Digital negatives and direct digital alt. process print exposure

I've never made it a secret that I dislike inkjet digital negatives. If you think they're great, that's fine. But my own experience with them has always been disappointing. Maybe it's just me, maybe it's Maybelline...IDK.

In any case, sufficient reason to keep on the lookout for alternatives. My own frustration combined with exciting developments in laser and digital imaging technology have formed the inspiration to write a blog about all this. The purpose of the piece is to offer a brief overview of the different methods to get from a digital on-screen image to an alt. process print. For much of the information, I'm indebted to @AndrewBurns who contributed loads of useful input in this thread but also in many emails, and who has provided effective illustrations for two crucial sections of the blog post.

What I hope to achieve with this blog is the following:

• Offering inspiration to those who want to embark on some DIY-fueled experimentation
• Triggering responses that point out how I/we overlooked important aspects
• Encourage people to come up with alternative methods, solutions etc. that help solve the puzzle

So it's really intended to spark discussion and to act as a small step on a longer road, that hopefully helps us to improve on what's presently possible with inkjet technology.

PS: this is not to poop on inkjet. It's great, really, if only for its accessibility for us amateur/home users. But it has its limitations, as I think we all realize, that's all.

Here's the link to the blog post:

Feel free to comment - here in this thread, in the comments section on my blog, or by contacting me through private messaging, email etc.

G'day, I own several LVTs and different film recorders. I can point you to several projects that have been done over the years that are not covered by your blog.

Andrews efforts are interesting however are not on the same level as some of others I have seen.

C

calebarchie said:

I can point you to several projects that have been done over the years that are not covered by your blog.

Click to expand...

Please do! LVT's are indeed an area I've only looked at superficially. I'd love to hear/read about projects that have explored this angle more in-depth. I'm aware of one person running a print shop in the US who built his own machine for exposing the film negatives; the name evades me at the moment, but perhaps you know who I mean. I think he also prints color carbon or otherwise maybe dye transfer. I know I've read a brief description of the machine he built which included some photos, but no critical details on control systems etc.

koraks said:

Please do! LVT's are indeed an area I've only looked at superficially. I'd love to hear/read about projects that have explored this angle more in-depth. I'm aware of one person running a print shop in the US who built his own machine for exposing the film negatives; the name evades me at the moment, but perhaps you know who I mean. I think he also prints color carbon or otherwise maybe dye transfer. I know I've read a brief description of the machine he built which included some photos, but no critical details on control systems etc.

Click to expand...

Yes that would be Mr. Browning and his dye transfer lab. There was a PDF that had all these details but I don't think it's available any longer..

PM me, I can try collate a list however many of these might be lost to internet history

Ah yes, I thought it was James Browning, but couldn't find the pdf I distinctly recall having looked at not too long ago (maybe a year or so?) I guess maybe it's still online, but no longer linked to on his website.

I've pm-ed you - thanks so much in advance!!

I have made negatives via Pos/Neg method of the 80's , I have made Lambda Laser negatives , and of course I make inkjet negatives. I have quite a bit of practical work experience with all
three and have observations of each .

Thanks @Bob Carnie, feel free to share - I always read your posts and observations with great interest. They've always been very inspiring to me! Indeed, when I wrote the bit about the LVT/film recorder, I also had your work with Lambda-exposed negatives in mind, which you've mentioned earlier here and elsewhere.

Great blog post, lots of info! Yes my experimentation has been fairly limited, but I only started trying to make alt-process prints of any kind about 6 months ago! I actually have one pdf from James Browning "Dye Transfer Materials" which goes more into the emulsion and coating aspects but does have one picture and a few lines talking about his laser exposure machine. The whole process seems very involved, but I suppose he was doing it as a business so it makes sense.

Thinking a little more about continuous tone vs. halftone. I know that generally for multiple transfer work in particular there's a desire for halftone negatives because they're inherently insensitive to exposure (tone being set by the density of dots rather than local transmission of light means that it should be impossible to over-expose the image). And also inkjet digital negatives are actually somewhat half-tone, just with limited density and resolution, which makes them difficult to get repeatable linear results from.

