Pages: [1]   Go Down

Author Topic: getting a better manually read profile  (Read 7886 times)

sesshin

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 94
    • http://
getting a better manually read profile
« on: January 28, 2009, 03:25:58 pm »

I've recently downloaded Bill Atkinson's rgb1728 profile targets and have been using that with an i1 Pro and Profile Maker's Measure Tool to profile papers on my Epson 7880. The resulting scan is attached below.

Obviously there are some parts where the color transition is kind of choppy and not as smooth as the original target image. Is this indicative of a bad profile? and if so, what is the best way to go about making a better one?

I remember reading somewhere that some people use Excel to average out a profile, but I'm not entirely certain on how to do that. Obviously I would love to just do one reading and have that work but will do multiple readings if necessary.
Logged

Wayne Fox

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4237
    • waynefox.com
getting a better manually read profile
« Reply #1 on: January 28, 2009, 05:57:27 pm »

Quote from: sesshin
I've recently downloaded Bill Atkinson's rgb1728 profile targets and have been using that with an i1 Pro and Profile Maker's Measure Tool to profile papers on my Epson 7880. The resulting scan is attached below.

Obviously there are some parts where the color transition is kind of choppy and not as smooth as the original target image. Is this indicative of a bad profile? and if so, what is the best way to go about making a better one?

I remember reading somewhere that some people use Excel to average out a profile, but I'm not entirely certain on how to do that. Obviously I would love to just do one reading and have that work but will do multiple readings if necessary.

To me that looks pretty typical.  What I look for on that screen is any glaring problems. It never looks as smooth as the actual target - perhaps it's a limitation of the monitor.  You've sent some raw color numbers to the printer, and the results are what they are.  

I evaluate the data by making the profile and print some reference prints ... Bill has several on his download page, and I have another one I use that I believe came from Digital Dog's site.  You can also take a look at the gamut volume, black density and graph in a program like ColorThink.  I personally don't have the expertise to completely understand what this tool can do, but I do know that if my gamut volume looks substantially off  (for example it is quite a bit lower than the same paper on a 7800), or if the black level seems too high (PK papers on a 7880 should be in he 3 to 5 range), or if the graph looks choppy with a lot of weird peaks and valleys, then I know either I didn't print a good target or I didn't get good readings out of Measure Tool.


Logged

sesshin

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 94
    • http://
getting a better manually read profile
« Reply #2 on: January 28, 2009, 08:36:06 pm »

Evaluating the profile shape its definitely larger than the stock profile and no glaring errors. So that's good. I'll print some test images and see if there is anything that stands out.

I was just thinking the act of scanning by hand could possibly introduce some errors compared to using something like the isis, but in practice I don't know how noticeable this actually is.
Logged

Ryan Grayley

  • Full Member
  • ***
  • Offline Offline
  • Posts: 217
    • RGB Arts Ltd
getting a better manually read profile
« Reply #3 on: January 29, 2009, 04:12:54 am »

Quote from: sesshin
Evaluating the profile shape its definitely larger than the stock profile and no glaring errors. So that's good. I'll print some test images and see if there is anything that stands out.

I was just thinking the act of scanning by hand could possibly introduce some errors compared to using something like the isis, but in practice I don't know how noticeable this actually is.

I find that soft proofing is a useful intermediate step for checking and comparing profiles. I check both Perceptual and Relative Colorimetric rendering (with and without BPC) and I turn on the gamut warning each time. I have spotted a few oddities this way. This approach has saved time and money.

I had also heard about using measurement averaging with Excel and after a short while I managed to figure it out.  This is a lot cheaper than buying the PM Measure Tool and dongle!

Cheers,

Ryan
« Last Edit: January 29, 2009, 04:17:30 am by Ionaca »
Logged
Ryan Grayley BA IEng MIET ARPS
RGB Arts Ltd, London, UK

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
getting a better manually read profile
« Reply #4 on: January 29, 2009, 06:42:39 am »

Quote from: Ionaca
I find that soft proofing is a useful intermediate step for checking and comparing profiles. I check both Perceptual and Relative Colorimetric rendering (with and without BPC) and I turn on the gamut warning each time. I have spotted a few oddities this way. This approach has saved time and money.

I had also heard about using measurement averaging with Excel and after a short while I managed to figure it out.  This is a lot cheaper than buying the PM Measure Tool and dongle!

Cheers,

Ryan

The old version of Measure Tool will average without a dongle. The Digital Dog distributes it on the CD that comes with his book. I can send it to you if you want a copy. It beats the hell out of working with Excel.

Cheers,

Jeff
Logged
Cheers,
 Jeff  www.jeff-grant.com

Ryan Grayley

  • Full Member
  • ***
  • Offline Offline
  • Posts: 217
    • RGB Arts Ltd
getting a better manually read profile
« Reply #5 on: January 29, 2009, 01:32:02 pm »

Quote from: Jeff-Grant
The old version of Measure Tool will average without a dongle. The Digital Dog distributes it on the CD that comes with his book. I can send it to you if you want a copy. It beats the hell out of working with Excel.

Cheers,

Jeff

Are you referring to Andrew Rodney's book 'Color Management for Photographers'?
If yes, then I think I will pick up a copy.

Thanks,

Ryan
Logged
Ryan Grayley BA IEng MIET ARPS
RGB Arts Ltd, London, UK

Wayne Fox

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4237
    • waynefox.com
getting a better manually read profile
« Reply #6 on: January 29, 2009, 06:00:50 pm »

Quote from: Ionaca
Are you referring to Andrew Rodney's book 'Color Management for Photographers'?
If yes, then I think I will pick up a copy.

