Resolution: Epson color inks Vs K7 Piezograph

[sanking]

I recently purchased a Piezograph K7 Selenium set and plan to change over my Epson 7600 soon from the Epson color inks to the Piezography set. One of the main reasons I purchased the Piezography set was to set up a dedicated printer for making digital negatives in anticipation that it would give better resolution than the Epson color inks. As we know, the maximum resolution of Epson printers is about 7 lp/mm, based on the fact that there is no improvement in sharpness in printing at file resolution of more than 360 ppi.
Before I make the change over I would like to print a resolution chart with the Epson color inks that I can later print with the Piezography set to compare results.

My question is, can someone point me to a good resolution chart for this purpose that I can download to use for the test?

Sandy King

[ric_kb]

these don't download, but are the "standard" charts mentioned often, used rarely...

http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productid=1790

this is a link to a PS version

http://photo.net/learn/optics/USAF1951.ps

richard.

[Ron Marshall]

Scroll down for the chart and instructions for use and interpretation:

http://www.clarkvision.com/articles/printer-ppi/index.html

[sanking]

Hi Ron and Richard,

Thanks for the links. I found a couple of useful test charts and am now printing my test prints with the Epson UC ink set.

Sandy

[Kirk Gittings]

FWIW, I recently (like three weeks ago) had a large digital print made from a scanned 4x5 b&w negative for a museum show. The service I use for large images runs both Epson and Piezography inks so I had identical prints made. I had seen first hand the very impressive resolution comparison tests that Cone had done and was curious to see if it was noticeable in my actual prints. In this case it wasn't. However this image did not have allot of fine detail. While I greatly preferred the tonalities of the Piezography inks, there was no apparent resolution advantage with this image that was visible even on very close inspection with reading glasses.

[sanking]

FWIW, I recently (like three weeks ago) had a large digital print made from a scanned 4x5 b&w negative for a museum show. The service I use for large images runs both Epson and Piezography inks so I had identical prints made. I had seen first hand the very impressive resolution comparison tests that Cone had done and was curious to see if it was noticeable in my actual prints. In this case it wasn't. However this image did not have allot of fine detail. While I greatly preferred the tonalities of the Piezography inks, there was no apparent resolution advantage with this image that was visible even on very close inspection with reading glasses.


You make a good point about the limitation of the original. The limit to resolution will always be the weakest link in the chain. In this case the fact that your image did not have a lot of fine detail would mask any difference between Epson and Piezography inks, if indeed they exist.

I had a similar experience in printing some of the resolution targets I downloaded from the net. It at first escaped my sense of logic that they were all file sized at 300 dpi, which essentially makes them useless for my purpose since even the Epsons inks are capable of ar least 360 dpi (7 lp/mm) of real resolution.

Fortunately one of the links (http://photo.net/learn/optics/USAF1951.ps) provided a file that I was able to rasterize to increase effective resolution so I am printing the tests again with the Epson UC ink set.

Sandy

[sanking]

Here is my first report on resolution with the Epson 7600. The tests were made printing with QTR, the file set to 1440 ppi, 720 ppi and 360 ppi. I used the Enhanced Matte Cool profile from the UC list of profiles, and set the printer to 2880 dpi and bi-directional printing.

Results were something of a surprise. From habit I have always set my file size at 720 at the print size but many of my friends who use Epson printers have told me that there is no difference in image quality with the file set at 360 ppi or 720 ppi.

That turned out to be wrong in my tests. Results with the file set at 360 ppi indicate a resolution of 7 lp/mm, just what I expected. However, with the file set at 720 ppi there was a significant increase in resolution to nearly 12 lp/mm. No additional improvement was seen with the file size at 1440 ppi.

In a few days I will change the printer over to the K7 Piezography ink set and repeat the tests.

Sandy

[Kirk Gittings]

Interesting, I look forward to the rest of your tests.


Fortunately one of the links (http://photo.net/learn/optics/USAF1951.ps) provided a file that I was able to rasterize to increase effective resolution so I am printing the tests again with the Epson UC ink set.

How does rasterizing increase effective resolution of the target?

[paulr]

Results with the file set at 360 ppi indicate a resolution of 7 lp/mm, just what I expected. However, with the file set at 720 ppi there was a significant increase in resolution to nearly 12 lp/mm. No additional improvement was seen with the file size at 1440 ppi.

Were there any differences that would be visible in a print, looked at without magnification?

I print at 720, just for the hell of it ... at least if the file has that much resolution in the first place. But the only differences I've been able to see have been in aliasing of certain slightly diagonal lines. Sometimes this is barely visible at 360 but never at 720.

My experiments have been on photorag, which may also be a limiting factor ... not sure.

[sanking]

Interesting, I look forward to the rest of your tests.



