Digital BW, The Print

Keith,

I agree with you about the statement as written and always found it
misleading. Since the goal is to successfully simulate continuous tone all
you need is a number of tones greater than the minimum required to reach
that point. Like resolution, at a certain point, adding more doesn't result
in visible improvement.

I can at least see the idea 1000 shades of gray coming from the concept of 4
inks times 256 shades of gray. I agree with Tyler that this is theoretically
possible if the 4 inks differ in hue or chroma, and the driver is a 4
channel RIP processing a CMYK file. The actually number of colors would be
somewhere in between 256 and 1024 since the inks are partitioned over
different density ranges.

This is at the heart of the issue with grayscale inks. One of the things
that makes a toned silver print so wonderful is that it is not just shades
of gray but that it varies in hue as it varies in luminance. That is why I
am so fond of the MIS-VM and Selenium PiezoTone inks. They give you a
variation in color as they change in density. This is important because it
makes it easier for our eye to distinguish between different print
densities. Our eyes seem to be incredibly sensitive to slight color changes
near the neutral point which makes it difficult to get a truly neutral
print. What we want to do is to take advantage of that fact to enhance our
prints.

Lets say we can only distinguish X shades of gray but we want to represent
more then X shades in out prints. We can do this by using subtle variations
in the color of the shades and put the eye's fine color sense to work for
us.

For instance the Selenium PT inks appears to be a red-warm black and three
dilutions of a blue-cool dark gray. While I doubt there are 1024 "colors" in
the print, I think that are  more than the 256 you would get from a mono-ink
print. Remember that regardless of what type of file we send to the printer
it is still a 4 channel CMYK, 6 channel CcMmYK or  a 7 channel CcMmYKk
printer. I would guess that each channel receives at least an 8-bit data
stream from the driver.

The Lysonic E small gamut inks are another approach to this technique and I
would like to see something similar done with pigment inks. In theory the
2200 should be able to work in this direction as well and I will be
interested to see what people can achieve with it.

Martin Wesley

http://www.borderless-photos.de/guests.html



----- Original Message -----

From: "Editor P.O.V. Image Service" <editor@...>
To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
Sent: Friday, August 02, 2002 10:21 PM
Subject: Re: [Digital BW] Do inkjets dither or not?


> Going back to the original...
>
> rmcooke wrote:
>
> >
> >In the August 2002 issue of the Digital Imaging Digest (which is
> >written by the folks at the PMA - Photo Marketing Association) they
> >cover quad tone printing among others and in the article they mention
> >Jon Cone's Piezography system.  They attribute the following quote to
> >someone from Cone's company:
> >
> >""Unlike the Iris process, PiezographyBW does not use a dithered
> >pattern. It is as close to silver properties as possible.
> >PiezographyBW's perceived resolution is greater than the human eye. It
> >renders more than 1,000 gray values in 8-bit mode.
> >
> I think the problem here is that if Jon is making prints with 1000
> luminance values he's gone totally nuts..  He must be marketing prints
> to CCDs, PMTs, and other sensors..
>
> > The new
> >PiezographyBW Pro24 renders more than 4,000 gray values in 16-bit mode."
> >
> >
> >
>
> Even more ludicrous..
>
> I think the fault probably lays with the publication in that the writer
> likely misunderstood Jon..
>
> There is no conceivable reason on God''s green earth to use 4000
> greyscale values, unless you like proving you can create things the eye
> cannot perceive..  In that case I would stop at 257 and be done with
> it.. LOL
>
> I would be willing to bet that Jon meant 4000 distinct RGB values on a
> print, when rendered from a 256 greyscale original...  That "more
> information is presented thereby to the human eye than one might get
> from a simple greyscale image"...  That would make sense..
>
> The statement as it is.. well.....  it's nonsense..
> Keith
>

I must be nuts for continuing this, clearly it's going nowhere. I'm going to describe this and you can make of it what you 
want.
I have a particular 16 bit Photoshop grayscale tiff, it has 32,076 levels.
...Slight divergence...
--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <darkroom@i...> wrote:
snip
> > Note that any 16-bit image saved out of Photoshop will have, at a
> > maximum, 32769 unique levels per channel, not 65536 as one might
> > expect...
> This is simply silly.
I suggest you bring that up with Adobe.