Thanks,

Ryan

I have that book ... glad to know that. I've got a project to do which needs averaging, will find that CD.

 Most dongles for ProfileMaker don't enable that feature as well.  The package I bought for around $2k doesn't allow me to average or make my own targets.
Logged

Pat Herold

  • Full Member
  • ***
  • Offline Offline
  • Posts: 161
getting a better manually read profile
« Reply #7 on: January 29, 2009, 09:16:05 pm »

If I understand the OP correctly, you are trying to judge the quality of the profile by printing the target through the profile you made?  Steve Upton had a color management "myth" about that a while ago:

http://www.colorwiki.com/wiki/Color_Management_Myths_11-15
Logged
-Patrick Herold
  Tech Support,  chromix.com

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
getting a better manually read profile
« Reply #8 on: January 29, 2009, 11:20:48 pm »

Quote from: Ionaca
Are you referring to Andrew Rodney's book 'Color Management for Photographers'?
If yes, then I think I will pick up a copy.

Thanks,

Ryan
That's the one.
Logged
Cheers,
 Jeff  www.jeff-grant.com

neil snape

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1447
    • http://www.neilsnape.com
getting a better manually read profile
« Reply #9 on: February 01, 2009, 04:41:19 am »

The i1 Pro definitely does introduce variables. The most important variable is the temperature of the lamps (tungsten versions) which change their values with delta T.  IF you wanted  the highest level of precision, you'd need to run through half a chart or more just to warm up the lamp. You can test this with the MEasure tool or any other method with a display of manual patch readings, or strip read with delta E differences over the same strip.
The amount of flare or adjacent patches with the i1 can largely influence the values. Any cockle or differences in the height from the i1 to the media also greatly influence the readings.
The amount of diffusion also changes if you are using the old ruler and not protecting the surface of photo surfaces (well maybe matte too). I've seen as much as delta 3 L* on blacks when they are scratched.

So yes you can average them, but take your time in the initial readings first.  Often I compare before averaging which tells me if I wavered in the passes. The new ruler is much more reliable BTW.
Logged

Wayne Fox

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4237
    • waynefox.com
getting a better manually read profile
« Reply #10 on: February 01, 2009, 02:35:22 pm »

Quote from: neil snape
The i1 Pro definitely does introduce variables. The most important variable is the temperature of the lamps (tungsten versions) which change their values with delta T.

Does this mean there are some non-tungsten  (perhaps LED) versions that are better?

If so, any clue which revision?  I have an A and B, my dealers rev. D seemed to perform much better than either of mine.  Not sure what is different about it.
Logged

neil snape

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1447
    • http://www.neilsnape.com
getting a better manually read profile
« Reply #11 on: February 01, 2009, 02:46:42 pm »

Quote from: Wayne Fox
Does this mean there are some non-tungsten  (perhaps LED) versions that are better?

If so, any clue which revision?  I have an A and B, my dealers rev. D seemed to perform much better than either of mine.  Not sure what is different about it.


I've only had the A and B revs. I think the rev D was supposed to be a white LED. I know it is in the more recent HP Z printers, and some other X-Rite devices. The IsIs is a little different as it has a UV emitting LED as well to make for dual readings.
The LEDs are narrow band, and supposedly very stable, yet filtration is a challenge to correlate to a broad spectrum according to what I have read.

So if the Rev D is indeed an LED, then yes it should be more stable, yet unfortunately the inter-device agreement with other spectros perhaps even more out of tolerance. Brings up the question is which one is the right one>
Logged

Czornyj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1948
    • zarzadzaniebarwa.pl
getting a better manually read profile
« Reply #12 on: February 02, 2009, 06:36:19 am »

Quote from: neil snape
I've only had the A and B revs. I think the rev D was supposed to be a white LED. I know it is in the more recent HP Z printers, and some other X-Rite devices. The IsIs is a little different as it has a UV emitting LED as well to make for dual readings.
The LEDs are narrow band, and supposedly very stable, yet filtration is a challenge to correlate to a broad spectrum according to what I have read.

So if the Rev D is indeed an LED, then yes it should be more stable, yet unfortunately the inter-device agreement with other spectros perhaps even more out of tolerance. Brings up the question is which one is the right one>

Very interesting!

I have i1 rev. D - any idea of how can I check what's inside? Is it safe to unscrew the housing and watch what's there, or may it cause some misalignment of the optical layout of the device?
Logged
Marcin Kałuża | [URL=http://zarzadzaniebarwa

neil snape

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1447
    • http://www.neilsnape.com
getting a better manually read profile
« Reply #13 on: February 02, 2009, 07:00:55 am »

Quote from: Czornyj
Very interesting!

I have i1 rev. D - any idea of how can I check what's inside? Is it safe to unscrew the housing and watch what's there, or may it cause some misalignment of the optical layout of the device?
No don't attempt any changes as the alignment of the optics is at stake.

Just google Rev D and LED to see if they are what we think or not. The light is much closer to D50 than the tungsten bulbs. IF the light looks like a mini HMI then it's LED.
Logged

Czornyj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1948
    • zarzadzaniebarwa.pl
getting a better manually read profile
« Reply #14 on: February 02, 2009, 09:12:39 am »

Quote from: neil snape
The light is much closer to D50 than the tungsten bulbs. IF the light looks like a mini HMI then it's LED.

Thanks - it was definetly a better, simplier and safer idea to observe the emitted light. Unfortunately, there's a tungsten bulb - it's to yellowish for a LED source.
Logged
Marcin Kałuża | [URL=http://zarzadzaniebarwa
Pages: [1]   Go Up