more advantageous is some patches just outside of a perfect grey and you'll notice that i1Profiler actually adds these in if you choose a number of patches that isn't a multiple of x^3, otherwise the Atkinson Profiles and the numbers i1Profiler generates are are pretty much identical. The Atkinson generated profiles are 16 bits and i1Profiler is generating 8 bit profiles but I don't see that this has any practical advantage because regardless of the amount of precision your reference has, you're still measuring with the same amount of precision and when compared to the reference it's going to be a certain amount off. It's all relative in that sense. Now, if i1Profiler is incapable of using 16 bit reference data then using the Atkinson profiles can actually introduce errors into your profiles. I asked whether or not i1Profiler could handle 16 bit reference data and I never got a straight answer.

Toss the Atkinson profiles aside, i1Profiler has a (theoretically... there's fine print) MUCH better method of generating charts and that's the optimization feature which measures the response of your printer, and then generates a custom set of patches based upon those measurements. Sounds pretty slick but as far as large format inkjet printers are concerned (I'm sure that there are other printing technologies where this does help), it makes no discernible difference in a print, at least when you're working with targets that start with 800 or more patches.

My recommendation is to just use the i1Profiler generated patches. If you want to get those patches that don't conform to the even distribution of colors then set the patch count to x^3 - 1 (so 728, 999, 1330, 1727 patches, etc.). I'd encourage you to experiment though. There's no magic formula for what's going to work best in all cases because everyone's needs/equipment/aesthetic values, etc are different.

Cheers, Joe

[Alan Goldhammer]

I've tested the addition of a greyscale test wedge for use in i1Profiler and I came to the conclusion that 1) in most cases it makes little if any difference and 2) sometimes it makes things worse.  What's sometimes a bit more advantageous is some patches just outside of a perfect grey and you'll notice that i1Profiler actually adds these in if you choose a number of patches that isn't a multiple of x^3, otherwise the Atkinson Profiles and the numbers i1Profiler generates are are pretty much identical.

Cheers, Joe

I don't have i1Profiler but have been making profiles using ArgyllCMS and you do have the ability to set up various combinations of greyscale wedges.  I also haven't seen much of an effect if I go from a small set to a much larger set (and of course the more grayscale patches you include the fewer color patches are printed assuming a fixed number of target pages).

[digitaldog]

I can't figure out why the Atkinson profiles get as much attention as they do.

With all due respect to Bill who’s nothing short of a genius (for color management as well as other accomplishments), I have to say that tests I did early on showed that a target generated from i1P of the same patch numbers provided slightly better quality than using Bill’s. The new generator is absolutely something X-rite color scientists did a really great job upgrading. So roll you own.

[Schewe]

With all due respect to Bill who’s nothing short of a genius (for color management as well as other accomplishments)...

I think we need to understand Bill's targets within the context of why he made them. Back when the Epson 9600 camera out, the No Color Adjustment part of the Epson driver was very non-linear. Using the targets of the day from X-rite and Greytag failed to produce a really good profile (in Bill's mind at least). So he spent a summer doing targets and testing the profiles and came up with the 918 target as a good compromise between accuracy and smoothness of the profiles he made. It also help jumpstart addition development of profiling targets as well as gave a wakeup call to Epson to fix the No Color Adjustment colors–which they did in the next print, the x800 series.

It would not be at all surprising that i1Profiler can now generate really good targets since they are building on top of what Bill has done.