Running HP inks through an Epson printer

[MHMG]

The colour inks would be to add colour (the a and b values in the Lab system) to the black-and-white luminance layer. Naturally, you'd want the colours to all be as durable as possible, hence the HP inks. Custom Cone pigment inks, or the Symphonic system by American Inkjet, would be other options, but no-one's really done any durability testing on those.

Well, cyan and magenta contribute to the L* values significantly, yellow not so much, so as you add color pigments you have to back down on the UCR/GCR with the gray and black inks or you'd overshoot the desired L* values. As for ConeColor and Symphonic pigmented ink sets, I now have a couple samples printed with ConeColor in test (ID#s 277 and 278 in the database, compare to 276 and 279 which use Epson OEM ink on same media). Submitted by an AaI&A member that uses it. Testing is still in early stages, but ConeColor magenta seems to be weaker than Epson's Magenta by a significant margin. Yellow may be a little worse, too, but magenta is the bigger issue with the ConeColor set.  No AaI&A member that I'm aware of uses the Symphonic Inks by American Inkjet. I'd certainly be willing to test, but no one has offered to submit, and I don't know personally know anyone who actually uses the Symphonic inks.


cheers,
Mark
http://www.aardenburg-imaging.com

[shadowblade]

Well, cyan and magenta contribute to the L* values significantly, yellow not so much, so as you add color pigments you have to back down on the UCR/GCR with the gray and black inks or you'd overshoot the desired L* values.

It's a custom inkset, so you'd need to create a custom RIP for it through a series of test prints.

What about backing off on the full-strength cyan and magenta and using the light cyan and light magenta instead, with the Cone inks creating the L value (or Brightness) and the colour inks essentially 'tinting' it to achieve the right hue and saturation?

As for ConeColor and Symphonic pigmented ink sets, I now have a couple samples printed with ConeColor in test (ID#s 277 and 278 in the database, compare to 276 and 279 which use Epson OEM ink on same media). Submitted by an AaI&A member that uses it. Testing is still in early stages, but ConeColor magenta seems to be weaker than Epson's Magenta by a significant margin. Yellow may be a little worse, too, but magenta is the bigger issue with the ConeColor set.

I saw those, which is why I'm a little hesitant on the ConeColor inks. The JonCone Studio inks, as far as I know, are different, and custom-made.

No AaI&A member that I'm aware of uses the Symphonic Inks by American Inkjet. I'd certainly be willing to test, but no one has offered to submit, and I don't know personally know anyone who actually uses the Symphonic inks.

A pity - I saw a print using the Symphonic system the other day, next to the same photo printed using HP and Epson printers, as well as a Fujiflex print, and it looked fantastic. Sharp, saturated and wide-gamut. It would be very interesting to know how durable these inks are.

But, as with anything else that's printed using Roland printers, they aren't as common, due to the price of the printer!

[MHMG]

Btw, pigment-dye hybrid sets are pretty easy to pick out in light fade testing. As Shadowblade noted, if the the dye component fades more quickly then the remaining pigment "backbone" proceeds to fade at a much slower rate, so a rate bifurcation in the I* metric response (or delta E, or density) is a key to identifying this hybrid behavior. I've heard it said many times by many people that Epson Utrachrome pigments are suspected to contain dyes as well. If they do, it can't be much because they don't give a bifurcated fade response.  A bifurcation might be hard to detect for the PK and MK black, because the PK and MK are throttled by the driver so that they only get used in the the dmax areas of the print. The excess black ink at that level gives sacrificial protection properties even to dyes such that we wouldn't easily detect the dye-pigment hybrid bifurcation behavior for the max black image areas. As for the rest of the image, I've seen no evidence in fade testing to suggest that Epson is spiking the other Ultrachrome pigments with more fade prone dyes, certainly not to any extent that produces a faster initial fade rate for the Ultrachrome pigmented ink sets. The Claria dye ink set on the other hand, notably the yellow dye, does have a very non linear fade response, but it's well noted by Epson technical literature as a synthesized dye with quasi pigment-like behavior, so no surprise there. The Claria black dye is also highly subject to significant hue shift during light fading when used for full gray scale printing with a custom RIP like QTR, but again, Epson doesn't use the Claria black in this way in its OEM driver functionality, so it's not the customers' weak link in Claria dye print longevity as it would otherwise be compared to the Claria yellow dye performance. Perhaps most surprising is that if we allow fading to go to more advanced stages of fade that typical consumers might still accept, the Claria dye performance, IMHO, is superior to the Epson Ultrachrome yellow. The Claria yellow slows down in its fade rate after about 25% yellow density loss while the Ultrachrome yellow keeps on fading at it's initial rate. This is why published consumer-oriented fade tests for Claria ink rate so favorably in comparison to Epson Ultrachrome inks. Early stage yellow fade performance is significantly worse, but at later more advanced stages of fading the Epson yellow pigment fade catches up and then ultimately surpasses the Claria yellow dye in total elapsed fade.

best,
Mark
http://www.aardenburg-imaging.com

[MHMG]

It's a custom inkset, so you'd need to create a custom RIP for it through a series of test prints.

What about backing off on the full-strength cyan and magenta and using the light cyan and light magenta instead, with the Cone inks creating the L value (or Brightness) and the colour inks essentially 'tinting' it to achieve the right hue and saturation?

I saw those, which is why I'm a little hesitant on the ConeColor inks. The JonCone Studio inks, as far as I know, are different, and custom-made.

