QTR-Quadtone RIP

Thanks for answering,
Joost
I�m still having
some problems with the User�s Guide. I even consulted some papers from
the competition trying to understand (to me) oscure concepts :-).
 
As you hint,
there are some errors in my numbers. This could be of interest to novices like
me, so I post it:
 
The example
uses the EPSON Stylus Photo 2200, TWO gray inks
 
�For Ultrachrome inks, the
measurements above are all that is required. For
QuadTone inks, this process is
repeated for each lighter ink, comparing it to
the next darker ink,
calculating its density relative to that ink and then
converting it to a density
relative to black.�
 
For my R2880 I used K, LK and
LLK inks. 
 
My mistake: I calculated  LK relative to K and LLK relative to LK when I
should have done both LK and LLK relative to K. I got it (hopefully) right when
I read the paper �QuadToneRIP 2.5 Calibration�, with an example w/ 3 inks. Everything is there but I missed it.
 
OK now my curves look much
better.

 
BUT:  in the example (QuadToneRIP 2.5 Calibration) we get:

LK = 32.7% of K 
and 
LLK= 38% of LK 
We make an interpolation and
get LLK=12.4% of K
 
Why don�t we just look at the matching
patches in K and LLK directly (like we do w/ K and LK) instead of doing
first mathematics w/ LK and LLK? Is there a difference?

Regards



      

[Non-text portions of this message have been removed]

I was confused about this when I was first trying to do my own profiles,  
also. The reason you compare each ink to the next darker shade and then  
multiply to get the percentage relative to the darkest shade (K), is that  
with more ink shades the lightest ones fall into the last few percent of  
the K ink where it would be extremely hard to make comparison judgments  
visually or even through measuring devices. With two or three inks it's  
not so difficult, but when you get four, six, or seven inks, the lightest  
shades are so light that you can hardly see them. To try to compare them  
directly with the darkest ink would mean that the darkest ink would not be  
covering the paper. You would have ink dots with lots of white space  
around them to get an apparent density equivalent. That is the way "black  
only" printing works. You get the lightest shades by printing very sparse  
dark black ink dots. It makes for a grainy appearance in the lighter tones  
of the picture. Some people like the look, others opt for more shades of  
inks so that at the lighter picture tones, you are still covering the  
paper and this results in a much smoother tonality. The printer driver  
used, such as QTR, still has to figure out how much of each ink shade to  
use at any given picture tone. The algorithms use each ink's density  
relative to the darkest ink to calculate that. Each printer/paper/ink  
combination reacts differently so you have to develop separate curves for  
each combination. The process starts with you inputting the maximum amount  
of each shade of ink to be used and the relative density of each ink. QTR  
then calculates the amounts of each ink to use at each of the possible 256  
picture values. Because of all the variables involved, this result is  
hardly ever perfect. That is why you then have to print out a step wedge,  
measure the results, and then feed the data back into the program where it  
goes through another process to linearize the output. The internals of how  
it does all this are unknown to me. Jon Cone of InkJetMall has his own  
proprietary way of creating these curves which, he says, results in even  
smoother transitions. I don't know how his process works either.  This  
turn out to be a longer post than I had intended. I hope it might be of  
some help for folks trying to understand what is going on. I am also still  
in the process of trying to learn about it all.

Steve

On Sun, 01 Feb 2009 10:32:29 -0800, jorge caruso <rabsanito@...>  
wrote:

> Why don\u2019t we just look at the matching
> patches in K and LLK directly (like we do w/ K and LK) instead of doing
> first mathematics w/ LK and LLK? Is there a difference?



-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

Hi Steve

thanks very much for your posting, i found it VERY helpful and am glad that it was "long", as you said :-)

>This  turn out to be a longer post than I had intended.


Jorge




________________________________

From: Steve and Ann Taylor <taylorfamily3688@...>
To: QuadtoneRIP@yahoogroups.com
Sent: Monday, February 2, 2009 1:22:44 AM
Subject: Re: MAYBE of interest to other novices was  Re: [QuadtoneRIP] Linearization Fails


I was confused about this when I was first trying to do my own profiles, 
also. The reason you compare each ink to the next darker shade and then 
multiply to get the percentage relative to the darkest shade (K), is that 
with more ink shades the lightest ones fall into the last few percent of 
the K ink where it would be extremely hard to make comparison judgments 
visually or even through measuring devices. With two or three inks it's 
not so difficult, but when you get four, six, or seven inks, the lightest 
shades are so light that you can hardly see them. To try to compare them 
directly with the darkest ink would mean that the darkest ink would not be 
covering the paper. You would have ink dots with lots of white space 
around them to get an apparent density equivalent. That is the way "black 
only" printing works. You get the lightest shades by printing very sparse 
dark black ink dots. It makes for a grainy appearance in the lighter tones 
of the picture. Some people like the look, others opt for more shades of 
inks so that at the lighter picture tones, you are still covering the 
paper and this results in a much smoother tonality. The printer driver 
used, such as QTR, still has to figure out how much of each ink shade to 
use at any given picture tone. The algorithms use each ink's density 
relative to the darkest ink to calculate that. Each printer/paper/ ink 
combination reacts differently so you have to develop separate curves for 
each combination. The process starts with you inputting the maximum amount 
of each shade of ink to be used and the relative density of each ink. QTR 
then calculates the amounts of each ink to use at each of the possible 256 
picture values. Because of all the variables involved, this result is 
hardly ever perfect. That is why you then have to print out a step wedge, 
measure the results, and then feed the data back into the program where it 
goes through another process to linearize the output. The internals of how 
it does all this are unknown to me. Jon Cone of InkJetMall has his own 
proprietary way of creating these curves which, he says, results in even 
smoother transitions. I don't know how his process works either.  This 
turn out to be a longer post than I had intended. I hope it might be of 
some help for folks trying to understand what is going on. I am also still 
in the process of trying to learn about it all.

Steve

On Sun, 01 Feb 2009 10:32:29 -0800, jorge caruso <rabsanito@yahoo. com> 
wrote:

> Why don�t we just look at the matching
> patches in K and LLK directly (like we do w/ K and LK) instead of doing
> first mathematics w/ LK and LLK? Is there a difference?

-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera. com/m2/
    


      

[Non-text portions of this message have been removed]