Pre-press Award: Compress for Success

by Christopher J. Feola

When you're trying to build a pagination system, the devil is in the details. Sure, pagination allows you to move a story from the top to the bottom of a page in a flash to make room for breaking news. But those tiny things--typos, for example--that could be fixed in the flash of a razor on the paste-up boards now require reprinting the entire page.

And that can take a long time.

Too long, in the case of the Uniontown Newspapers. Editors were reduced to staring at their screens for 10 or 15 minutes each time they printed a page, according to Chief Photographer Charles S. Rosendale, and that was unacceptable.

Rosendale and his colleagues, including Data Processing Manager William L. Palya, first followed conventional--and expensive--industry wisdom and installed an OPI server. OPI can greatly speed operations by allowing paginators to work with small, low-resolution images while storing the high-res images. Since the high-res images are by far the largest elements on a page--one color photograph uses as much memory as thousands of inches of type--getting them off the page turbocharges pagination.

However, using the conventional OPI scenario requires the last high-resolution images to be pushed across a typically strained network at deadline. The OPI server must swap the low- for high-resolution images and then transport the large files to a raster-image processor.

Rosendale says there was an even bigger problem with his OPI--It didn't work.

"We went through, I believe, 16 updates of the software," he says. "We couldn't get it to run for 24 straight hours without crashing. And it always crashed on deadline."

Needless to say, this was even less acceptable than the slow printing times. While trying to deal with OPI, says Rosendale, the team came up with another idea--Why not compress the images? A photo crunched with the Joint Photographic Expert Group's compression engine can be one-tenth to one-one hundredth the size of the original image.

JPEG achieves its high compression ratio through the use of a technique called lossy compression. High-resolution color images take up so much memory because they use enormous amounts of data to describe each dot in the image, allowing the use of millions of different colors. Lossy compression schemes toss out what are called the "least significant bits," in the process, cutting down, say, 10,000 possible shades of blue to 100.

The approach involves saving a Photoshop image in EPS format, with DCS on and JPEG encoding. This is a new sub-menu available in Photoshop 3.0 or higher. The preview is placed on the Quark page and, when printed, the workstation processes the elements in under five minutes. A Postscript Level 2 RIP will decompress the JPEG images on the fly.

The change has provided an enormous speedup for paginators, and concerns about image quality proved groundless--The JPEG process turned out to have more resolution than needed for newspaper printing. "People recommend using the maximum resolution setting. We are using medium," Rosendale says. "We can't see any difference."

NAA Pre-press Manager Thomas Croteau learned about the JPEG technique during a visit to Uniontown and now teaches it in his consulting work. "Several other newspapers have started using this approach," he says. "Many clients have told me that even if I did nothing else for them, this one tip would justify the cost of my services."

Today Rosendale is happy, but the copy desk isn't. Says Rosendale, "Now they're complaining that they don't have time anymore for a smoke and a coffee!"

Christopher J. Feola is a regular contributor to TechNews.

Sources

Thomas Croteau, NAA, 11600 Sunrise Valley Drive, Reston, Va., 20191. E-mail, crott@naa.org; phone, (703) 648-1213; fax, (703) 648-1333.

William L. Palya and Charles S. Rosendale, Uniontown Newspapers Inc., 8-18 E. Church St., P.O. Box 848, Uniontown, Pa., 15401. Phone, (412) 439-7500; fax, (412) 439-7528.


TechNews Volume 2, Number 4: July/August 1996
Return to July/August Home Page
Return to TechNews Topic Index

©1997 Newspaper Association of America. All rights reserved.