Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: The effect of using more i1Profiler patches  (Read 3612 times)
bill t.
Sr. Member
****
Offline Offline

Posts: 2693


WWW
« on: May 15, 2013, 11:30:30 AM »
ReplyReply

Here are graphs of profiles obtained from 1831 and 1218 patch targets, printed at the same time on adjacent sections of the same canvas, and read back to back.  Old style i1ProPhoto puck, latest i1Profiler software. 

The wireframe is the profile I got with a 1831 patches, the solid hull is the profile obtained from 1218 patches.

There are some really significant differences, particularly in those blue and violet colors that are so elusive on matte media.  The 1831 patch profile does a lot nicer job on transparent blue skies, and you can see it on prints.

So, now I don't know what to think.  I assume the 1831 patch profile is truer and better because it looks that way on prints.  But the 1213 patch profile also yields very nice prints, but is clearly second best in the graded blue sky department.  At any rate, I was surprised to see such a difference.  Of course, I've only got two samples here.  Maybe I need to knuckle down and try this a few more times.  But even on this example I read each target twice, and there are essentially no discrepancies between reading sets on the same targets, because their graphs overlay perfectly.

Do you iSis jockeys see similar differences?

So maybe the moral is...you can benefit from lots of patches, and it may be very worth your while to use targets with with upwards to 2000 patches.
Logged
keith_cooper
Sr. Member
****
Offline Offline

Posts: 442



WWW
« Reply #1 on: May 15, 2013, 05:27:37 PM »
ReplyReply

Yup, my standard RGB target for the iSis is ~3000 patches on an A3+ sheet.

I've not seen much benefit from over 2000, but since I'm printing on an A3+ sheet, I might as well fill it up with patches :-)

During i1Profiler testing I did try out the 2nd patch set refinement process a few times, but starting with sets of ~1700 I couldn't see much difference, and have always assumed it would be of more use if I was starting with smaller sets (and didn't have an iSis ;-).

I don't do profiling for a living, so my data set is relatively small, but recent results on the iPF6450 I'm testing, and an Epson R2000 have produced some very nice prints from the 3k patch profiles...
Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 8990



WWW
« Reply #2 on: May 16, 2013, 08:36:37 AM »
ReplyReply

It's not so much the number of color patches, but where in color space they are made. I can tell you that a custom 1728 patch target from i1P produces slightly better visual results than a 1728 patch target generated from ProfileMaker Pro. But the differences are rather small. IF you own i1P and you're making custom RGB targets, let it generate the patches. For CMYK, I always stick with the ECI2002.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Rhossydd
Sr. Member
****
Offline Offline

Posts: 1917


WWW
« Reply #3 on: May 16, 2013, 09:13:47 AM »
ReplyReply

There's been past discussion about this before here (sorry I can't quickly find the references) and the number of patches itself doesn't always seem to be an indicator of performance.
In i1Profiler there seems to be some 'sweet spot' numbers of patches that deliver particularly good results. I think 2033 was a particular favourite. It's actually quite interesting to work through the numbers of patches possible and watch the changes in tonal distribution, especially the numbers of grey patches that the patch set includes.

I've spent some time testing these and agree with Andrew that i1P only has small advantage over PMP5 for the same patch sets, but I also think that using patch sets generated from i1Profiler helps PMP5 deliver it's best performance.
Logged
JohnAONeill
Full Member
***
Offline Offline

Posts: 111



WWW
« Reply #4 on: June 19, 2013, 05:22:42 PM »
ReplyReply


Hi Keith

I'm looking to improve my rgb profiles from ones created with the i1pro standard IT8 two page a4 targets. Would it be impolite to ask if I could get a copy of 3000 patch targets? Do you find much improvement in greyscale neutrality when using these profiles? I have colorport 2 for importing new target data. Would this work with your 3000 patch ones?

Best regards
John


Yup, my standard RGB target for the iSis is ~3000 patches on an A3+ sheet.

I've not seen much benefit from over 2000, but since I'm printing on an A3+ sheet, I might as well fill it up with patches :-)

During i1Profiler testing I did try out the 2nd patch set refinement process a few times, but starting with sets of ~1700 I couldn't see much difference, and have always assumed it would be of more use if I was starting with smaller sets (and didn't have an iSis ;-).

I don't do profiling for a living, so my data set is relatively small, but recent results on the iPF6450 I'm testing, and an Epson R2000 have produced some very nice prints from the 3k patch profiles...
Logged
keith_cooper
Sr. Member
****
Offline Offline

Posts: 442



WWW
« Reply #5 on: June 20, 2013, 02:28:09 AM »
ReplyReply

Hi Keith

I'm looking to improve my rgb profiles from ones created with the i1pro standard IT8 two page a4 targets. Would it be impolite to ask if I could get a copy of 3000 patch targets? Do you find much improvement in greyscale neutrality when using these profiles? I have colorport 2 for importing new target data. Would this work with your 3000 patch ones?

Best regards
John

No problem - they are targets for use with an iSis, but I can export the patch sets if it's of use - mail me directly at the Northlight site (just me there, so the contact address is fine) and that will ensure it gets into my 'to-do' list when I'm in the office.

The previous comment about the tonal distribution is an interesting one, I've watched the generated patch range change as you increase the patch count one by one, and certain numbers reach a maximum grey patches before jumping back to all colour (almost like watching electron shells filling up as you go up the periodic table :-)

Using one of these 'maximum grey' profiles on the Epson R2000 gave me very reasonable B&W prints - a genuine surprise given the ink set
Logged

JohnAONeill
Full Member
***
Offline Offline

Posts: 111



WWW
« Reply #6 on: June 20, 2013, 06:18:15 PM »
ReplyReply

Thanks Keith

I sent an email through your contact page today. I only have the i1pro version 1 with i1 match software. Also have colorport 2 trial version and measure tool. Will any of these work with the larger targets?