However in the case of the LCD exposed images I'm currently making, the LCD mask is almost like a perfect continuous tone negative. The pixels are very accurately lined up with no gaps/overlaps, the density of each pixel follows a very linear and repeatable curve which can be accurately controlled and with very high contrast range. So as long as the amount of exposure is also very accurately controlled (probably by including a UV sensor between the LEDs and LCD to integrate exposure) I don't see why the results wouldn't be linear and repeatable enough for multi-transfer or colour prints despite not being entirely insensitive to exposure like half-tone negatives would be.

Original Neg direct contact print- Original Neg enlarged print- Original Neg to Ortho pos to enlarged negative to print- Scan Neg to direct Lambda laser print- Direct from digital camera to Lambda laser print- Inkjet neg from scan or digital file......
I have been lucky enough to do all of the above , I cannot speak about LVT other than I purchased LVT negatives to print in the late 80's to mid 90's and have seen Salgado's Genesis silver prints from LVT negs
I have also been lucky enough to have access to Epson , Creo Eversmart, Imocan and now Phase one scanner systems.

I point this out as one can see there is a lot of technology and practice involved in making fine art prints and my whole career has been involved in gallery quality printing projects, an commercial photo labs since 1976. In this time I have heard many , many boasts about how one system is better than the other. I feel most of this boasting is due to someone trying to qualify to themselves their financial purchase in the gear or methods of work.

A direct contact print from original negative is as sharp as a direct digital print on same material from a good digital camera system.
An enlarger print can be out of focus or soft due to sloppy workmanship.
An digital file or print can be brutalized by an incompetent technician.
the argument goes on and on and on .
I am just as happy looking at an image on the wall and not considering how it was made but rather enjoying the image for itself, or a group of images that tell me a good story.
I think all methods have there strengths and weakness and we all gravitate to the methods that we like.

For me I like working from digital files in PS and then making expressive gum prints over palladium, for me this is the most liberating method of creating prints in colour and black and white, but its not for everyone. If you ask Drew he will tell you direct enlarger prints using large format C41 film to Fujiflex colour RA4, if you asked MAS when with us it was ultra large camera film direct contact to Azo paper, if you ask Ed Burtynsky he would say inkjet print from Phase One Camera system. If you ask Sarah Moon she would say colour carbon from digital file, Magnum photographers would say 35mm film enlarger to silver.

I am not on a rant here but want to state early in this thread that all methods are good for the right person and their perceived audience of their work.

It takes a very , skilled person to be able to clearly state when looking at the same image in various print method s , which one is which - I tried this years ago on a very large sample size of people (over 300) and the only person who was definitively able to tell me which one was which was Les Mclean who many may remember here on this site, it was an absolute shock to me and helped me decide that digital was ok to work with.

not sure if this helps Korak .

Totally agree @Carnie Bob - while my experience of course is not nearly as extensive as yours, I recognize all of the things you say. All means are valid, some will appeal more to some people, other means will work better for others. Whichever means you choose, it takes time, experimentation and reflection to get the best results from it.

Currently, I'm going through the motions with inkjet again, but the negatives I get from it (so far only test charts for linearization) just don't appeal to me that much. They're sort of OK - but I feel I'm spending way too much time poking around with a photospectrometer. There's only so much time I'm willing to spend going round in circles ('iterations') on that front. It doesn't feel very productive, even though my insight is growing and the results are improving. I feel it's just not going to cut it.

I may bite the bullet and get some imagesetter negatives made to see how those work for me. I'd really rather not, though - mostly because of the risk I'd like them too much...

A few things I noticed re digital negs , lambda negs , imagesetter negs.

On matt surface a lambda silver neg and inkjet digital neg - it is impossible to notice a difference in print quality if you are viewing the print on a wall from close distance.- I have viewed thousand of prints
in the same room as other well known print processes.
On glossy silver surface , I noticed a slight bleed through the white border which introduce a bit of tonality on what is suppose to be white borders. But at proper viewing distance I feel inkjet negs do the job.

Regarding image setting negatives - they are quite popular with the colour carbon crowd as the dot pattern holds highlight detail better than con tone inkjet negatives ( I think because there is nothing to hold onto when processed out in hot water) - this is the reason I went to gum process as hot water is not needed and I feel the highlights hold much better.
As well I was shown a portfolio of PT PD prints by a very well known book publisher and these prints were small and in all prints I could see the stochastic pattern in all the prints which turned me off but
a good friend of mine makes spectacular colour carbon prints in the range of 22 x 28 size using image setter images and because of the natural viewing distance the stochastic pattern is not visible in his work unless you are of course a print sniffer and go within inches of a print which btw I find a quite comical habit amongst our tribe of photographers.