The Cone Carbon Sepia (now called Carbon K7) is the only full carbon inkset produced by Jon, AFAIK. As such, it's incredibly lightfast and this is well demonstrated in the AaI&A database. But this Carbon K7 longevity performance is in a class by itself and thus not to be confused with all the other monochrome Piezography shades that must add colored pigments to neutralize the warm full carbon hue. These tints are most likely incorporating the Cone magenta pigment now showing weakness in the full Color set. Magenta colorant dropping out first leaving full carbon backbone in the print explains the typical initial green shift we see in the testing of the other Piezography ink sets followed by a slowing down and stabilization in the overall fade rate after that (in other words, a classic fade rate bifurcation due to mixing of a less stable colorant with a more stable one to create the ink's desired initial hue and chroma values). Hence, the more neutral piezography sets are only moderately light fast, not a disaster by any means, but one is nevertheless better off with HP, Canon, and Epson OEM black inks as far as light fade resistance is concerned when trying to print neutral or near neutral B&W. Even Epson ABW mode with much magenta and yellow added to create a heavy sepia image tint still outperforms the majority of the Piezography ink sets except the full carbon K7.  Please note, I'm remarking on light fade performance.  Point is already well taken and duly noted that the 6 and 7 channel Piezography image quality is a very good incentive in its own right for discerning printmakers to want to use the Piezography ink systems, but the printmaker must nevertheless weigh the perceived nuances in image quality against the moderate but not great light fade resistance of many of the Piezography blends.

cheers,
mark
http://www.aardenburg-imaging.com

[shadowblade]

The Cone Carbon Sepia (now called Carbon K7) is the only full carbon inkset produced by Jon, AFAIK. As such, it's incredibly lightfast and this is well demonstrated in the AaI&A database. But this Carbon K7 longevity performance is in a class by itself and thus not to be confused with all the other monochrome Piezography shades that must add colored pigments to neutralize the warm full carbon hue. These tints are most likely incorporating the Cone magenta pigment now showing weakness in the full Color set. Magenta colorant dropping out first leaving full carbon backbone in the print explains the typical initial green shift we see in the testing of the other Piezography ink sets followed by a slowing down and stabilization in the overall fade rate after that (in other words, a classic fade rate bifurcation due to mixing of a less stable colorant with a more stable one to create the ink's desired initial hue and chroma values).

That's what I suspected. In other words, the coloured pigments quickly fade away, leaving the pure carbon print behind as a permanent print that essentially never fades. One may as well have just printed with the pure carbon inks in the first place.

The thing is, carbon nanoparticles, like all other nanoparticles, can produce different hues, depending on the exact size of the particle. By grinding the carbon to different sizes, you could produce a range of pure carbon inks varying in hue from very warm to very cool, and they'd all be as stable as each other. Similarly, using nanoparticles of gold, silver sulfides and selenides, and other chemically-unreactive and UV-resistant materials, it's possible to produce every shade and hue from bright red, to greens and blues, all the way to deep violet - and even infrared and ultraviolet 'colours', if you wanted to.

I'd say that, for archival and ultra-stable inks at least, nanoparticle 'pigments' relying on the quantum properties of light and interactions with small particles of chemically- and UV-inert, non-catalysing substances, are the future, rather than traditional chemical pigments.

Hence, the more neutral piezography sets are only moderately light fast, not a disaster by any means,

I guess it depends what standards of lightfastness you're using. Certainly, they'll quickly change colour, to that of a pure carbon ink. But, once there, it will stay that way forever - the final image isn't going anywhere.

but one is nevertheless better off with HP, Canon, and Epson OEM black inks as far as light fade resistance is concerned when trying to print neutral or near neutral B&W. Even Epson ABW mode with much magenta and yellow added to create a heavy sepia image tint still outperforms the majority of the Piezography ink sets except the full carbon K7. Please note, I'm remarking on light fade performance.  Point is already well taken and duly noted that the 6 and 7 channel Piezography image quality is a very good incentive in its own right for discerning printmakers to want to use the Piezography ink systems, but the printmaker must nevertheless weigh the perceived nuances in image quality against the moderate but not great light fade resistance of many of the Piezography blends.

Is it?

As far as I am aware, the problem is the lightfastness of the coloured pigment component in the neutral and cool inks. The black carbon pigment component of the inks is more durable than anything else out there. So, instead of using HP/Canon/Epson black inks, would it not be more sensible to continue using the ultra-permanent carbon Cone inks, toning them to a more neutral colour using the ultra-stable HP inks, rather than using the OEM inks alone? That way, the colour would still slowly shift to the warm black of pure carbon (but much more slowly than the Cone non-pure-carbon inks), but, unlike with the OEM inks, once they turned warm black, they would stay that way forever without fading.

Maybe, for neutral or cool prints, the answer isn't so much in changing the ink as it is in changing the paper. A custom paper, made with 100% cotton plus titanium dioxide or baryta particles incorporated into the paper itself (not just as a fragile layer on the surface) would have a much cooler white tone than standard non-OBA papers, and would give the overall image a cooler or more neutral tone, even using the same pure-carbon pigments.


***

The other reason for trying to put HP inks through an Epson print head is that, although the Vivera inks are demonstrably better than the Epson Ultrachrome inks, the thermal inkjet print process is quite inferior to Epson's piezo-inkjet system. The piezo system gives much finer control of ink droplet volume, allowing Epson heads to produce as many as seven different dot sizes (in Mimaki and Roland printers), allowing for smooth tonality, expanded gamut (particularly in the light tones) and true 1440dpi output all the way into the highlights - something that just isn't possible using thermal inkjet technology. Not only would putting Vivera inks through a piezo head, if it were possible, give you greater permanence than Epson inks, but, in theory at least, it should significantly improve the colour gamut of HP inks.