How does rasterizing increase effective resolution of the target?

Kirk,

Perhaps it did not. Looking at the original file again in Preview it looks like the resolution might have been there to begin with. However, the first time I opened it in Photoshop the resolution was quite low. Then I opened it again at the original file size, but at 5000 ppi, and I got a message, "Rasterization" in process, and it took a long time for the file to open. The fact that this is a high contrast resolution target with black lines on white background may have something with how it can open at different resolutions. I don't know for sure, perhaps you could download the file and see what is going on.

Sandy

[sanking]

Were there any differences that would be visible in a print, looked at without magnification?

I print at 720, just for the hell of it ... at least if the file has that much resolution in the first place. But the only differences I've been able to see have been in aliasing of certain slightly diagonal lines. Sometimes this is barely visible at 360 but never at 720.

My experiments have been on photorag, which may also be a limiting factor ... not sure.

There would probably not be a visible difference in looking at the comparison prints, which are 10X14" in size, at normal viewing distance of 10-12 inches since 7 lp/mm is already beyond the threshold of human vision. However, I am pretty sure the increased resolution would make a difference with some finely detailed subjects are larger print sizes.

Also, my goal here is to use the printer with the K7 Piezography ink set to print on Pictorico for making digital negatives for the carbon transfer process, which potentially is as sharp or sharper than the best silver gelatin printing. In other words, within limits carbon transfer can capture all of the detail in a negative, as can silver in most cases. I suspect that the printer and media (Pictorico) will still be the limit to resolution with carbon transfer, but if I can get up to 12-24 lp/mm of real resolution on the print that would be a significant improvement over the Epson with its color ink set.

And I would again emphasize that I am making the tests with QTR, not the Epson printer driver. The reason is that QTR, or some other RIP software, must be used with the Piezography ink set.

Sandy

[Tyler Boley]

the results between 360 and 720 do not surprise me, there is a dearth of misinformation out there on the inner tubes...
My assumption about this particular testing is that due to your specific needs, stepping up the performance may turn out more critical. Since the printing is for a neg, the mechanical nature of it's production, particularly in the densest areas which normally become our shadows and least visible, now become your delicate highlights, most visible.

The problem with placing judgments on comparisons in general is criteria and context needs to be applied. Which is better? Well, better for what? By what criteria?

It will be very interesting to get your eventual impression of results in a final fine print.
Tyler

[Vertex Ninja]

Kirk,

Perhaps it did not. Looking at the original file again in Preview it looks like the resolution might have been there to begin with. However, the first time I opened it in Photoshop the resolution was quite low. Then I opened it again at the original file size, but at 5000 ppi, and I got a message, "Rasterization" in process, and it took a long time for the file to open. The fact that this is a high contrast resolution target with black lines on white background may have something with how it can open at different resolutions. I don't know for sure, perhaps you could download the file and see what is going on.

Sandy

If the file is a vector graphic, then it is resolution independent until rasterized.

Thanks for the info regarding Epson's resolution. I've always used QImage to interpolate my files to 720 if the resolution of the file exceeded 360 at any given print size. The prints always looked slightly sharper(720 vs 360) on very close inspection, but I've never done a true resolution test like you.

These links are what got me to originally try the "finest detail" setting and I noticed it did make a slight difference, particularly with a loupe:p .
http://www.ddisoftware.com/qimage/quality/
http://ddisoftware.com/tech/printers/printer-resolution-some-tests-for-you-to-do/10/?wap2

[sanking]

If the file is a vector graphic, then it is resolution independent until rasterized.

Thanks for the info regarding Epson's resolution. I've always used QImage to interpolate my files to 720 if the resolution of the file exceeded 360 at any given print size. The prints always looked slightly sharper(720 vs 360) on very close inspection, but I've never done a true resolution test like you.

These links are what got me to originally try the "finest detail" setting and I noticed it did make a slight difference, particularly with a loupe:p .
http://www.ddisoftware.com/qimage/quality/
http://ddisoftware.com/tech/printers/printer-resolution-some-tests-for-you-to-do/10/?wap2

Thanks for the explanation about vector graphics, and for the interesting links. I am fairly certain that is the type of file I was dealing with since it had to be rasterized for a specific resolution.

Sandy

[Bob McCarthy]

The extension of this thread I'd personally like to read, is the impact the substrate has on resolution. I can see how a hard substrate (Pictorico) can yield sharpness, but how much does the impact of a glossy Baryta compared to a matte photorag impact the ability to print very fine detail!!

I fully agree with printing the file at 600 or 720 (printer dependent) does improve sharpness. Seen it with my own tests.

