I recently had the chance to benchmark the DVDO iScan VP30 and commended it on its performance and feature standpoint. The VP30 added a few new features to the previous iScan line, including 10 bit proprietary scaling, HDMI switching, and some new tweaks to further improve the overall experience. I personally thought this would be the last thing I saw from DVDO until their follow-up unit came out (except firmware updates of course!). Little did I know that the folks at Anchor Bay Technologies (ABT) were working on their own proprietary de-interlacing solution to replace the already well received Silicon Image SIL504 that has been the core of the DVDO line for quite some time. This new solution is for SD material only, but the folks at DVDO are almost done with their HD de-interlacing solution that they plan on premiering later this year in a new video processor. The SD de-interlacing card is available as an optional upgrade to existing VP30 owners or can be added to any new orders. The upgrade comes in the form of a daughterboard that attaches to the main video board inside.
Since I already had a VP30 running in my system, DVDO just sent me out the new daughterboard and directions on how to install it. Installation is extremely easy and doesn’t require much more than a screwdriver to open up the VP30. The board comes with two standoffs that attach to pre-existing holes in the main board, as well as a connector that snaps into a socket on the board. Installation is similar to the optional SDI board that can be ordered for the iScan HD or VP30.
Once installed, the processor starts using this new board for all of its SD de-interlacing duties, and the results are superb. The SIL504 previously used is already an excellent de-interlacer, but it is a design that has been around almost as long as DVD. With new de-interlacing solutions upping the ante in cadence detection and diagonal line processing, the DVDO line was due for an upgrade. This new de-inerlacer takes the DVDO to a higher level of performance and easily matches or exceeds the performance of any video processing solution I’ve seen yet for SD material.
This product brings up some new issues though. The Benchmark has always been a very intensive set of tests for SD video processing, and even now the majority of video processing solutions still have trouble with it. However, with the recent introduction of video processing solutions like the new ABT chip, the bar has been raised. Processing that we never really considered once is now achievable, and with that, the testing needs to change. We at Secrets will be adding new tests to the Benchmark over the next few months that will push the limits of video processors and incorporate the capabilities of the high-performance units out there such as the ABT design. Look for a complete revamp by years end!
Running the ABT through our de-interlacing tests almost seemed pointless after awhile. The previous SIL504 had no issues, so I couldn’t imagine the ABT solution having any either. This was exactly the case. The chip performed flawlessly in every test, including 2-2. So instead of just leaving it at that, I decided to venture on to some harder test material. This included the Silicon Optix HQV Benchmark DVD and its mixed cadences tests and diagonal line torture tests. I also took a look at ABT’s own Video Reference Series DVD that will be shipping very soon. Like the HQV disc, the VRS DVD includes not only test material designed to show the difference between the ABT video processing and competitors but also real world material to show you how it correlates to everyday viewing. This test disc was created by Stacey Spears and Don Munsil who were the founders of the DVD Benchmark here at Secrets! On top of that, our very own Brian Florian created many of the test patterns and graphics for the disc. I highly recommend this disc for anyone looking for great test material to put his or her DVD player or video processor through the ringer.
One of the big differences I found with the ABT board compared to the SIL504 was diagonal line processing. This is a feature that was exclusive to the Faroudja based video processing chips for a long time. Using the HQV test disc and some of the Faroudja test patterns available I was able to compare the ABT to both Faroudja and Silicon Optix’s Realta processing. The ABT did a phenomenal job in every respect, matching the DCDi processing, and actually exceeding the performance of the HQV processing. Using real world material, the three solutions performed on a very similar level. The HQV chip in my reference Denon DVD-5910 had some difficulty with some of the Faroudja test material, but the ABT did just fine. Some of the more difficult material required me to change the de-interlacing mode of the ABT chip, but the results were well worth it.
The HQV solution still performed a bit better with mixed cadences, including those commonly found in foreign animation. It would also lock onto high detail images with mixed cadences a bit faster. The ABT was better at holding the cadences though, once it locked on. This was evident using HQV’s reference DVD. The Super Speedway clip is a bit of a torture test for high frequency detail, and the HQV couldn't hold the image intact through their reoccurring loop. The ABT locked on to the material and held it through the loop. The only other chip I’ve seen that can do that is the Gennum VXP in my reference SSP, the Anthem Statement D2.
Conclusions
The ABT-102 SD de-interlacer is a welcome addition to the iScan line and continues DVDO’s reputation for delivering top of the line performance at very attractive prices. The daughterboard is being offered at a significant discount for early adopters, but will be going up in price after awhile, so if you are interested, get one now. The ABT-102 is every bit as good as any other de-interlacer I’ve seen, and exceeds most of them in diagonal line processing and mixed cadences. Congratulations to Anchor Bay Technologies and DVDO, and we hope to see more from this chip in other hardware soon!
