Ad
Ad
Ad
Pages: « 1 [2]   Bottom of Page
Print
Author Topic: EyeOne Printer Profiling (iPF5000)  (Read 8444 times)
John Hollenberg
Sr. Member
****
Offline Offline

Posts: 765


« Reply #20 on: December 14, 2006, 09:27:47 PM »
ReplyReply

Quote
PC.  V2.0.0 of the plug-in.
[a href=\"index.php?act=findpost&pid=90616\"][{POST_SNAPBACK}][/a]

Rats!  Have you called Canon support to see what they have to say about this?

--John
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2108


« Reply #21 on: December 14, 2006, 11:57:42 PM »
ReplyReply

Quote
It's also helpful to know that we can, in addition to using Match and MeasureTool, also use MonacoPROFILER.

One thing I forgot to mention is that you can use a single set of measurements obtained with MeasureTool to build profiles with either Match, ProfileMaker, or MonacoPROFILER. This is nice because you don't have to re-measure the targets.

For example, let's say you got an Eye-One Photo package (which comes with Match) and also have MonacoPROFILER. You can use MeasureTool as described above to read a target then use the renaming trick to build a profile with Match. But then you can use the same measurement data (without re-reading the targets) to build a profile with MonacoPROFILER. Then when it comes to printing a particular image, you can soft-proof in PS using both profiles and pick the one that gives the most pleasing rendition for that particular image.

Eric
Logged

Dale Allyn
Full Member
***
Offline Offline

Posts: 225


WWW
« Reply #22 on: December 15, 2006, 12:02:50 AM »
ReplyReply

That's good to know, Eric. Thank you. It seems that there's a lot more versatility to be had here than I was aware of.

Cheers,

Dale
Logged

Dan Wells
Sr. Member
****
Offline Offline

Posts: 334


WWW
« Reply #23 on: December 15, 2006, 12:25:02 AM »
ReplyReply

For what it's worth, I've had no problems printing targets. I use the TIFFs that Bill Atkinson provides, open them in Photoshop and send them out the plugin. Other than that I once forgot to turn off color management when printing a target (very little quite as useless as a profiling target printed with a profile for another paper), they have all printed fine... As you can see I'm not attempting to use any profiling software either to generate the target (Bill Atkinson has done that), or to print it (Photoshop to Canon plugin).

My next big experiment is the Canon plugin with Photoshop CS3 - I'm downloading the beta from Adobe right now (it just showed up).


                                                 -Dan
Logged
John Hollenberg
Sr. Member
****
Offline Offline

Posts: 765


« Reply #24 on: December 15, 2006, 12:36:19 AM »
ReplyReply

Quote
For what it's worth, I've had no problems printing targets. I use the TIFFs that Bill Atkinson provides, open them in Photoshop and send them out the plugin. Other than that I once forgot to turn off color management when printing a target (very little quite as useless as a profiling target printed with a profile for another paper), they have all printed fine... As you can see I'm not attempting to use any profiling software either to generate the target (Bill Atkinson has done that), or to print it (Photoshop to Canon plugin).

Yes, but are you having any problems printing through the plugin using the profile that you made after measuring the 4096 patch target?  That was the problem brought up on page 1 of this thread by another poster.

--John
Logged
David White
Sr. Member
****
Offline Offline

Posts: 272



WWW
« Reply #25 on: December 15, 2006, 01:31:43 AM »
ReplyReply

Quote
Rats!  Have you called Canon support to see what they have to say about this?
[a href=\"index.php?act=findpost&pid=90619\"][{POST_SNAPBACK}][/a]

On the list for the next call.  It's really a low priority since I don't think there is much to be gained with profiles that large.
Logged

David White
Andrew Larkin
Jr. Member
**
Offline Offline

Posts: 85


« Reply #26 on: December 15, 2006, 04:48:08 AM »
ReplyReply

Quote
On the list for the next call.  It's really a low priority since I don't think there is much to be gained with profiles that large.
[a href=\"index.php?act=findpost&pid=90648\"][{POST_SNAPBACK}][/a]

I independently found the technique for getting i1Match to handle larger charts about 2 years ago and have been developing new test charts since using my own generator software.

My current chart is 5100 patches over 3 x A3 pages.

There is a world of difference between a profile based on a 918 patch chart and one based on a much larger number of patches (2000, 4000, 5100, whatever).

