Opinions please: Lightjet at 200 dpi?

An old thread I posted a year and a half ago describes a way to create a printer test file which will check a printer's ability to print at various pixel line widths:

 

http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=004PUN

 

Basically for a 300 ppi printer like the Lightjet, it is quite easy in Photoshop to create line widths of 1, 2, 4, 8 pixels at different orientations and with different color combinations. Printing the test image on various Lightjets at different labs will show that between 240 ppi and 300 ppi the Lightjet does not always register correctly. Basically it is nearing its capability which is why it is specified to print at about 300 max.

 

I'm one that does not believes the 300 ppi number is the highest ppi for printers that the human eye can differentiate levels of sharpness. It's just a crude point where it is obvious. When someone makes a good 600 ppi printer with papers of the future, we will see.

What a funny thread from a printers viewpoint. Looks like a few folks here have learned that printers are abit less sharp than a drum scan :)

FWIW the older Lightjet, I forget the model number right now, will run at 400 dpi as well as

305, but most labs won't bother to do the extra calibration. I have gotten

www.reedphoto.com to run theirs at 400 dpi for me.

 

Comparing maximum lp/mm on digital vs. optical is a little misleading.

 

On an optical print MTF falls off gradually--you want 8-10 lp/mm not because you're

actually going to notice detail at that level, but becauase high max resolution is a good,

easily measured proxy for high MTF at 2 lp/mm, where it actually matters. A 4 lp/mm

print loses very little usable detail vs. a 10 lp/mm print; it's the difference in 2 lp/mm MTF

that's really the distinction between a print that's merely sharp, and a print that grabs you

by the throat and shakes you.

 

But on a digital print, competent sharpening (neither to much nor too little) combined with

a good original with excess resolution can give you 100% MTF very nearly to where you

stop resolving lines pairs. Thus a 4 lp/mm digital print CAN be the equal of a 8-10 lp/mm

optical print for all but loupe-viewing purposes.

 

Scott: there is a good reason to overscan to resolve grain/dye clouds. If you just scan at

the final size, where the pixel is not much smaller than the "grain", the "grain" aliases out

to larger, but lower contraast, blotchs. A high-res scan followed by a simple downsample

does the same thing.

 

Grain removal software does poorly in this case, generally taking out meaningful detail

along with the "grain", because they're at similar spatial frequencies.

 

But if you do a gonzo-rez scan, resolving the grain at a much higher spatial frequency

than the finest image detail you want to save, and let the grain-removal software chew on

it for an hour or so, when you finally downsize to output resolution you have a much

smoother image, with very little "grain"-artifact blotchiness left.

<p>"<i>When the LightJet is working at it's highest resolution, it is making 12 lines per millimeter. That's really, really, really....small. Like 3/10,000 of an inch??</i>"</p>

 

<p>More like 3/1,000 = 0.003 inches, which matches the 305 per inch resolution number.</p>

Troy:

 

If you see scan lines, their LJ isn't well calibrated.

 

Another problem is that you are actually dueling with the LJs soft/firmware interface. The highest input resolution (model 5000) is Res16, 16 dpmm or about 406 dpi. The LJ5000 will also accept Res12, 12 dpmm (about 305dpi) and 200 dpi (Res7.9). In the later two cases, the printer upsamples to 406 dpi. Whether 305 or 200 dpi input upsamples better varies between images and within an image.

 

I am told that the newer LJ430 prefers an even 300 dpi or 200 dpi despite the superiority of the metric system.

Steve:

 

You forgot film and paper flatness!

With our old INK JET poster printers; the 4 heads; ie CMYK are adjusted in both X and Y; with respect to each other. Unless the heads are aligned; the ultimate "sharpness" is lower. <BR><BR>Glossy paper; of high quality; gives a tight sharpness. <BR><BR>Artsy fartsy linen materials radically drop the ultimate resolution of the printed inkjet image.<BR><BR>Low quality papers allow ink to enter the paper core; and the micro image smear; and resolution drops.

Glenn: I believe you're mistaken about the LJ5000 upsampling

everything to 406. When I've gotten my stuff run at 406 they said

they had to run separate calibration for that resolution,

something that wouldn't be an issue if it was running 406 all the

time.

 

Certainly imagesetters, which are the same premise as a

Lightjet (photosensitive material exposed by laser) can be run at

a variety of real resolutions, but can only write a given number of

lines per minute, so higher resolutions are slower.

 

Which means it makes sense that the lab wants to run the

Lightjet at 200, and may in fact tell the lower level types that's all

that's possible--it gives them more throughput, the lifeblood of a

fixed-cost intensive venture like a Lightjet.

"When the LightJet is working at it's highest resolution, it is making 12 lines per millimeter. That's really, really, really....small. Like 3/10,000 of an inch??"

 

More like 3/1,000 = 0.003 inches, which matches the 305 per inch resolution number.

 

You are absolutely correct. I did the calculation rapidly in my head - I put the decimal point in the wrong place - maybe I really should use that calculator I own? Thanks.

Roger:

 

You may be right... I had been told by a Cymbolic rep (before Oce swallowed them) that everything was upsampled. But the LightJet documentation is unclear. It mentions the upsampling algorithms used, 16 sample bicubic, but it also clearly shows faster printing times at lower resolution. That would agree with your suggestion of a fixed print time per line. It may upsample along a line? Agree totally with your comments about importance of MTF at 2 lp/mm. This is why relatively low resolution digital cameras can produce surprisingly "sharp looking" results.

 

 

Troy:

 

Perhaps the most important agreement throughout this thread is the importance of the digital flow before the LightJet. Skill of the scan operator, correct preservation of highlights and shadows, careful adjustment of tonal curve, careful resampling and careful sharpening at output resolution are all essential to achieving the quality result that the LightJet can produce. In my experience, few labs really have the talented staff to do this. I have been very happy with results from Reed (Denver) and WestCoastImaging, but less pleased with others. But even with the best, you will be getting their interpretation of your image. Ultimately, I spent the time to learn to do these steps (sans scanning) myself, and now send preflighted data directly to the printer. It requires lots of time, monitor calibration hardware and software, and some investment in good software, but what comes out of the LightJet (or Chromira or Epson 9600) is now exactly what I want and expect.

Glenn: Well, I think everything is upsampled (if needed) to the physical resolution the

machine is running at, it's just that there's more than one choice of physical resolution.

"1500-2000dpi is the maximum you want to screw with scanning LF or MF film if you can get away with it because beyond that all you are doing is gaining better resolution of fuzzy dye clouds."

 

Scott has made this statement over and over again, and it clearly demonstrates his lack of knowledge when it comes to scanning. His Epson 1600 can not resolve at the dye cloud level dispite how many times he posts to the contrary.

Well, Joseph Holmes scans at about 2000 dpi, and West Coast Imaging recommends scanning about 2200-2300 dpi, so regardless of Scott's personal scanning equipment or skills, he is not too far off on this one.

Those numbers are for 4x5. For MF, scanning up to 3000 dpi is not unreasonable.

Steve:

 

You wrote:

 

"For a 20x24 inch print from 4x5, at 305 dpi for the LightJet, you need an image that is 2288 ppi - this number even accounts for resampling by the printer"

 

At 2288 dpi, I get about 8555 x 10629 pixels, assuming a typical 95mm x 118 mm usable image area for 4x5 film. Seems to me that can make about a 28 x 35 inch print at 305 dpi?