So a lot of variables play into things.

Thanks for sharing your experience @Carnie Bob - this is very valuable.

My primary purpose with these digital negatives is color carbon at this point. While I'm quite happy with monochrome carbon from in-camera negatives, I can't quite let go of the idea of trying my hand at color. I'm primarily interested in a photo-realistic result with predictable/controllable color rendition, and this means that I need to have a system that is reasonably consistent. At the same time, it's a hobby for me, so I can't really justify gigantic expenses to the effort. I'll have to cut corners somewhere, so it's a matter of finding the best compromise.

Final paper surface does matter a lot in how detail renders and how tonality ends up looking. I acknowledge that and it echoes my experience.

Carnie Bob said:

Regarding image setting negatives - they are quite popular with the colour carbon crowd as the dot pattern holds highlight detail better than con tone inkjet negatives ( I think because there is nothing to hold onto when processed out in hot water) - this is the reason I went to gum process as hot water is not needed and I feel the highlights hold much better.

Click to expand...

Sort of, I think, although I don't believe that carbon highlights are necessarily more delicate than gum. Hardened gelatin is pretty robust, even in a hot water bath. It just takes a lot of practice to get it right in either process. With carbon, it's easier with dichromate sensitizer as it somehow makes for a more robust gelatin matrix that more easily survives wet processing. DAS sensitizer is much more challenging as the gelatin matrix is weaker and more easily breaks down, and there's less opportunity for contrast control as well. I've done extensive experimentation on this (here's one blog I wrote about it: https://tinker.koraks.nl/photograph...of-das-carbon-with-continuous-tone-negatives/ But I've touched on the subject in several other writings as well) in order to get where I am now. Based on that I can say that it definitely is possible to render delicate highlights from continuous tone negatives with DAS carbon - but it's not necessarily easy.

Translating that effort to a process where color separations are also used and the different layers need to be balanced carefully makes me hesitant to go a (quasi-)continuous tone route with this. Otherwise I would have just plunged into inkjet head-on. But there's just too compelling an argument against it in terms of consistency and robustness of the process in the face of myriad process parameters that in the real world tend to shift about ever so slightly.

So it boils down to a choice of where to put in the effort. With (quasi-)continuous tone inkjet negatives, the effort is spent on extreme process control; basically we're talking about a Six-Sigma approach towards alt. processing: reduce process variations at all stages to the very margins of the bell curve. That's one way. The other is to somehow find a way to harness the power of halftone screens, which revolves around finding a way to generate high-quality screen negatives in the first place, and controlling a printing process in such a way that it'll reliably image tiny little dots.

They're different challenges and sometimes require a fundamentally different approach, although they overlap in many places at the same time.

Thinking more on getting half-tone exposures using an LCD screen. The best easily and cheaply available LCD screens are around 1000 to 1200 dpi, which is just about good enough for a nice FM halftone, so the real remaining issue is getting the light source collimated well enough that there's little to no dot-gain/blurring despite the relatively large distance between the LCD and tissue, AND getting the light source powerful enough that you can lose 90% of it to transmission losses and still have reasonable exposure times.

I was watching a video on youtube the other day about a hobbyist trying to DIY manufacture silicon chips using photolithography, which is very similar to alternative-process printing except on a tiny scale:

As part of his process he has to tackle exactly the same issue, namely to make a relatively large area, powerful and highly collimated UV light source. One interesting trick he uses is a stack of two orthogonal layers of LCD screen 'privacy film'. The privacy film is something you put on your computer screen to reduce its visible angle so that people can't look at your screen from the side, and it does this by only letting light through if it's travelling perfectly perpendicular to the film. The result is the film will only allow (nearly) perfectly collimated light through and blocks rays trying to leave at different angles, so you would lose some of your light but you could guarantee that everything going through was collimated, even with a fairly short distance between the light and LCD screen.