[MHMG]

I guess I should have been more specific in my earlier remarks about how Cone appears to be neutralizing the warm carbon ink set in order to give customers ink sets with cooler hues. Although paper chemistry does contribute to the final monochromatic hue imparted by a particular ink set, the very warm brown hue of the Carbon K7 ink still needs to be neutralized by blending a blue color component. Cone's ink formulator could do that by finding a satisfactory blue pigment or by combining both cyan and magenta pigments in appropriately blended quantities. Cone appears to be doing the latter in his more neutral piezography ink sets. A practical reason for taking the latter approach is that it is cost effective for a small company if magenta and cyan encapsulated pigmented ink batches are already in production.  So, now the image printed with such a cyan and magenta neutralized hue will fade towards greenish gray first as the magenta fades preferentially faster and while the cyan is fading at a much slower rate. But the magenta is moderately light fast, so this early stage of deterioration does not occur as quickly as you seem to believe.  Hence, your the reasoning that the image may eventually fade to the base hue imparted by highly stable carbon backbone is correct in theory, but it definitely won't happen in any time frame to make it a useful outcome. There will be a very long period of time where hundreds of megalux hours of more light exposure will be required to burn out both cyan and magenta in order to reach that full carbon backbone. Image Information content will be largely intact throughout this process, but the artist's original intent (i.e., the color of the monochromatic print he or she chose) will be compromised significantly as the aging process unfolds.  And by the time that much light exposure takes place, light, heat, humidity, and air pollution will also have taken its toll on the paper, even the best cotton papers with the most inert whitening and sizing agents available.

It's important not to confuse thermodynamics with kinetics in discussions like this. Thermodynamics deals with equations of state, in other words an initial state and a final state, but not the time it takes for the reaction to take place.  For example, the field of thermodynamics predicts that my wife's diamond ring will eventually revert to carbon black, but reaction kinetics tells us it will take hundreds of millions of years at temperatures and pressures where humans routinely keep their jewelry. I don't think we need to worry about diamond rings losing their shine any time soon.  Similarly, I don't envision a succession of museum curators documenting the fading curve of a print made with a Cone neutral ink set all the way to where only the pure carbon pigment remains.  What will be seen within a generation or so of cool-toned Piezography prints on display are ones that have acquired a less desirable greenish-gray hue and thus deviating to varying degrees from what the artist originally intended.

cheers,
Mark

[shadowblade]

I guess I should have been more specific in my earlier remarks about how Cone appears to be neutralizing the warm carbon ink set in order to give customers ink sets with cooler hues. Although paper chemistry does contribute to the final monochromatic hue imparted by a particular ink set, the very warm brown hue of the Carbon K7 ink still needs to be neutralized by blending a blue color component. Cone's ink formulator could do that by finding a satisfactory blue pigment or by combining both cyan and magenta pigments in appropriately blended quantities. Cone appears to be doing the latter in his more neutral piezography ink sets. A practical reason for taking the latter approach is that it is cost effective for a small company if magenta and cyan encapsulated pigmented ink batches are already in production.  So, now the image printed with such a cyan and magenta neutralized hue will fade towards greenish gray first as the magenta fades preferentially faster and while the cyan is fading at a much slower rate. But the magenta is moderately light fast, so this early stage of deterioration does not occur as quickly as you seem to believe.  Hence, your the reasoning that the image may eventually fade to the base hue imparted by highly stable carbon backbone is correct in theory, but it definitely won't happen in any time frame to make it a useful outcome. There will be a very long period of time where hundreds of megalux hours of more light exposure will be required to burn out both cyan and magenta in order to reach that full carbon backbone. Image Information content will be largely intact throughout this process, but the artist's original intent (i.e., the color of the monochromatic print he or she chose) will be compromised significantly as the aging process unfolds.  And by the time that much light exposure takes place, light, heat, humidity, and air pollution will also have taken its toll on the paper, even the best cotton papers with the most inert whitening and sizing agents available.

Looking at the test results, it seems that the neutral and cool blacks fade to an unacceptable, sickly-green hue with just a few megalux hours of light exposure. Certainly not something to rely on...

It's important not to confuse thermodynamics with kinetics in discussions like this. Thermodynamics deals with equations of state, in other words an initial state and a final state, but not the time it takes for the reaction to take place.  For example, the field of thermodynamics predicts that my wife's diamond ring will eventually revert to carbon black, but reaction kinetics tells us it will take hundreds of millions of years at temperatures and pressures where humans routinely keep their jewelry. I don't think we need to worry about diamond rings losing their shine any time soon.  Similarly, I don't envision a succession of museum curators documenting the fading curve of a print made with a Cone neutral ink set all the way to where only the pure carbon pigment remains.  What will be seen within a generation or so of cool-toned Piezography prints on display are ones that have acquired a less desirable greenish-gray hue and thus deviating to varying degrees from what the artist originally intended.

cheers,
Mark

I don't see a succession of curators doing it either. With the print on display, it looks like the green shift will only take a few years. Sure, it will take much longer - maybe a few decades to a century of display - for the green shift to turn to a more pleasant carbon tone. By then, they'd probably have thrown out the print due to the unacceptable colour shift...

May as well print it in carbon in the first place - and, if you want to tone it, use colour inks on top of it that have better permanence and better fade characteristics (i.e. so that magenta doesn't go first and turn the print a nasty green colour).

