Digital BW, The Print

I've review the all group messages without finding the reply.
Given a printing resolution, a fixed dot size (to make it simple), I 
guess the only way ink is limited is at the dithering level (e.g. less 
dots printed). Is that correct ?

Olivier

Olivier wrote:
> I've review the all group messages without finding the reply.
> Given a printing resolution, a fixed dot size (to make it simple), I 
> guess the only way ink is limited is at the dithering level (e.g. less 
> dots printed). Is that correct ?
> 
> Olivier

Yes, depends where it falls. Could be less dots, distribution 
of the dots, smaller droplet size, a lighter ink in the 
inkset. In that order I think but there are many different 
media around that will not all cope with only dot number 
changes (and that's limited too if you want to keep some 
smoothness). In RIPs you can change the point where darker 
inks are replaced by lighter inks and Epson must have that 
choice too when making the paper settings for the driver.

Ernst

-- 

                    --
           Ernst Dinkla


www.pigment-print.com
(         unvollendet         )

> 
> Yes, depends where it falls. Could be less dots, distribution 
> of the dots, smaller droplet size, a lighter ink in the 
> inkset. In that order I think but there are many different 
> media around that will not all cope with only dot number 
> changes (and that's limited too if you want to keep some 
> smoothness). In RIPs you can change the point where darker 
> inks are replaced by lighter inks and Epson must have that 
> choice too when making the paper settings for the driver.
> 
> Ernst
> 

Very clear, Ernst as usual. In fact, it's in the 95-100% range I was 
wondering relating it to Dmax. 

Again at a given dot size (forgetting VDS here), ink set... I'm 
assuming dot gain will be the ultimate factor in ink limtiation, 
meaning on one hand the paper on the other hand the dpi resolution. 

Say paper is given again, I'm left with dpi and a direct relation to 
larger dpi-smaller fixed dots (at 2880). I've seen you repeatedly 
mentionning smaller dots lead to higher relative dot gain (unless 
wrong) : so where do we stand in terms of reaching higher Dmax. We're 
better off with small drop or larger ones ? 

View like this I'd go for higher dpi and smaller dots. Basically, I'm 
still stuck with what best dpi reso to use despite hundreds of posts 
on the issue and Roy's mentionning 1440 is very fine vs 2880 makes me 
wonder, because the 95-100% will probably use larger dots and put 
together with your saying smaller increase dot gain and density make 
me thing Dmax can/should be increased with smaller dots.

I'm sure I've got something wrong here...

Olivier

Olivier wrote:
>> Yes, depends where it falls. Could be less dots, distribution 
>> of the dots, smaller droplet size, a lighter ink in the 
>> inkset. In that order I think but there are many different 
>> media around that will not all cope with only dot number 
>> changes (and that's limited too if you want to keep some 
>> smoothness). In RIPs you can change the point where darker 
>> inks are replaced by lighter inks and Epson must have that 
>> choice too when making the paper settings for the driver.
>>

> In fact, it's in the 95-100% range I was 
> wondering relating it to Dmax. 
> 
> Again at a given dot size (forgetting VDS here), ink set... I'm 
> assuming dot gain will be the ultimate factor in ink limtiation, 
> meaning on one hand the paper on the other hand the dpi resolution. 

I have made a step wedge for QTR with some checks on bleeding.
Usually on good papers the Dmax stops increasing before the
bleeding gets ugly. Looks like the chemists have an idea how
to keep the print characteristics of the ink under control.
> 
> Say paper is given again, I'm left with dpi and a direct relation to 
> larger dpi-smaller fixed dots (at 2880). I've seen you repeatedly 
> mentionning smaller dots lead to higher relative dot gain (unless 
> wrong) : so where do we stand in terms of reaching higher Dmax. We're 
> better off with small drop or larger ones ? 

You can bet that Epson engineers fix the relation between the
different dotsizes on an average dotgain. So not a 1:2:4
relation on dot area but a 1:2:4 relation on actual density
measured in tests on several papers. You run into a different
problem with smaller dots, the placing of the smaller dots
gets far more irregular than they are with larger dots. The
differences between smaller droplet contents is higher than
between larger droplets contents. The pro printers almost
always had larger droplets than the desktop models and Epson
mentioned that consistency of droplet size was more important.
So you either choose for a more precise dot placement but
slightly rougher screen or you go for a finer dot with more
fluctuation in the density and in detail. The Dmax will not be
better at 100% when there's no white left. The control of the
95-100% range isn't better with smaller dots either.

There's a good piece of PR on the new Canon FINE heads here:

http://consumer.usa.canon.com/pixma/press/FINE_TG_0909.pdf

Propaganda but I think they mention some important issues.

