Welcome to OnyxTalk - Global Support Community, please Login or Create an account to get full access to the forums.
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5

Profile's Average Error

Offline Digital_Mike Posted 02-16-2015 - 02:51 PM
Post: #1
Member
3 Posts
Member

Onyx Version:
Production House 10
Printer(s):
Many

Greets all,

I have color management knowledge going back to around 2000 or so (Colorblind) but have admittedly always been stymied by Onyx's implementation. Fast forward to today where I have taken over for the person who (among many things) handled the color management.

Now, we do some scanning here and our clients certainly arent the pickiest but I challenge myself to nail the color every time. That said, the profiles that were created here, in everyday use, look pretty darn good. We recently changed ink in our GT350 flatbed and so its my turn to play color manager. Basically I used all the previous ink limits and the settings he used to build his profile and simply re-outputted the MM, Lins and ICC targets and re-read them. Easy Peasy....kinda.

So, when testing mine against his, I used Media Manager's Gamut Report to see how I fared. His profiles were typically 16.5-17 regardless of substrate. Mine was a solid 16. I notice mine are more open than his as well as a touch lighter in the half and 3/4 tones, and generally I like the look of mine.

So that brings me to my question. 16.o in Onyx terms seems high. and as the documentation states, the target should be a delta of 5 or less-ish. So, despite things looking pretty darn good, are y'all hitting deltas of 5 or less?
Back to top Find Quote
Offline Correct Color Posted 02-24-2015 - 10:07 AM
Post: #2
Frequent Flyer
128 Posts
Frequent Flyer

Onyx Version:
10
Printer(s):
Epson

Mike,

First off, I wouldn't worry too much about the Onyx Gamut Report. I suppose it's measuring a difference between the total color gamut of whatever spots are in its library, vs. the particular gamut it's analyzing, but for all I can see, it never says so anywhere. And without some frame of reference, a value isn't a whole lot of good.

As far as a delta e of under five, that's an indication of how well you've done reproducing a specific color, or how close a profile of a device printing in a certain state can be expected to reproduce a certain color. An average across a profile might give you some clue of the overall gamut size of a profile, but it's such a clumsy measurement that it's not going to tell you much at all.

Overall, the GT350 is pretty much a mid-sized gamut device. And just like any other device in its range, there are colors it can hit spot on, and others it's going to miss by a pretty wide margin, regardless of how well it's profiled. The trick is to get all out of it you can.

And the thing is, it's creating a machine state that's going to determine of you get a larger-gamut profile. Nothing else.

And the only machine-state variable you've changed here is the ink.

So any differences between "his" profiles and "your" profiles are strictly the difference between the new and old ink. As long as you leave all the inking-related variables as they are, then you're really just updating existing profiles, you're not creating new ones.

Mike Adams
Correct Color
Back to top WWW Find Quote
Offline Digital_Mike Posted 02-25-2015 - 03:28 PM
Post: #3
Member
3 Posts
Member

Onyx Version:
Production House 10
Printer(s):
Many

Good to know. I guess the "Average Error" being a value of 16 got me tweaked. As I said, the results are quite good using the profile so I guess I expected better.

Thanks Mike
Back to top Find Quote
Offline Mikulas Posted 05-04-2015 - 05:26 PM
Post: #4
Banned
55 Posts
Banned

Onyx Version:
10.2
Printer(s):
L26500, W8400+Onyx / Mimaki JV4, JV5 +TexPrint

Hi Mike.

I´m fresh new here, so a bit later answer.

1/ measuring dE has much more reason inside the printer´s gamut.
If you don´t have your specific "color checker" then maybe start with something simple like Gretag-Macbeth Color Checker. It´s colors are made to be inside any average printer gamut. Printing and measuring such target tells you much more about your profile´s accuracy than some "large library spot color test". In facts, such test is only a kind of "gamut mapping" report. It´s no real test at all, because it does not measure and compare real colors!! It only does some mathematical comparisons. The biggest dE is caused by gamut mapping outside of the printer´s gamut.

2/ if you are after accuracy I suggest a different comparison: when profiling - repeat you measurement 2-4x and try to compare the data sets (PM5 is very good, but Excell can also do it). From my experience - this is the point where you can save some accuracy. Most resources suggest averaging 2-3 measurements, but if one of them is realy bad (dE>10) you will never get an accurate result. After comparing them, leave out the worst data sets and do the averaging only from the best readings.
Back to top Find Quote


Possibly Related Threads...
Thread: Author Replies: Views: Last Post
  using our output profile as standard workflow profile in Photoshop? jeffburris 5 4,686 02-27-2008 09:18 AM
Last Post: M.A.C.


User(s) browsing this thread
1 Guest(s)
myRepono - Website & mySQL Database Backup Service
Community software by MyBB