> If this is the case, then simply save it as TIFF, and
> you should have no loss problems.
Wrong, not if a Photoshop TIFF, same problem.

> Anyway, even if this does happen, it is
> irrelevant to this discussion.  A 16 bit grayscale file has the ability to
> have 64k possible values per pixel, period.  If Photoshop clips the file to
> 15 bits, then it isn't a 16 bit file.
I assume that would be technically correct.
 
> It also appears that Bruce is talking about RGB files.
Wrong

To continue-
I convert my high bit gray file to a high bit CMYK file per Dan Culbertson's method. This gives me a four channel high bit file 
with each channel nearly identical to my original gray. There is some slight loss, I now 
have 28,871 levels in each channel, however the end points are intact and they appear identical to my original Gray on the 
monitor, relationships are intact.
Temporarily working with one (Y) channel, via levels or curves we pull 25% to 100%. So the entire range of the channel, from 
0 - 100%, is filled with the top 25% tones of our original file. The Levels utility now 
tells me I've lost some tones, I'm left with 7,540 in that channel. If I now send this channel to a driver that won't mux it up, 
and print it to one ink of an Epson, I'm using the full range of that output system's capability 
for one ink.
However, we have an ink there who's maximum density only comes to 25% of our output system/paper/ink combo's 
maximum density capability. It is also monochromatic.
I still have three channels left, each with 28,871 levels of image to manipulate and send to three more inks of varying 
densities to take over at 25% and continue on down the scale. With the severe curves 
necessary to properly partition, each channel will probably wind up with something more like 7000 levels, and each is 
given it's own part of the scale. Clearly I wind up with a lot of levels at the end of it. Even if I 
have to convert to 8 bit for the printer to accept it I will have more than 256, and none of it has anything to do with hue or 
chroma in this particular discussion. These channel's levels, because of the 4 differing inks 
and correct partitioning, wind up serial, not parallel, to a large degree.
How many levels do we actually wind up with as ink on paper? I don't know. Is it all "real" or is some of it noise? Maybe, but 
it is all tonal information derived from the original file. Do we need this much info, can 
we see it? I don't know that either, and that's not what I was addressing.
This is hypothetical, the inks are never really 25%, 50%, 75%, 100%. We have channel/ink overlap to deal with for smooth 
transitions, and different paper and ink combinations require one to deviate from this 
ideal in different ways. But this is one of the main reasons quad tones look so good, we are assigning a full scale to an ink of 
limited density, resulting in a long scale on paper before you even move on to the next 
ink. You have to look at his as a system (which even includes the paper), not just a file.
How the Piezo driver accomplishes this is probably different. It may not actually convert the file to four channels in the 
background, but it certainly assigns levels from the original single channel file to four different 
inks in a very similar fashion. Jon actually taught these methods of manual quads before he offered the R9 as Piezography, 
and the R9 was commercially available before his agreements with them. They offered 
it as a straight colorsync compatible CMYK export module and also a quad version. It's clearly a four channel driver.
So now that my testosterone is burned off, I need to get back to making prints with thousands of tones.
Tyler

Tyler,

Before responding to your post, I have one simple question for you.

If the file only contains 256 levels of gray, and IF (BIG IF) what you claim
is true, that standard Piezo is capable of 1000 tones, where do the extra
744 levels of gray come from?  How, exactly, are they derived?

