Deinterlacing

This player uses the Genesis FLI2200 chip, and as expected did well on deinterlacing. There is only one "Auto" mode, unlike the similar range of Panasonic players which have two (Auto1 and Auto2). The single Auto mode on the Denon is equivalent to Auto1 on the Panasonic players, and uses the FLI2200 exclusively in cadence-reading mode. This works well for most material with the exception of NTSC 2-2 pulldown (30fps progressive) material, which is relatively uncommon. In PAL mode, the FLI2200 will generally recognize 2-2 properly (though we did not test any PAL material with this player).

The Basics

The 1600 had excellent core video measurements. Black and white video levels are within spec at 99.3 IRE. The player is capable of reproducing below black picture information. The YC delay in the player is right at 5ns, which is a pass. (The RP82 was around 2ns.) The frequency response is slightly less flat than the Panasonic RP82, and has a very minor rolloff at 10MHz.

No part of the image is cropped. It has all lines on the top and bottom as well as all samples on the left and right sides of the image.

The layer change takes around 1.75 seconds. Not the best we have seen but not terrible. The overall response is average and we rated that a 3, which is a sharp contrast to some of Denon's other players like the 2800, which have some of the fastest layer changes and UI responsiveness we've seen.

Overall, the 1600 delivers an excellent picture and has deinterlacing to match. The responsiveness is only average, but this our only significant gripe. We recommend this player highly.

This player receives the Secrets Recommended 480p award.

This player uses the Genesis FLI2200 chip, and as expected did well on deinterlacing. There are three deinterlacing modes on this player, but only two are worth using: Auto 1 and Auto 2. Auto 1 uses the FLI2200 to analyze the picture cadence and make decisions about film or video mode on the fly. However, the FLI2200 doesn't have a good NTSC 2-2 pulldown mode (used for 30 fps progressive material), so to improve the performance of 2-2 material, Panasonic added Auto 2, which forces the FLI2200 into film mode when it sees a 2-2 progressive flag pattern in the MPEG stream. There is also a video mode but strangely enough if you force it into video mode, DCDi™ is disabled.

Given the relative dearth of 2-2 titles in the US, Auto 2 is unlikely to add much value here. If you know that a particular title is 30 fps progressive, then it might be worth trying Auto 2, but in general we'd leave it on Auto 1. In Europe, however, 2-2 pulldown is very common, so Auto 2 might be a good default. The FLI2200 chip will auto-detect 2-2 pulldown if the source is PAL, so it might not be necessary to use Auto 2 even then. With Auto 2 the user should be watchful for combing artifacts, and switch to Auto 1 if they rear their head.

The Basics

The XP30 appears to be identical in terms of video performance to the RP82. It is just as good in the core video department as it is in the deinterlacing department. It looks like Panasonic has really done things well with their latest crop of DVD players. Unlike the RP82, the XP30 does not offer DVD-A, which may or may not matter to you.

Black and white video levels are within spec, with white at 99.7 IRE. The player is capable of reproducing below black picture information. The YC delay in the player is less than 5ns, which is within spec. The frequency response is virtually flat and virtually identical to the RP82.

No part of the image is cropped. It has all lines on the top and bottom as well as all samples on the left and right sides of the image.

The layer change takes around 1.75 seconds. Not the best we have seen but not terrible. The overall response is on average and we rated that a 3. If we could choose one area for Panasonic to improve on, it would be here. We have become spoiled by the awesome response of Meridian and Denon.

While the XP30 delivers one of the most accurate images we have seen, combined with top-notch deinterlacing, it can be improved on. To start with, we hope that Panasonic can speed up the layer change and response on future players. That said, we certainly recommend this player highly. It's not a perfect substitute for the RP-82, as it doesn't have DVD-A, but as a video player it seems essentially identical.

This player receives the Secrets Recommended 480p award.

This player uses the Genesis FLI2200 chip, and as expected did well on deinterlacing. There are three deinterlacing modes on this player, but only two are worth using: Auto 1 and Auto 2. Auto 1 uses the FLI2200 to analyze the picture cadence and make decisions about film or video mode on the fly. However, the FLI2200 doesn't have a good NTSC 2-2 pulldown mode (used for 30 fps progressive material), so to improve the performance of 2-2 material, Panasonic added Auto 2, which forces the FLI2200 into film mode when it sees a 2-2 progressive flag pattern in the MPEG stream. There is also a video mode but strangely enough if you force it into video mode, DCDi™ is disabled.

Given the relative dearth of 2-2 titles in the US, Auto 2 is unlikely to add much value here. If you know that a particular title is 30 fps progressive, then it might be worth trying Auto 2, but in general we'd leave it on Auto 1. In Europe, however, 2-2 pulldown is very common, so Auto 2 might be a good default. The FLI2200 chip will auto-detect 2-2 pulldown if the source is PAL, so it might not be necessary to use Auto 2 even then. With Auto 2 the user should be watchful for combing artifacts, and switch to Auto 1 if they rear their head.

The Basics

The RP82 is just as good in the core video department as it is in the deinterlacing department. How is it that they can get it right in a $229 player while some of their competition can’t for 2 and 3 times the price? We recommend that all of their competitors pick up an RP82 and see what a DVD image should look like!

Black and white video levels are within spec, with white at 99.7 IRE. The player is capable of reproducing below black picture information. The YC delay in the player is less than 5ns, which is within spec. The frequency response is virtually flat.

No part of the image is cropped. It has all lines on the top and bottom as well as all samples on the left and right sides of the image.

The layer change takes around 1.75 seconds. Not the best we have seen but not terrible. The overall response is on average and we rated that a 3. If we could choose one area for Panasonic to improve on, it would be here. We have become spoiled by the awesome response of Meridian and Denon.

While the RP82 delivers one of the most accurate images we have seen, combined with top-notch deinterlacing, it can be improved on. To start with, we hope that Panasonic can speed up the layer change and response on future players. That said, we certainly recommend this player highly. It is clearly the best value in DVD players as of this writing.

This player receives the Secrets Recommended 480p award.

Deinterlacing

The 2900 uses the Silicon Image SiI504 chip, which as usual passed essentially every deinterlacing test with flying colors. The only real situation where it doesn't compare quite as well to the best players we've tested is that it doesn't have a chroma filter to alleviate the effects of the ICP (Interlaced Chroma Problem, which we documented in our Chroma Upsampling article, and which is different than the CUE itself).

Because the SiI504 preserves all the chroma detail, you see the effects of ICP very clearly on certain discs that are video sourced. This is not a bug, but it does make us wish Silicon Image would take the excellent chroma filter that was developed for the iScan Ultra and incorporate it into the next version of the SiI50x chip.

The Basics

The video frequency response has a small dip around 4 MHz, but is otherwise basically flat out to 7 MHz, and then begins a gentle rolloff. The rolloff isn't enough to soften the image severely.

Our initial unit had several lines cropped from the bottom of the image, but Denon sent a firmware fix that addressed it. Any Denon dealer should be able to provide the fix
The layer change is the fastest we've ever seen. On our test disc, which is intentionally designed to be the worst layer change a player will ever need to deal with, the break was invisible. We watched it several times to be sure, and we could not see any kind of stutter or hiccup at all.

Black and white levels are well within spec, as is chroma delay. Responsiveness navigating through menus and skipping chapters is lightning fast.

This player uses a Mitsubishi MPEG chip from the same series that Pioneer has used for several years in their players. Denon told us they've been working with Mitsubishi to fix the chroma bug and they've largely succeeded. We see no evidence of chroma bug at all on 3-2 or 2-2 normally flagged film material. Unfortunately, the chip doesn't handle the 3-2 alternating progressive flag issue properly, and alternates every other frame between interlaced and progressive upsampling. This actually looks worse than just staying in interlaced mode (though only by a little) because of the constant flicker as the upsampling mode changes. Several scenes in Monster's Inc show this problem clearly, and since all recent Disney releases have the problem, we imagine there are many more movies that will not look their best in this area. Denon has looked into this issue and apparently they don't believe a firmware fix is possible for it. Still, we are glad that Denon took the time to fix the problem for a wide range of movies. Toy Story looks great on this player, as does Moulin Rouge and many other films that show the chroma bug clearly on other players.

Overall we think this player has excellent progressive video quality marred by one significant flaw - the 3-2 alternating flag chroma upsampling issue. Whether that flaw is a deal-breaker is really a personal choice. For people looking for a combo SACD/DVD-A player with top-notch progressive video, this is competitive with the Yamaha DVD-S2300 player we looked at a few months ago. It exceeds it in some areas, like chroma delay, and doesn't do as well in a couple of areas mentioned above. With the 2900, Denon hasn't quite made a unit that passes every one of our tests, but they've come pretty close.

This player uses the SiI504, but adds an extra twist - it has a flag-reading mode as well. This could potentially be useful in the rare situations where the SiI504 isn't going into film mode, but you know the disc has good flags. The Denon's flag-reading mode is a very simple one, which doesn't handle the common "alternating progressive flag" problem.

The names of the modes could be better - "Mode 1" is the cadence-reading mode, and "Mode 2" is the flag-reading mode. In the US, most of the time Mode 1 is going to do the best job. If you have the player in Mode 1 and a film looks soft, it might be worth trying Mode 2. If you have the player in Mode 2 and start seeing lots of combing, switch it back to Mode 1.

The Basics

In addition to the deinterlacing mode switch, there is another video mode option that lets you select between "auto," "film," and "video." This makes it sound like a deinterlacing mode switch, but it's not. From what we can tell it changes two things. chroma upsampling and pause type.

Pause type is pretty simple: “Auto” uses the progressive_frame flag to determine whether to do frame or field pauses. (A frame pause shows the whole frame, while a field pause takes a single video field and scales it up, which looks pixellated and soft.) In "film" mode, all pauses are frame pauses, which can cause the image to look combed if the frame you're pausing on has movement between fields. "Video" mode causes all pauses to be single fields. We can't imagine why anyone would want video mode.

