Product Review - DVDO iScan Plus Line Doubler - March, 1999
Stacey L. Spears
DVDO iScan Plus Line Doubler
Scan Rate: 31.5 kHz
Inputs: One Composite Video, Two S-Video
Outputs: VGA 15-Pin (RGB + H/V Sync/ YUV)
Size: 1 1/8"H x 5 1/4"W x 8 1/8"D
Weight: 2 1/2 Pounds
MSRP: $699 USA
DVDO Inc., 1129 Dell Avenue, Campbell, California 95008, Phone 408-364-3836; Fax 408-364-6736; Web http://www.dvdo.com; E-Mail email@example.com.
Follow-up – April – 2000
Here is my much overdue follow-up to the DVDO iScan Plus line doubler. When we last left off, our hero the iScan was just a pre-production prototype that only had the RGB outputs active.
After meeting with Paul Wolf at this years CES, I decided it was time to stop waiting for another product from DVDO and find out just how the current production model compared to the prototype I had played with last year. It’s hard to believe that is has been an entire year since I had last played with the little box. I was only allowed to play with the prototype for a mere 7-days. I have spent much more time with the full production unit.
A look back the prototype
Lets start by addressing some of the complaints I had about the pre-production unit. Video games, you love them and I love them, but last March, the iScan did not appear to love them as much. Shortly before the iScans started to ship, I received an email from Dale Adams, one of DVDOs founders that they had developed and additional algorithm to deal with single field sources like video games. I am here to say that they indeed deed do something because both my Dreamcast and N64 do not have produce the ugly effects that the original prototype did. Though, if you have a Dreamcast I would recommend their VGA adapter.
Another problem I noticed with the prototype was tearing on the Snell & Wilcox test pattern. This was very evident on the prototype with tearing happening in all 4 corners of the ball. The tearing is 99.999% gone. I give such a strange number because on a new special test DVD, the Microsoft WHQL test DVD, there is a slight tearing, everyone once in a while on it’s moving wedge pattern. It does not happen all the time, it does a stellar job on this true torture test.
A new home
When I did the initial review, I had to use the RGB outputs and I fed those outputs straight into the Dwin HDP-500 projector on a 100” 4:3 screen. The Dwin’s sweetspot is really closer to 768p on a 4x3 screen and 720p on a 16x9 screen. So I opted to use the iScan on the Toshiba TW40X81. The Toshiba has a built-in line doubler, but it does not do any type of film motion detection. (3:2 pulldown)
I have spent the last several weeks using the iScan with the Toshiba and it was not until I went back to the internal doubler that I realized just how great the iScan really is. Using the TVs internal doubler products a much softer picture and the depth that I had grown accustomed to on ZDTV was gone. Something else I should mention when using the iScan with the TW40X81. You have to use one of the DTV inputs and when you do, those annoying gray bars are replaced by black bars.
I was happy when DSS began broadcasting the local Seattle channels on channels 974-977. I did have many friends who told me that their cable looked much sharper than the DSS channels. This is the complete opposite of what I had experienced at my house. The cable channels are not only less sharp, but are filled with noise. Removing the iScan had the effect of making the local DSS channels to appear softer and have more noise, still not as much as cable, but enough to draw attention to itself. The iScan is a godsend to the DSS local channels!
I also fed ReplayTV into the iScan. I had been wondering how this was going to look because Replay digitizes everything that comes in using an MPEG2 encoder. I was happy, no relieved, that the iScan and Replay played together very nicely.
There is some strange video artifact that I get with the Replay and iScan on material recorded from my noisy cable. When the picture is really noisy, I get these green vertical lines in areas of the picture that contain intense red information. I never see this on anything recorded off of DSS, just cable. Usually shows like the Outer Limits, which I pickup off of UPN.
If you read my review of the Dwin TranScanner you will remember that I found the de-interlacing artifacts from sporting events to be distracting. They happen much less and are only occasionally noticeable on the iScan with the Toshiba TV.
What about progressive DVD?
When the iScan debuted, progressive DVD players where still vaporware. While I am not sure if the legal issues have been fully resolved, players calling themselves progressive are now available on the market. Just how does the iScan compare to these players? I only have the Toshiba SD-5109 on hand at the time of writing, so I will compare them.
First, the 5109 only offers DVD in a progressive format, it does nothing for any other formats you may have such as VHS, ReplayTV, TiVo, or DSS. Second, films are stored on DVD as fields so either the DVD player or the iScan will have to de-interlace the picture.