[MHMG]

Looking at the test results, it seems that the neutral and cool blacks fade to an unacceptable, sickly-green hue with just a few megalux hours of light exposure. Certainly not something to rely on...

I'm pleased that the Aardenburg lightfade database has allowed you to make an informed decision about various printer/ink/media combinations. Helping printmakers make appropriate ink and media choices that meet both their initial image quality and print permanence expectations over time has always been the goal of this crowd-sourced testing program. I would never have been able to acquire the variety of print samples now represented in the database without direct help from the digital printmaking community.

cheers,
Mark
http://www.aardenburg-imaging.com

[shadowblade]

I'm pleased that the Aardenburg lightfade database has allowed you to make an informed decision about various printer/ink/media combinations. Helping printmakers make appropriate ink and media choices that meet both their initial image quality and print permanence expectations over time has always been the goal of this crowd-sourced testing program. I would never have been able to acquire the variety of print samples now represented in the database without direct help from the digital printmaking community.

cheers,
Mark
http://www.aardenburg-imaging.com

Definitely - and, as you say, this sort of research, by individuals rather than large institutions beholded to corporations, is never going to attract funding, nor will reliance on donations and subscriptions come up with the funds required. I detailed another internet-forum-based microfinance idea here - let me know what you think about its feasibility. After all, it's work that every photographer, bar short-cycle news journalists, will have a vested interest in.

Now, I just need to find myself a cheap/outdated/second-hand Epson printer that I can use for proof-of-concept testing with HP inks... if it works, I'd then move testing to a Roland to put together individual inksets and custom RIPs - but there's no point in wasting money on the Roland in the first place if the ink doesn't even go through the print head.

[MHMG]

I detailed another internet-forum-based microfinance idea here - let me know what you think about its feasibility. After all, it's work that every photographer, bar short-cycle news journalists, will have a vested interest in.

I do appreciate your suggestion, and I do think somebody could make it work quite successfully, but I'm not that person. I'm just an imaging scientist by training doing print preservation research because I care about this topic for my own printmaking efforts and wanted to find a way to share my research with others. That the Aardenburg testing protocol and the science of the I* metric is second to none is where I find the most satisfaction in this research. However, to also wear a full time marketing hat in order to promote the project beyond what a field of dreams "build it and they will come" notion naively believes is possible requires that one make marketing and promotion pretty much a full time job. If I tried to tackle such a marketing effort, I wouldn't have any time left over to do the research. For me to hire someone to do this job on my behalf takes seed money I don't have.

The Aardenburg digital research program continues... just at a pace that I can handle largely as a volunteer effort and grateful for whatever donations and sample submissions come along to add to the body of the published work.


best,
Mark
http://www.aardenburg-imaging.com

[shadowblade]

I do appreciate your suggestion, and I do think somebody could make it work quite successfully, but I'm not that person. I'm just an imaging scientist by training doing print preservation research because I care about this topic for my own printmaking efforts and wanted to find a way to share my research with others. That the Aardenburg testing protocol and the science of the I* metric is second to none is where I find the most satisfaction in this research. However, to also wear a full time marketing hat in order to promote the project beyond what a field of dreams "build it and they will come" notion naively believes is possible requires that one make marketing and promotion pretty much a full time job. If I tried to tackle such a marketing effort, I wouldn't have any time left over to do the research. For me to hire someone to do this job on my behalf takes seed money I don't have.

The Aardenburg digital research program continues... just at a pace that I can handle largely as a volunteer effort and grateful for whatever donations and sample submissions come along to add to the body of the published work.


best,
Mark
http://www.aardenburg-imaging.com

Well, I think it's definitely doable as a forum effort, and very worthwhile - pretty much every photographer here has a vested interest in it. The point is, it's a group effort - obviously your strength is research, rather than fundraising, marketing or project coordination, which would take time and effort away from the research we're trying to fund anyway. So, other people can put the project together, organise printing, do the desktop publishing work, collect information (names of photos and photographers, descriptions to accompany each photo, high-resolution versions of the photos themselves), run the contest for entries across multiple forums, while pouring the final proceeds towards Aardenburg. I'd be more than happy to coordinate such a project, if there were sufficient interest out there (whether the interest is in funding ink permanence research, or just in producing a LuLa photo-book). It's basically the only way such research will get funded, short of a private benefactor with an interest in photography. Anyone care to introduce Bill Gates or Warren Buffett to photography?

Regarding your research methods, I note that your testing grid contains a lot of standard colours, but no 'pure' samples of each of the inks used by a printer. Would it not be worth having an extra square for each 'pure' ink (i.e. an Epson Ultrachrome K3 on matte test chart would have extra squares of MK, LK, LLK, C, LC, M, LM, Y) in order to have raw data about the performance of each individual ink, so that we can see what the strong points and weak points of each inkset are? For instance, yellow is Epson's well-known weak point, while magenta is Cone's weak point. So, why not replace Epson's yellow with Cone's yellow and build a custom RIP around that, or use Epson's magenta with ConeColor inks? We're not really sure what the weak and strong points of other inksets are - if we did, we'd have good information for building custom inksets that far surpass OEM inksets.