My thought is that you could take one of those 100W COB UV LED lights people like Calvin use with the 60-degree lens it's sold with, use another fresnel lens to take the 60-degree beam and largely collimate it, and then two layers of the privacy film to guarantee that only the collimated rays get through to the LCD screen. You'd probably still need to apply some kind of vignette correction like I'm already doing, but hopefully it would be far less of an issue as the LED dies are all so close together.

AndrewBurns said:

two orthogonal layers of LCD screen 'privacy film'.

Click to expand...

Ah, that's creative. It's kind of a micro-scaled alternative to a large collimator grid - a.k.a. a honeycomb on a studio strobe! It'll also have the same disadvantage: it works by blocking out most of the light, allowing only collimated light to pass. What's crucial in the video I think is that the light source is really a stack of materials:
1: Diffusor (ground glass)
2: Brightness Enhancer Film (2 layers)
3: Two perpendicularly stacked layers of privacy enhancement film.
It's around the 30-minute mark.

(Edit: about the blocking of light, he mentions this in the video where he guesstimates the transmission of each modifier layer as 80-85%. Given 5 layers of modifiers, this would make total transmission something like 30% - which actually isn't too horrible (what he didn't mention is what the actual photon yield of the light source in a modern EUV semicon litho system is, LOL!)

AndrewBurns said:

100W COB UV LED lights people like Calvin use with the 60-degree lens it's sold with

Click to expand...

I got one underway as it happens, to play with. I was doing some experiments with inkjet negatives and that required better collimation as well, so I'm going to see if the 60 degree lenses do any good. I don't have any of the privacy screen stuff though. Maybe I should get some.

But frankly, I'd be inclined to replicate the design used in the video instead. So instead of relying on powerful COB LEDs, just use an array of smaller LEDs (i.e. a panel, like you're using now) and then work to collimate that. This means that the LED array will have to be approximately the same size as the final print size, but that isn't too much of an issue. UV LED panels are pretty cheap, especially if you can get away with just 395nm UV (365 costs more, but is also available).

On second viewing of the video it doesn't seem like the collimation he's getting out of his light source is that great based on the video footage, but it looks like he also has the light source a fair distance from his mask, which should improve the quality of the light at the expense of wasted power (again, similar to Calvin's setup).

AndrewBurns said:

it looks like he also has the light source a fair distance from his mask

Click to expand...

Yup, in a column/chimney that basically acts as a collimator in itself. I bet that a significant part of the collimation results from this. Still, the tricks with the filter materials will add, too, but I suspect that a lens array like you have on your light source is pretty much as effective (and more efficient).

Yes the main issue I encountered with the distributed LED lens array setup I've got was uniformity, but in theory that's only important for continuous tone negatives/masks. In theory if you had well collimated light and a fine enough resolution to do half-tone masks then light uniformity doesn't matter at all, as long as you expose at least long enough that the dimmest area is fully exposed.

My own quick tests of printing a dithered image went poorly, but that was with a 30-degree diffusing film in the light source causing huge amounts of dot gain (my pixels are about 50 micron squares, but are probably several hundred microns from the paper, so every degree counts). On the one hand I'm now interested to try the dithered image again with the diffusing film removed, on the other hand now that I have my continuous tone printing working well I'd rather not touch it! And anyway the 50 micron pixels are probably too large for a good print.

Also as a side note, the DLP projector evaluation kit mentioned in the video is this one: https://www.ti.com/tool/DLPDLCR471TPEVM

4k resolution which is pretty good but the minimum focal distance is quite large so it would need the projector lens modified or replaced. Also it's $1k USD, which is actually extremely cheap for a 4k DLP projector but also a fair bit for experimentation purposes...

Edit: Looks like they have a 1080p one for only $300 https://www.ti.com/tool/DLPDLCR230NPEVM

AndrewBurns said:

In theory if you had well collimated light and a fine enough resolution to do half-tone masks then light uniformity doesn't matter at all, as long as you expose at least long enough that the dimmest area is fully exposed.

Click to expand...

No, that's not correct in most cases. Even when working with halftone negatives, you still aim for a certain target density especially if you're doing separations. Uneven illumination is disastrous in that case; this includes any color work. Moreover, it'll show up in most processes since dmax is often not a firm endpoint - it's the top of the h/d curve and there's still some differentiation going on there, or even reversal (silver halide emulsions).
Long story short: no, you can't really get away with uneven illumination. It'll always be a problem one way or another.

