I've been working to my Epson 7880 to make film to print process color. Along the way i've come across some strange interesting findings.
Since i would like the Epson to do more than just print film, i decided to start off with the standard 8 inks, which includes the standard photo black. While the film is much less dense than imagesetter film, in my exposure tests, it has shown to be more than dense enough to be used to expose high quality screens.
The Epson REALLY likes to drink ink! Even though i have hardly printed a single thing in color, my color inks are about 1/2 empty. I can only assume they are getting drank during the cleaning cycles the printer decides for itself, that it needs to do.
The Epson printer uses paper type to define the ink flow of the printer. In my testing, i observed the epson would throw away pixels in order to compensate for the paper selection. "Single Weight Matte Paper (line drawing)" was the best paper setting i found for 720x720 and 1440x720, which should be more than ample for screen printing.
I've tested the old epson drivers inside ghostscript and it appears that epson switched to a new image command when the epsons moved past 360 DPI. The old epson ghostscript drivers used the old commands, which when the file is changed to high resolution, the printer fails to print anything.
About the same time epson changed the command set, they also stopped publishing the command references. This is probable the reason no new epson drivers have been released for ghostscript.
Someday somebody may write a ghostscript device for a modern epson? The unfortunate problem is that without documentation, the driver would most likely only function on the printer the printer the driver is designed to work with. For example, my epson 7880 is typically roll fed, but its baby brother, the 1800/1900 is most often sheet fed. Without some manufacturer documentation, it is very difficult to translate anything produced for one epson to work for any other epson.
The free nature of ghostscript combines with the copyrights ensnaring ghostscript precludes any profit being made on anything developed involving ghostscript eliminates any profit motive for development. Even should someone develop a driver and get it working for more than just themselves, the would then need to interface with the GNU community to get the code released to the world.
How many hours of your time would you spend trying to give something away free??
I'm left to wonder why anyone would ever release a ghostscript driver for an modern epson printer.
Just like any rip, no matter how free the rip, it will not automatically calibrate the film to any specific press. I fear that any free rip will be greated by lazy people complaining, "why do i need to calibrate?" "i'm looking for free software, why should i need to work???"
At the price of "free". it seems to me, it seems like far more trouble than it would be worth? Right???
fred
__________________
A day late, a dollar short, so it goes.
It's throwing away pixels when you print black and white halftones or when you print full-color prints depending on what paper selection you pick? I'm a little confused (easily done, mind you). Are you trying to calibrate ghostscript with the halftone output from your 7880 printer so that the halftones are accurate, as far as lines per inch and dot size being accurate to the percentage of black that it's supposed to represent? I presume most users are just using ghostscript either as a free interpreter for their inkjet printer, or as one for a printer for which no rips are available (such as the HP9800 I use). I think that in general, most printers are looking for a reasonably accurate reproduction of their file, since there are so many variables in textile printing (especially manually) that unless the line count/dot shape/dot size was so far off that to go to the trouble of trying to calibrate ghostscript to a particular printer would be much more trouble that it's worth, as you mentioned.
I know by your other posts that your graphics background is far and away more involved than most who are here, and that you do quite a bit of testing to get consistent, repeatable results, but if you're trying to reengineer ghostscript for one particular printer, you've probably answered your own question –– why bother. In general, and I'll assume you're doing this anyway, most fiddle with the print dialog box options as regarding ink density, lightness or darkness, resolution, and media, then apply them to whatever brand of film we happen to be shoving through the printer, until those reasonably accurate reproductions I mentioned earlier are arrived at. One set of printer settings will not necessarily work with every brand of transparency film. On my 9800, I started using Filmsource's regular non-waterproof film and was pleased with the results. I switched to Screener's Choice Chromopaque waterproof film, and had to back off the density and amount of black, changed to Plain Paper instead of the High-Gloss Professional Paper setting I had been using, and get much better, much denser blacks, in addition to the quick drying. I think every different make of film would necessitate changes. All this from the standard ink cartridges that come with the HP.
Now, all that said, if I was going to use an Epson (I have one sitting unused on a table), I'd buy (I bought) a dedicated rip for that printer. I had FastRIP on mine, and it worked great until the printhead went south, I think from using a chip resetter too much as the other colors went empty from printhead cleanings. Then my bulk ink cartridge electronics crapped out a week after the new printhead. I sold the FastRIP and dongle to someone who needed it. The Epson has sat for almost a year unused. What technical differences there are between the output from the FastRIP/Epson combo and my current HP9800/Ghostscript combo, are, to my eye, negligible, and I've placed the same file output from both side-by-side. For printing t-shirts, any difference there is probably inconsequential, given the other variables encountered in printing on jersey material, as mentioned before.
OR (as I mentioned, I'm confused) are you trying to output a true process color PRINT on paper (involving halftones, rosettes, etc.) from the Epson like you would get from an offset printing press, as opposed to the output the Epsons normally put out (looks something like a diffusion dither as opposed to halftones)?
Epsons Folklore 
