QTR-Quadtone RIP

Dear List,

as a newcomer to QuadToneRIP I am trying to create profiles for my 
printer-paper combination (Epson R2400-Hahnemuhle FineArt Pearl 285 
g/m^2), and I am afraid not to understand some basic concepts, like the 
<Black Boost> setting in the Curve Creator.

Preface.
As probably all of you experts I found out that for having the Curve 
Creator producing a valid linearization, the <Default Ink Limit> should 
be set at some 5% below the limit at which the grey steps become visibly 
undistinct. The linearization procedure usually involves the inversion 
of the grey level transfer matrix. Obviously the inversion is performed 
on the basis of the "literal" densitometer values without taking care 
for a possible non-regularity of the matrix. As soon as the ink steps 
become undistinct the matrix becomes irregular. Truncation on values 
below this point corresponds to a mathematical brute force 
regularization. The program code so given, this is the only possibility 
at the moment. In my printer-paper case I choose to truncate at a K-ink 
level of 45% leading to a luminosity L* of 17.6. L* then slowly taperes 
off to reach the lowest level of 5.5 at the 75% K-ink level. One might 
discuss shifting the <Default Ink Limit>, but that might be paid by an 
excessivly uncertain linearization. Thus, being forced to put the black 
end of the gray scale below "real" black, the <Black Boost> option is 
most valuable - but...

Problem.
Linearization with above parameter set leads to an acceptably smooth 
InkLevel-L*-function. When I set <Black Boost> to zero then a 
5%-increase in the InkLevel translates to a consistent L*-difference of 
around 4.1. However, if I go through the linearization with <Black 
Boost> set at 75% the L*-difference between 5% patches increases to 4.5, 
with the last step from 95% to 100% up at 6.5. Thus <Black Boost> steeps 
up the whole curve, so shifting the middle gray tones to darker values - 
not exactly what I intended. It seems to me, that in this way the <Black 
Boost>=75%-curve practically corresponds to a curve obtained with a 
<Default Ink Limit> of 75% and <Black Boost> of zero. Up to now it was 
my understanding that boosing the blacks is intended to increase dMax 
and should limit is action to the dark black image tones, but leaving 
the middle grays unaffected.

Is my understanding wrong? Is somebody suggesting the alternative 
approach of linearizing with <Black Boost> set to zero and only after 
linearizing setting the <Black Boost>-level to a higher value, for 
example 75% ? What do the long time experts think?

I am most grateful to learn from you!


Franco Laeri

franco.laeri@...

--- In QuadtoneRIP@yahoogroups.com, Franco Laeri <franco.laeri@...> wrote:
> 
> as a newcomer to QuadToneRIP I am trying to create profiles for my 
> printer-paper combination (Epson R2400-Hahnemuhle FineArt Pearl 285 
> g/m^2), and I am afraid not to understand some basic concepts, like the 
> <Black Boost> setting in the Curve Creator.
>  In my printer-paper case I choose to truncate at a K-ink 
> level of 45% leading to a luminosity L* of 17.6. L* then slowly taperes 
> off to reach the lowest level of 5.5 at the 75% K-ink level. 

If your 17.6 L* to "5.5" L* level is correct, it doesn't seem a very
gradual taper. It represents going from an approximate density of
d'log 1.62 to one of d'log 2.22. That's considerable.

> Up to now it was 
> my understanding that boosing the blacks is intended to increase dMax 
> and should limit is action to the dark black image tones, but leaving 
> the middle grays unaffected.

That's my understanding also.

> 
> Is my understanding wrong? Is somebody suggesting the alternative 
> approach of linearizing with <Black Boost> set to zero and only after 
> linearizing setting the <Black Boost>-level to a higher value, for 
> example 75% ? What do the long time experts think?

I'm no expert Franco but it seems to me that you would want to use or
not use the black boost from the very beginning step in curve
building. Once you have your maximum useable d'max as set by limit
plus any boost you might choose, linearizing should space out the
steps appropriately. I don't know if the black boost function was
intended as a means of increasing ink lay down at the dense end for
inks and papers requiring extra large amounts of black or if it is
only really intended as a control for slight variation to gain best
d'max after determining the coarse limit setting or perhaps some other
purpose entirely. Roy might have some input on that.

In my case,I prefer to begin with 100% limit and zero boost and only
use the boost if it can contribute something to the d'max after I have
coarse set the limit and before linearizing. That may be all wrong
from other's perspective but it's what I currently do. I'd be
interested in other's thoughts as well.