The other thing that changes with the video mode switch is chroma upsampling. On the few other players that allow selectable upsampling, “Auto” was clearly the way to go. However, on the Denon “Auto” mode causes it to (wrongly) use interlaced upsampling on "alternating progressive flag" material like Monsters Inc. Amazingly enough, "film" mode is not forced progressive upsampling. It's the same as “Auto” mode, except that it fixes the "alternating progressive flag" problem. On video material, it continues to use interlaced upsampling, even in "Film" mode. Frankly, this doesn't make sense to us. Most people are going to leave the setting on “Auto”, which means they're going to be seeing the Chroma Upsampling Error on a wide variety of popular films. And if they switch to "film," which we certainly recommend, they'll get combed frames when they pause on certain discs.

We don't think they should have offered the Auto chroma upsampling mode. As far as we can tell, "film" is the correct chroma mode for all situations, and passed all of our chroma upsampling tests. If there is material that it's incorrect on, we don't know what it is.

The initial unit we received had the chroma error. However, we applied a firmware fix Denon sent us, and it did fix the problem. Current units at retail supposedly already have this fix or a new, fixed MPEG decoder that has an equivalent fix in hardware. Even though it is correct, it was not the best upsampling we have seen. We would describe it as a little blocky, but it looks far better than the best wrong upsampling (which we would say is the Toshiba).

White is virtually perfect at 99.26 IRE and black is the same at 0 IRE. Both Y-Pb and Y-Pr have a pretty large delay. We estimate it is close to 1 pixel, which is 37ns at 480p. (74ns at 480i) They can probably correct this by making a register change inside of the Analog Devices Video Encoder. The frequency response on this player is also hot. There is no roll-off but rather a steady climb up. At 10 MHz the player is up 1 dB.

There are 0 lines cropped from the top, 2 from the bottom, 5 from the left, and 1 from the right. The disparity between left and right pixel cropping is not great, in our opinion, because it causes the image not to be centered. It's a tiny thing, but one that we think should be easy to get right.

The layer change is very fast, clocking in at around 0.25 seconds. We rate the overall performance a 5. This player just smokes when navigating a disc.

In general, this is a good performer. It has excellent video quality, excellent deinterlacing, and performs well. It's heavy, for those who like that kind of thing, and it's not just lead weights. The Y/C delay and the quirky video mode switch are the only disappointments.

Deinterlacing

The deinterlacing performance and issues were essentially equivalent to the DVD-3800, also covered in this shootout. Please refer to the deinterlacing section of that review, as the performance and modes were identical.

The Basics

The same issues we found with the "video mode" switch in the DVD-3800 menus were also present on the DVD-9000. Please read the "Basics" section of that review for more on those quirks.

White is virtually perfect at 100.52 IRE and black is the same at 0 IRE. Both Y-Pb and Y-Pr have a pretty large delay. We estimate it is close to 1 pixel, which is 37ns at 480p. (74ns at 480i) They can probably correct this by making a register change inside of the Analog Devices Video Encoder. The frequency response looks good.. The player is down .23 dB at 10 MHz.

There are 0 lines cropped from the top, 2 from the bottom, 5 from the left, and 1 from the right.

The layer change is very fast clocking in at around .25 seconds. We rate the overall performance a 5. This player just smokes when navigating a disc.

This player is using the high end 14-bit / 108 MHz video encoder from Analog Devices. Even so, the output was not as clean as the Krell DVD Standard. This goes to show that as good as the DAC is, the analog anti-imaging filters that come later are just as important.

In general, this is a good performer. It has excellent video quality, excellent deinterlacing, and performs well. It's heavy, for those who like that kind of thing, and it's not just lead weights. The Y/C delay and the quirky video mode switch are the only disappointments.

The video section of this player is based on the same core design as the Panasonic RP-82, which is a good thing. There are some changes, however. The analog video section is different, which means that it measures somewhat differently. The DVD drive had to be changed to allow for reading SACDs, which have a slightly different physical format than other DVD-style media. Also, the whole audio output section is a different design, thought we don't cover audio output in the shootout.

Deinterlacing

This player is based on the RP-82, and thus shares its excellent deinterlacing as well as a few quirks relating to its two main deinterlacing modes. Rather than repeat the same information, we recommend you read our deinterlacing writeup for the RP-82, in this same shootout report.

The Basics

As we mentioned, this player does not have the chroma error. It is based on the same great MPEG decoder that the Panasonic RP82, RP91, and H2000 use.

White measures at the bottom end of our passing range at 98.14 IRE. Black is exactly where it belongs. The DVD player is also capable of reproducing below black picture information.

The frequency response is rather rolled off. At 10 MHz the video is down 2.11 dB. The image from this player is just a little softer than the rest, though this is not necessarily wrong. Y-Pb is greater than 5ns while Y-Pr is less than 5ns. It is worth noting that the Pb and Pr channels are delayed in the opposite directions, which increases the visibility of the chroma delay. There are 0 lines cropped from the top and bottom and 0 samples cropped from the left and right, which is excellent.

This player has the cross-color suppressor in the FLI2200 on, so you may see some flicker on saturated colors like those found in Toy Story.

The layer change on this player takes approximately 1.75 seconds. Not the best but in the range we consider a pass. Overall we rated this player a 3 in performance.

In general this is a very good player from a video standpoint, though the chroma delay is disappointing. The flaws, however, are relatively minor, and for the most part this is equivalent in video quality to the Panasonic RP-82, which is one of the best players we've seen. Certainly this player is head and shoulders better than any of the other universal players we've looked at.

Deinterlacing

This player uses the Genesis FLI2200 chip, and as expected did well on deinterlacing. There are three deinterlacing modes on this player, but only two are worth using: Auto 1 and Auto 2. Auto 1 uses the FLI2200 to analyze the picture cadence and make decisions about film or video mode on the fly. However, the FLI2200 doesn't have a good NTSC 2-2 pulldown mode (used for 30 fps progressive material), so to improve the performance of 2-2 material, Panasonic added Auto 2, which forces the FLI2200 into film mode when it sees a 2-2 progressive flag pattern in the MPEG stream. There is also a video mode but strangely enough if you force it into video mode, DCDi™ is disabled.

Given the relative dearth of 2-2 titles in the US, Auto 2 is unlikely to add much value here. If you know that a particular title is 30 fps progressive, then it might be worth trying Auto 2, but in general we'd leave it on Auto 1. In Europe, however, 2-2 pulldown is very common, so Auto 2 might be a good default. The FLI2200 chip will auto-detect 2-2 pulldown if the source is PAL, so it might not be necessary to use Auto 2 even then. With Auto 2 the user should be watchful for combing artifacts, and switch to Auto 1 if they rear their head.

The Basics

The performance of the CP72 on the basic video measurements was nearly identical to the RP82, with two exceptions: there was slightly more chroma delay and the frequency response was different.

The frequency response isn't as flat as the RP82 (or the RP62), though not so rolled off as to make the image look unpleasantly soft. There is a bit of peaking in the middle frequencies, which could make this player appear sharper or more detailed. At 6 MHz, the RP82 was up 0.17 dB while the CP72 is up 0.44 dB.

The chroma delay was slightly out of spec at around 10ns with both channels delayed the same amount in the same direction. We measured two RP82s and the XP30 at 2-3ns.

Black and white video levels are within spec at 99.3 IRE. The player is capable of reproducing below black picture information.

No part of the image is cropped. It has all lines on the top and bottom as well as all samples on the left and right sides of the image.

The layer change takes around 1.75 seconds. Not the best we have seen but not terrible. The overall response is on average and we rated that a 3. If we could choose one area for Panasonic to improve on, it would be here. We have become spoiled by the awesome response of Meridian and Denon.

Overall, this is a very good player and very similar to the RP82. While the Y/C delay was just outside the EIA spec, we still think the player is an excellent value and worth considering.

Deinterlacing

Overall deinterlacing performance was excellent, as the player uses the Faroudja-designed FLI2200.

This has the same deinterlacing modes as the RP-82, and rather than repeat the lengthy comments on when to use the different modes, we recommend you read the deinterlacing notes in our RP-82 writeup in this shootout report.

The Basics

This player does not have the Chroma Upsampling Error (CUE). It is based on the same great MPEG decoder that the Panasonic RP82, RP91, and H2000 use.

White measures virtually perfect at 99.68 IRE. Black is exactly where it belongs. The DVD player is also capable of reproducing below black picture information.

The frequency response is very good. At 10 MHz the video is down .76 dB. Y-Pb is greater than 5ns while Y-Pr is less than 5ns. It is worth noting that the Pb and Pr channels are delayed in the opposite directions, which increases the visibility of chroma delay. There are 0 lines cropped from the top and bottom and 0 samples cropped from the left and right.

This player has the cross-color suppressor in the FLI2200 on, so you may see some flicker on saturated colors like those found in Toy Story.

The layer change on this player takes approximately 1.75 seconds. Not the best but in the range we consider a pass. Overall we rated this player a 3 in performance.

In general this is a very good player, though without really anything specific to recommend it over the Panasonic player that it is based on, and the chroma delay is disappointing. The flaws, however, are relatively minor, and for the most part this is equivalent to the Panasonic, which is one of the best players we've seen.

This player uses the FLI2200, and had performance similar to other players with that chip. Recovery time from video to film mode was too high for us to give full marks, and NTSC 2-2 pulldown wasn't handled, which isn't a huge issue for the USA. This player does offer the ability to enable NTSC 2-2 pulldown, but when enabled we saw field inversion on our test disc. They do not let you invert the fields manually so we don’t know if that would have solved our problem. Otherwise, it passed with flying colors.