I'm still seeking the holy grail of ink/printer/paper for max black, max sharp, with a matte ink setup (to avoid overspraying).

bob

[sanking]

OK, I did a comparison test with the Piezograhy K7 Selenium Ink Set installed, printing with QTR, and sharpness was very good, slightly better than with the Epson UC inks when printing with 2880 dpi and Unidirectional printing. The improvement was not as dramatic as I had hoped, but at this point I am getting a solid 14 lp/mm on Epson enhanced matte, which is more than twice the threshold of human resolution so probably good enough. This may not be the best comparison media for me since my goal is to print on Pictorico for digital negatives and that surface may have slightly better resolution than paper, or at least this specific paper. However, it seems that the change in printer driver from the Epson to QTR accounts for more of the improvement in resolution than the change from Epson UC to Piezography K7.


Sandy

[Bob McCarthy]

QTR should get the nod for any B&W printer using an epson printer.

I agree with the idea QTR can yield really nice results with any epson ink set. The very first print I made using the 7600 (photo black) with QTR and Harmon Baryta paper really blew me away. That was a 2 black setup as you know. But the highlights were not there, mostly paper white with little ink coverage. But the blacks were "deep" and well beyond my expectation based on using the Epson driver up to that point.

What I think(?) is happening with the 6 or 7 ink systems, is the L-LL-LLLK ink's are giving full coverage and not purely relying on the paper and a darker grey "dither".

I'm more attracted to the MIS approach (actually Roark C-6), but they don't have a solution. I'm trying to find a compatible paper which givs a great black.

bob

[Tyler Boley]

increased resolution discussed in this particular thread is not due specifically to an ink change, or a driver change. It's the combination of utilizing many shades of monochromatic ink to their greatest advantage. Not only does this require inks designed specifically for that purpose, but a driver that can be customized to use the inks in that particular manner.
If QTR were using the OEM inks, building hues using the inks as the Epson driver does, it would offer no greater resolution. That said, there are ways to use QTR with the OEM inks to slightly increase resolution with monochrme printing. However, throwing the multi K inks in with properly set up QTR curves of course then extends performance a bit more. The opposite is not true, it would be difficult to use the Epson driver with K7 inks to approach the quality of the K7 QTR setups.
More than a decade of using multi K inks, with drivers to utilize them, for this reason amongst others, we are still attempting to explain this.
At risk of opening up the criticism I took when first posting this in 2008, here is a link to some work I did on the topic. At the time it was to show the work done here to optimize quad printing, but you can see the step up using the K7 setup as well-
http://www.custom-digital.com/2008/09/bw-print-quality/

All this is well known and utilized amongst the fine art B&W inkjet community. I thought this thread might be uniquely interesting because Sandy, an expert alternative printmaker, might be an ideal judge of moving that world of printmaking forward using niche technology as well. I still don't think we're demanding enough from those supplying our tools in this new marketplace. We're being supplied with tools that are good enough, but see some masterwork of a short 10 or 20 years ago, and it's clear what we're in danger of leaving behind.
Tyler

[sanking]

increased resolution discussed in this particular thread is not due specifically to an ink change, or a driver change. It's the combination of utilizing many shades of monochromatic ink to their greatest advantage. Not only does this require inks designed specifically for that purpose, but a driver that can be customized to use the inks in that particular manner.

Tyler

Tyler,

Very good point about the role played by the importance of the many shades of gray in increasing resolution. I realize that the particular type of resolution chart I used in the comparison test, i.e. high contrast line target, is less than ideal for comparing sharpness of the Epson UC inks+Epson driver and Piezography K7 inks and QTR driver. But even with the limitations I learned something from the tests.

And thanks for the link to your tests. Very interesting results.

Sandy

[Tyler Boley]

I wouldn't begin to question your results. I think what will be more interesting, if and when you get to it, is how well it functions as a neg and whether it contributes to that process in the final print.
Tyler

[sanking]

I wouldn't begin to question your results. I think what will be more interesting, if and when you get to it, is how well it functions as a neg and whether it contributes to that process in the final print.
Tyler

Actually I was just questioning my own methodology for the resolution part of the test. I think the possible smoothness possible from multiple shades of gray is far more important than resolution itself.

On that score, what I have found so far is that only four of the seven inks in the K7 Selenium set have much impact on printing carbon transfer with digital negatives. The UV blocking density of the other three when printed on Pictorico is very low, so low in fact that I am considering mixing different shades of strong and weak UV blockers in the K7 set to create intermediate shades to enhance overall smoothness. Course, if I do this the printer will probably be useless for making regular inkjet prints.

Sandy

[Vertex Ninja]

I wouldn't begin to question your results. I think what will be more interesting, if and when you get to it, is how well it functions as a neg and whether it contributes to that process in the final print.
Tyler

Yes I agree. I'm wanting to get into alternative printing and am very interested in your results. Please keep us posted!

