Datacolor User to User Support Group.

Hi,
I've been using the Spyder3 colorimeter to generate profiles for my 
Epson R1800 with a newly installed InkRepublic CIS with Image 
Specialists ink. I'm using the 729 patch target with extended grays. 
I would develop the profile, print two DataColor test prints (the 
colored one with the children and a B/W with the tower, boat and B/W 
strip at the bottom), and try to decide if the grays had a color 
cast. I would then try to "tweak" the profile to get to a neutral 
gray. After a while I had many test prints, many profiles and many 
opinions from different family members as to what was gray and what 
wasn't. Time for a different approach.

I went back to the original unadjusted profile and printed just the 
test strip across the top of the paper. I then measured the gray box 
with the colorimeter and converted the LAB values to RGB using this 
tool http://www.colorpro.com/info/tools/convert.htm. I got RGB values 
of 116.7, 107.9 and 108.4 compared to PhotoShop's measured 109.4 for 
all three. This confirmed the red cast that I saw earlier so I 
started to move the sliders eventually winding up at Red/Cyan -4, 
Mag/Green +4, Yel/Blue 0. This setting gave 109.64, 108.76, 109.32 
for the neutral gray box as measured by the colorimeter on a test 
print.  Pretty darn good in my opinion.

I then printed a test image that has a GM test box of 24 colors and 
shades of gray. Some colors were very close, others were not. That's 
when I found out that the LAB to RGB conversions vary depending on 
which converter I use and what setting. (Heavy Sigh)

So, Can I use the colorimeter to spaot measure and edit in this way? 
If yes, how should I convert the Colorimeter Lab output to RGB so I 
can decide how to adjust the sliders? Should I just try for a neutral 
gray and assume that the colors will be ok?

One other question. Is there any way to go directly to the profile 
editor after launching Spyder3 without rebuilding the profile?

Thank you very much,

Bob Schoner

..........Learn to speak Lab. It makes clear statements about neutrality
without an RGB conversion: zero values in a* and b* means neutral; positive
a is red, and so on around the color wheel.........interesting. Is there
something somewhere that tells me what the shift is with various A8 B8
results? (green, red blue etc)

Thank you

Gaetan Dery

On 5/11/08 7:31 AM, "CDTobie@..." <CDTobie@...> wrote:

>  
>  
> 
> 
> In a message dated 5/10/08 5:18:53 PM, rschoner@... writes:
> 
> 
> So, Can I use the colorimeter to spaot measure and edit in this way?
> If yes, how should I convert the Colorimeter Lab output to RGB so I
> can decide how to adjust the sliders?
> 
> Learn to speak Lab. It makes clear statements about neutrality without an RGB
> conversion: zero values in a* and b* means neutral; positive a is red, and so
> on around the color wheel.
> 
>  Should I just try for a neutral
> gray and assume that the colors will be ok?
> 
> When you are adjusting it is for the neutrals and near neutrals where such
> tiny variations are visible. So yes, just work on the neutrals. But the
> spectro and software have already tried to produce neutral for you already. If
> you have a printer with at least two gray inks, then printing neutral images
> will be much more precise than it will with printers that mix their grays from
> color inks.
> 
> One other question. Is there any way to go directly to the profile
> editor after launching Spyder3 without rebuilding the profile?
> 
> Spyder3Print does not edit profiles, so there is no way to "go directly
> there". What it does is adjust build parameters from measurement sets, so it
> requires the measurement set(s) to do this.
> 
> C. David Tobie
> WW Product Technology Manager
> Digital Imaging & Home Theater
> Datacolor Inc.
> CDTobie@...
> www.datacolor.com/spyder3
> 
> 
> 
> **************
> Wondering what's for Dinner Tonight? Get new twists on family favorites at AOL
> Food.
>       (http://food.aol.com/dinner-tonight?NCID=aolfod00030000000001)
>  
>

--- In colorvision_group@yahoogroups.com, Gaetan Dery <cgdery@...> wrote:
>
> 
> ..........Learn to speak Lab. It makes clear statements about neutrality
> without an RGB conversion: zero values in a* and b* means neutral; positive
> a is red, and so on around the color wheel.........interesting. Is there
> something somewhere that tells me what the shift is with various A8 B8
> results? (green, red blue etc)



Go to the files section of this user forum and look for "Lab Space Graphic". Its s nice LAB 
diagram.