There are several advantages to perform the de-interlacing inside the DVD player vs an external box such as the iScan. One you void two stages of D-to-A and A-to-D conversion, which could degrade the picture. The DVDs themselves contain flags that explain how to re-assemble the picture for film or video where as the iScan has to use complex algorithms to perform the same task. And to make things worse, those flags are not always encoded properly; just take a look at Fargo and Titanic.
I did several direct comparisons of the 5109 and the iScan. Each time I did prefer the 5109, it produced a picture that appeared to have more depth. The 5109 also had slightly cleaner and more saturated colors. This could be in part that the iScan only accepts S-Video.
Does anyone know why the iScan does not offer component inputs? Do you remember my review of the Dwin TranScanner? I had complained about these green horizontal lines on dark scenes. Well, the problem was in the decoder used in the TranScanner, it was a Samsung decoder. DVDO uses the same decoder and they discovered the problem in the development of the iScan. They decided it was better to not offer component input than to offer a flawed component input. (I did not use the component input on the TranScanner until it was fixed.) The problem was eventually fixed in a later rev of the chip, but it was too late for DVDO. The DVDO actually contains the later rev of the chip, but it came too late for them to really take advantage of it.
If you already own a DVD player, it might not be economically feasible for you to buy a new player. The iScan sales for about the same price as the SD-5109, but if you have more sources, the iScan would be a smarter investment. Then again, if you are like me and want the best of both worlds, the Toshiba TVs have enough inputs for both.
Associated Equipment used for this review:
- Stacey Spears -
A New Era
The iScan Plus is a new line doubler product from DVDO Incorporated. It is the first, really affordable outboard line doubler made available to consumers. Although the retail price after May 1, 1999 is $699, if you pre-order before May 1, you can get it for $599.
DVDO, Inc. is a privately-held startup company that designs, manufactures, and markets chip and system-level solutions for the next generation of Digital Television and Digital Video electronic products. The target market spaces for DVDO's products include Digital Video Disc (DVD), Digital Television (DTV), Direct-to-Home Digital Satellite Television, Home Theater, Automotive, and Airline Entertainment applications. DVDO is the creator of the proprietary Image Enhancement Engine Technology, a new high-performance and cost-effective solution for chip-level Digital Video processing and enhancement. DVDO has a total of fifteen U.S. and International patents pending on this new technology.
Dale Adams and David C. Buuck initially founded DVDO, Inc. in August of 1997. Other founding members include Laurence A. Thompson and Cheng Hwee Chee. The company is based in Campbell, California, and the owners are all ex-Apple employees.
This is a "first look" analysis based on a pre-production sample (Sample # 3). As soon as Secrets receives a production sample, our review will be updated (we will provide a link back to this page). The only thing that should change between now and then is the inclusion of our test results on the Y-Pr-Pb outputs (aka YUV, Component, Y-Cr-Cb). For our first tests, I used the RGB+Sync output connection to a Dwin HDP-500 projector and various NEC 17 monitors.
Since this is initial, I will only be discussing the performance of the DVDO with a DVD player and DSS, along with a short comparison against the Dwin TranScanner (TS) and the Mpact2 3D Fusion PC Card.
The follow-up review will also examine laserdisc and VHS VCR performance, a more in-depth look at the TS and Mpact cards, and the component outputs of the DVDO when connected to a current market HDTV set. I will also put its internal comb filter through the wringer.
Beauty is not just skin deep
The photo of the iScan Plus in the product literature, available on their web site, does not really represent the size and look of the unit. Actually, the iScan looks like an old Hayes modem. It is very small and lightweight. It has seven LEDs on the front: one power, three input sources, and three processing modes (Video, Graphic, and Film). There is also a switch for priority that selects which input you want to use, 1, 2, or 3. This switch feels a bit fragile to me (especially prone to abuse by excited consumers and magazine reviewers).
The back of the unit has an AC adapter jack for DC-in, one composite and two S-Video inputs, an RGB/Component selector switch, and a 15-pin VGA output connector. You will need a 15-pin-to
RGB + H/V Sync adaptor to connect the output of the DVDO to your projector. The DC supply is from a wall wart, suggesting that future upgrades are possible, like the Audio Alchemy PS-1 did.
I like the inclusion of the mode LEDs. These automatically indicate mode changes, such as switching from video to film mode. During the opening of VE, where the swelltone logo is shown, the iScan was in Graphics mode. It would then switch to Video mode as the program started. The montage of images resulted in constant switching between video and film recognition (video and film differ in the way two interlaced fields align with each other to represent a single frame).
With the TS, I never really knew when it was processing film or video, but with the iScan, it is indicated by the LEDs.