Regards
Duane

> I am most grateful to learn from you!
> 
> 
> Franco Laeri
> 
> franco.laeri@...
>

Dear Duane,

thank you for passing your advice that I appreciate very much!
Well, the taper from L* 17.6 to 5.5 refers to black levels between 45% 
and 75 %... Let me contemplate the dilemma from an other perspective:

When printing an image I am primarely interested to obtain a print that 
corresponds to the monitor image as close as possible. As the nature of 
the monitor image (self-luminous) is different from the paper print 
(reflecting ambient light) and the response of the human visual system 
is different for various light levels, one has to carefully consider the 
conditions under which the comparison should be performed. Condition 1: 
Matching the color temperature; condition 2: Matching the light levels - 
I try to match the light levels that pure white produces on the retina. 
Given similar color temperatures - they can be approximately evaluated 
with a spectrophotometer -  usual values for a monitor are around 120 
cd/m^2. That is not particularly bright as anyone who has tried to work 
with the laptop in the garden experienced... So, for a meaningful 
comparison the print should be inspected at a correspondigly low light 
level. The problem then is that at this light level the range of black 
levels that the human eye is able to discern is limited. Depending of 
the level of eye-internal scattering (-> age of the person) depper 
blacks than d'log around 2 are hardly resolved. As a consequence I tend 
to regard a grey level distribution in the print as ideal that matches 
the one I observe at the monitor. I obtained this with Curve Creator 
settings <ink level> 45% and <Black Boost> zero and corresponding 
linearization: QuadTone RIP does a perfect job! However, when such 
prints are inspected at higher light levels - in an extreme case in 
bright sunlight - one recognizes the lack of depth in the deep blacks. 
Although the middle gray distribution is still acceptable, I then would 
like to see deeper blacks. It is for this case that I hoped that setting 
<Black Boost> would help.
However, initiating the linearization with <Blach Boost> at 75% shifts 
the middle grays to dark values. Values that might be acceptable under 
bright light illumination, but when inspected at 120 cd/m^2 do not 
correspond to the gray level distribution shown by the monitor. So my 
question is, did you or somebody found a way out of this dilemma.

Many thanks for comments; regards
Franco


dlruckus schrieb:

> --- In QuadtoneRIP@yahoogroups.com, Franco Laeri <franco.laeri@...> wrote:
>> as a newcomer to QuadToneRIP I am trying to create profiles for my 
>> printer-paper combination (Epson R2400-Hahnemuhle FineArt Pearl 285 
>> g/m^2), and I am afraid not to understand some basic concepts, like the 
>> <Black Boost> setting in the Curve Creator.
>>  In my printer-paper case I choose to truncate at a K-ink 
>> level of 45% leading to a luminosity L* of 17.6. L* then slowly taperes 
>> off to reach the lowest level of 5.5 at the 75% K-ink level. 
> 
> If your 17.6 L* to "5.5" L* level is correct, it doesn't seem a very
> gradual taper. It represents going from an approximate density of
> d'log 1.62 to one of d'log 2.22. That's considerable.
> 
>> Up to now it was 
>> my understanding that boosing the blacks is intended to increase dMax 
>> and should limit is action to the dark black image tones, but leaving 
>> the middle grays unaffected.
> 
> That's my understanding also.
> 
>> Is my understanding wrong? Is somebody suggesting the alternative 
>> approach of linearizing with <Black Boost> set to zero and only after 
>> linearizing setting the <Black Boost>-level to a higher value, for 
>> example 75% ? What do the long time experts think?
> 
> I'm no expert Franco but it seems to me that you would want to use or
> not use the black boost from the very beginning step in curve
> building. Once you have your maximum useable d'max as set by limit
> plus any boost you might choose, linearizing should space out the
> steps appropriately. I don't know if the black boost function was
> intended as a means of increasing ink lay down at the dense end for
> inks and papers requiring extra large amounts of black or if it is
> only really intended as a control for slight variation to gain best
> d'max after determining the coarse limit setting or perhaps some other
> purpose entirely. Roy might have some input on that.
> 
> In my case,I prefer to begin with 100% limit and zero boost and only
> use the boost if it can contribute something to the d'max after I have
> coarse set the limit and before linearizing. That may be all wrong
> from other's perspective but it's what I currently do. I'd be
> interested in other's thoughts as well.
> 
> Regards
> Duane
>  
>  
>> I am most grateful to learn from you!
>>
>>
>> Franco Laeri
>>
>> franco.laeri@...
>>
> 
> 
> 
> 
>  
> Yahoo! Groups Links
> 
> 
> 
>

Hello again Franco.
I see what you are getting at in your question but I'm not sure that
many of us approach things in quite that way. We all want the maximum
darkest tone we can achieve for black in our prints irrespective of
what the viewing conditions will be. From that point on it becomes
necessary to actually alter the tonal relationships to suit any given
viewing environment if one is trying to maintain a specific "look" for
an image over the range of conditions you describe. I don't believe
that it is possible to overcome the physical realities that prevent
making a sort of universal print for all viewing levels.

In regard to the monitor image, it's my understanding that it is, and
should be, calibrated to a specific environment itself and it's view
won't be valid outside of that criteria either.

A gentleman, C D Tobie, who posts on this and the black and white
print group fairly often, most probably could address your question
with a much more technical explanation than I.