<http://f1.grp.yahoofs.com/v1/MEInSF4hNobXX9OTjFL8yuf3vv7Draj9pb1QmirVKRRWXAkwn
UtjGi2Azue4eqAj2-SC_orJToED-NoozJD6Dw/Lab%20Space%20Graphic.png>

Hi,

Thanks for getting back.

My problem seems to be that I have difficulty translating the LAB 
values to adjustments for the RGB sliders. This gets compounded by 
changes in one color affect the other colors.  For example I profiled 
Epson Glossy paper with Image Specialists Ink. The original reading 
for the gray patch was  LAB 49.45, 4.10, -0.06 or RGB of 115.6. 
105.8, 108.28. SInce this had a pinkish cast I started reducing the 
cyan/red slider.  After several iterations and a complete re-start I 
wound up with slider adjustments of Cyan/Red +2, Magenta/Green +2, 
Yellow/Blue -4.  This gave measured values of LAB 47.93, 0.26, -0.07 
and RGB 104.7, 104.2 104.4 essentially neutral and confirmed by 
different prints; colors seem to be very close, if not right on.  At 
this point I am happy.

It may not be possible, but maybe the Spyder software could have LAB 
correction sliders as well as the conventional RGB sliders.

Thanks for your help.

Bob Schoner

--- In colorvision_group@yahoogroups.com, CDTobie@... wrote:
>
> 
> In a message dated 5/10/08 5:18:53 PM, rschoner@... writes:
> 
> 
> > So, Can I use the colorimeter to spaot measure and edit in this 
way?
> > If yes, how should I convert the Colorimeter Lab output to RGB so 
I
> > can decide how to adjust the sliders?
> > 
> Learn to speak Lab. It makes clear statements about neutrality 
without an RGB 
> conversion: zero values in a* and b* means neutral; positive a is 
red, and so 
> on around the color wheel.
> 
> >  Should I just try for a neutral
> > gray and assume that the colors will be ok?
> > 
> When you are adjusting it is for the neutrals and near neutrals 
where such 
> tiny variations are visible. So yes, just work on the neutrals. But 
the spectro 
> and software have already tried to produce neutral for you already. 
If you 
> have a printer with at least two gray inks, then printing neutral 
images will be 
> much more precise than it will with printers that mix their grays 
from color 
> inks.
> > 
> > One other question. Is there any way to go directly to the profile
> > editor after launching Spyder3 without rebuilding the profile?
> > 
> Spyder3Print does not edit profiles, so there is no way to "go 
directly 
> there". What it does is adjust build parameters from measurement 
sets, so it 

> requires the measurement set(s) to do this.
> 
> C. David Tobie
> WW Product Technology Manager
> Digital Imaging & Home Theater
> Datacolor Inc.
> CDTobie@...
> www.datacolor.com/spyder3
> 
> 
> 
> **************
> Wondering what's for Dinner Tonight? Get new twists on family 
> favorites at AOL Food.
>       
> (http://food.aol.com/dinner-tonight?NCID=aolfod00030000000001)
>

Hi,

Thanks again for the follow up.

First, the easy question: What Printer?  This is an Epson R1800 with 
an InkRepublic Continuous Ink System and Image Specialists inks. 
Although it has two black inks, Photo Black and Matte Black, I think 
it only uses one ink at a time depending on the paper/media setting. 
So, I guess that makes it a "one black" printer. It has Cyan, Yellow, 
Magenta, Red and Blue for the colors. No more Photo Cyan or Photo 
Magenta.

Now for the embarassing part. For some unknown reason the engineer in 
me decided to reproduce the process.  I went back to the original 729 
patch plus extended grays data set and printed the test chart with no 
correction. Measured gray was similar to the first (48.7, 3.24,-.01 
vs 49.45, 4.1 -.06) but when I applied the correction of +2, +2, -4 I 
got a gray reading of 48.53, 2.88, 1.87--pretty far from neutral 
gray.  So, I went through the process again ( 18 iterations) and 
wound up with corrections of -5, +4, -2 and a measured gray of 48.62, 
0.21, -.32. I have not yet printed full test imges with this new 
profile.