The Basics

Most of the time, this player does not have the Chroma Upsampling Error. Krell is using the same LSI MPEG decoder that is found in the Proceed PMDT and several Sony DVD players like the 7700 and 9000. Unlike the Sony players, the Krell does not have the chroma bug with "standard" 3-2 encoded DVDs. The problem the Krell has is on DVDs encoded with alternating progressive flags, see image below. You will have to look close to see it, sorry it is so small. This is a simulation of what it looks like. It is much larger when viewed in person. You will probably not see it on the 480p out because of the filtering in the FLI2200, but if you feed the 480i into an external processor as we did, it will become obvious. If the external processor is also using the FLI2200, then it will probably masked. In this case, they switch upsampling methods every frame, which results in a noticeable and annoying flicker. The FLI2200 will mask this on the 480p output, but if you plug it into an external processor, it's obnoxious. This is something that could be corrected, but would require some work.

When casually looking at the sharpness pattern on Avia, this DVD player provided the least amount of ringing. The Avia pattern is a severe test in this area. This suggests that the analog anti-imaging filters used after the DACs are exemplary.

The DVD Standard is using a 54 MHz DAC and it outperformed the Denon DVD-9000, Pioneer 47Ai, and Sony DVP-NS999ES, which are all using 108 MHz DACs. This goes to show just how important the analog video section is. The higher sampling DACs are supposed to make analog filtering easier, but clearly a good job can be done with a lower-frequency DAC. (It is a form of upsampling, just like that used for audio DACs.)

The white level on the Krell was just about perfect and well within spec. The DVD Standard also had the flattest frequency response of all players tested in shootout 3. The DVD Standard is also capable of reproducing below black picture information.

However, not all is perfect on this player, at least not yet. While Y-Pr was below 5ns, Y-Pb was >30ns. The Kenwood from Shootout 2 suffered from the same problem and we believe that, like the Kenwood, this can be solved with a firmware update.

Krell has provided at least one firmware update since we first received the player, which exposed the chroma enhancement controls in the FLI2220. Since the chroma enhancement features are controversial (we prefer them set low or off), this update was in our opinion a good thing.

There is a menu option called animation mode. We believe this is supposed to toggle the cross-color suppression (CCS) feature off and on, but currently it does not do anything. The CCS is set to on at the moment. With it in this position you will see flickering on some saturated reds like those found in Toy Story.

Pixel cropping on this player is severe. 13 samples are cropped from the left side of the image, which we think is well beyond any standard of reasonableness. Aside from the fact that picture information is being lost unnecessarily, the fact that it's all from one side shifts the image noticeably.

Next to the WMC PC, this had the fastest layer change of any DVD player we have seen. It even nudged out the Denon and Meridian in terms of speed. We estimate it takes around .1 seconds under the worst case scenario.

We gave the DVD Standard a 4.5 in response time. It would have received a 5, but we felt the next-chapter skip was not as fast as the Denon or Meridian.

This is overall a generally good player, verging on excellent. The pixel cropping is disappointing, and we hope that a future update will provide a working animation mode, a fix for the chroma upsampling on alternating flag material, and a fix for the Y-Pb delay problem. In a player at this price point, we think it's only reasonable to expect these details to be done right.

This player uses the Faroudja-designed FLI2200 chip for deinterlacing, and thus has excellent overall performance in this area. Like almost all players using the FLI2200 it recovers from video mode to film mode too slowly to get full points, and it doesn't go into film mode on NTSC 2-2 pulldown. This feature is available in the FLI2200 but it is often always disabled because of it not being reliable. In the rare case we have seen it enabled, it ends up inverting the fields. You can also control this, but shouldn’t it be automatic?

The Basics

We are not sure which MPEG encoder this player is using. Denon tells us it is not the ESS, but we also know it is not using an MEI like the Denon 1600. This player does suffer from the chroma error, but it is masked to a degree if you are using the 480p outputs, by the FLI2200.

Black is fine at 0 IRE. White is low at 96.04 IRE. The frequency response on this player rolls off sharply. It is down 1 dB at 4 MHz. It is down 2 dB at 7.16 MHz. And finally it is down 3.24 dB at 10 MHz. This resulted in a very soft image. We wonder if the anti-imaging filters were optimized for 480i instead of 480p, which would explain the steep roll-off.

On the plus side, the YC delay was virtually perfect coming in at under 2ns. There are 0 lines cropped from the top and bottom and 0 samples cropped from the left and right.

The layer change takes around 1.75 seconds to complete. The player also falls into the average range and we rate its overall performance as 3.5.

This is a generally good player, with a lot to recommend it. The fairly steep frequency roll-off and the Chroma Upsampling Error are the only significant concerns.

This player uses the Genesis FLI2200 chip, and as expected did well on deinterlacing. There are three deinterlacing modes on this player, but only two are worth using: Auto 1 and Auto 2. Auto 1 uses the FLI2200 to analyze the picture cadence and make decisions about film or video mode on the fly. However, the FLI2200 doesn't have a good NTSC 2-2 pulldown mode (used for 30 fps progressive material), so to improve the performance of 2-2 material, Panasonic added Auto 2, which forces the FLI2200 into film mode when it sees a 2-2 progressive flag pattern in the MPEG stream. There is also a video mode but strangely enough if you force into video mode, DCDi™ is disabled.

Given the relative dearth of 2-2 titles in the US, Auto 2 is unlikely to add much value here. If you know that a particular title is 30 fps progressive, then it might be worth trying Auto 2, but in general we'd leave it on Auto 1. In Europe, however, 2-2 pulldown is very common, so Auto 2 might be a good default. The FLI2200 chip will auto-detect 2-2 pulldown if the source is PAL, so it might not be necessary to use Auto 2 even then. With Auto 2 the user should be watchful for combing artifacts, and switch to Auto 1 if they rear their head.

The Basics

White is actually a little low and below spec. The player is capable of reproducing below black picture information. The YC delay is longer than 5ns, which is outside of spec. The frequency response looks good.

While no lines are cropped on the top or bottom, there are 5 samples cropped from the left and 6 cropped from the right.

The layer change takes around 1.75 seconds. Not the best we have seen but not terrible. The overall response is on average and we rated that a 3. If we could choose one area for Panasonic to improve on, it would be here. We have become spoiled by the awesome response of Meridian and Denon.

All in all, not a bad player, but it is not quite up to the level set by the RP82 (or even its predecessor the RP56). However, it's a cheaper player, and readily available in more stores than the RP82. At its price point, it's a perfectly fine player.

Deinterlacing

This player uses the new Sigma Designs all-in-one chip, which has basic flag-only film detection, combined with a simple vertical filter for video mode deinterlacing. As with all chips of this type, it works well on film-based content when the flags are standard, but when the flags are non-standard or the content is not from a film or other progressive source, the image is much softer than it needs to be. In the worst cases, when the progressive flag is set to "true" improperly, the player will comb constantly. Its resistance to bad edits was poor, with 5 total combs in the Big Lebowski montage.

On the plus side, it handles the common "alternating progressive flag" encoding problem well, correctly treating it as progressive content.

The Basics

Black level on the analog output is incorrect at 7.6 IRE instead of the standard 0 IRE. There is no switch we could find to adjust it to 0 IRE black. White level is low, at 97.3 IRE, which was too low to pass. Both Y-Pb and Y-Pr have delay greater than the EIA standard of 5ns, again only on the analog output. The player does pass blacker-than-black in progressive mode, though we should note that because the player only does 7.5 IRE black, we were unable to check blacker-than-black at 0 IRE.

The video frequency response on the analog output of this player is astonishingly rolled off, actually down a whopping 13 dB at 10 MHz. Considering that DVD goes all the way to 13.5 MHz, that's pretty bad. The rolloff was consistent even in the higher resolutions. Through the DVI connection, of course, there is no rolloff.

There are 0 video lines cropped from the top and bottom, and 0 video samples on the left and right, which is excellent.

The layer change is fast, at around 0.5 seconds, and responsiveness was 2.5, which is about average.

The scaling and DVI features of this player are what people are buying it for, and those features work well, as long as you use them together. The analog outputs have too many problems for us to recommend the player for use via an analog connection, but through a DVI connection the clarity is just stellar, with no chroma delay and no rolloff. The scaling engine in the player is excellent, producing clean output with no obvious jaggies or moire even on difficult material like the Avia resolution patterns.

The player is not without its quirks, however. In our limited time working with it, we found that while it has a zoom that is useful for zooming in on non-anamorphic letterboxed discs, the zoom snaps back to normal when you hit the chapter next button, and the subtitles turn off while it's zoomed in. These are clearly firmware quirks, and perhaps can be corrected, but they are nevertheless annoying.

Another minor issue is that the player always window-boxes 4:3 discs for you, even if your display has aspect ratio controls of its own. This is non-optimal for people with digital displays, as they often only sample 720 times across a scan line when taking in 480p content. With the window-box bars on, there are more than 720 sample intervals across the screen, and thus picture information gets lost. You can clearly see this on our Sony VW10HT test projector. We put the player in 16x9 mode, then viewed a 4x3-encoded resolution chart. Three fourths of the horizontal resolution was lost because the projector wasn't sampling more than what it thought it "needed" to do to capture the full 480p signal. If we put the player in 4x3 mode so it wouldn't add the window-box bars, then window-boxed using the display's aspect controls, the full resolution was visible. The player could really use a mode where it doesn't try to do any aspect control of its own, for people who have a display that can do a better job of aspect control than the player can. This is not an issue when using the DVI output.

Finally, we noted that when the player is in 16x9 mode, whenever subpictures (like menus and subtitles) are displayed on screen for 4x3 content they look very odd, like they've been scaled using a simplistic algorithm. The effect is jarring. The menus on Avia are a good example of this phenomenon because Avia contains both 4x3 and 16x9 menus. Put the player in 4x3 mode, and the effect disappears. We have seen similar behavior on other DVD players.