Also, I'm not sure if this is your area or not, but do we have any information on the solvent chemistry of each of the ink manufacturers? Given that inkjet printing involves layering lots of dots from lots of different inks together, it's important for the solvents used in different inks to be compatible with each other. This isn't a problem if you're just sticking with the OEM inkset, but, if you mix-and-match for better results, you'd have find combinations that work with each other. How much would it cost to carry out gas chromatography on each of the solvents out there (Epson K3, Ultrachrome HDR, Cone (I would assume the Piezography and ConeColor inks use the same solvent), Symphonic, Canon Lucia, HP Vivera and others) to work out what exactly is in them? If we knew the solvent makeup of each manufacturer, it would be easy enough to precipitate out the durable yellow pigments in Cone or HP ink and suspend them in an Epson solvent mixture, to run through an Epson head with the other Ultrachrome inks...

Incidentally, have you done any testing on the Roland Eco-Sol Max and Eco-Sol Max 2 solvent inks?

[MHMG]

Regarding your research methods, I note that your testing grid contains a lot of standard colours, but no 'pure' samples of each of the inks used by a printer. Would it not be worth having an extra square for each 'pure' ink (i.e. an Epson Ultrachrome K3 on matte test chart would have extra squares of MK, LK, LLK, C, LC, M, LM, Y) in order to have raw data about the performance of each individual ink, so that we can see what the strong points and weak points of each inkset are? For instance, yellow is Epson's well-known weak point, while magenta is Cone's weak point. So, why not replace Epson's yellow with Cone's yellow and build a custom RIP around that, or use Epson's magenta with ConeColor inks? We're not really sure what the weak and strong points of other inksets are - if we did, we'd have good information for building custom inksets that far surpass OEM inksets.

Firstly, thank you for you wilingness to get involved with the Aardenburg effort. I appreciate it, and indeed have had a few volunteers helping me with the day to day testing over the past few years, but help with the fundraising challenge has never been offered before.

Re: testing of pure colors rather than blended colors, the vivid Macbeth color patches such as C,M,Y, R, G, B call out "pure enough" ink usage that we don't miss out on how these inks would behave if used as pure color channels. For example, it is why we've learned that Epson's yellow is the weak link, that Cone's got some weakness in magenta as well, and that HP has amazingly balanced fade amongst all of it's Vivera pigmented inks such that lightness and contrast drop gently while overall color balance is being nicely maintained. I've seen no evidence yet that the Aardenburg 30 patch color target is missing anything important, so adding more patches would certainly give us more nuanced results to look at but the conclusions drawn would be pretty much the same.  Additionally, pure color channels would have to be driven with special RIPs in many instances since the OEM drivers don't give us access to said channel purity. However, it's the blends where most of the bad interactions occur, so one really wants a test that produces the patches exactly the way any printmaker would print them, ie. with the intended driver or RIP provided by the printer manufacturer or selected by the enduser. NO deviation from standard practice.  Hence, I designed a test method that avoids the need for special calibration procedures in order to print the target (as other light fade testing methods require). Had I elected to use special RIPs to study the inks in more isolated manner, the member-submitted sample making would be unworkable and we'd lose more information about today's inks and papers than we've already gained by this crowd-sourced effort.  Lastly, the lighter blends appearing in the 24 Macbeth color set also neatly exercise the issue of cross-over between low concentration and high concentration color channels. For example, HP's first effort at in improved dye set several years back used a magenta that tested well by itself but exhibited catalytic fading not with yellow but with cyan. This catalytic fade problem was nicely detected in the AaI&A blue and purple patches (i.e, patches that blend both cyan concentrations with both magenta concentrations). My vindication that I'd found the problem came later when HP retrofit a new magenta+ dye into it's line of desktop printers. They were aware of the problem too, even though the testing lab they routinely used for marketing "certification" used a target which provided pure color ink patches but no blend colors for red, green, blue, etc, and hence did not factor this weak link into the fade rating of the original ink set.

Shadowblade, I admire your enthusiasm and I realize you've asked more questions as well, but if I keep going, other forum users will be put to sleep

cheers,
mark
http://www.aardenburg-imaging.com

[Ernst Dinkla]

***

The other reason for trying to put HP inks through an Epson print head is that, although the Vivera inks are demonstrably better than the Epson Ultrachrome inks, the thermal inkjet print process is quite inferior to Epson's piezo-inkjet system. The piezo system gives much finer control of ink droplet volume, allowing Epson heads to produce as many as seven different dot sizes (in Mimaki and Roland printers), allowing for smooth tonality, expanded gamut (particularly in the light tones) and true 1440dpi output all the way into the highlights - something that just isn't possible using thermal inkjet technology. Not only would putting Vivera inks through a piezo head, if it were possible, give you greater permanence than Epson inks, but, in theory at least, it should significantly improve the colour gamut of HP inks.

Part of the excellent fade resistance properties of HP Vivera pigment prints is in the way the ink mixing is controlled by the HP Z media presets. If you put a microscope on a Z3100/3200 matte print you will see there is no composite grey mix in either color or grey mode print. Build with 11 inks that get a 100% black generation and clever N-color hue substitution.  To get a similar control of even more inks asks for a lot of skill in all the disciplines involved; the custom inkjet printer system + a RIP that can handle that inkset + the profiling that can cope. I mean in PRACTICE. The HP engineers in Barcelona were not amateurs. How the gamut would improve with that custom system has to be explained to me.