I can not figure out where I messed up. I thought I took meticulous 
notes on each change.  Maybe I mixed up the original data set ( I 
have data sets for Ilford Pearl as well as Epson Glossy); maybe I 
left out extended grays, I just don't know. But, the first profile 
that I can't reproduce, makes very nice prints. The second iteration 
seems more intuitively correct since the red is decreased as opposed 
to increased in the first iteration.

One observation from this process--the measurements change with time 
and location. The largest variations come just after the chart is 
printed; a drying effect I guess.  But, I do notice a change 
overnight. One example: 47.57, 4.39, 0.54 went to 48.77, 3.69, 1.60.  
Do you have a feel for what acceptable measurement variation is?  
Your suggestion to back up the measurement charts with extra paper is 
very good. I notice a variation if I do not back up the paper. I also 
notice some variation within the measurement patch, I guess that's 
the printer/ink.

I tried the PhotoShop ACV curve correction process but did not have 
good luck. I used the R channel since the image had a pinkish cast 
but the error was so slight a small adjustment fixes the SoftProof 
but the print is not correct. I tried assigning the printer profile 
to the test image and then "auto correct" with the curves layer but 
that's disastrous. I guess that's why I asked about a LAB correction 
option in Spyder.  BTW, I find the SpyderProof "exagerates" the 
errors/correction compared to PhotoShop or QImage soft Proof.

Sorry for the long post but, in spite of the effort involved, it is 
still fun.  Just wish I knew more.

Thanks,

Bob Schoner

--- In colorvision_group@yahoogroups.com, CDTobie@... wrote:
>
> 
> In a message dated 5/11/08 9:04:04 PM, rschoner@... writes:
> 
> 
> > My problem seems to be that I have difficulty translating the LAB
> > values to adjustments for the RGB sliders.
> > 
> There aren't really 'RGB' sliders, only color sliders (which, 
admittedly are 
> not labeled as 'positive a star' or the like <G>)
> 
> >  This gets compounded by
> > changes in one color affect the other colors.
> > 
> Thats the way life works... 
> 
> 
> 
> >   For example I profiled
> > Epson Glossy paper with Image Specialists Ink. The original 
reading
> > for the gray patch was  LAB 49.45, 4.10, -0.06 or RGB of 115.6.
> > 105.8, 108.28. SInce this had a pinkish cast I started reducing 
the
> > cyan/red slider.  After several iterations and a complete re-
start I
> > wound up with slider adjustments of Cyan/Red +2, Magenta/Green +2,
> > Yellow/Blue -4.  This gave measured values of LAB 47.93, 0.26, -
0.07
> > and RGB 104.7, 104.2 104.4 essentially neutral and confirmed by
> > different prints; colors seem to be very close, if not right on.  
At
> > this point I am happy.
> > 
> > It may not be possible, but maybe the Spyder software could have 
LAB
> > correction sliders as well as the conventional RGB sliders.
> > 
> Unfortunately people would take them too literally, and expect 
units of the 
> exact size of Lab value units. We can't know exactly how much 
change, in Lab 
> units, an adjustment will make, as that will vary with factors 
beyond our 
> control. You are the first to ask for this, what I hear more often 
is a request to 
> make our units line up exactly with Photoshop values... again not 
really 
> practical for similar reasons. So the recommendation is: if you 
want to use 
> Photoshop RGB units, or see your adjustments visually on screen, 
build a Photoshop RGB 
> curve set and once that is doing what you want, import it into 
Spyder3Print 
> and apply to your profile.
> 
> You never mention what printer you are using, but its unusual with 
a 'two 
> grays' printer to have results after profiling that are as far off 
neutral as 4 

> a* units. Is this by any chance a one gray, or no grays, printer?
> 
> C. David Tobie
> WW Product Technology Manager
> Digital Imaging & Home Theater
> Datacolor
> CDTobie@...
> www.datacolor.com/Spyder3
> 
> 
> 
> 
> 
> **************
> Wondering what's for Dinner Tonight? Get new twists on family 
> favorites at AOL Food.
>       
> (http://food.aol.com/dinner-tonight?NCID=aolfod00030000000001)
>

Hi,

I bought the R1800 because I do mostly, if not all, color.  I went 
after the b&W as part of the profiling process because it was 
something I could measure easily and from a consistent patch on a 
test print.  I assumed that making the B&W good it would improve the 
shadow and light colors. This may not be a good assumption.  BTW,I 
think the B&W is quite acceptable at this point.