> View like this I'd go for higher dpi and smaller dots. Basically, I'm 
> still stuck with what best dpi reso to use despite hundreds of posts 
> on the issue and Roy's mentionning 1440 is very fine vs 2880 makes me 
> wonder, because the 95-100% will probably use larger dots and put 
> together with your saying smaller increase dot gain and density make 
> me thing Dmax can/should be increased with smaller dots.


At the highest resolution the number of droplet sizes get
less. The evolution we have seen  with increased resolution is
a smaller minimum droplet size, more droplet sizes, more grey
inks. For a good reason as just making droplets smaller isn't
working.

> I'm sure I've got something wrong here...

I'm sure it is more complicated than we both think it is.

Ernst
-- 

                    --
           Ernst Dinkla


www.pigment-print.com
(         unvollendet         )

> I have made a step wedge for QTR with some checks on bleeding.
> Usually on good papers the Dmax stops increasing before the
> bleeding gets ugly. Looks like the chemists have an idea how
> to keep the print characteristics of the ink under control.

Yes, I experienced +/- the same : except EEM (and I don't know why), 
all VFA, USFA, HPR and so-called fine art papers "hold" ink pretty 
well. In fact, I'd have expected a little bleeding in total black 
areas (99-100) would have provided total coverage while retaining 
Dmax, hence my question. 
 
> You can bet that Epson engineers fix the relation between the
> different dotsizes on an average dotgain. So not a 1:2:4
> relation on dot area but a 1:2:4 relation on actual density
> measured in tests on several papers. 

Which turns for us, users, uncontrollable. But that a fair technical 
proposal from Epson, since I'd assume the driver then turns to be a 
nice piece of complexity.


You run into a different
> problem with smaller dots, the placing of the smaller dots
> gets far more irregular than they are with larger dots. The
> differences between smaller droplet contents is higher than
> between larger droplets contents. 

Ok, you have a point. With the new x800 generation, do you feel 
(finger in the air) that 1)this physical precision has improved, 2) 
the new K3 resin-coated formulation plays a big and different role in 
dot gain, overprint effects...What makes me think this is that the 
OEM calibration Colorbase can only be used in 2880dpi whichever media 
type.

I'm sorry I can't get the translation of a "droplet content". I'm 
French.

The pro printers almost
> always had larger droplets than the desktop models and Epson
> mentioned that consistency of droplet size was more important.
> So you either choose for a more precise dot placement but
> slightly rougher screen or you go for a finer dot with more
> fluctuation in the density and in detail. The Dmax will not be
> better at 100% when there's no white left. The control of the
> 95-100% range isn't better with smaller dots either.

That partly replies the above, but you might refer to the x600 
printers (?).

> There's a good piece of PR on the new Canon FINE heads here:
> 
> http://consumer.usa.canon.com/pixma/press/FINE_TG_0909.pdf


I've read through it, but will properly study it cool headed.

 
> At the highest resolution the number of droplet sizes get
> less. The evolution we have seen  with increased resolution is
> a smaller minimum droplet size, more droplet sizes, more grey
> inks. For a good reason as just making droplets smaller isn't
> working.

I'd have thought that all those elements precisely call for smaller 
drops. Basically, if the purpose is smoothness, maximum shades 
reproduction (grey values), detailed prints and media coverage,... 
one has to combined a precisely built head (your link) to be capable 
of as small as possible dots (for smoothness and dotless prints) 
dithered by a first-class driver (for the best trade off between 
value/detail) with a variable density ink set (for maximum shades 
reproduction) with ink formulation that combines both right pigment 
load and eventually low dot gain.

I've seen K3 lacking a L-LLK, and somehow lacking (a bit) Dmax, so 
I'm at a stage where I wonder whether on principle highest dpi 
setting for smallest drop possible would eventually better the print 
at all levels without generating issues like dot registration, drop 
size... I also count on a reasonnably long overlap and some 
significant extra K blackboost to smooth the print yet give good 
Dmax. 

Does that all make sense or not worth trying ?

Olivier

Olivier wrote:

> You run into a different
>> problem with smaller dots, the placing of the smaller dots
>> gets far more irregular than they are with larger dots. The
>> differences between smaller droplet contents is higher than
>> between larger droplets contents. 
> 
> Ok, you have a point. With the new x800 generation, do you feel 
> (finger in the air) that 1)this physical precision has improved, 2) 
> the new K3 resin-coated formulation plays a big and different role in 
> dot gain, overprint effects...What makes me think this is that the 
> OEM calibration Colorbase can only be used in 2880dpi whichever media 
> type.
> 
> I'm sorry I can't get the translation of a "droplet content". I'm 
> French.

1/ another grey ink added does wonders, no I don't think that 
precision became much better, speed was more important, there 
must a trade off between the two  2/ Yes, I think the 9800 
range could be better for gloss printing in the first place 
and isn't as good for matte as the 9600 range was. At least we 
start to think that over here. The 9800 is also harder to 
adapt to third party inks. However with a  RIP it must be 
possible.