After unpacking this compact unit, I plugged it into my computer's NEC monitor. Using the S-Video input, I fed the iScan with my Panasonic DVD-L10 Palm Theater, and then tossed in my trusty copy of Video Essentials DVD. I wanted to see how the iScan handled the Snell & Wilcox (SW) zone plate pattern.
While the SW was playing, the DVDO mode stayed in Video the entire time. The SW ball actually switched between 24 and 30 fps during the clip. There seemed to be an artifact in the four corners of the ball. This artifact can also be seen in the diagonal line box under the 400 text at the top of the pattern. When I paused the video, about every second they appeared then disappeared. Even though the mode never changed to film, it did lock onto the 24 fps and displayed the ball correctly. When the ball switched to 30 fps, it handled that as well.
I played the same material on the Mpact2 progressive decoder card with a DVD-ROM drive. The Mpact produces no artifacts that I can see. In fact, the Mpact does a better job on this than the TS! However, when the pattern switches to 30 fps, the Mpact gets tripped up a bit, so it does not handle 30 fps video very well. This is a stress pattern designed to cause havoc on any video processor. In the final analysis, what matters is how well it looks when watching real program material.
Next, I set the contrast and brightness on the NEC monitor. After that, I loaded up the NTSC color bar pattern to see how well the iScan does at chroma decoding. To keep the price of the iScan at a reasonable level, they provide no way of adjusting color intensity, tint, sharpness, contrast, or brightness. However, contrast and brightness adjustment can and should be done from your display device. There is no way to adjust the color decoder inside to compensate for your video source, which all provide a different output level. When I looked at the color bars through a blue filter, both color and tint were slightly off. This level will vary with every source including DVD, LD, DSS, VHS, etc. With that said, the levels were not that far off. There was not the usual red push you find with the chroma decoders found in consumer TVs. (The colors were not as saturated as with the TS or Mpact.)
The picture appears to be just a tad soft. I examined both the horizontal and vertical resolution using the multiburst [Click here to see pattern] on VE [Frame: 46619] or A Video Standard (AVS) [Frame: 5060, 14384, 50790, and 50816], and the sweep pattern [Click here to see pattern] from VE [Frame: 46622]. There was no apparent loss of resolution. There was also no sign of over enhancement, which some video processors have lots of. DVDO is also not using any form of edge enhancement.
It would be nice if they could add a control for color intensity (saturation) and tint. I know that they have done everything possible to keep the costs low. If I had the option of adding two features to this box, the first would be component inputs to bypass the chroma decoder which would put back the chroma resolution lost in the conversion to NTSC (S-Video and Composite), and the second would be color/tint controls.
I next played the montage of images found on Video Essentials. This part looked spectacular! One of the more difficult scenes, the football stadium pan, was handled even better by the iScan than by the TS. Everything was locked into place, just slight movement at the top of the bleachers.
The Big Show
I connected the iScan to the Dwin HDP-500. This was with a 2.0 gain 100 diagonal 4:3 Da-lite screen. Because the primary video processor on this system has been the TS, I had to use a memory that has not been touched in a while. The convergence was still good at the 31.5 kHz setting.
I had lots of help from Evan Upchurch here, since the Dwin HDP-500 and Da-lite screen belong to him. He also provided lots of video sources.
I decided to watch the intro to "The Flintstones" live action movie. The picture was softer than I am used to seeing with the TS. At 1/10th the cost, I was OK with that. At 0:56 (0:56-1:02) into the movie, the Slate and Company building panned in. This scene has several windows that were flickering madly. I thought, "wow, the motion logic was not doing so hot on this scene." After a few minutes into the film, I realized that I had the DVD player set for a 4:3 TV.
Switching the player to 16:9 mode required a different projector memory. This one was out of whack! Convergence was off and the centering was messed up. I was able to re-center the picture and do some basic convergence that made the picture watchable, but the edges were slightly obscured. I replayed that same sequence, and the flickering was totally gone! I re-tried the entire scenario using the TS, and the results were the same, flicker in 4:3 and clean in 16:9. Playing the same scene on the NEC monitor did not have this effect - it is so small you can hardly tell - but blown up on a 100 screen, it is very obvious. Not only was the flicker gone when using the 16:9 mode, but the picture was also sharper. This is a great scene to show just how important that extra 33% of resolution provided by 16:9 enhancement is (DVDs that are anamorphic). DVDs that are not 16x9 enhanced will still appear to be much softer. The iScan is slightly softer than the TS and Mpact, but it does the job well.
