Printing bitmap?

Hi everyone,

First, I haven’t been printing for months (either inkjet or carbon) and won’t be in the next years.
But I try to keep an eye on the inkjet world, PiezoDN, etc… for the day I have a lab again.
And think of ANYTHING that could help me be happy with digital negatives that day.

I was wondering if anyone has an experience using Icefields FM halftone files (bitmap) or equivalent with an Epson printer.
Long ago, I had a conversation with S. Herron who told me some photographers liked to use Icefields and their printers to print color photographs.
I believe a PostScript RIP (and PostScript printer?) is needed, but that’s something I don’t know about (I’m not a digital imaging expert…).

What I have in mind is to print Icefields true stochastic patterns and if possible reprint in registration the same original file re-converted (different stochastic patterns each time) until I get enough D-max.

My questions are:

  • What is required to print Icefields FM halftone with an Epson printer (let’s say a 3880)? PostScript?
  • Will the printer actually fire dots where dots are in the Icefields file (what I want) or will it reproduce the “haltone image” with its own Epson dither patterns (what I don’t want) after some conversion?
  • Could QTR play a role here (I think not)? Or Guttenprint?
  • Walker, any progress with your flatbed printer project (your Epson teaser #113515)?

Thank you all for your attention and help.


What is required to print Icefields FM halftone with an Epson printer (let’s say a 3880)? PostScript?

No need to do halftone unless you are needing to print Gravure in which case the microweave and dither of QTR or the Epson driver will do a random stochastic just fine when you use a single MK channel (Don Messec can talk more in depth on that.) PiezoDN does its best to replicate tradition continuous done (aka smooth silver) negs.

Will the printer actually fire dots where dots are in the Icefields file (what I want) or will it reproduce the “haltone image” with its own Epson dither patterns (what I don’t want) after some conversion?

If you make a stochastic bitmap in Photoshop out of your file at something like 720 or 1440 dpi and print at 2880x1440 dpi, the printer will do a pretty good job at this. Otherwise you need a more true halftone rip like AccuRIP (or Icefield though I’ve never heard of it) or StudioPrint.

Could QTR play a role here (I think not)? Or Guttenprint?

As described in Q1.

Walker, any progress with your flatbed printer project (your Epson teaser #113515)?

I can neither confirm or deny. :wink:

Thank you Walker.

Icefields is not a RIP. It simply converts grayscale or color files into stochastic bitmap, primarely for imagesetters. But better than AGFA or Heidelberg (Tod Gangler’s opinion).
Alas, I think Icefields and support are not available anymore.
Last time I had a discussion with the developper, a CUPS driver came out as a possible solution to drive Epson from Icefields without a RIP. Of course, I have no experience with CUPS…

“the Epson driver will do a random stochastic just fine”
Not my opinion. Things would look like some aquatint, but they don’t. Epson does not a trulely random job.
Jon Cone told me once: “there is no random in anything printing”.
And Roy Harrington: “Dithering is often referred to as stochastic or random placement. It’s not truly
random - or mathematically random”.
That can’t be discussed!
We’re talking negatives of course; you won’t see what’s not random in a positive inkjet print.
I agree that printing only one channel or some channels with low ink limits will look pretty random to a certain point. But at that point D-max is too low.
That’s actually when the idea of printing several passes (is this the correct english term?) in registration came.
Then I fear some moiré or electonic noise, because Epson doesn’t print in a true stochastic way. I suspect its algorithms will make the same job everytime one prints the same file. Icefieds will generate differents files every time you hit “convert”, with different dots placements. No moiré. Feeding could make some resonace and banding, but this can be adjusted.

I have tried AccuRIP a little. I could be wrong, but it looks like the halftone is converted for Epson to be able to print it, using its dithering again… I see it as an Epson reproduction of a halftone image.
Like printing PS’s bitmap: when a lot of ink is needed, Epson’s patterns show for sure.
If you print bitmaps with the Epson driver, some conversion(s) must take place.
What I’m looking for, is to truely bypass Epson’s dithering, which could be impossible.
I’ll investigate CUPS.
Maybe you could show some direction to a novice?

Thanks again.


Dan Burkholder writes about the use of Icefields in his book "Making Digital Negatives for Contact Printing. I have the 2nd edition, published in 1999. It’s chapter 8 - Output: Getting Ready for the Imagesetter. If you can’t find the book, and are interested, I could scan the chapter for you. It’s too much to type out.

I have a vague memory of a place in Santa Fe that made negatives this way. I can’t remember the name of the business, but Peter Ellzey worked there. I also think that Stan Klimek was one of the people who made negatives this way back in those days. Stan recently moved back to Santa Fe so I’m probably be seeing him soon, since I live not too far away in Taos. I remember seeing some of Stan’s large Pt/Pd prints for a client back in the early oughts made from imagesetter negs. They were quite beautiful.

StudioPrint has true random stochastic dither at the printhead nozzle level
if this is what you actually need.

You would not be able to do channel channel dot placement registration

But single channel stochastic using matte black ink and studioprint, yes.
Pricey rip however.


Hello Keith,

Thank you for you offer. I bought this book in 2001 and made a few carbon prints with Burkholder’s colorized negatives method. Already fighting with Epson’s pattern then…
Icefields must work great with imagesetters. Tod Gangler has professionaly made carbon prints with such negatives that are described as beautiful. I live in France and never was lucky to see one so far…
Unfortunately, the imagesetter technology is dying (might be dead?) and finding a service bureau is nearly impossible in my area. I also wish to make my negatives at home, preferably without chemistry, and at reasonable costs. I bet on inkjet years ago and still stick to it.
Also, imagesetters make dot negatives and I’m looking for continuous tone negatives. So why trying to print bitmap! In my mind, multiple-pass-printing stochastic bitmap with moderate UV blocking ink should make something close to contone. In theory, dots should accumulate and overlap randomly and generate tones without artefacts. I even see ink dots here as silver grains in films. Meanwhile, I see chances this wouldn’t work. It’s all about the absolutely-true-random-stochastic placement of dots, that could be a fantasy.


Thank you for your precisions about StudioPrint. “True random stochastic dither at the printhead nozzle level” sounds good. I feel some little hope there! Still have to build a flatbed printer…


That said, I really don’t think you need this for carbon. There are a few people currently using PiezoDN quite happily with/for carbon. Because all the channels print, this gets rid of automation lines (digital printer artifacts).



You may or may not remember I have purchased and started to try PiezoDN. But much too briefly as I moved then in a place that needs a lot of repair before I can even plug a printer. A lot…
It has been very frustrating to have to brutally stop as some tools were still shipping to me (cartridges, inks, etc…), but I must say the very first results were promising. Still, I’m not sure about zero artifacts; I believe it’s in the printer’s gene. Maybe the way I carbon print is less forgiving than other’s, and I should reformulate?

I understand my questions could be nearly provocative in a place that offers great tools for making inkjet negatives.
I don’t doubt PiezoDN is amongst the best solutions, and that’s where I’ll start again when possible.
Meanwhile I feel the need to expore things, which I’m sure you understand. Multiple layers inkjet printing is an old fantasy and a long term project, if it ever comes to life. I’m gathering infos and try to educate myself; all I can do at this time.



We are working on an updated driver (for both PiezoDN and Piezography) that changes the dither, microweave, dot size, dot placement, vertical resolution, etc.

On small formats it brings the resolution to 5760x2880 and on 17+ printers is gets to 2880x2880. I think you may be interested in this. We will release privately to the PiezoDN group as beta first.