In the end, while we are impressed with the clarity that DVI brings, we're not ready to lose the excellent deinterlacing of the best chipsets in order to get it, and we weren't happy with the interface quirks mentioned above. For those who just want an inexpensive way to feed a digital display with a DVI connection and aren't as picky about deinterlacing and quirky UI, the Bravo D1 may well fit the bill.

Deinterlacing

This is the first player we've looked at that uses Panasonic's new all-in-one MPEG decoder/deinterlacer combo chip, and it's a big step down in many areas from their previous line of players, which used the Faroudja-designed Genesis Fli2200 chip. This chip appears to use a hybrid flag/cadence reading algorithm and not a very good one. It fails the chapter stops test and fails the bad-edit test with 5 combs, which is not good at all. Like all the other all-in-one chips thus far, the video deinterlacing is not motion-adaptive.

The Basics

The biggest surprise in this player is that the new MPEG decoder has the common chroma upsampling error (CUE) on 3-2 material (but amazingly not on 2-2 material). This is just astounding, because Panasonic has always passed this test with flying colors, starting with their very first players. Just when many other decoder manufacturers are starting to ship chips without this flaw, Panasonic takes a massive step backwards. To say that we're disappointed is an understatement.

Chroma delay is more than 5 ns, which is outside the EIA standard for consumer-grade video. There is significant pixel cropping on three sides of the image.

On the plus side, the black and white levels are within spec, and the player passes below-black content on the progressive output. The video frequency response is very flat with a small rolloff at the higher frequencies, which is perfectly fine.

If this player is representative of the rest of Panasonic's line this model year (and we think it is, based on some informal testing of other players in the line) then Panasonic has moved in the wrong direction. Given the shaky deinterlacing and the chroma bug, we can't really recommend this as a progressive video player.

Deinterlacing

Things haven't changed much since last we looked at a Pantera-based player. On the plus side, this player handles alternating-progressive-flag material like Monsters Inc. properly, which is good. The chipset is still completely flag-based, and thus has no way of identifying film material other than the flags on the disc. The video deinterlacing is still not motion-adaptive in any way. It looked to us like the video deinterlacing was defaulting to simple single-field interpolation (or "bob"). The chipset has excellent scaling and upsampling, so on some material, the bob algorithm wasn't as obvious. But in general, we'd have to call the deinterlacing bare-bones.
There are also no user adjustments to allow you to tweak the deinterlacing modes, as there are on the JVC, Malata, and other Pantera-based players. Normally we think the user shouldn't have to adjust the settings from disc to disc to get a good picture, but given the quirks of this player and the potential for very ugly results on certain discs, we'd have preferred at least the option of forcing video mode. In general, though, most people would be better off switching the player to interlaced mode if a disc is combing too often.

The Basics

Like all other Pantera-based players we've evaluated, this player has no chroma bug (CUE) at all. The ICP is not really a problem through the progressive output because the very soft video deinterlacing cuts both luma and chroma resolution in half, effectively hiding the chroma issues as a side effect.

The white level was just a tiny bit high at 103 IRE, just outside our "pass" criterion. The inter-channel delay was nearly perfect at <2 ns delay on all channels, which meets EIA professional standards. There were 3 samples cropped on both the left and right side of the image, but none on the top or bottom. The player does pass blacker than black on the progressive output.

Video frequency response was impressively flat with a mild rolloff near the top, which is perfectly fine. The layer change was middle of the pack at 1 second on our test disc, and the responsiveness was only so-so.
Overall this is a decent player, and actually one of the better ones we've looked at that uses this chipset. The one major downside is the deinterlacing, which doesn't work well for anime, obscure titles, video-sourced titles, and essentially anything other than big-name Hollywood titles (and not well on some of those).
 

This player uses the new Sigma Designs all-in-one chip, which has basic flag-only film detection, combined with a simple vertical filter for video mode deinterlacing. As with all chips of this type, it works well on film-based content when the flags are standard, but when the flags are non-standard or the content is not from a film or other progressive source, the image is much softer than it needs to be. In the worst cases, when the progressive flag is set to "true" improperly, the player will comb constantly. Its resistance to bad edits was poor, with 5 total combs in the Big Lebowski montage.

On the plus side, it handles the common "alternating progressive flag" encoding problem well, correctly treating it as progressive content.

The Basics

Black level on the analog output is incorrect at 7.6 IRE instead of the standard 0 IRE. There is no switch we could find to adjust it to 0 IRE black. White level is low, at 97.3 IRE, which was too low to pass. Both Y-Pb and Y-Pr have delay greater than the EIA standard of 5ns, again only on the analog output. The player does pass blacker-than-black in progressive mode, though we should note that because the player only does 7.5 IRE black, we were unable to check blacker-than-black at 0 IRE. However, like Toshiba players, we have passed this player on the blacker-than-black test.

The video frequency response on the analog output of this player is astonishingly rolled off, actually down a whopping 13 dB at 10 MHz. Considering that DVD goes all the way to 13.5 MHz, that's pretty bad. The rolloff was consistent even in the higher resolutions. Through the DVI connection, of course, there is no rolloff.

There are 0 video lines cropped from the top and bottom, and 0 video samples on the left and right, which is excellent.

The layer change is fast, at around 0.5 seconds, and responsiveness was 2.5, which is about average.

The scaling and DVI features of this player are what people are buying it for, and those features work well, as long as you use them together. The analog outputs have too many problems for us to recommend the player for use via an analog connection, but through a DVI connection the clarity is just stellar, with no chroma delay and no rolloff. The scaling engine in the player is excellent, producing clean output with no obvious jaggies or moire even on difficult material like the Avia resolution patterns.

The player is not without its quirks, however. In our limited time working with it, we found that while it has a zoom that is useful for zooming in on non-anamorphic letterboxed discs, the zoom snaps back to normal when you hit the chapter next button, and the subtitles turn off while it's zoomed in. These are clearly firmware quirks, and perhaps can be corrected, but they are nevertheless annoying.

Another minor issue is that the player always window-boxes 4:3 discs for you, even if your display has aspect ratio controls of its own. This is non-optimal for people with digital displays, as they often only sample 720 times across a scan line when taking in 480p content. With the window-box bars on, there are more than 720 sample intervals across the screen, and thus picture information gets lost. You can clearly see this on our Sony VW10HT test projector. We put the player in 16x9 mode, then viewed a 4x3-encoded resolution chart. Three fourths of the horizontal resolution was lost because the projector wasn't sampling more than what it thought it "needed" to do to capture the full 480p signal. If we put the player in 4x3 mode so it wouldn't add the window-box bars, then window-boxed using the display's aspect controls, the full resolution was visible. The player could really use a mode where it doesn't try to do any aspect control of its own, for people who have a display that can do a better job of aspect control than the player can. This is not an issue when using the DVI output.

Finally, we noted that when the player is in 16x9 mode, whenever subpictures (like menus and subtitles) are displayed on screen for 4x3 content they look very odd, like they've been scaled using a simplistic algorithm. The effect is jarring. The menus on Avia are a good example of this phenomenon because Avia contains both 4x3 and 16x9 menus. Put the player in 4x3 mode, and the effect disappears. We have seen similar behavior on other DVD players.

In the end, while we are impressed with the clarity that DVI brings, we're not ready to lose the excellent deinterlacing of the best chipsets in order to get it, and we weren't happy with the interface quirks mentioned above. For those who just want an inexpensive way to feed a digital display with a DVI connection and aren't as picky about deinterlacing and quirky UI, the Bravo D1 may well fit the bill.

Deinterlacing

Pioneer has appeared to make some improvements with their deinterlacing compared to the previous 47A. It no longer combs on subtitles and it is able to deal with encoding errors on subtitles.

They also appear to have fixed the strange combing problem on the Galaxy Quest "bonus features" menu. At least we were not able to reproduce it on any of the universal players based on the Pioneer electronics.

But in general, this suffers from the same basic problems that most flag readers have, and some that most don't have. It has a cadence analysis algorithm, so it does OK on film material that is flagged as video, but it failed our bad edit test with the worst score in the shootout (equaled only by the other players based on the Pioneer design).

The Basics

The bad news is that the MPEG decoder still suffers from the Chroma Upsampling Error. It appears to be just as bad as previous Pioneer players. Pioneer players are on the severe end of the Chroma Upsampling Error scale.

The white level on this player appeared to be pretty dead on and the over all frequency response also appeared to be pretty good. The 47Ai is also capable of passing below black picture information.

The frequency response on this player is nice and flat with a small roll-off at 10 MHz.

The component timing is outside of spec. While you can fix level differences among players with the picture controls, there is virtually nothing you can really do with the timings of the component signal. While there is a YC delay adjustment in the menus, it is too coarse to correct the problem. One option is to get different lengths of cable for each video signal (Y, Pb, and Pr) but that has its own set of problems.

The layer change takes around 1.25 seconds to complete under the worst case scenario. We rated the overall response time of the player a 3, which is a hair more responsive that the previous DV-47A.

There are 3 lines cropped from the bottom of the image. This is unacceptable in our opinion. There is no technical reason to crop the top or bottom of the image.

The most positive thing we can say about this player is that it's the best Pioneer player we've looked at. Sadly, that really isn't enough. This player should be better, given its price and its THX certification. Given that you can now get a universal SACD/DVD-A player from Yamaha with excellent video, we can't recommend this player at all.

Deinterlacing

This player uses the deinterlacing built into the Ziva5 to create the progressive image and not any Sony algorithms. It appears to make decisions by analyzing the overall pattern of the flags and not by blindly following each flag literally. Its biggest blind spot is that it appears to treat all 480p/30 (aka 2-2 pulldown) sources as video. It also doesn't have motion-adaptive video deinterlacing, which isn't great.