I would not be afraid of HP Vivera pigment inks in an Epson. Paul Roark has used the HP PK Vivera + mixes in Epson desktop models with even finer droplet heads than the Epson wide formats have. No issues at all as I understand it. Whether the Epson wide format printing with the 3.5 picoliter droplets (only size used at the highest resolutions!!!!) is that much better in image quality than the Canon wide format 4 picoliter or HP Z 4/6 picoliter droplet has to be seen. The latest range of Epson wide formats often show inconsistency in print results with even the OEM inks. Users that have to rely on the highest resolution settings/slowest speeds to get rid of banding/waves, even on papers that can not deliver the image quality related to those resolution settings. Image quality has a lot to do with the dithering/weaving algorithms, the resampling/print sharpening functions and above all good inkjet paper coatings. There are no published MTF resolution numbers for coating/ink/printer combinations but my gut feeling is that the best are already reached beneath 600 PPI input and anything above that is not recorded in the print anymore. I might be wrong but even then I rather use a slightly lower image quality printer that delivers a consistent quality at both 300 PPI and 600 PPI input, today and tomorrow, without extensive cleaning steps in the morning or between print runs. And allow an easy exchange of one head to get that consistency again when it fails.

About uncoated papers mentioned in another thread here; There is an interesting discussion on the Yahoo digital B&W forum where users of alternative photographic processes write that even the papers from 500 years old mills changed in the past decades due to paper process changes. What worked one day for a platinum print could not be repeated when changes at the factory were made for all the batches later on. That will not tell us either whether inkjet coated papers or uncoated papers will last longer but at least raises some questions about the aging of the mill's papers then and now.

Paul Roark has done some fine jobs with MIS Eboni MK inkjet inks on uncoated Arches papers that showed good fade resistance. The Eboni MK is in itself the most neutral pure carbon black MK ink of all discussed here, more neutral than the HP MK too.


--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.

[shadowblade]

Part of the excellent fade resistance properties of HP Vivera pigment prints is in the way the ink mixing is controlled by the HP Z media presets. If you put a microscope on a Z3100/3200 matte print you will see there is no composite grey mix in either color or grey mode print.

Which is consistent with test results showing minimal fading in the HP black inks. The lack of faster-fading colour pigments in HP's pure grey shades certainly helps with their fade resistance.

But HP's colour inks, even without any black or light black, also show better permanence than colour inks from other manufacturers. So it's not just due to the black/grey.

Build with 11 inks that get a 100% black generation and clever N-color hue substitution.  To get a similar control of even more inks asks for a lot of skill in all the disciplines involved; the custom inkjet printer system + a RIP that can handle that inkset + the profiling that can cope. I mean in PRACTICE. The HP engineers in Barcelona were not amateurs. How the gamut would improve with that custom system has to be explained to me.

Tonality would improve because, instead of using just two densities of black ink in a single dot size, you'd have six or seven different densities of black ink, available in three (Epson) to seven (Roland, ?Mimaki) different dot sizes. Whatever the quality of the RIP, there's only so much tonality you can achieve in the luminance channel if you only have two inks and one dot size to interpolate lighter colours from.

This also applies to colour inks - more dot sizes available for colour inks means that lighter colours can be simulated by using just as many dots as darker colours, but smaller dots, rather than a few widely-separated large dots.

When you're using black inks for L values, and coloured inks primarily to control a and b values, this translates to smoother gradients, particularly when you're dealing with gradients of brightness and colour saturation, with colour hue remaining largely the same (e.g. skies, or anything printed in monochrome).

Gamut would also increase because, with the ability to place lots of small dots (both coloured and ultra-light black) rather than a fewer number of larger dots, one can get much closer to the white of the background paper without actually becoming white.

Epson's K3 and HDR inksets have also shown that, with variable dot size, you can exceed the gamut of HP's 12-ink system with just 8 or 10 inks.

Naturally, this would require capable RIP software, but it's not insurmountable - custom RIPs can do better than OEM RIPs even with standard inksets.

I would not be afraid of HP Vivera pigment inks in an Epson. Paul Roark has used the HP PK Vivera + mixes in Epson desktop models with even finer droplet heads than the Epson wide formats have. No issues at all as I understand it.

This is good to hear - it's what I suspected, based purely on the mechanisms by which the various print heads work, but it's good to know that someone else has done it successfully. Maybe I can skip the testing on a low-end Epson pigment printer, and move straight to the Roland...

Whether the Epson wide format printing with the 3.5 picoliter droplets (only size used at the highest resolutions!!!!) is that much better in image quality than the Canon wide format 4 picoliter or HP Z 4/6 picoliter droplet has to be seen. The latest range of Epson wide formats often show inconsistency in print results with even the OEM inks. Users that have to rely on the highest resolution settings/slowest speeds to get rid of banding/waves, even on papers that can not deliver the image quality related to those resolution settings. Image quality has a lot to do with the dithering/weaving algorithms, the resampling/print sharpening functions and above all good inkjet paper coatings. There are no published MTF resolution numbers for coating/ink/printer combinations but my gut feeling is that the best are already reached beneath 600 PPI input and anything above that is not recorded in the print anymore. I might be wrong but even then I rather use a slightly lower image quality printer that delivers a consistent quality at both 300 PPI and 600 PPI input, today and tomorrow, without extensive cleaning steps in the morning or between print runs. And allow an easy exchange of one head to get that consistency again when it fails.

I normally print at 300ppi on HP, and 360ppi for Epson. I usually print large (large enough that upsampling is required even from 80MP files), and am definitely not convinced either that, at these sizes, 600/720ppi files give any better results than 300/360ppi files. Doubly so for papers with any sort of texture.

Banding/waves and consistency with some Epson printers may or may not be an issue, but Epson print heads are also used in Mimaki, Roland and other printers - even if certain Epson printers and RIPs may not be ideal, the proof of concept (running HP inks through Epson heads) is still there, and other printers utilising Epson heads may be able to give better results. Roland printers, at least, have a self-maintenance schedule that keeps the heads running and clog-free so long as the machine is turned on.

