We judged printing quality, both monochrome and color, with NSTL's standard output pages, which include text, graphics, and photographic images (
see the figure
). We set each printer in the mode that produced the best-quality output. In the monochrome tests, we pushed each printer's resolving power with fine horizontal and vertical lines, printed both black-on-white and the reverse. In the line squeeze test, for example, we tested resolution by gauging how closely together two lines can print without merging. Among the text-based tests, we determined the limits of legibility with decreasing font size. Higher print resolution confers an advantage with text printing, especially with finer font sizes.
Among the graphics tests, we measured a printer's ability to produce a complete range of gray scales. We judged monochrome-imaging quality with a photographic image, looking for even shading in flat areas and the ability to keep detail in both highlights and shadows. High resolution helps with detailed photographic images, too. We also measured how close a laser printed to the edges of the paper and how well it centered the print area
on the page.
The color-quality pages contained a mixture of text, color photographic images, and color graphics. We printed some of these pages from Adobe Photoshop 3.0 and CorelDraw 6.0, while others were raw PostScript files. We judged pages subjectively using a jury of experienced NSTL staff members, but we also used such objective criteria as color range and fine line detail.
Performance and the Rest
The NSTL performance tests measure how fast a printer can produce a number of basic document types: raw text, bit-mapped graphics, monochrome graphics, color graphics, and fonts. Speed testing also included both PostScript and Printer Control Language (PCL) pages. We tested speed at a printer's standard resolution (600 dpi, except for the 300-dpi HP Color LaserJet 5M) and did not use high-quality modes for performance testing. Timing started when a job was initiated and ended when the last page dropped into the printer's output tray.
Raw text tests a printer's raw engine
speed as this test doesn't involve significant processing time. The various graphics tests include bit-mapped images with custom fonts and screen shots, and complex lines with fills, curves, and gray scales. The font test requires the printer to create Times and Helvetica fonts in regular and boldface in 30-point sizes, stressing the printer's processor.
We tested printing speed over a NetWare LAN (10Base-T) consisting of a single server and client. Printing originated from a Windows 95 system (the client) using printer drivers supplied or recommended by each vendor. We disabled all print servers, spoolers, and buffers during testing.
Important factors affecting our usability score are the installation process, which ranged from involved to downright difficult with these color lasers; ease of driver installation and network setup; and the quality of the manuals. We develop the features score based on our assessment of the utility of such features as networking capabilities, PCLs, driver support fo
r different OSes, paper capacities, and warranty.
photo_link (91 Kbytes)
Among other pages, the NSTL Printer Quality Test has a color page with mixed cont
ent (1), color graphics (2), and text and fine lines (3).
Test Specs