The Basics

Given that all previous Sony players other than the 7000 have had the Chroma Upsampling Error, we were pleasantly surprised when we viewed chapter 4 of Toy Story. Gone was the Chroma Upsampling Error we are so used to seeing from Sony DVD players. The most surprising part is that it is missing from this $599 package and yet it still exists in their new flagship standalone player.

This fact alone made us peel the top back on the chassis and much to our surprise; this DVD player is now using the LSI Ziva5. Early Sony products like the 7700 and 9000 used an older LSI MPEG decoder, which was before LSI acquired C-Cube.

This MPEG decoder was not free of problems however. The Toy Story menu (which is 2-2 or 30fps progressive) still showed the chroma error. And there is another problem: the chroma is shifted down vertically from the luma when the player uses the interlaced chroma upsampling algorithm. It is shifted at least one pixel, perhaps more, which is visible and not good. We have seen this previously in the Samsung 721, which was the first DVD player to use the new Ziva5. We told C-Cube about it, but so far we haven't seen a fix.

The shifted chroma on interlaced content is annoying, but it's really obvious when the flags alternate between interlaced and progressive. At the end of our bad edit test on Big Lebowski, the chroma appears to bounce up and down on screen. We have provided a sample image below so you can see the effect. We've since seen it on a variety of material, though this particular image is among the best for showing it. If you look at the star on the right, you will see it pulsate up and down.

This was the slowest player we used in the shootout -- painfully slow at times. The layer change on this player took around 4 seconds. We rated the overall performance as a 1.5. If responsiveness is important to you, look elsewhere.

We did not take any video measurements on this player due to limited time with this unit.

Overall, we're happy that Sony has finally released a player without the Chroma Upsampling Error, or at least without the most common form. And given the implied price of this player as part of a whole theater package, we think it's an OK value. If you want the best performance, though, we recommend looking elsewhere.

The 7700 was the direct replacement for the 7000 and it was a step backwards in the performance department for Sony. It was also the first Sony player to have the Chroma Upsampling Error, something the 7000 did not have.

Sony replaced the 7700 with the 9000. In the build quality department, the 9000 was better built than either the 7700 or 7000, but it still suffered from the Chroma Upsampling Error. It had some impressive video performance but wasn't so good in the deinterlacing department.

It appears the new 999 is the replacement for the 9000, but just like the 7700 did not meet the bar set by the 7000, the 999 does not meet the bar set by the 9000.

Deinterlacing

Sony is no longer using the gmVLX1A they used on the 9000, which is generally good. We believe they are using some form of their DRC technology. Mostly it works better than the gmVLX1A, but there are some new quirks, including some that we didn't see in the previous players like the 700P.

There are a couple of adjustments to the deinterlacer available, but we left both controls in their default position as we don’t expect users to have to adjust the deinterlacer for each film.

Overall, the deinterlacing was OK on this player. It failed the bad edit test, but did somewhat better than some other players in the shootout.

The biggest surprise was the Super Speedway test. This disc is encoded as film and the 3-2 pulldown signature is just fine. The scene we use to test contains a lot of fine detail and the test was introduced to check the bad edit settings in the FLI2200 from Sage. We never really expected any other player to fail this test and we were surprised that Sony dropped to video during the pan of the stadium. The 999 did eventually switch back to film mode, but it should never have been in video to begin with. There are at least 40 frames spent in video, which shows up as a moiré pattern in the bleachers.

We suspect that the player has some technology similar to the bad edit detector in the FLI2200, which detects things that look like combing and drop to video mode when they occur. If so, it does not appear it has been fine-tuned yet. In fact, when we looked further we found several occurrences of this problem. This can also be seen in the film ET. When ET comes out of the closet, the window blinds in the background break up into moiré. This same behavior could be found on More Tales of the City, it would constantly switch between film and video mode. It was very distracting. This appeared to be a common theme on this player with every disc we tried. If you did not know it was not supposed to flicker, shimmer, or be filed with moiré you might not know it was a problem.

The Basics

The 999 still suffers from the Chroma Upsampling Error, which is unusual considering Sony has a lower priced Home theater in a Box that has resolved this problem. Why would Sony use an inferior MPEG decoder in their supposed flagship product?

The layer change was slower than average taking around 2 ľ seconds on our disc. The overall performance was also a bit slower than normal, but far from the slowest in the bunch. We rated it a 2.5 in terms of overall responsiveness.

Out of the box, the image appeared to be very punchy. We quickly noticed that most whites were clipped on our display. We pulled contrast way down from where it had been and the detail in the whites returned. Once we got it on the scope we clearly saw why the whites where clipped and the image was so punchy. We measured the top of the ramp at 106 IRE, which is very hot. This was a lot hotter than all of the other DVD players. If you put side by side and did not bother to adjust any picture controls, the Sony would appear to have more snap and be brighter than the other player. However this is easily corrected by pulling contrast down to where it belongs.

This is a very large error. Was it simply a mistake on their part or a gimmick to make it stand out on the showroom floor? We don’t know, but we do think people have a right to expect standard voltage levels from an expensive flagship player. Amazingly, this player actually has several video presets and they actually change the video levels, but none of them is correct.

On the good side, the component timing on this player was the best of the bunch and was pretty much dead on. It was the only player in shootout 3 to be under 2ns.

On the bad side, this was one of only two players that was not capable of reproducing blacker than black picture information. All below black information was clipped. This causes problems with the shadow detail, as there are a lot of films with below-black information in the darkest details on screen. All of those details become flat even black and look unnatural with the black is clipped. There is no technical reason to do it, and no limitation of the DACs they chose; it's just poor attention to details.

The frequency response of the player was very good overall. This player is using the new 14-bit / 108 MHz video DAC from Analog Devices and this one part costs more than some Apex DVD players.

There was 1 line cropped from the bottom of the image and 1 sample cropped from the right.

In general, we think Sony can do better than this. Especially at this price point, it's just not competitive with other players that get more core details right. The shimmering problem with high detail scenes is alone annoying enough to keep us from recommending this player.

Deinterlacing

The Cirrus Logic chip had the same issues as most of the other combined MPEG decoder/deinterlacers we've seen - it's a flag reader, and failed most of the deinterlacing tests that involve flag patterns other than the standard NTSC 3-2 film pulldown.

Amazingly, it wouldn't go to film mode on material that was encoded with the extremely common "alternating progressive flag" problem unless we forced it into film mode. There's no excuse for this. There must be hundreds of films with this problem, including many huge hits like Titanic, Austin Powers, and Monsters Inc. In fact, all of the players passed this test in shootout 1. We assumed that every chipset was designed to deal with this problem properly, so we stopped running the test in shootout 2. If we hadn't (just for grins) run the test again on this DVD player, we might never have found out that there are still players that fail to recognize this issue. Clearly, we can't assume that every chipset manufacturer has done their homework.

The Basics

What do you expect when you pay $69 for a DVD player? The biggest surprise for us was that this player does NOT have the Chroma Upsampling Error CUE). That pretty much throws down the gauntlet - if they can make a player with no chroma error for $69 retail, there's no excuse for any other manufacturers not to fix it.

White on this player is extremely low at 72 IRE, instead of the standard 100 IRE. You will really need to crank contrast up on your display, which will of course cause all the other inputs to have horrible white clipping, unless you can dedicate a separate memory on your display just for this DVD player. This player is one of two players in shootout 3 that is not capable of reproducing below black picture information. We used to think that below black was only used for standard PLUGE test patterns for setting black level, but it turns out that hard clipping the below-black information causes odd artifacts in shadow areas. Many films have shadow details encoded below black. When the player clips them off, they're lost for good and no tweaking of the display can cure it.

The video on this player is soft! We wonder if the anti-imaging filters were optimized for 480i instead of 480p, which would explain the steep roll-off. The player is down 1 dB at 6 MHz. It is down a little over 2 dB at 8.36 MHz. And finally it is down 3.56 dB at 10 MHz.

The layer change is ultra fast and we estimate it takes around .1 seconds. However the player is a bit schizophrenic on overall performance. Sometimes it jams from one menu to the next while other times it takes seconds. It also takes a long time to load a DVD. We rate the overall performance as a 2.

Both the Y-Pb and Y-Pr YC delay are also greater than 5ns. There are 2 pixels cropped from the top, 1 from the bottom, 5 from the left and 0 from the right.

Overall this player is somewhat better than we expected, given the price. Still, we'd certainly recommend spending a few dollars more to buy something that gets more of the details correct.

Deinterlacing

We're usually unimpressed by the deinterlacing in the all-in-one MPEG decoder/deinterlacers. So far, they've always used the MPEG flags to decide whether to go into film mode, which doesn't work very well for a wide variety of material. And they never have motion-adaptive video deinterlacing, which makes a big difference on video material and those times when the player has to drop out of film mode.

This player (and chipset) was a pleasant surprise. The film mode decision making is based on cadence, not flags. And there is a simple motion-adaptive video deinterlacing algorithm. It's not a FFLI2200 or SiI504, but we are happy to see Zoran moving to the next level.

The video deinterlacing ultimately fails our "motion adaptive" tests because we require per-pixel motion adaptive algorithms for a passing grade. This one appears to change its algorithm in areas of the screen larger than a single pixel but smaller than one quarter of the screen. We couldn't tell precisely how big the areas involved were. Again, while this isn't the gold standard, it's a step in the right direction, and vaults them over all the other all-in-one chipsets, at least in deinterlacing.

The Basics

The bad news: like all Toshiba DVD players before it, it still has the Chroma Upsampling Error. Toshiba wrote the DVD spec so how is it they have the problem? They should be leading by example. We spoke with Toshiba at Cedia prior to the Vaddis V models and they had hinted that this would be resolved, looks like they failed to resolve it. It is not nearly as bad as Sony or Pioneer, because the Zoran uses excellent filters to do the upsampling, but wrong is wrong.