About uncoated papers mentioned in another thread here; There is an interesting discussion on the Yahoo digital B&W forum where users of alternative photographic processes write that even the papers from 500 years old mills changed in the past decades due to paper process changes. What worked one day for a platinum print could not be repeated when changes at the factory were made for all the batches later on. That will not tell us either whether inkjet coated papers or uncoated papers will last longer but at least raises some questions about the aging of the mill's papers then and now.

The manufacturing process and exact makeup of the paper may be a different from 500 years ago, but the basic product is still the same - matted cellulose fibres, derived from cotton, firstly soaked in an internal sizer and the fibres then coated with an external sizer to prevent water absorption - and has proven stability, even where exact manufacturing processes and chemistry have varied worldwide. After all, Arches watercolour, Indian khadi and Japanese washi papers are all made very differently, but have all proven to have great longevity. One cannot say the same about the layer of polyvinyl alcohol binder and silica particles which make up the inkjet receptive layer.

Paul Roark has done some fine jobs with MIS Eboni MK inkjet inks on uncoated Arches papers that showed good fade resistance. The Eboni MK is in itself the most neutral pure carbon black MK ink of all discussed here, more neutral than the HP MK too.

Is it pure carbon (like the Cone Carbon inks), or is it carbon mixed with another pigment (like the Cone Neutral and Selenium inks)? If it's pure carbon and more neutral than the Cone Carbon inks, I suspect this would be because of different particle size as compared to Cone.

Of course, fade resistant tests with pure carbon inks may not be the best test of the longevity of inks on uncoated paper (which is, of course, quite separate from the longevity of the paper itself) - after all, pure carbon is essentially fade-proof. Tests using coloured inks may be far more telling. I suspect it would be much more related to the UV stability of the pigments than tests on coated, microporous paper - without the huge surface area of coated papers, fading due to atmospheric contaminants should be relatively negligible.

[Ernst Dinkla]

With the highest print resolutions set on all three brands wide formats you will get the smoothest gradations, the best image detail, the least banding. In that setting they all use droplets of about 4 picoliter: 3.5 for the Epsons, 4 for the Canons, 4-6 for the HPs. No larger droplets are used. If you go to lower print resolutions the Epson starts to throw in larger droplets next to the 3.5 picoliter size, 11 picoliter and larger. It does that for speed. With 360 nozzles per channel there is no other choice to compete with Canons etc that have >2000 nozzles per channel that can print in similar firing rates the Epson nozzles can handle. The Canons and HPs still squirt their 4 or 4-6 picoliter droplets at the lower resolution settings. Epson often boasts about its droplet size variation, it had more meaning when the competition could not make droplets as small as Epson's minimum droplet size.

The HP Z3100/3100 use 4 black/grey inks in their best matte art paper media presets and 3 black/grey inks on photo gloss papers. In color and B&W mode of the driver. Yes, you could go the piëzography route for better gradients but it will not improve gamut and you need a black generation as good as HP's media presets do it to keep similar fade resistance.

Gamut increases due to droplet size will be very limited if the ink opaqueness remains the same. I very much suspect that Epson's larger gamut even with a smaller inkset is mainly the result of more transparency in the inks/pigments they use. Let us say a step into the direction of Claria, the Epson "Dye". In that case subtractive color mixing improves. Related to that is a change in gamut with bleeding/dotgain on dots as the white background diminishes in that process and the transparency increases at the dot's edge. Some articles exist on what happens then. Smaller droplets have another ratio on dot area/circumference that should increase dotgain if the paper coating allows that, there is less ink to absorb per dot so it might as well work out differently. Important is dot addressing which does not improve the smaller the droplet is. So the variable dot size has in my opinion nothing to do with the Epson gamut quality but the dot addressing may benefit of it. But again; at the highest resolutions only the smallest droplet is used.

The Eboni is a pure carbon and of the pure carbon inks the most neutral. It is true that particle sizes have an influence on the color of the carbon but I would guess that different processes to produce the carbon particles have an influence too. MIS gets its base inks from Image Specialists. Check Paul Roark's work anyway as it has been done with a good brain and eye for what matters and a lot of trial and error, where the discussion here is very much theoretical. Check some Aardenburg Lab measurements of the MK black inks where HP MK ink and Eboni ink are used on the same paper.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.

[shadowblade]

With the highest print resolutions set on all three brands wide formats you will get the smoothest gradations, the best image detail, the least banding. In that setting they all use droplets of about 4 picoliter: 3.5 for the Epsons, 4 for the Canons, 4-6 for the HPs. No larger droplets are used. If you go to lower print resolutions the Epson starts to throw in larger droplets next to the 3.5 picoliter size, 11 picoliter and larger. It does that for speed. With 360 nozzles per channel there is no other choice to compete with Canons etc that have >2000 nozzles per channel that can print in similar firing rates the Epson nozzles can handle. The Canons and HPs still squirt their 4 or 4-6 picoliter droplets at the lower resolution settings. Epson often boasts about its droplet size variation, it had more meaning when the competition could not make droplets as small as Epson's minimum droplet size.

Fair enough - I see how that could work... I was under the impression that Epson printers (and Iris, and Roland, and Mutoh) threw out dots of varying sizes onto the same print, improving tonality in the highlights using smaller droplets, by being able to render light colours using many small drops of ink, instead of larger drops separated by wide gulfs of paper white.

The HP Z3100/3100 use 4 black/grey inks in their best matte art paper media presets and 3 black/grey inks on photo gloss papers. In color and B&W mode of the driver. Yes, you could go the piëzography route for better gradients but it will not improve gamut and you need a black generation as good as HP's media presets do it to keep similar fade resistance.