However, this difference is very much dependant on the printer driver and the printer model.

In my case, I developed the new profiles to address a chronic problem I had with profiling my Epson 4000.  Getting usable colour out of a printer is pretty easy, even with a 918-patch chart.  The problem is getting the gray balance right at the same time.

The native Epson 4000 space (that is, what you get when you print with all management turned off) produces grays that aren't within cooee of being right with a heavy green cast.

Off-the-shelf test charts basically put test points (patches) uniformly throughout the RGB coordinate "space".  In the case of the 4096-patch Atkinson chart, this is basically 16 values of red with 16 values of green with 16 values of blue.

In the case of a 918-patch chart, it is closer to 9 values of each RGB, plus some augmentation around the corners and edges.  This basically puts the patches at around 28 units apart or worse.

These control points are simply incompetent at trying to control the wayward gray axis of the Epson.

I have profiled only one Canon printer - a consumer grade 8500.  The test charts through that printer/driver were vastly better as far as giving basic grays before profiling.  This guaranteed neutral grays, but I then have to wonder whether the printer/driver really was stepping out of the way for colour management and therefore to what degree the available gamut was being compromised.

I would be interested in knowing what you get with the iPF5000 printing a gray-scale test chart with all colour management off (as you would for a profiling test chart).

If the chart does show significant colour cast in the "native" grays, then a chart with additional patches surrounding the gray axis will produce noticable improvements.

The result I now get from the Epson 4000 as far as monochrome prints are as neutral and even as I have ever wanted.

So, run the simple test: if a gray scale test with all colour management disabled produces significant colour cast, then you really should try to use a much larger profiling chart.

The test of any profile chart is how well it produces gray tones as well as the colours.

Andrew
Sydney
Logged
Dan Wells
Sr. Member
****
Offline Offline

Posts: 334


WWW
« Reply #27 on: December 15, 2006, 10:41:30 AM »
ReplyReply

No problem getting prints based on the new profile, either. The 4096 patch profile is about 1.8 megabytes, while the 918 patch profile is a bit under a megabyte, in case profile size has anything to do with the other poster's problems with the plugin crashing.

                          -Dan
Logged
David White
Sr. Member
****
Offline Offline

Posts: 272



WWW
« Reply #28 on: December 15, 2006, 11:49:52 AM »
ReplyReply

Quote
No problem getting prints based on the new profile, either. The 4096 patch profile is about 1.8 megabytes, while the 918 patch profile is a bit under a megabyte, in case profile size has anything to do with the other poster's problems with the plugin crashing.
[a href=\"index.php?act=findpost&pid=90704\"][{POST_SNAPBACK}][/a]

At time time I tried the 4096 patch profiles I must have had something wrong in the system due to all the software installs and removals I do.  I recently did some major system housekeeping so I tried it again and it is now printing with the 4096 patch profile.  My sizes are different with the Default size profile being 2.003MB and the Large size being 2.757MB.  This could be due to using different profiling packages.  I am using ProfileMaker 5.07.
Logged

David White
JeffKohn
Sr. Member
****
Offline Offline

Posts: 1671



WWW
« Reply #29 on: December 15, 2006, 12:04:21 PM »
ReplyReply

Quote
I independently found the technique for getting i1Match to handle larger charts about 2 years ago and have been developing new test charts since using my own generator software.
Is this software you wrote available?
Logged

David White
Sr. Member
****
Offline Offline

Posts: 272



WWW
« Reply #30 on: December 15, 2006, 12:20:47 PM »
ReplyReply

Quote
I would be interested in knowing what you get with the iPF5000 printing a gray-scale test chart with all colour management off (as you would for a profiling test chart).

If the chart does show significant colour cast in the "native" grays, then a chart with additional patches surrounding the gray axis will produce noticable improvements.

The result I now get from the Epson 4000 as far as monochrome prints are as neutral and even as I have ever wanted.

So, run the simple test: if a gray scale test with all colour management disabled produces significant colour cast, then you really should try to use a much larger profiling chart.
[a href=\"index.php?act=findpost&pid=90668\"][{POST_SNAPBACK}][/a]

My ipf5000 prints a gray-scale test chart with no visual hints of a color cast.