Given that the Toshiba clearly has the best upsampling of the players that have the error, we wondered if the bug was still visible at normal viewing distances. We did a test using chapter 4 on Toy Story. We had to backup 20’ before we could no longer see the bug. This was with the image set to 65” diagonal. We are pretty sensitive to the bug, but it does illustrate our core point: even the best wrong chroma upsampling still looks pretty bad.

The scaling engine in the Vaddis V is excellent. It is every bit as good as the scaling engine in the Mediamatics Pantera II and perhaps better. The aspect controls that Toshiba exposes, however, are not as well implemented as they are in the Panasonic RP91 with the auto selection and pre-defined modes, but with a couple of button pushes, you can get the mode you want.

The 480p output has 7.5 IRE black setup and we could not find a way to set it to 0 IRE, which is the standard for progressive video. There may be a way to do it, but if so we couldn't find it in the menus or the manual, so in the end we failed it on black level. While it passed below-black at 7.5 IRE, we have no idea if it passes it at 0 IRE. In the past we have seen DVD players that pass BTB at 7.5 IRE and not at 0 IRE. We are going to give Toshiba the benefit of the doubt here and pass BTB.

The white level on the 3800 is a little low at 97 IRE and outside of the requirement we have set. Black is also a little low at 7.14 IRE, which is not bad. The frequency response is virtually ruler flat. Both Y-Pb and Y-Pr are beyond the 5ns tolerance set forth by EIA-770. There are 0 lines cropped from the top and bottom and only 1 sample cropped from each side, which is very good.

The layer change under worst-case conditions takes around 2.8 seconds to complete. All in all the player is pretty slow at navigating a DVD. Toshiba is also pretty strict at handling "Operation Prohibit" codes, which means that if a DVD says you can’t skip past it, you can’t skip past it! Other players are a bit more lenient with some of the grayer areas (and a few of the less gray areas), and frankly we wish Toshiba would lighten up.

Overall, this is a good player in lots of areas. The low score is perhaps misleading, because it failed many of the tests by fairly small margins. The lack of standard 0 IRE black is frustrating, but practically speaking most people will be able to handle it fine, perhaps even easier than the "correct" 0 IRE level. (This is because of limitations in the way most televisions are designed, but that's a rant for another article.) The biggest flaw in our opinion is the Chroma Upsampling Error. To the best of our knowledge, Zoran knew about the error. Yet they didn't fix it, which just boggles our minds.

Deinterlacing

We're usually unimpressed by the deinterlacing in the all-in-one MPEG decoder/deinterlacers. So far, they've always used the MPEG flags to decide whether to go into film mode, which doesn't work very well for a wide variety of material. And they never have motion-adaptive video deinterlacing, which makes a big difference on video material and those times when the player has to drop out of film mode.

This player (and chipset) was a pleasant surprise. The film mode decision making is based on cadence, not flags. And there is a simple motion-adaptive video deinterlacing algorithm. It's not a FLI2200 or SiI504, but we are happy to see Zoran moving to the next level.

The video deinterlacing ultimately fails our "motion adaptive" tests because we require per-pixel motion adaptive algorithms for a passing grade. This one appears to change its algorithm in areas of the screen larger than a single pixel but smaller than one quarter of the screen. We couldn't tell precisely how big the areas involved were. Again, while this isn't the gold standard, it's a step in the right direction, and vaults them over all the other all-in-one chipsets, at least in deinterlacing.

The Basics

The bad news: like all Toshiba DVD players before it, it still has the Chroma Upsampling Error. Toshiba wrote the DVD spec so how is it they have the problem? They should be leading by example. We spoke with Toshiba at Cedia prior to the Vaddis V models and they had hinted that this would be resolved, looks like they failed to resolve it. It is not nearly as bad as Sony or Pioneer, because the Zoran uses excellent filters to do the upsampling, but wrong is wrong.

Given that the Toshiba clearly has the best upsampling of the players that have the error, we wondered if the bug was still visible at normal viewing distances. We did a test using chapter 4 on Toy Story. We had to backup 20’ before we could no longer see the bug. This was with the image set to 65” diagonal. We are pretty sensitive to the bug, but it does illustrate our core point: even the best wrong chroma upsampling still looks pretty bad.

The scaling engine in the Vaddis V is excellent. It is every bit as good as the scaling engine in the Mediamatics Pantera II and perhaps better. The aspect controls that Toshiba exposes, however, are not as well implemented as they are in the Panasonic RP91 with the auto selection and pre-defined modes, but with a couple of button pushes, you can get the mode you want.

The 480p output has 7.5 IRE black setup and we could not find a way to set it to 0 IRE, which is the standard for progressive video. There may be a way to do it, but if so we couldn't find it in the menus or the manual, so in the end we failed it on black level. While it passed below-black at 7.5 IRE, we have no idea if it passes it at 0 IRE. In the past we have seen DVD players that pass BTB at 7.5 IRE and not at 0 IRE. We are going to give Toshiba the benefit of the doubt here and pass BTB.

The white level on the 5700 is virtually perfect at 99.68 IRE. Black is also virtually perfect at 7.56 IRE. Virtually perfect in the sense that the player has setup on all the time and there is no way to remove it. The frequency response, while not as flat as the 3800, is very good. Both Y-Pb and Y-Pr are beyond the 5ns tolerance set forth by EIA-770. There are 0 lines cropped from the top and bottom. There is 1 sample cropped from the left and 3 from the right.

The layer change under worst-case conditions takes around 2.8 seconds to complete. All in all the player is pretty slow at navigating a DVD. Toshiba is also pretty strict at handling "Operation Prohibit" codes, which means that if a DVD says you can’t skip past it, you can’t skip past it! Other players are a bit more lenient with some of the grayer areas (and a few of the less gray areas), and frankly we wish Toshiba would lighten up.

Overall, this is a good player in lots of areas. The low score is perhaps misleading, because it failed many of the tests by fairly small margins. The lack of standard 0 IRE black is frustrating, but practically speaking most people will be able to handle it fine, perhaps even easier than the "correct" 0 IRE level. (This is because of limitations in the way most televisions are designed, but that's a rant for another article.) The biggest flaw in our opinion is the Chroma Upsampling Error. To the best of our knowledge, Zoran knew about the error. Yet they didn't fix it, which just boggles our minds.

Deinterlacing

Like the Pioneer it is based on, at its core this is a flag-reading player, though it does have a cadence-reading solution internally that it falls back on when the progressive flag is set to "false." It doesn't seem to have much, if any bad edit detection at all, as its performance on our test was just terrible.

The Basics

Just like the Onkyo and Pioneers, this player suffers from the Chroma Upsampling Error (CUE) in all cases. The visibility of the error is fairly severe compared to some other players like the latest Toshibas.

Also like the Onkyo and Pioneers, the white level is incorrect. However while the Onkyo is too high the Marantz is too low. (Interestingly, it is at exactly the same level as the Pioneer DV-45A.)

The frequency response looks pretty decent on this player. It has just a slight roll-off.

The component timing is also within spec for both Y-Pb and Y-Pr. The player is also capable of reproducing below black picture information.

There appears to be 4 pixels cropped from the bottom of the image. We feel that any amount of pixel cropping from the top or bottom is unacceptable. There is no technical reason for it - the 480p spec allows for 480 visible lines.

The layer change takes around 1.25 seconds to complete under the worst case scenario. We rated the overall response time of the player a 2.5, which is below average.

Overall, we were unimpressed with this player for video playback. There are just too many things wrong or below average in performance. That it is THX certified reduces further the credibility of THX as a certification standard.

Deinterlacing

The deinterlacing on this player, like the Pioneers and the Marantz, is poor, especially at this MSRP. There really is no other way to describe it. It is, at the core, a flag-reading player with a cadence-reading solution for sequences where the progressive flag is not set. However, because it has little or no bad-edit detection, the player combs on tons of material.

The Basics

The downside of using another company's platform as a starting point for a player is that it inherits all of the flaws of the original. In this case, because of its lineage it suffers from the Chroma Upsampling Error, badly.

In its default position the white is around 102.5 IRE. This is just outside of our window of passing. It is possible to get it back down to a correct level with the player's own adjustments, though most consumers will not have the scope required to get this correct. Below is a screen shot of where we set the white level and sharpness to obtain the correct white level and flattest frequency response. We do want to note that we saw a bit of contouring when we started to lower the white level, which suggestions this is being done in the digital domain and may in fact be causing problems.

The mid section is a bit hot, but not outrageous.

The component timings for both Y-Pb and Y-Pr were within spec and this player is capable of reproducing below black picture information.

There were 4 lines cropped from the bottom of the image. We feel that any amount of pixel cropping from the top or bottom is unacceptable. There is no technical reason for it - the 480p spec allows for 480 visible lines, and 476 is 4 too few.

The layer change takes around 1.25 seconds to complete under the worst case scenario. We rated the overall response time of the player a 2.5.

When used with the Sony 10HT, there are faint vertical lines on-screen, probably as a result of outside interference leaking into the ground. We were able to get rid of this by pulling the luma cable out just until the shield was no longer touching. The Sony S9000ES had what looked like the same issue and Sony actually came up with a fix for it. Perhaps the same fix would work for this player.

As with its close siblings from Marantz and Pioneer, we just weren't impressed by this player, especially at its price point. That it is THX certified reduces further the credibility of THX as a certification standard.

Deinterlacing

We were amazed to find that this player is a step backward in deinterlacing performance from the DV-37 and DV-38A (which didn't have the best deinterlacing to begin with). The 45A appears to be a pure flag reader, with none of the cadence analysis the 47Ai has. What is even more surprising is that this player does not even have motion-adaptive video deinterlacing, which all of the Pioneer progressive players have had (with the exception of the 434, which doesn't deserve to be called progressive).

The Basics

Like all Pioneer players thus far, the 45A suffers from the Chroma Upsampling Error, and it's very visible. Enough said.

