Pages: [1]   Go Down

Author Topic: Repeatability of results with i1 vs io  (Read 6786 times)

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Repeatability of results with i1 vs io
« on: August 21, 2008, 04:29:31 am »

As I scan targets manually with my i1, there is some variation if I scan the same target a second time. Is this normal with the i1, and will the same thing happen if I use an io to do the scanning?

I have noticed when scanning that Measure Tool will occasionally allow some pretty weird colours through and have been rescanning whenever I see it but a few still get through.

Cheers,

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

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2115
    • Web
Repeatability of results with i1 vs io
« Reply #1 on: August 21, 2008, 08:53:56 am »

Try running Eye-One Diagnostics to make sure your instrument is ok.

You will always get some variation due to measurement noise/error. The only question is whether the differences are sufficiently small.

Are you scanning row by row using the ruler? Are you scanning too quickly or too slowly?
Logged
Eric Chan

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Repeatability of results with i1 vs io
« Reply #2 on: August 21, 2008, 09:46:22 am »

As Eric said, there will always be some variations reported in MeasureTool but they should be well below a deltaE 2000 of 1.

Also make sure you have several sheets of the same paper behind the target!
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Repeatability of results with i1 vs io
« Reply #3 on: August 22, 2008, 12:24:34 am »

Andrew and Eric, many thanks for the replies. I have run the diags and it is fine. Most differences are below 1. What got me going was a set that had a few above 1. It was then that I started wondering why I would get such a result.

Funnily enough, the last scan that I did was with the three pages from the 1728 Atkinson target sitting on the table. It was the cleanest scan that I have got.

I would love to get some guidance on whether folks average, and, if so, how many passes is a good starting point.

Thanks for humouring a newbie. This is quite a learning curve.

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

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Repeatability of results with i1 vs io
« Reply #4 on: August 22, 2008, 09:17:52 am »

The i1 should measure about 100 patches per second in scan mode. So, keep that in mind when scanning. Not too slow, certainly not too fast.

To give you an idea of another instrument, here's a 800 patch target run through my iSis three times and reported from ColorThink:

Here's a report of the iStar run through the iSis three times:

dE Report

Number of Samples: 832

Delta-E Formula dE2000

Overall - (832 colors)
--------------------------------------------------
Average dE:   0.05
   Max dE:   0.33
   Min dE:   0.00
StdDev dE:   0.03

Best 90% - (748 colors)
--------------------------------------------------
Average dE:   0.04
   Max dE:   0.09
   Min dE:   0.00
StdDev dE:   0.02

Worst 10% - (84 colors)
--------------------------------------------------
Average dE:   0.12
   Max dE:   0.33
   Min dE:   0.09
StdDev dE:   0.04
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Repeatability of results with i1 vs io
« Reply #5 on: August 24, 2008, 08:06:02 am »

Quote
The i1 should measure about 100 patches per second in scan mode. So, keep that in mind when scanning. Not too slow, certainly not too fast.

To give you an idea of another instrument, here's a 800 patch target run through my iSis three times and reported from ColorThink:

Here's a report of the iStar run through the iSis three times:

dE Report

Number of Samples: 832

Delta-E Formula dE2000

Overall - (832 colors)
--------------------------------------------------
Average dE:   0.05
   Max dE:   0.33
   Min dE:   0.00
StdDev dE:   0.03

Best 90% - (748 colors)
--------------------------------------------------
Average dE:   0.04
   Max dE:   0.09
   Min dE:   0.00
StdDev dE:   0.02

Worst 10% - (84 colors)
--------------------------------------------------
Average dE:   0.12
   Max dE:   0.33
   Min dE:   0.09
StdDev dE:   0.04
[a href=\"index.php?act=findpost&pid=216645\"][{POST_SNAPBACK}][/a]

Thanks Andrew. I'm now a proud owner of a copy of Color Think Pro and have even reported my first bug. I get the following from two scans of the Atkinson 1728 target. I did three but one gives  me a delta over 1 so I have used the closest 2.


dE Report

Number of Samples: 1728

Delta-E Formula dE94

Overall - (1728 colors)
--------------------------------------------------
Average dE:   0.11
    Max dE:   0.68
    Min dE:   0.00
 StdDev dE:   0.06

Best 90% - (1554 colors)
--------------------------------------------------
Average dE:   0.10
    Max dE:   0.18
    Min dE:   0.00
 StdDev dE:   0.04

Worst 10% - (174 colors)
--------------------------------------------------
Average dE:   0.24
    Max dE:   0.68
    Min dE:   0.18
 StdDev dE:   0.07

--------------------------------------------------
My numbers appear to be about twice yours. Any idea whether I need more practice, or if this is what I should expect? What I have noticed is that the first 2 rows are the hardest to scan. Both take a lot of goes while most others are easy.
« Last Edit: August 24, 2008, 08:06:46 am by Jeff-Grant »
Logged
Cheers,
 Jeff  www.jeff-grant.com

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Repeatability of results with i1 vs io
« Reply #6 on: August 24, 2008, 09:55:50 am »

Quote
My numbers appear to be about twice yours.


Your numbers seem fine. you'll NEVER got an EyeOne Pro to behave like an iSis (it's a $5K device).
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Repeatability of results with i1 vs io
« Reply #7 on: August 24, 2008, 06:00:12 pm »

Quote
Your numbers seem fine. you'll NEVER got an EyeOne Pro to behave like an iSis (it's a $5K device).
[a href=\"index.php?act=findpost&pid=216940\"][{POST_SNAPBACK}][/a]
Thanks Andrew. That's what I needed to know. I still don't know what is going on with the Atkinson 1728 patch target. I have printed the 3 page version and can scan it easily with the exception of the first two rows. I can scan the Xrite 918 patch without problems. Perhaps I should just go off and practise a bit more but it's only those two rows that present a problem.
« Last Edit: August 25, 2008, 08:51:31 am by Jeff-Grant »
Logged
Cheers,
 Jeff  www.jeff-grant.com

Jeff-Grant

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 568
    • http://www.jeff-grant.com
Repeatability of results with i1 vs io
« Reply #8 on: August 26, 2008, 10:08:20 pm »

Quote
Thanks Andrew. That's what I needed to know. I still don't know what is going on with the Atkinson 1728 patch target. I have printed the 3 page version and can scan it easily with the exception of the first two rows. I can scan the Xrite 918 patch without problems. Perhaps I should just go off and practise a bit more but it's only those two rows that present a problem.
[a href=\"index.php?act=findpost&pid=217009\"][{POST_SNAPBACK}][/a]
I'm responding to my own post to close the loop, hopefully. Having spent a fair chunk of time fruitlessly scanning the 1728 patch target and always having trouble with the first two rows, I came to the conclusion that the gap from paper white to paper white (1st patch) was too thin for the i1. I carefully thickened the line with a ruler and fine point pen and can now scan these rows without problems.

Is there a flaw in this approach. I'm about to build a profile from my latest measurements. Fingers crossed.
Logged
Cheers,
 Jeff  www.jeff-grant.com
Pages: [1]   Go Up