Cheers
John
Logged
keith_cooper
Sr. Member
****
Offline Offline

Posts: 442



WWW
« Reply #7 on: June 21, 2013, 03:00:43 AM »
ReplyReply

Ah - therein lies a problem... I did once see a hack for using big targets with i1match, but it didn't work for me. The target file I sent is created by i1Profiler, so also optimised (to some extent) to being used with that software.

The only non-expensive solution is ArgyllCMS, but your tolerance of command line software would need to be somewhat greater than mine to explore ;-)
Logged

Alan Goldhammer
Sr. Member
****
Offline Offline

Posts: 1631


WWW
« Reply #8 on: June 22, 2013, 05:33:31 AM »
ReplyReply

The only non-expensive solution is ArgyllCMS, but your tolerance of command line software would need to be somewhat greater than mine to explore ;-)
LOL, yes it's a bit of a drudge using Argyll but the flexibility is incredible.  I suppose that someone could write a GUI shell for paper profiling the same way that one has been written for display calibration but it would be a chore.  I get very good profiles using 1848 patches with an i1Pro on standard letter size paper.  I include a 51 step B/W patch set which provides flexibility in printing B/W if I'm not going to use the Epson ABW driver.  Argyll can easily be used in conjunction with Roy Harrington's QTR software to generate ABW profiles for Epson printers.

I did look at some profiles using larger patch sets but did not see a whole lot of difference in the profiles and certainly resulting test prints were identical.

Alan
Logged

keith_cooper
Sr. Member
****
Offline Offline

Posts: 442



WWW
« Reply #9 on: June 22, 2013, 08:02:21 AM »
ReplyReply

Thanks Alan  - I'll admit to being a bit spoilt from having all the i1Profiler kit ;-)

Argyll is still definitely on my 'must get round to looking at again' list...
Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 8990



WWW
« Reply #10 on: June 22, 2013, 11:38:32 AM »
ReplyReply

To add another possible product into the mix, I just started using a new profiler from ColorLogic called Copra. I just got tired of all the issues with i1Profiler and needed a backup solution. The RGB profiles are just is good as i1P but I'm doing most of the work building CMYK press profiles. There's one tiny area of color space (blues) where i1P is a bit cleaner, but other wise, feeding the same data to both products, Copra is producing hands down better output, especially in terms of neutrals. I have no association with the company (IOW, I paid for the software) but I'm pretty happy there's some competition in this space. Their web page is pretty awful but you can download a demo. The limitation is you have to separate (run the profiles you build as demo) within it but that's a fair means of proving a way to test prior to purchase. Assuming you have a CGATs file, even the new XGRA format, Copra will accept it and build you a profile. For those that miss some of the analysis of MeasureTool, the product comes with ColorAnt and it's reporting is vastly more through than MT!
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Rhossydd
Sr. Member
****
Offline Offline

Posts: 1917


WWW
« Reply #11 on: June 22, 2013, 12:40:02 PM »
ReplyReply

The RGB profiles are just is good as i1P but I'm doing most of the work building CMYK press profiles. There's one tiny area of color space (blues) where i1P is a bit cleaner, but other wise, feeding the same data to both products, Copra is producing hands down better output, especially in terms of neutrals.
Out of interest is it just with CMYK profiles that you're seeing improvements ?
If you're seeing improvements in RGB work too, what impresses you ?
Quote
Their web page is pretty awful but you can download a demo.
But getting a purchase price took quite a search..... not cheap when found either Sad
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8990



WWW
« Reply #12 on: June 22, 2013, 03:59:06 PM »
ReplyReply

Out of interest is it just with CMYK profiles that you're seeing improvements ?
If you're seeing improvements in RGB work too, what impresses you ?But getting a purchase price took quite a search..... not cheap when found either Sad

Better gray balance by a large factor. Even better appearing contrast and depth with lower TAC which is huge. More control in setting black generation.

I was able to get just RGB/CMYK/Gray without Device Link for about $3K.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
JRSmit
Sr. Member
****
Offline Offline

Posts: 361


WWW
« Reply #13 on: July 04, 2013, 02:44:13 AM »
ReplyReply

Better gray balance by a large factor. Even better appearing contrast and depth with lower TAC which is huge. More control in setting black generation.

I was able to get just RGB/CMYK/Gray without Device Link for about $3K.
Andrew, with "appearing contrast and depth" do you mean it gives a more 3-D appearance to the image?

If so, then indeed very interesting. I had a discussion a while ago with a top illustrator on how to make in an illustration(2D) a round shape to appear as a ball (3D) to the obeserver. His reply in essence, "a fine intricate balancing of tone (light to dark) and saturation changes, especially in a relative small area(portion) of the shape".
The supplier for my printer profiles has just recently also purchased CopRa, is on holidays right now, will be interesting when he returns.
Logged

Fine art photography: www.janrsmit.com
Courses and workshops: www.centrumbeeldbeleving.nl

Jan R. Smit
digitaldog
Sr. Member
****
Offline Offline

Posts: 8990



WWW
« Reply #14 on: July 04, 2013, 08:55:26 AM »
ReplyReply

Andrew, with "appearing contrast and depth" do you mean it gives a more 3-D appearance to the image?

Kind of. With a lower TAC, one would expect to see less density and pop, that wasn’t the case. In the CMYK world, if you can use less ink and get as good (or better) results visually using a higher TAC, that's all good (ink or toner is expensive).
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad