DVDO iScan HD high-resolution video scaling engine Testing


I used a variety of test patterns, from both the Digital Video Essentials and Avia Pro test DVDs, to evaluate the DVDO iScan HD's performance in several operating modes (component in/out, component in/DVI out), and at both 480p and 720p output resolutions. Except as noted, all tests were run with 480i sources.

The iScan HD's overscan measured less than 1% all around with a 720p DVI output, and 2.5–3% with a 720p component output. The luminance response extended smoothly out to DVD's limit of 6.75MHz. Both red and blue chroma patterns were still visible (though slightly reduced in level) out to the 3.38MHz DVD limit from the DVI output, but the red channel was strongly rolled off at that frequency with a component output. (I suspect the BenQ projector's component input was at least partly responsible for this.)

The iScan HD turned in a mixed perfor-mance with the Faroudja test DVD, with only fair results (some jagged edges) on the moving pendulum, waving flag, and fast-motion hockey tests. It performed poorly on the video-overlaid-with-film test, but well on film-overlaid-with-video. It did not lock on to the 3:2 cadence on the unflagged pulldown test—though it did lock on quickly to the flagged test on the Snell & Wilcox zone-plate pattern from DVE. When I switched the Denon DVD-5900 DVD player to 480p component output (which uses that player's Faroudja deinterlacing chipset), the results on all of the above-mentioned Faroudja tests, not surprisingly, improved to excellent. But none of these test-pattern results turned up as problems on real-world DVDs, which looked consistently excellent on the iScan HD, program material permitting.—TJN