The Piezography inksets seem to blow away the HP inkset in black-and-white, though, and the HP inkset is already pretty much the best 'standard' inkset out there for black-and-white printing. Why would fade resistance be worse with pure-carbon piezography inks, if shades of grey are made up not of coloured inks, but just of different dilutions of carbon ink?

The Eboni is a pure carbon and of the pure carbon inks the most neutral. It is true that particle sizes have an influence on the color of the carbon but I would guess that different processes to produce the carbon particles have an influence too. MIS gets its base inks from Image Specialists. Check Paul Roark's work anyway as it has been done with a good brain and eye for what matters and a lot of trial and error, where the discussion here is very much theoretical. Check some Aardenburg Lab measurements of the MK black inks where HP MK ink and Eboni ink are used on the same paper.

So, possibly the MIS inks would be a better candidate for the various 'black' inks than the Cone Piezography inks? Or possibly split-toning them with darker MIS inks and lighter Cone inks for a dedicated black-and-white inkset. It's also useful to know that he's had success combining MIS inks with HP Vivera inks (in an Epson printer, no less) - this would indicate that the solvents used in each inkset are compatible with each other, and that the individual inks from each set can be combined to form a custom inkset without compatibility problems.


The other problem with HP printers is their very limited paper thickness (0.8mm) compared to other printers on the market. Epson can take at least 1.5mm without modification, while Rolands can be modified to take truly stupidly-thick media!

[Ernst Dinkla]

Desktop printers with droplet sizes down to 1.5 picoliter but with the same inkset do not show better gamut than their wide format brothers.

I did not write that the extra grey inks will harm fade resistance. In color mode you have to create good black generation to get rid of CMYetc composite mixes that could harm fade resistance of a print, that fact does not change with more grey inks.

Paul did mixing tricks with ink medium, surfactants, etc to get the properties right for Epson heads. All documented though.

Maybe there is a misunderstanding. I do not write the replies to advocate the use of a HP Z3200 printer or any other brand printer but I try to explain that your Super Customised Solution asks for a lot of skill to get all what you want: fade resistance, wide gamut, highest image quality. Compromises are made in OEM solutions and compromises are accepted by users with any of the OEM choices, we simply get not all in one package. Sometimes a wider gamut is aimed at and less fade resistance accepted as the penalty.  Yes, the Zs are compromised on paper thickness and loading sheets is not their forte either. Your SCS does not include an integrated spectrometer for calibration and profiling like on the Zs but I guess it will after this message.

I have to repeat Mark D Segal's last words in this thread: Thanks - it will be interesting to follow your results.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
July 2013, 500+ inkjet media white spectral plots.

[wattsies]

This has been a very interesting discussion, especially in relation to the piezography inks, which I have been using. As a monochrome inkset, they do produce lovely results, but I have been concerned about their lightfastness.  Although as I tend to store my prints in dark storage (at least one copy anyway), perhaps this is not such a big deal.

The HP design jet printers are said to produce very good monochrome output, at least in terms of neutrality.  Where's that 17" HP printer that I can fit in my room!

[wattsies]

I should also note that silver gelatin prints, that we thought were the holy grail of longevity next to the platinum print, are not performing so well themselves in Mark's testing. While the manufacturers probably don't make their commercial decisions on this basis, at least for those of us interested in permanence the digital era seems to be giving us more and more opportunities to move forward.

[shadowblade]

This has been a very interesting discussion, especially in relation to the piezography inks, which I have been using. As a monochrome inkset, they do produce lovely results, but I have been concerned about their lightfastness.  Although as I tend to store my prints in dark storage (at least one copy anyway), perhaps this is not such a big deal.

Pure-carbon inks, from any manufacturer, are no doubt completely lightfast and resistant to chemical attack, and will likely outlast the paper - this has been verified by Aardenburg for both Cone and MIS inks. The problem is the durability of inkjet-coated paper itself - there's just no real evidence on the longevity of a thin layer of highly-porous silica suspended in polyvinyl alcohol. There's no point in the image lasting 1000 years if the paper doesn't.

Even the inks that aren't pure carbon, but are largely carbon and are used to make black-and-white prints, will decay to pure carbon as the non-carbon pigments fade, without losing any information.

The HP design jet printers are said to produce very good monochrome output, at least in terms of neutrality.  Where's that 17" HP printer that I can fit in my room!

The Z3200 comes in a 24" version, which isn't all that much bigger!

I should also note that silver gelatin prints, that we thought were the holy grail of longevity next to the platinum print, are not performing so well themselves in Mark's testing.

Silver gelatin prints never performed that well. They needed toning to completion with sepia, selenium or gold, or a split-tone combination of the three (a 1:19 'archival' bath in selenium doesn't do the trick - try 1:5, where you get a colour shift even in the highlights, for proper protective toning) in order to hold up.

Also, I'd consider the holy grail of longevity to be the carbon print, not the platinum print, due to platinum's destructive effect on the paper base. Platinum images will last forever. The catalytic effect of platinum, however, will form acids from atmospheric gases that will destroy the paper - many century-old platinum prints are in perfect condition, with no fading, but the papers they are printed on (and, in albums, the papers, they are in contact with) have become brittle or turned orange due to this catalytic effect. This isn't helped by the fact that, due to the acidic nature of the printing process, a platinum print cannot be made on buffered paper (although the acid can be washed out and the buffer re-established once the print has been made).