Regards
Duane
 



--- In QuadtoneRIP@yahoogroups.com, Franco Laeri <franco.laeri@...> wrote:
>
> Dear Duane,
> 
> thank you for passing your advice that I appreciate very much!
> Well, the taper from L* 17.6 to 5.5 refers to black levels between 45% 
> and 75 %... Let me contemplate the dilemma from an other perspective:
> 
> When printing an image I am primarely interested to obtain a print that 
> corresponds to the monitor image as close as possible. As the nature of 
> the monitor image (self-luminous) is different from the paper print 
> (reflecting ambient light) and the response of the human visual system 
> is different for various light levels, one has to carefully consider
the 
> conditions under which the comparison should be performed. Condition 1: 
> Matching the color temperature; condition 2: Matching the light
levels - 
> I try to match the light levels that pure white produces on the retina. 
> Given similar color temperatures - they can be approximately evaluated 
> with a spectrophotometer -  usual values for a monitor are around 120 
> cd/m^2. That is not particularly bright as anyone who has tried to work 
> with the laptop in the garden experienced... So, for a meaningful 
> comparison the print should be inspected at a correspondigly low light 
> level. The problem then is that at this light level the range of black 
> levels that the human eye is able to discern is limited. Depending of 
> the level of eye-internal scattering (-> age of the person) depper 
> blacks than d'log around 2 are hardly resolved. As a consequence I tend 
> to regard a grey level distribution in the print as ideal that matches 
> the one I observe at the monitor. I obtained this with Curve Creator 
> settings <ink level> 45% and <Black Boost> zero and corresponding 
> linearization: QuadTone RIP does a perfect job! However, when such 
> prints are inspected at higher light levels - in an extreme case in 
> bright sunlight - one recognizes the lack of depth in the deep blacks. 
> Although the middle gray distribution is still acceptable, I then would 
> like to see deeper blacks. It is for this case that I hoped that
setting 

> <Black Boost> would help.
> However, initiating the linearization with <Blach Boost> at 75% shifts 
> the middle grays to dark values. Values that might be acceptable under 
> bright light illumination, but when inspected at 120 cd/m^2 do not 
> correspond to the gray level distribution shown by the monitor. So my 
> question is, did you or somebody found a way out of this dilemma.
> 
> Many thanks for comments; regards
> Franco
> 
>

Hi Franco,

As you've found experimentation is usually necessary -- it's unpredictable when
you get high ink amounts.

But I'd like to clarify the black boost parameter.  It's basically an
alternate ink limit
i.e. its the same units.  So it really only makes sense to have a
BOOST_K  >=  LIMIT_K  and I usually don't do much more than about 15%
higher than LIMIT_K.

Roy

On 10/22/07, Franco Laeri <franco.laeri@...> wrote:
> Dear List,
>
> as a newcomer to QuadToneRIP I am trying to create profiles for my
> printer-paper combination (Epson R2400-Hahnemuhle FineArt Pearl 285
> g/m^2), and I am afraid not to understand some basic concepts, like the
> <Black Boost> setting in the Curve Creator.
>
> Preface.
> As probably all of you experts I found out that for having the Curve
> Creator producing a valid linearization, the <Default Ink Limit> should
> be set at some 5% below the limit at which the grey steps become visibly
> undistinct. The linearization procedure usually involves the inversion
> of the grey level transfer matrix. Obviously the inversion is performed
> on the basis of the "literal" densitometer values without taking care
> for a possible non-regularity of the matrix. As soon as the ink steps
> become undistinct the matrix becomes irregular. Truncation on values
> below this point corresponds to a mathematical brute force
> regularization. The program code so given, this is the only possibility
> at the moment. In my printer-paper case I choose to truncate at a K-ink
> level of 45% leading to a luminosity L* of 17.6. L* then slowly taperes
> off to reach the lowest level of 5.5 at the 75% K-ink level. One might
> discuss shifting the <Default Ink Limit>, but that might be paid by an
> excessivly uncertain linearization. Thus, being forced to put the black
> end of the gray scale below "real" black, the <Black Boost> option is
> most valuable - but...
>
> Problem.
> Linearization with above parameter set leads to an acceptably smooth
> InkLevel-L*-function. When I set <Black Boost> to zero then a
> 5%-increase in the InkLevel translates to a consistent L*-difference of
> around 4.1. However, if I go through the linearization with <Black
> Boost> set at 75% the L*-difference between 5% patches increases to 4.5,
> with the last step from 95% to 100% up at 6.5. Thus <Black Boost> steeps
> up the whole curve, so shifting the middle gray tones to darker values -
> not exactly what I intended. It seems to me, that in this way the <Black
> Boost>=75%-curve practically corresponds to a curve obtained with a
> <Default Ink Limit> of 75% and <Black Boost> of zero. Up to now it was
> my understanding that boosing the blacks is intended to increase dMax
> and should limit is action to the dark black image tones, but leaving
> the middle grays unaffected.
>
> Is my understanding wrong? Is somebody suggesting the alternative
> approach of linearizing with <Black Boost> set to zero and only after
> linearizing setting the <Black Boost>-level to a higher value, for
> example 75% ? What do the long time experts think?
>
> I am most grateful to learn from you!
>
>
> Franco Laeri
>
> franco.laeri@...
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>