GhostScript RIP imagesetter negs on inkjet?

[jag2x]

Hi all,

I came across a method of printing actual halftones with the epson printer using a program called Ghostscript. I came across this on the t-shirt forums,
http://www.t-shirtforums.com/screen-printing/t27267.html
you need an account to see the attachments.
Have a look at page 8, it tells you how to do it in Illustrator if you dont have Corel.
You need Ghostscript and Ghostview
http://pages.cs.wisc.edu/~ghost/doc/GPL/gpl871.htm
http://pages.cs.wisc.edu/~ghost/gsview/get49.htm

From what I can understand, you work on your file in Photoshop, open it in Illustrator, print it via a HP printer (print to a file( a ppd file?) ), play around with the seperation of which colour you want, the line angle and also the lpi/dpi. Outputting it to a .ps file.

Then open the .ps file in Gsview. Then go to print it using your epson printer or any other printer for that matter.

I haven't done any extensive testing but it seemed to work well enough when I tried a lpi in illustrator as 60lpi and then at 200 lpi. As you can physically see the actual halftone effect, harder with the 200lpi.

I dont know how effective this is for digital negatives, as I am still trying to figure it all out myself.

Anyhow just putting this out there as this might produce the same effect you get from imagesetters negatives or laser printers on your inkjet printer.
Let me know how you go..
Cheers
Jacek

 

[Loris Medici]

Jacek, I don't get all this... In short, (a.) What's the point in printing halftone negatives from an inkjet printer that has ~ 12 l/mm (lines per mm) *hardware resolution - at max*? (12 l/mm -> 304.8 dpi...) and (b.) Why don't you do the separations / conversion to bitmap in PS, and also print from PS directly? and (c.) The whole point of using inkjet printers for digital negatives is their ability of doing stochastic (FM) screening. (= more apparent detail and tonal smoothness with less hardware resolution + no moire problems for multilayer prints...) Is there any advantage in using halftone (AM) screening - that I don't know about? (Even current halftone RIPs do hybrid AM+FM screening to my knowing...)

Shorthened / compact version -> (a.) + (b.) + (c.) = Why on Earth one would want to do that!?

 

[jag2x]

Hi Loris, You could be absolutely right.
I always had the impression for halftone screens you get a better tonality for the highlights and more apparent detail? So the opposite of what you said!

As this ghostscript utility would allow you to print just a dot of ONE LEVEL of DENSITY. Just like a laser printer and imagesetter does. So in extending it to work via an Epson printer my reasoning of allowing a halftone screen to print via this program allows for a one density value instead of the range of densities(ink levels).

I hope I am making myself clear as perhaps...just maybe.. I'm trying to make assumptions from my imagesetter experience to equate to Epson printers

You are also right you can make the seperation in PS.
In PS you could make the bitmap version and convert it to a halftone or a difusion dither and print it via PS.

Plus I'm unclear at what lpi a Epson printer can print to?

So to sum up I'm looking at this as a means to print one density value of a dot throughout the whole print, instead of a range of densitiy dots.

I think you have a Canon printer, I'm sure you can give it a go with Ghostscript, see if there is even an added advantage to printing with it...if you ever wanna give it a go
Cheers
J

 

[Loris Medici]

I have an HP 9180 which I'm very happy with what I can currently get. I wouldn't bother to try the Ghostscript route because I'm extremely sure it would give me way inferior results. (Please note that I have printed with 3600dpi/225lpi imagesetter negatives before, in other words, I have practical experience in the other side...) Testing the lpi of Epsons (or any other) is easy, just print negatives with 1x1, 2x2, 3x3 and 4x4 px checkerboard+line patterns at 300 or 360 ppi and print it with your preferred / sharp alt process; you'll understand what I mean...

Regards,
Loris.

 

[L Gebhardt]

An image setter has much higher resolution than an inkjet printer if you just compare the numbers. But the image setter is not continuous tone, so you must make the tones out of the black or white dots it can print. So your 3600dpi image setter must use many dots to simulate gray. For 8bit precision (256 color) you would need to use an 8x8 grid of dots to simulate 1 pixel, so the image setter is really more like a 450 ppi device (3600/8).

Your Epson can print around 450ppi as well in continuos tone, but I doubt it can get to 3600 dpi in black only mode as it would need to to emulate in image setter. In other words I don't see how you would be gaining anything by halftoning your image.

If you want higher resolution, say for enlarging from digital negatives, you need to find a continuous tone film recorder like the lightjet or LVT recorders. These will give you 2032 ppi images, and can look as good as a film original.

 

[Loris Medici]

Gebhardt, will make some corrections:

8x8 = 64... It should have been 16x16 = 256. Therefore, 3600 / 16 = 225lpi line screen - for 256 distinct tones... Not 450.

I haven't seen any Epson that was able to resolve 1x1px checherboard/line pattern printed @ 360ppi, only 2x2px - which makes a "confirmed" 1 / (2 / 360 x 25.4) ~= 7 l/mm (or 180ppi). Probably they will go up to something around 300ppi (~ 11-12 l/mm) and that's all. (Maybe newer 2880 / 3880 are slightly better, but I don't think it's by a large margin...)

They say Lambda's can go up to 400ppi (~ 15-16 l/mm)...

A confirmation: I saw some s/g contact (and enlarged - from 4x5") prints made with LVT negatives and they were gorgeous indeed...

Regards,
Loris.

P.S. I wouldn't call inkjet prints "continuous tone", they're just FM halftones with variable dot size...

 

[L Gebhardt]

Loris, you are correct on the 8x6 vs 16x16. I'm not sure where my mind had wandered too.

Printing a black and white test pattern (like the USAF 1951) chart it seems to come out around 7 line pairs/mm or (360 ppi) in one dimension and around 12 lp/mm (574 ppi) in the other. When printing images with lots of fine detail you can certainly see an improvement when you feed it 720ppi images vs down sampled to 360ppi on the 3880. In the real world (without a loupe) I can't tell the difference.

When I say the inkjet is continuos tone, I meant you feed it a continuous tone image and it does it's own internal halftoning. But the resolution numbers (such as 180 or 360) are real pixel dimensions and don't need to be converted like the image setter numbers. So I should have said "Your Epson can print around 450ppi as well in continuos tone terms".

 

[Loris Medici]

I think you've confused 8bits for 256 tones with 8x8 dot matrix.

OTOH, w/o doubting the numbers you provide; all I'm talking about is actual alt-process prints from digital negatives - that's the context, and it's the end result that counts really... I could resolve dots (elliptical) - at many many %K values - with cyanotype (by observing with a loupe), using imagesetter negatives printed 3600dpi hardware 225lpi screen resolution, whereas I couldn't resolve 1x1px checkerboard/line pattern printed at 360ppi file / 2880dpi hardware resolution (= 360lpi) with Epson inkjet negatives, with exactly the same emulsion, paper and lightsource... (I later checked with a 10x loupe and the lines/pattern also wasn't apparent optically!) One can blame cyanotype/watercolor paper combo for the failure - but I definitely know cyanotype can do better! (See prints made with in-camera negatives...) You can test your printer with s/g paper (no-one will doubt about the medium's resolution) and see it for yourself. Apparently, something happens during printing and decrease the final print resolution...

See this BTW. In short it says: "Use 360ppi image when 'Fine Detail' is unchecked in the driver, 720ppi when it's checked."

Regards,
Loris.