If you don't mind I think I will post my experience in the DPReview 
forum.  I think it is a positive experience.  Who knows, I may be 
able to tweak David Miller and Andrew Rodney again.
:-)


Thanks again,

Bob Schoner

--- In colorvision_group@yahoogroups.com, CDTobie@... wrote:
>
> 
> In a message dated 5/12/08 2:18:00 PM, rschoner@... writes:
> 
> 
> > Thanks again for the follow up.
> > 
> > First, the easy question: What Printer?  This is an Epson R1800 
with
> > an InkRepublic Continuous Ink System and Image Specialists inks.
> > Although it has two black inks, Photo Black and Matte Black, I 
think
> > it only uses one ink at a time depending on the paper/media 
setting.
> > So, I guess that makes it a "one black" printer. It has Cyan, 
Yellow,
> > Magenta, Red and Blue for the colors. No more Photo Cyan or Photo
> > Magenta.
> > 
> Okay, the R1800 is a great color printer, but absolutely not a 
black and 
> white printer. It has no gray inks at all, and all of its grays are 
generated from 
> color inks... and not evey light tone inks, full density color 
inks. That 
> would explain your issues...
> > 
> > 
> > 
> > One observation from this process--the measurements change with 
time
> > and location. The largest variations come just after the chart is
> > printed; a drying effect I guess.
> > 
> Yes, there are drying effects, but even with great patience and 
great care, 
> the R1800 is just not going to print great black and white images; 
any change 
> in the light source will cause significant changes is the percieved 
tone of 
> grays formed from color inks; so your B&W images will be green in 
one light and 
> pink in another.
> 
> >   But, I do notice a change
> > overnight. One example: 47.57, 4.39, 0.54 went to 48.77, 3.69, 
1.60. 
> > Do you have a feel for what acceptable measurement variation is? 
> > 
> While its still changing whole numbers, its not dry yet; but as I 
note above, 
> the R1800 is not going to produce gallery quality B&W no matter how 
exacting 
> you are.
> 
> > Your suggestion to back up the measurement charts with extra 
paper is
> > very good. I notice a variation if I do not back up the paper. I 
also
> > notice some variation within the measurement patch, I guess that's
> > the printer/ink.
> > 
> > I tried the PhotoShop ACV curve correction process but did not 
have
> > good luck. I used the R channel since the image had a pinkish cast
> > but the error was so slight a small adjustment fixes the SoftProof
> > but the print is not correct. I tried assigning the printer 
profile
> > to the test image and then "auto correct" with the curves layer 
but
> > that's disastrous. I guess that's why I asked about a LAB 
correction
> > option in Spyder.  BTW, I find the SpyderProof "exagerates" the
> > errors/correction compared to PhotoShop or QImage soft Proof.
> > 
> Probably because it factors in ink black and paper white. If you 
check those 
> boxes in Photoshop, results should be equivalent.
> > 
> > Sorry for the long post but, in spite of the effort involved, it 
is
> > still fun.  Just wish I knew more.
> > 
> > Well, you may know something now that you didn't previously. An 
R2400 would 
> produce gallery quality, visually stable B&W prints (as would a 
3800, a Canon 
> iPF5000 or a Pro9500) but the R1800 is aimed very much in the other 
direction: 
> glossy color prints. Which is not to say you can't improve your B&W 
results, 
> just that they won't ever be stable enough to do gallery quality 
B&W.

> 
> Sorry to be the bearer of bad news...
> 
> C. David Tobie
> WW Product Technology Manager
> Digital Imaging & Home Theater
> Datacolor
> CDTobie@...
> www.datacolor.com/Spyder3
> 
> 
> 
> 
> 
> **************
> Wondering what's for Dinner Tonight? Get new twists on family 
> favorites at AOL Food.
>       
> (http://food.aol.com/dinner-tonight?NCID=aolfod00030000000001)
>