[Peter De Smidt]

Very interesting stuff. I agree with Sandy that smoothness is a very high priority. The best digital negatives would probably be made with multiple densities of a color ink that gives the best combination of smoothness and uv blocking. Isn't that usually green?

[Tyler Boley]

well the new Warm Neutral K7 set is a bit to the olive, perhaps it would block a hair more than the Selenium set, which is sort of red brown. Of course transmissive hue may be slightly different than what we see on a print, particularly matte. Coincidentally I'm testing these inks right now, and on photo surfaces with the GO, the hue is stronger than the same ink on matte. It's the nature of the ink/coating interactions. Transmissive might be another different beast.
Part of the smoothness of the system comes from how the partitioning curves were designed for the particular inksets. I could not match that with the same set and StudioPrint, and attempting to start from scratch designing curves for QTR for an altered set specifically for negs would be a diversion at the least.
It might be worth running all that by Cone, he would have some thoughts for sure.
Of course, changing all this after Sandy's 7600 is all loaded up is a pain. I've been through it too many times...
Tyler

[sanking]

Very interesting stuff. I agree with Sandy that smoothness is a very high priority. The best digital negatives would probably be made with multiple densities of a color ink that gives the best combination of smoothness and uv blocking. Isn't that usually green?

Green is usually a good color for UV blocking and smoothness. However, if you are using green you are probably using the Epson driver. So even when you have the best blocking color you still have to apply a fairly hard curve to linearize output. Normally this is not a problem but if you apply a hard curve over an image file that has already seen a lot of tonal corrections there could be some posterization.

QTR solves the curve problem by linearizing output at a lower level, that of the inks themselves.

Sandy

[sanking]

Part of the smoothness of the system comes from how the partitioning curves were designed for the particular inksets. I could not match that with the same set and StudioPrint, and attempting to start from scratch designing curves for QTR for an altered set specifically for negs would be a diversion at the least.
It might be worth running all that by Cone, he would have some thoughts for sure.
Of course, changing all this after Sandy's 7600 is all loaded up is a pain. I've been through it too many times...
Tyler

I have someone helping me with this who has a loty of experience designing QTR profiles so I am optimistic that we will come up with a good one for my carbon transfer printing in a few days. But the bottom line is that in order to optimize the 7600 for digital negatives it would be necessary to replace three of the inks in the K7 set, and that would entail having to develop special profiles for printing on paper. So for now we are going to work with the four or five inks in the set that give good blocking. That would be shades #1, #2, #3 , #5, and possibly #6. Shade #4 is so light it is very difficult to even see the nozzle check on paper, and shade #7 is also too light to contribute any UV blocking to the profile.

Sandy

[Vertex Ninja]

Green is usually a good color for UV blocking and smoothness. However, if you are using green you are probably using the Epson driver. So even when you have the best blocking color you still have to apply a fairly hard curve to linearize output. Normally this is not a problem but if you apply a hard curve over an image file that has already seen a lot of tonal corrections there could be some posterization.

QTR solves the curve problem by linearizing output at a lower level, that of the inks themselves.

Sandy

This makes me wonder how the green and orange ink in the 4900 and other ultrachrome hdr printers will work for negatives. Maybe similar to what HP is doing with their 3200 but with the fine tuning of QTR. Your thoughts?

[sanking]

This makes me wonder how the green and orange ink in the 4900 and other ultrachrome hdr printers will work for negatives. Maybe similar to what HP is doing with their 3200 but with the fine tuning of QTR. Your thoughts?

I can not confirm it as fact from personal experience but I have heard from several different sources that the green in the Epson 7900 has good blocking and prints very smoothly with UV sensitive processes. I assume the 4900 uses the same ink set?

Sandy

[Vertex Ninja]

I can not confirm it as fact from personal experience but I have heard from several different sources that the green in the Epson 7900 has good blocking and prints very smoothly with UV sensitive processes. I assume the 4900 uses the same ink set?

Sandy

Yes, I believe so. Sorry for being slightly off-topic I was just wondering about it and hadn't found anything definitive.

[Peter De Smidt]

Green is usually a good color for UV blocking and smoothness. However, if you are using green you are probably using the Epson driver. So even when you have the best blocking color you still have to apply a fairly hard curve to linearize output. Normally this is not a problem but if you apply a hard curve over an image file that has already seen a lot of tonal corrections there could be some posterization.

QTR solves the curve problem by linearizing output at a lower level, that of the inks themselves.

Sandy

I wouldn't use the Epson driver. I'd try multiple dilutions of green in something like a 1400 using QTR.

[sanking]

I wouldn't use the Epson driver. I'd try multiple dilutions of green in something like a 1400 using QTR.

That might be the best way, but learning to profile QTR to linearize output with digital negatives is way more complicated than finding the right blocking density and making an .acv curve with the Epson driver.

Sandy