AndrewBurns said:

My own quick tests of printing a dithered image went poorly

Click to expand...

I've done some tests previously and am working on a new set presently. Resolution is poor, but the concept always works - it's just much easier to linearize and even the raw/unlinearized negatives are much closer to the correct response to begin with. However, dot gain and 'dot loss' (Calvin calls it 'blowdown') are problems that need to be sorted. I'm currently tackling that by simply increasing distance, but it makes the whole thing go slooooooowwww....so that's why the lens-covered COB LEDs are on their way.

AndrewBurns said:

the DLP projector evaluation kit

Click to expand...

Neat! A bit spicy to my taste at $1k though. Still, very nice to play with.

Btw, this is entirely subjective and by no means intended to discourage anyone from trying, but I notice that I find (subjective, personal, irrational etc.) the whole approach with LCD's or DLP's very unappealing. I find I prefer an actual film-based negative. I think the direct imaging methods you're exploring are probably the way of the future, objectively inherently better than anything else - but I just don't like the feeling of it.

Thanks Koraks for the detailed blog. I can add some points here, primarily regarding DLP UV projection imaging and some examples.

First, I know of someone modifying a UV imaging machine; for instance, this detailed project: 'https://docs.hackerfab.org/home/fab-toolkit/patterning/lithography-stepper-v2.1'. Their imaging target is photoresist (or UV lithography), but to me, it doesn't seem significantly different from photographic film.

There are many research papers on similar ideas, such as 'https://www.nature.com/articles/s41378-025-01131-x'. There are also some commercially available ultraviolet imaging products on the market, such as the NVM-F from Young Optics—a 4K (XPR) projector with a 15µm pixel size and 385/405/365nm wavelengths. It is used in areas including jewelry, dentistry, research, industry, and biology (from https://www.youngoptics.com/en/top-95-product81).

These DLP-based designs typically present several challenges:

Resolution and Stitching: High-resolution DLPs are expensive, and 'stitching' requires both precise mechanical positioning and a deep understanding of photoresist properties. I suspect that even with 4K, the results on 35mm film will still be unsatisfactory. But 4k for 8mm？

Wavelength and Lifespan: DLPs designed specifically for UV light are costly. Standard DLPs typically operate reliably at 405 nm and barely at 385 nm; however, manufacturers usually do not offer warranties for 365 nm usage. If blue light (405 nm or 450 nm) is acceptable, standard DLPs perform exceptionally well.

Optical Path Design: The optical path of a DLP is relatively straightforward—it is essentially equivalent to an enlarger lens, a macro lens, or a microscope objective, imaging one plane onto another at close range. The primary challenge arises when using wavelengths of 365 nm or shorter UV, as it becomes difficult to utilize the extensive library of existing commercial lenses for visable light. For scientific research, pixel size can typically reach submicron at 365nm, which means 36000 x 24000 resolution at 35mm film(of course, splicing is required.)

An advantage of DLP is its realtime-calibration capability, which doesn't require high light uniformity. Imagine projecting onto a uniformly reflective white board, taking a photo, sampling the brightness of white points, and then adding a grayscale mask to the projected image until the brightness of all white points in the image is consistent. A small program can complete the calibration.

Yezishu said:

I suspect that even with 4K, the results on 35mm film will still be unsatisfactory. But 4k for 8mm？

Click to expand...

In all honesty, I don't care. The blog I wrote and my own interest revolves around printing; not film recording. I get what you say if the aim is to record digital images onto film. I guess there's interest in that, too, but I've never seriously looked into it, nor am I planning to.

For printing, my main concern in the options mentioned is power = throughput/print speed.

koraks said:

In all honesty, I don't care. The blog I wrote and my own interest revolves around printing; not film recording. I get what you say if the aim is to record digital images onto film. I guess there's interest in that, too, but I've never seriously looked into it, nor am I planning to.

For printing, my main concern in the options mentioned is power = throughput/print speed.

Click to expand...