Regards,

Austin

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <darkroom@i...> wrote:
> Tyler,
> 
> Before responding to your post, I have one simple question for you.
> 
> If the file only contains 256 levels of gray, and IF (BIG IF) what you claim
> is true, that standard Piezo is capable of 1000 tones, where do the extra
> 744 levels of gray come from?  How, exactly, are they derived?

Austin, I honestly can't explain it any better than I did. I really think you have to think beyond your file, and see the entire 
system including the ink and paper.
How about this to think about?
256 level grayscale file not altered in any way. Printed with a hypothetical two ink printer.
Exact same information sent to two mono inks of different density. Is the tonal scale extended beyond the same thing 
printed with one ink?
With respect Austin, I think I'm done with this. I also think, with specific regard to the Piezography driver, someone from 
ConeTech would be far more qualified than I to discuss that system.
If my previous BS didn't communicate properly that's my shortcoming. If it did, but is flawed, I'd be interested in hearing 
how.
Tyler

Tyler,

> > Before responding to your post, I have one simple question for you.
> >
> > If the file only contains 256 levels of gray, and IF (BIG IF)
> what you claim
> > is true, that standard Piezo is capable of 1000 tones, where do
> the extra
> > 744 levels of gray come from?  How, exactly, are they derived?
>
> Austin, I honestly can't explain it any better than I did. I
> really think you have to think beyond your file, and see the entire
> system including the ink and paper.

I have, and that's why I claim it doesn't do it.  I also claim it simply
can't.  See my other post in response to Roy.

> How about this to think about?
> 256 level grayscale file not altered in any way. Printed with a
> hypothetical two ink printer.
> Exact same information sent to two mono inks of different
> density.

But that's not how it works.

> Is the tonal scale extended beyond the same thing
> printed with one ink?

Not necessarily.  You can create an infinite number of tones with ONE ink.
The use of multiple inks simply means you take up less area creating the
tones.

> With respect Austin, I think I'm done with this. I also think,
> with specific regard to the Piezography driver, someone from
> ConeTech would be far more qualified than I to discuss that system.
> If my previous BS didn't communicate properly that's my
> shortcoming. If it did, but is flawed, I'd be interested in hearing
> how.

I'll get an answer from Jon, I'm sure.  It may not be a very "forward"
answer though...but I'll do the best I can.

Austin

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <
darkroom@i...> wrote:
snip
> > How about this to think about?
> > 256 level grayscale file not altered in any way. Printed with a
> > hypothetical two ink printer.
> > Exact same information sent to two mono inks of different
> > density.
> 
> But that's not how it works.
Of course not, that wasn't the point
> 
> > Is the tonal scale extended beyond the same thing
> > printed with one ink?
> 
> Not necessarily.
I would say that, as ink on paper, it definitely is extended. And this 
would be from a single unmanipulated file simply by the use of inks, 
and this is only a very small and oversimplified part of the real 
overall pictue.
I've been told that the tonal extension achieved by that simple 
experiment is technically noise, but there we are, and it was done by 
adding no new "levels" to the actual file. Just done with ink on paper.

> I'll get an answer from Jon, I'm sure.  It may not be a very "forward"
> answer though...
I'm with you there,
Tyler

> > > How about this to think about?
> > > 256 level grayscale file not altered in any way. Printed with a
> > > hypothetical two ink printer.
> > > Exact same information sent to two mono inks of different
> > > density.
> >
> > But that's not how it works.
> Of course not, that wasn't the point
> >
> > > Is the tonal scale extended beyond the same thing
> > > printed with one ink?
> >
> > Not necessarily.
> I would say that, as ink on paper, it definitely is extended.

Hi Tyler,

Then you're missing the point of dithering.  You can represent ANY number of
tones by dithering, it's simply the area that it takes to do so...you simply
need to stand further away ;-)  Increasing the number of inks simply
decreases the area needed.  It can also be considered "resolution"...

> > I'll get an answer from Jon, I'm sure.  It may not be a very "forward"
> > answer though...
> I'm with you there,

So, help me out here...  Make a 256 Z pattern, and see if you can
distinguish 256 Zs...  What do you think of doing that???