Pardon, I'm Dutch so it will not be proper English either. 
Combien d'encre par une goutellete (had to look up that word 
:-) 3.5 picoliter could be 3 or 4 picoliter.

> The pro printers almost
>> always had larger droplets than the desktop models and Epson
>> mentioned that consistency of droplet size was more important.
>> So you either choose for a more precise dot placement but
>> slightly rougher screen or you go for a finer dot with more
>> fluctuation in the density and in detail. The Dmax will not be
>> better at 100% when there's no white left. The control of the
>> 95-100% range isn't better with smaller dots either.
> 
> That partly replies the above, but you might refer to the x600 
> printers (?).

To all the pro printers in time compared to the 
consumer/desktop models of the same generation.

>> At the highest resolution the number of droplet sizes get
>> less. The evolution we have seen  with increased resolution is
>> a smaller minimum droplet size, more droplet sizes, more grey
>> inks. For a good reason as just making droplets smaller isn't
>> working.
> 
> I'd have thought that all those elements precisely call for smaller 
> drops. Basically, if the purpose is smoothness, maximum shades 
> reproduction (grey values), detailed prints and media coverage,... 
> one has to combined a precisely built head (your link) to be capable 
> of as small as possible dots (for smoothness and dotless prints) 
> dithered by a first-class driver (for the best trade off between 
> value/detail) with a variable density ink set (for maximum shades 
> reproduction) with ink formulation that combines both right pigment 
> load and eventually low dot gain.

I think you see it as a theoretical concept. I guess the Epson 
heads have their limitations as well. So is the hardware and 
software that drives them. The Canon pages I gave the link too 
may be exaggerating the limitations but they are revealing 
some basic requirements for further developments. For all the 
three inkjet head technologies. It is also interesting that 
Canon goes for a 16 bit driver with this technology. This 
could go two ways, at last there is the hardware that can 
represent the 16 bit smoothness, color accuracy and/or they 
need the 16 bit to get it all controlled.

> 
> I've seen K3 lacking a L-LLK, and somehow lacking (a bit) Dmax, so 
> I'm at a stage where I wonder whether on principle highest dpi 
> setting for smallest drop possible would eventually better the print 
> at all levels without generating issues like dot registration, drop 
> size... I also count on a reasonnably long overlap and some 
> significant extra K blackboost to smooth the print yet give good 
> Dmax. 
> 
> Does that all make sense or not worth trying ?

For one reason or another it has always been hard to develop a 
continuous tone printing technology that has a high Dmax + 
consistency in output + sharp detail. That was so with 
Collotype, Screenless Offset, Conventional Rotogravure, 
Woodbury Type to add an odd one as the last. For process 
control you see that they don't push the pure concept but add 
better controllable technology and make it a hybrid process 
like the semi halftone in modern rotogravure shows. I think 
that inkjet right now is very close to the ideal that the 
other processes tried to achieve. With a hybrid of all tone 
forming elements known in printing as one only couldn't 
achieve that goal. Your infinite small droplets has it 
limitations as described. Best example that they tried before 
is the Iris printer with its continues one size droplet stream 
(though then with larger droplets). It had wonderful smooth 
gradations, good color but lacked sharp detail and the printer 
needed a lot of maintenance for consistent work. It has been 
replaced by the Epson head technology. The successor of the 
last could be Canon's new heads. It can pump a lot of small 
droplets to the right place, so they are trying again.

Ernst
                    --
           Ernst Dinkla


www.pigment-print.com
(         unvollendet         )

> 
> 1/ another grey ink added does wonders, no I don't think that 
> precision became much better, speed was more important, there 
> must a trade off between the two  2/ Yes, I think the 9800 
> range could be better for gloss printing in the first place 
> and isn't as good for matte as the 9600 range was. At least we 
> start to think that over here. The 9800 is also harder to 
> adapt to third party inks. However with a  RIP it must be 
> possible.
> 
> Pardon, I'm Dutch so it will not be proper English either. 
> Combien d'encre par une goutellete (had to look up that word 
> :-) 3.5 picoliter could be 3 or 4 picoliter.

1/ yes, which also comes down to how many dilutions is enough. 
Apparently, 3 can achieve the goal. A L-LLK is missing, so a more 
precise partitionning and dithering is probably needed. 2/ Just K3 
alone + QTR might get me there e.g. glossy prints. If not I'll wait 
for Glossy Piezotone inks.

Regarding droplet content (excellent French BTW and we're not hat far 
apart) you mean say a 20pl has a higher probability of being close to 
20 than a 4 to be 4 (for instance 19.5 vs 3.5)?

> The successor of the last could be Canon's new heads. It can pump a 
>lot of small droplets to the right place, so they are trying again.
> 

I really need to dig into that pdf that I printed and Canon site... 

Olivier