There are many influencing factors here, and it largely depends on the specific use case. For example, the Young Optics NVM-F 4K has a pixel resolution of 3840 x 2160 and can provide a 50 µm resolution over an area of 192 mm x 108 mm. Its typical output optical power for 385 nm or 405 nm can reach 6 W, which translates to an UV intensity of 289 W/m² (or 0.289 mW/mm²).

As for how much exposure time is required for a specific process (whether using film, photo paper, or other alternative technologies) at this UV intensity—or how many steps are needed based on the graphic size and whether higher resolution (which means a smaller single-exposure area) is required—it all depends on the specific application. However, this represents the current high-end UV DLP projection modules(Or roughly the upper limit of a self-assembly).

Yeah, I'm aware of the tradeoff. If you look at the experimental outcomes so far of the few people who have been publishing about their experiments. So far those observations seem to be in line with what I just said.

The potential is great, the technology is exciting and conceptually it's beautiful, but in reality, there's some distance to cover yet if direct digital print to a UV-sensitive alt. process is intended. Perhaps the most feasible route would be to not just work on the exposure method, but also on the chemistry. I.e. depart from DAS carbon and cyanotype and look for chemistry that's inherently faster/more efficient. I think there's a lot to be said for that if the intent is to move to some form of direct digital exposure; after all, you'd not be bound to a particular heritage once you cut ties by using a digital exposure method.

Whether I personally find that a worthwhile direction to pursue...I don't know. The tech is interesting, but so far I don't feel motivated to put effort in it, despite the fact that I put plenty (way too much) time in embedded systems and electronics. When it comes to printing, I'm too much of a traditionalist, at least as long as it's 'alt process'. Inkjet is lovely, too, but I see very little reason to somehow try to improve that. Canon & co seem to be doing a pretty swell job there as it is.

koraks said:

Yeah, I'm aware of the tradeoff. If you look at the experimental outcomes so far of the few people who have been publishing about their experiments. So far those observations seem to be in line with what I just said.

The potential is great, the technology is exciting and conceptually it's beautiful, but in reality, there's some distance to cover yet if direct digital print to a UV-sensitive alt. process is intended. Perhaps the most feasible route would be to not just work on the exposure method, but also on the chemistry. I.e. depart from DAS carbon and cyanotype and look for chemistry that's inherently faster/more efficient. I think there's a lot to be said for that if the intent is to move to some form of direct digital exposure; after all, you'd not be bound to a particular heritage once you cut ties by using a digital exposure method.

Whether I personally find that a worthwhile direction to pursue...I don't know. The tech is interesting, but so far I don't feel motivated to put effort in it, despite the fact that I put plenty (way too much) time in embedded systems and electronics. When it comes to printing, I'm too much of a traditionalist, at least as long as it's 'alt process'. Inkjet is lovely, too, but I see very little reason to somehow try to improve that. Canon & co seem to be doing a pretty swell job there as it is.

Click to expand...

I believe this is what have been done on silver process: enhancing sensitivity, expanding spectral response, and improving the usable contrast range to make exposure more manageable. Some alternative processes—such as photopolymers for photoresists, vesicular microfilm, and xerography—also evolved due to industrial utility, others did not for various reasons. The definition of an ‘blueprint’ has evolved from the Cyanotype (requiring roughly 2000–5000 W/m² for 1s exposure) to Diazo printing (around 60–80 W/m² for 1s exposure), and finally to xerography, which requires far less. For some industrial imaging applications—such as exposing diazo photoemulsions on screen printing for PCB fabrication or textiles printing(a halftone example), or 3D printing with photopolymers—DLP, LCD, and laser scanning are all effective. But as we hobbyists do those alt processes for just interest, as you say, historical tradition may over practicality, digital exposure methods may need more development.

Yezishu said:

finally to xerography

Click to expand...

Yes, although that's no longer a pure photochemical process, so conceptually a really major deviation from the original concept. However, I do agree with you that we could see it as a logical progression / technological evolution, in which case I'd add that the latest/most recent step in that chain is also inkjet as this is currently replacing xerography.

Of course, the key question is, as you also allude to, what the purpose of the technology is. In the case of hobby printmaking, much of that purpose is in the process itself, and thus, the sensible technological trajectories are often not the ones that will be preferred.

Still, it's interesting to see/read how there's crossover between different technologies and practices. I do personally find that interesting to 'passively consume', so to say.