When I was watching in 4:3 mode, I could still see scan lines using the HDP-500. This projector really needs to be scanned at a higher rate. More than tripling, but less than quadrupling. When I switched to 16:9, I had to un-squeeze the picture. Doing this actually eliminated the remaining visible scan lines. I have the same effect on my Toshiba TW40F80 TV. When I watch anamorphic DVDs, scan lines are not that noticeable. When I watch a non-anamorphic transfer and zoom in, scan lines are very obvious.
If you read my review of the TS, you will remember that I saw motion artifacts while watching DSS. Sporting events were the worst. The iScan has similar artifacts, though not at severe or as often. They happen mainly on quick transitions. For example, during the LA Lakers drubbing of the Orlando Magic basketball game, when a player would jump for a shot, he would appear to be sliced by the scan lines. Or to be more technical, horizontal line structure would become visible.
The iScan picture is softer than the TranScanners, and I saw motion artifacts during transitions between scenes (especially black to white). It appeared that one field was being shown before the entire pair of fields (two fields make up one frame) had been buffered. This resulted in readily apparent horizontal lines. That being said, the iScan's performance with standard DSS video was superior to the TranScanner, which showed stutter with vertical motion and blurring with horizontal motion.
To be fair, whenever I viewed the Faroudja, it was always with DVD movies whose content was a digital transcription from a film. I have never seen DSS TV broadcasts through the Faroudja, but both the TS and iScan do display some motion artifacts with this source.
I was surprised at the poor performance of the iScan with video game material. The text in this Nintendo64 image shows an anti-aliasing effect that is extremely apparent. (This same effect can be seen on still frames during Video Essentials that contain text.)
The line structure is clearly visible and emphasizes the jaggedness of diagonal lines. If you look at the bears backpack, you can see that the lines have been seemingly doubled and poorly anti-aliased. Look at the bottom of the picture frame and you can see an example of a line that has been averaged with the one above it. This effect was much more visible in the moving image. Closeup views of the backpack and text are shown in the two graphics below.
The TranScanner had artifacts of its own, but they were not as obvious as those induced by the DVDO. The Phillips DVX-8000 actually did the best job with video games.
The iScan is to high-performance video what the Dac-in-the-box (Audio Alchemy) was to high-performance audio. It does a good job de-interlacing film material without introducing serious artifacts, and it's inexpensive. It also has good film mode detection. Many video processors get this wrong, but the Faroudja, Dwin, and now, the DVDO, do it right.
Being able to obtain this type of quality for such a low cost is going to cause a few rifts in the video industry. This new competition should help drive the costs of other such products down.
One of the differences between high-performance audio and video is the cost it would take to move to the next level of performance. In the audio world, it was much easier to move up the chain. In the video world, it was a HUGE step . . . as in $10,000.
I do need to remind you that in order to use the iScan, you will need a display device that can scan at 31.5 kHz. Typical consumer TVs only scan at 15.75 kHz and will not be able to use the iScan. But, you can use it on your computer, and of course, with any projection TV that handles this scanning rate. An interesting combination would be the DVDO and one of the less expensive ($3,000) LCD front projectors, capable of SVGA resolution.
I will forgive weaknesses if I can get nearly equivalent performance for roughly 10% of the price. There has been much hype on the Internet lately about the iScan Plus, and I can honestly say that it is a component the video world has been waiting for.
Associated Equipment used for this review:
- Stacey L. Spears -
Thanks for the review on our new iScan product. Overall, this was a very favorable review. We definitely appreciate your comments and critique and are pleased with what you've written about the iScan.
As you say in your review, this was a pre-production unit. Key areas that will change from the reviewed unit to the full production units are a) Addition of YPrPb output; b) Significant improvement in the detection of film sources, particularly with noisy and poor quality sources; c) Improved handling of poor quality video signals, particularly from VCRs; d) Increased immunity to noisy video sources, reduced noise in the video output, and overall video noise improvements.
There are several issues you raised or problems you noted which will be (or already have been) fixed in the production iScans. These are a) The processing artifacts around the border of the Snell & Wilcox zone plate test pattern will be gone; b) The color/tint level will be adjusted to be properly set up. There still won't be an adjustable control, but the default setting will be correct for a source that has the correct levels. Overall output levels (black level, white level) will also be adjusted to be correct (they're a bit off on the unit you reviewed).
The video processor in the iScan was optimized for natural image sources, not video games or other similar low resolution single-field sources. We certainly made some design decisions in our video processor which result in better processing of natural images, but perhaps poorer processing for video-game type images. This is something we'll be taking a closer look at, but at this point I can't promise that the production iScan will show any real improvement over the pre-production unit you looked at.
Neil D. Newman
© Copyright 1999 Secrets of Home Theater & High Fidelity
Return to Table of Contents for this Issue.