Austin

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <
darkroom@i...> wrote:
> So, help me out here...  Make a 256 Z pattern, and see if you can
> distinguish 256 Zs...  What do you think of doing that???

Oh lordy ( I think they say that in North Dakota), I think it's time 
for an Austin's Zees file, would you care to prepare it?
Much more rerwarding to go make a beautiful Piezo print of a favorite 
image. I'll bet good money you would agree, no matter how many tones it 
has resolved.
By the way, I have great apreciation for your pure sense of 
photographic art, it's very valid. This relates to a previous thread, I 
also apreciated that you have room for aproaches that are opposed.
Tyler

Hi Tyler,

> > So, help me out here...  Make a 256 Z pattern, and see if you can
> > distinguish 256 Zs...  What do you think of doing that???
>
> Oh lordy ( I think they say that in North Dakota), I think it's time
> for an Austin's Zees file, would you care to prepare it?

I don't know how to use Photoshop...remember ;-)

> Much more rerwarding to go make a beautiful Piezo print of a favorite
> image.

Of course, but that won't "resolve" this debate, now will it?

> By the way, I have great apreciation for your pure sense of
> photographic art, it's very valid. This relates to a previous thread, I
> also apreciated that you have room for aproaches that are opposed.

Why, thank you...but that's not going to get you out of your sworn bound
duty to help us all out and create a 256 Z file now...please... ;-)

Austin

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <
darkroom@i...> wrote:
> Of course, but that won't "resolve" this debate, now will it?
Better minds than I will be required.

> Why, thank you...but that's not going to get you out of your sworn bound
> duty to help us all out and create a 256 Z file now...please... ;-)
Oops, I just remembered, I have a lot of laundry to do.
Tyler

my sworn duty? Did you get me drunk?

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <
darkroom@i...> wrote:

> duty to help us all out and create a 256 Z file now...please... ;-)

I got it, if you want it. Turned out to be easy to derive from the 
other.
Tyler

> > duty to help us all out and create a 256 Z file now...please... ;-)
>
> I got it, if you want it. Turned out to be easy to derive from the
> other.
> Tyler

Hi Tyler,

OF COURSE I WANT IT!!!  Did you print it out, and if so, what was your
result?

Please email it to me at austin@......

Damn, now we're getting somewhere ;-)

Regards,

Austin

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <
darkroom@i...> wrote:
> OF COURSE I WANT IT!!!  Did you print it out, and if so, what was your
> result?
I'm at home today, Epsons are at the studio. When I get there tomorrow 
there will be these very strange people thinking I should be doing 
things for them instead of this stuff. How odd and boring.
It'll have to wait I'm afraid.
> 
> Please email it to me at austin@d...
Can't get the full address on the yahoo site, cam you email it me be by 
clicking on my address shown above? (I'm assuming you're looking at 
this on yahoo). Then I can reply and send it to you directly rather 
than through ya#*@%.
Do you have stuffit expander?
Tyler
> 
> Damn, now we're getting somewhere ;-)
Yeah, Mars
Tyler

--- In DigitalBlackandWhiteThePrint@y..., "tboleyyh" <tyler@t...> 
wrote:
> --- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <
> darkroom@i...> wrote:
> > OF COURSE I WANT IT!!
by the way, the relationships are the same as the other. Upper left is 
0, 1, Z=2, 3, 4.
Next square is 5, 6, Z=7, 8, 9.
Next square 10, 11, Z=12, bla bla.
Each square is 5 levels.

We will be holding a meeting on Thursday night where, we will drink 
copious amounts of Benedictine & Brandy,
then hold open discourse as to how many cows can dance on the head of a 
pin, what exactly will be considered dancing
and whether Gernseys will be allowed as cows because they are not 
black&white (as being germaine to the list)

See you all there.

J.Z.