I have to wonder about the validity of this test.  Depending upon the algorithms that the driver and printer firmware use, perhaps this only shows how well the printer can produce a gray-scale chart using the gray and black inks and the neutrality of the gray and black inks.
Logged

David White
Andrew Larkin
Jr. Member
**
Offline Offline

Posts: 85


« Reply #31 on: December 15, 2006, 05:48:08 PM »
ReplyReply

Quote
Is this software you wrote available?
[a href=\"index.php?act=findpost&pid=90719\"][{POST_SNAPBACK}][/a]
Sorry, no.

I wrote the program to do the patch shuffling required to allow the Eye One to operate in strip mode.  It would take a week or two of effort to turn it into something usable to anyone else and I just can't spend that time.

Since then the need for the program has been obsoleted to some degree by new methods of distinguishing patches in a test chart being available via the "demo mode" PM5 package.  It is possible to generate charts that don't need to be shuffled now by using black/white edges to the patches.

Andrew
Logged
Andrew Larkin
Jr. Member
**
Offline Offline

Posts: 85


« Reply #32 on: December 15, 2006, 06:26:11 PM »
ReplyReply

Quote
My ipf5000 prints a gray-scale test chart with no visual hints of a color cast.

I have to wonder about the validity of this test.  Depending upon the algorithms that the driver and printer firmware use, perhaps this only shows how well the printer can produce a gray-scale chart using the gray and black inks and the neutrality of the gray and black inks.
[a href=\"index.php?act=findpost&pid=90725\"][{POST_SNAPBACK}][/a]
Oh, believe me the test is valid!

I fully expected the result you have received the very first time I printed a "native" gray scale for the Epson 4000.  When I saw the heavy colour cast, my reaction was "What the?"!  How could it be that outputting gray RGB values could be using anything other than gray inks?  I still don't know.

Given the result you have found, the 4096-patch Atkinson chart should work just fine.

If you want to see how much adjusting is going on in a profile to get the grays, try this:

1. Either loadup a gray-scale or create a gray gradient image in photoshop.  Make a note of the working colour space for the image.
2. Use "Convert to Profile" with your printer profile selected.  You can now see what RGB values are being used to reproduce your grays by looking at the Info window and dragging the mouse over the image.

To really see what is going on, try using "Assign to profile" to assign the image back into its original working space.  Comparing this image with the original will show how far the RGB values are being shifted to give the right output.

It really may not even be a problem on the newer Epson K3-based printers either.  But on the UltraChrome printers it is very much an issue.

One "professional" profile maker in Australia is in the habit of instructing their clients to print the test charts for the Epson WITH printer-driver colour management enabled on the grounds that it gives better results.  And they are right - the driver does most of the correction with the newly-developed profile making finer corrections over the top.  You can get some nice neutral grays using this method with a 918 chart.  The cost, however, is reduced colour gamut.

Andrew
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2108


« Reply #33 on: December 15, 2006, 09:15:34 PM »
ReplyReply

Under Bill Atkinson's color profile downloads (see his web site), there is a Profile Targets FAQ. In this FAQ Bill describes in detail how his older 9600 (regular Ultrachrome) had "lumpy" non-linear behavior and was far from gray-balanced. This is the behavior that Andrew describes with his 4000. I have seen this personally with my 2200. The K3 printers (and their associated drivers) are significantly more linear and gray-balanced and therefore make it easier to create good profiles for them. With the older Ultrachrome printers, it does help to use more patches (e.g., 4096) to help correct the gray balance.

With Bill's targets freely available online for several different measuring devices, several different patch sizes, and several different sheet size configurations, I no longer see the need for using custom software to generate yet more targets. Just browse his profile targets and pick what you need.

Eric
Logged

David White
Sr. Member
****
Offline Offline

Posts: 272



WWW
« Reply #34 on: December 15, 2006, 09:56:00 PM »
ReplyReply

With the Canon ipf5000 there is virtually no difference between the values obtained after converting a gray-scale step chart between a 4096 and a 1472 patch profile.  It's kind of like putting high octane gas in a car designed for regular gas.  It doesn't hurt, but it doesn't help and you could spend your resources better elsewhere.

Agreed, that with the older printers, profiles generated from very large test charts probably helped quite a bit, but defintely not with my 5000.
Logged

David White
Pages: « 1 [2]   Top of Page
Print
Jump to:  

Ad
Ad
Ad