The white level is low on this player at 96.88 IRE. There really is no valid reason for the levels being this low. The video encoders used by Pioneer all come from Analog Devices and can be set pretty close to perfect, at least within the +/- 2 IRE window we use for ranking. You may need to bring contrast up when using this player.

The frequency response in this player is hotter than any other player. It looks like they have really emphasized the mid band frequencies. At first glance, this will make the 45A appear sharper than other players. Once you have properly adjusted the sharpness control on your display, this difference will go away.

The component timing is outside of spec. The 45A does reproduce blacker than black picture information, which is at least one nice thing we can say about the player.

There was 1 sample cropped from the left and 3 from the right, which we consider acceptable.

We get accused of being biased against Pioneer constantly, which we're not. Many of us at Secrets of Home Theater & Hi Fi are fans of many Pioneer products, including their audio gear, TVs, and laserdisc players. But it's hard to be fans of their DVD players when they produce players like this. It's not a terrible player, but given that several cheaper players are better overall, we just can't recommend this one, at least based on progressive video quality.

This is a unique product because the configuration we tested is only one of many possible configurations, and the final results will vary based on the hardware used. Even the software the PC manufacturer uses makes a difference, because Windows XP Media Center doesn't include a DVD decoder. Instead it relies on whatever decoder is installed on the machine, though it does integrate it into the Media Center UI.

For our setup we used an nVidia 4600Ti video card and the NVDVD decoder, which was developed in-house by nVidia. From our testing, we believe the Ti offers the best deinterlacing within the nVidia family, and NVDVD 2.22 offers the best decoding available right now. ATI video cards also do well when using the Catalyst 2.4 drivers, with the exception of the 9700. As of this writing the video drivers for the 9700 were not ready for prime time.

The machine we used was provided by the eHome team at Microsoft. The display drivers we used for testing were version 40.80, which at that time were unsigned drivers. 41.09 are the most recent drivers available from nVidia. You can also purchase NVDVD 2.22 directly from the nVidia website for $39.95.

Deinterlacing

The new nVidia hardware and software is supposedly designed to offer two features that had been previously unavailable on an HTPC: cadence-based film mode analysis (instead of flag-based), and motion-adaptive deinterlacing. Unfortunately, we were unable to see cadence analysis in action. No matter how we set the drivers and settings, we still got results that clearly pointed to flag reading.

We did get motion-adaptive deinterlacing working, and it's a great step in the right direction for handling video content. However, what we saw still didn't look nearly as good as the best motion-adaptive video deinterlacing in chips such as the Silicon Image SiI504 or Genesis FLI2200, or even Sony's DRC chips. The test patterns we used didn't look quite right, as though one field was dominant and the other recessive. We suspect that the algorithm nVidia is using uses one field as the primary source for the final frame, and mixes in some of the other field in order to improve the vertical resolution. But they appear to not be giving each field an equal share. One field would generally be dimmer in the final mix, and the other field brighter, as though the temporal filter coefficients weren't 50/50. In addition, on the Video Essentials Zone Plate it just wouldn't give us a full-resolution image, even on the film-mode section. By our standards, that's a fail.

Since we weren't able to get cadence reading working, we had to fail the player on essentially all the "unusual flag" tests.

The Basics

The interface was smooth and refined, with generally well designed menus and graphics. Navigating the menus to get to the various functions like DVD playback was easy. The remote felt very much like a DVD remote, and had most of the standard buttons. There were some odd exceptions. For example, the remote has only one menu button, labeled "DVD menu," whereas a DVD player always has two, labeled "menu" and "top menu." After some experimentation, we found that the "DVD menu" button corresponds to "top menu," which is the button on a DVD player that navigates to the topmost menu on the disc. We found this incredibly annoying, as we're used to using the normal "menu" button to return to the last menu visited. If you're several levels deep in the menu system, play some video, realize you've made a mistake, and want to go back to the menu you were just looking at, Media Center has no way to do that.

In addition, the remote has no "subtitle" button, "audio" button, or "angle" button. These are all standard DVD buttons, and on some discs, you need to use them to access features. Media Center has an incredibly convoluted and non-standard way of selecting subtitles and audio tracks that we are sure would not work with some discs.

To select a subtitle or audio track, one must navigate to the Media Center menu, go several menus deep into the options, and then select a subtitle language from a huge list of essentially every available language that a DVD could use. To get to German, for example, you have to go through Armenian, Azerbaijanian, Basque, etc. by pressing a button many many times. You get the same set huge of languages to choose from no matter what languages are available on the disc. Who thought this was a good idea? As a method of setting your "preferred" subtitle language, it makes some sort of sense. As a primary method of changing subtitles, it's terrible. And what happens if a disc has three or four "English" audio tracks, as many often do? There isn't any way to distinguish between them. You can, in most cases, change audio tracks from the DVD's own menus, but if you're used to starting the movie and hitting the "audio" button to get to a commentary track or switch to the 5.1 audio, Media Center doesn't offer that as an option.

The end result is that Media Center feels almost like a standalone DVD player, but the omissions and deviations from the DVD standards are frustrating. Still, it's by far the best out-of-box UI for an HTPC that we've ever seen. For most people, these menu issues will be minor annoyances; the core UI functionality works fine.

The overall picture quality from the hardware and NVDVD was excellent. Like other HTPCs, there is no pixel cropping and extremely clean video with almost no visible ringing or noise. Of course like most video cards, there are no anti-imaging filters other than basic EMF and RMI because a PC video card must run at several different scan rates and you can only really design a filter for one scan rate. The chroma upsampling in nDVD is the best we have seen from any DVD player.

As a DVD player, Windows XP Media Center (with NVDVD decoder) worked well, in fact better than any HTPC we've looked at so far. The remote gives you most of the feel and usability of a standalone player, which has been a sore point for many people who are uncomfortable with using a PC as a media hub. The deinterlacing still isn't at the level of the best standalone players, but it's come leaps and bounds in the last year or so.

Deinterlacing

The iScan Ultra is based on the Silicon Image SiI504 chip, and like other players and deinterlacers that use that chip it had stellar deinterlacing results. There were no significant flaws that we could find. It just works well nearly all the time.

The Basics

The Ultra is loaded with new features, and represents a substantially more complex and flexible product than the three previous models. There are six inputs instead of three -- two each of composite, s-video, and component. There is a passthrough port for HDTV signals. The front panel has a much nicer interface for adjusting parameters, and there are more parameters to adjust, including two levels of sharpness control.

The most interesting addition is a chroma filter, which almost completely fixes the Chroma Upsampling Error as well as the interlaced chroma problem we cover in our DVD chroma article. We found that it worked as intended, removing essentially all traces of chroma upsampling artifacts, even from players with incredibly strong artifacts. With DirecTV, which like all other MPEG-based video sources suffers from the same artifacts as DVD players, the improvements were also significant and worthwhile.

The Ultra also has a full set of aspect controls, including zooming letterboxed content, a feature sorely missed from the previous models (and missing from the Focus Enhancements CS-1).

We didn't do any evaluation of the basic video quality measurements on the scope, but we did look at key test patterns like resolution and Y/C delay, and saw nothing out of the ordinary. Resolution loss was minimal. The 6.75 MHz pattern on Avia was still visible, though very slightly washed out and a bit more aliased than normal. Y/C delay was on the money. Default sharpness was neither too high or too low.

Overall, we give this box a strong thumbs up. If you don't need scaling, but would like to have solid 480 line deinterlacing, we don't know of a better choice.

Deinterlacing

The deinterlacer at the heart of the CS-HD is the SiI504 from Silicon Image, and like most every product using that chip it passed all of our tests well. The CS-1 had a problem with the high-detail scene in Super Speedway, but we didn't see that with the CS-HD. Like the CS-1, this unit offers a "film bias" mode for deinterlacing that we don't recommend at all.

Like the CS-1, the CS-HD doesn't hide the interlaced chroma problem (ICP) with a chroma filter. We continue to think it would be a great enhancement. The iScan Ultra and the Faroudja products have shown the way, and we hope every other manufacturer will get on the bandwagon soon.

The Basics

There are plenty of inputs on the back panel, which makes it useable for most people as the core video switcher in their setup. There are 2 composite, 2 S-Video, and 2 component interlaced inputs. These will accept any 480i material for deinterlacing and scaling. There is also a 480p input for VGA and component. The progressive inputs bypass the SiI504 deinterlacer and just go through the scaler.
Two output formats are available: VGA (RGB+HV) or component (YPbPr). Pre-programmed output resolutions include all the standards: 1080i, 720p, 480p, 800x600 and others. There were all the aspect ratio controls that one might need for various input types.

The user interface is simple to navigate and can be controlled from the front panel or the supplied remote. There are simple input and output switches that choose the source and output. There are controls for the deinterlacer on the setup menus as well as picture controls including contrast, brightness, color, and tint.

Overall, this is a user-friendly scaler that offers excellent deinterlacing and scaling, at a very friendly price point.
 

We knew nothing about this unit before it was sent to us to evaluate. Given its very low price point, we had very low expectations. Those expectations were largely shattered.

Deinterlacing

This box uses the Trident DPTV chip for video decoding, deinterlacing, scaling, and reencoding. Essentially, it's a box wrapped around a single chip, with a bit of extra analog circuitry and power supply. The deinterlacing on this box is superb. It's the first shipping deinterlacing chip we've seen that challenges the supremacy of the Silicon Image and Genesis/Faroudja chips. Take a look at the test results - it passed everything. We tried a few torture tests that we don't normally run on deinterlacers because we suspect they aren't fair. It passed those as well. Judged purely on deinterlacing performance, this is an amazing value. It doesn't have any kind of diagonal processing like Faroudja's DCDi, and it doesn't have a chroma filter to help with the CUE and ICP, but that's perhaps picking nits.

The Basics

We weren't able to get this device on the scope, as we evaluated it late in the shootout process. Most of the core functionality was excellent, with a few exceptions we'll cover in a moment. The A/D conversion was very good, in fact better than most analog scalers we've looked at (though the iScan Ultra comes very close). We could not see any resolution loss at all on the Avia resolution wedge, and the 6.75 MHz pattern was clean. The scaling was similarly clean, and better than the Focus Enhancements CS-1, and even better than any HTPC we've looked at. In both cases we should note that the improvements are relatively subtle. It's not that the quality "blows away" the competition, it's just very surprising to see this level of quality in such an inexpensive product.

The first negative issue that stands out is that the 480p output has completely incorrect voltages. Both whites and blacks were crushed. They are so wrong that we were unable to correct them with the settings on the box, or the controls on the display. We think that perhaps the box is incorrectly using the RGBHV voltages for 480p output, even though it was set for YPbPr at the time. This is just a guess. 720p, 1080i, and all the computer resolutions all had at least nominally correct voltages. Given that this is a deinterlacer/scaler, not many customers will be buying it to use the 480p output, but it's still disappointing.

The second issue, and most problematic, is that the box introduces a large and visible chroma delay, and the delay is different for each channel, in some cases in opposite directions. This is very disappointing, as it's a very visible flaw and one that could be eliminated with careful engineering. The delay is different at 1080i than it is on 720p, so it's on the output stage and not the input. The fact that it changes at different output resolutions tends to suggest that it's an internal flaw in the chip, but we don't know that for sure.

From a usability standpoint, the box is functional, but doesn't have as many inputs or bells and whistles as more expensive scalers. It has only a three inputs, one each of composite, s-video, and component, so most people would not be able to use it as a core switcher. The interface is all on screen, and controlled by a perfectly functional remote. It has brightness, contrast, color, and hue controls, and an aspect ratio control that can do a 16x9 shrink for playing anamorphic DVDs on a 4x3 display device. There is no control for zooming letterbox movies on a 16x9 display, or windowboxing 4x3 movies on a 16x9 display. When the device is in YPbPr mode, the aspect ratio control is disabled, and only the 480p, 720p, and 1080i resolutions are available. If the output is changed to RGBHV, then only "computer" resolutions (640x480, 800x600, 1024x768, and 1280x1024) are available, and the aspect ratio control allows one to change between "normal" and "16x9 shrink" modes.

Build quality is perfectly adequate. It didn't seem flimsy, but clearly you don't get massive power supplies and mahogany panels for this price.

In the final analysis, this is an impressive item for the price, marred by some disappointing flaws. If the chroma delay and incorrect voltages were fixed, it would be an astounding value. We are looking forward to seeing other products that make use of this chipset, as they've come out of the gate strong.

Deinterlacing

The deinterlacer at the heart of the CS-1 is the SiI504 from Silicon Image, which is an excellent product and passed all the key tests in auto mode with one small hiccup on the High Detail test pattern. Things get a little uglier in the film bias mode. We highly recommend you do not use film bias.

The CS-1 doesn't hide the interlaced chroma problem, but then nothing really does other than Faroudja and the DVDO iScan Ultra. We wouldn't call this a deal-breaker. If they release a new model next year, though, we would hope it would have a chroma filter.

The Basics

While the deinterlacing is handled by the SiI504, as mentioned above, we are not sure which scaling engine they are using. It comes with several pre-defined scan rates, and with their software you can create your own.

Because we didn't get our evaluation unit until after we had run our video quality testing, we did not measure any of the video characteristics of this product, like Y/C delay and frequency response.

There are plenty of inputs on the back panel, which makes it useable for most people as the core video switcher in their setup. There are 2 composite, 2 S-Video, and 2 component interlaced inputs. These will accept any 480i material for de-interlacing and scaling. There is also a 480P input for VGA and component. This is great for those of you who want to use your X-Box or other high resolution gaming systems and want to use the native resolution of the game. The progressive inputs bypass the SiI504 deinterlacer and just go through the scaler.

Two output formats are available: VGA (RGB+HV) or component (YPbPr). Pre-programmed output resolutions include all the standards: 1080i, 720P, 480P, 800x600 and others. Those looking for more flexibility can hook the CS-1 right into a PC via a RS-232 connection. With the PC interface, one can customize resolutions and synch signals as well as aspect ratios. Like most units that use the SiI504 chip there is some aspect ratio control. Users can choose between an anamorphic source or a 4:3 window box for full screen material. Unfortunately there are no zoom functions for letterbox material, which surprised us greatly.

The user interface is simple to navigate and can be controlled from the front panel or the supplied remote. There are simple input and output switches that choose the source and output. There are controls for the deinterlacer on the setup menus as well as picture controls including contrast, brightness, color and tint.

There were some visible scaling artifacts when looking at the WHQL wedge. Horizontal lines (vertical resolution) contained patterns of darker lines. It isn't terrible scaling, but we've certainly seen better.

Overall this is a very user friendly scaler that offers excellent deinterlacing and decent scaling.

Deinterlacing

The Vision Pro uses the Silicon Image SiI504 chip for deinterlacing, and as usual it passed most of our tests fine. The device came to us with "film bias" turned on, and in general we found that it caused more problems than leaving it off. "Video bias" is just the normal SiI504 mode which switches automatically from video to film as needed, and we found that turning on film bias just increased the amount of combing on video material.
There was one unexpected result in that it failed the bad edit test with two combs. We aren't sure what caused the problem, but given that any number of other DVD players and scalers that use this chip can pass this test, we have confidence that the engineers at Lumagen and/or Silicon Image can track down the problem and get it fixed.

The Basics

This box has a lot of inputs, and all of them use the superior BNC connectors. Getting all your equipment hooked up may involve lots of BNC adapters (though it comes with a lot of cables and adapters as standard accessories), but the peace of mind that comes from using high-quality bayonet connectors is worth it for a high-end system. There are eight inputs and two pass-throughs, which is more than enough for most people.
There are a lot of menu options available, allowing you to tweak just about every feature you could ever want to tweak. Resolutions, cropping, scan rates - all are adjustable. The aspect controls work as you would expect them to. The remote worked fine; it's arguably not a perfect button layout, but there aren't that many buttons. Most high-end users will be programming the functions they care about most into their universal remote anyway.
If we could add only one thing to this unit it would be a chroma filter like the one on the iScan Ultra. It's not a complicated filter to implement - a simple 3-tap tent filter will do the trick. Once you get used to not ever seeing the chroma bug (CUE) on screen, switching to something without the filtering is like a glass of cold water in the face.

In general, we were impressed with the Vision Pro. Lumagen has packed an enormous amount of flexibility into this unit, and have come out of the gate with a strong contender. Assuming they fix the bad edit issue, it should compare favorably to any other high-end scaler out there.

We knew nothing about this unit before it was sent to us to evaluate. Given its very low price point, we had very low expectations. Those expectations were largely shattered.

Deinterlacing

This box uses the Trident DPTV chip for video decoding, deinterlacing, scaling, and reencoding. Essentially, it's a box wrapped around a single chip, with a bit of extra analog circuitry and power supply. The deinterlacing on this box is superb. It's the first shipping deinterlacing chip we've seen that challenges the supremacy of the Silicon Image and Genesis/Faroudja chips. Take a look at the test results - it passed everything. We tried a few torture tests that we don't normally run on deinterlacers because we suspect they aren't fair. It passed those as well. Judged purely on deinterlacing performance, this is an amazing value. It doesn't have any kind of diagonal processing like Faroudja's DCDi, and it doesn't have a chroma filter to help with the CUE and ICP, but that's perhaps picking nits.

The Basics

We weren't able to get this device on the scope, as we evaluated it late in the shootout process. Most of the core functionality was excellent, with a few exceptions we'll cover in a moment. The A/D conversion was very good, in fact better than most analog scalers we've looked at (though the iScan Ultra comes very close). We could not see any resolution loss at all on the Avia resolution wedge, and the 6.75 MHz pattern was clean. The scaling was similarly clean, and better than the Focus Enhancements CS-1, and even better than any HTPC we've looked at. In both cases we should note that the improvements are relatively subtle. It's not that the quality "blows away" the competition, it's just very surprising to see this level of quality in such an inexpensive product.

The first negative issue that stands out is that the 480p output has completely incorrect voltages. Both whites and blacks were crushed. They are so wrong that we were unable to correct them with the settings on the box, or the controls on the display. We think that perhaps the box is incorrectly using the RGBHV voltages for 480p output, even though it was set for YPbPr at the time. This is just a guess. 720p, 1080i, and all the computer resolutions all had at least nominally correct voltages. Given that this is a deinterlacer/scaler, not many customers will be buying it to use the 480p output, but it's still disappointing.

The second issue, and most problematic, is that the box introduces a large and visible chroma delay, and the delay is different for each channel, in some cases in opposite directions. This is very disappointing, as it's a very visible flaw and one that could be eliminated with careful engineering. The delay is different at 1080i than it is on 720p, so it's on the output stage and not the input. The fact that it changes at different output resolutions tends to suggest that it's an internal flaw in the chip, but we don't know that for sure.

From a usability standpoint, the box is functional, but doesn't have as many inputs or bells and whistles as more expensive scalers. It has only a three inputs, one each of composite, s-video, and component, so most people would not be able to use it as a core switcher. The interface is all on screen, and controlled by a perfectly functional remote. It has brightness, contrast, color, and hue controls, and an aspect ratio control that can do a 16x9 shrink for playing anamorphic DVDs on a 4x3 display device. We didn't see a control for zooming letterbox movies on a 16x9 display.

Build quality is perfectly adequate. It didn't seem flimsy, but clearly you don't get massive power supplies and mahogany panels for this price.

In the final analysis, this is an impressive item for the price, marred by some disappointing flaws. If the chroma delay and incorrect voltages were fixed, it would be an astounding value. We are looking forward to seeing other products that make use of this chipset, as they've come out of the gate strong.