Denon - DVD-5910

MPEG Maker: ESS MPEG Model: Vibratto Deinterlacer Maker: Silicon Optix Deinterlacer Model: Realta HQV     MSRP: $3500.00 Website: http://www.usa.denon.com

Denon has clearly established themselves as a leading force in DVD technology. No other company has pushed the envelope as far as they have in features, design, or execution. Like any company, they have had their share of hiccups, but that is almost to be expected when you are at the bleeding edge.

The DVD-5910 is their latest flagship offering and the followup to the DVD-5900, but actually has little in common with its predecessor. I would almost say it has more in common with the DVD-9000.

This is just about a completely new video player from the ground up. The only thing I could find that was the same as in the 5900 is the ESS MPEG decoder. And that is one of my sore spots. The ESS chip seems to be getting worse in regards to chroma performance. Even with the advanced chroma filters this player has, it barely passes our tests. I really hope Denon moves past this ESS chip and finds one with better performance in this area. Chips from Sigma Designs, Mediatek, and Mediamatics do just this. Because the filters do mask the problem (although just barely), this player gets a borderline score for our chroma test for 3-2 alternating flags. This means the player passes these tests but only with the filters. The nice thing is, this player doesn’t suffer from the flicker problem like the DVD-2900 and previous Mitsubishi MPEG decoder players did.

The first breakthrough feature on this player is definitely the inclusion of Silicon Optix’s new chip; the Realta HQV. This is, in a nutshell, a new chip based on the Teranex video processor. For those not familiar with the Teranex, it is a $60,000 commercial video processor that has long been considered the de facto video processing standard. This chip offers true 10-bit processing end to end and is one of the only chips on the market that will do true 1080i to 1080p inverse telecine de-interlacing. While this chip does offer a full scaling capability, Denon has chosen to only use the chip for its I/P conversion and video post processing. As a de-interlacer, the Realta is the best I’ve seen. It combines the features I love about both the Faroudja and Silicon Image chips and doesn’t have any of their drawbacks.

This chip recovers from a cadence break faster then the Faroudja and stays in film mode longer. It also has incredible video de-interlacing capabilities. I called up the famous Faroudja flag waving in the wind test to see how it would do. Previously only Faroudja based de-interlacers with DCDi technology would render this flag perfectly with no obvious jaggies. The Realta performed every bit as good though. I also tried some of my harder tests, including a clip from the recent Blue Man Group DVD-A The Complex. This disc features a music video with Dave Matthews and always causes problems for de-interlacers. Comparing the Realta with the Faroudja solution revealed far less combs and more stability in the image. This chip also doesn’t exhibit any of the macroblocking issues associated with the latest Genesis FLI-23xx based video chips. The only area that the Faroudja solution eclipsed the Realta processing was with diagonal lines. The Faroudja showed less jaggies with extreme diagonals then the Realta. I know Silicon Optix is still tweaking their performance in this area and I look forward to seeing what advancements they’ll make in this regard.

On the video post processing side, this chip is really quite impressive. Denon has incorporated some of the noise reduction features this chip offers. Normally, I am not a big fan of these filters, as they usually tend to hurt the image more then help it, but in this case, the benefits are quite obvious. Turning the filters on, and using material that was pretty bad in terms of artifacts, definitely cleaned the image up without adding any unnecessary ringing or softness. This was readily apparent using the Silicon Optix demonstration DVD. I would still only recommend this for badly authored DVDs and only at the end users discretion. Since Denon offers five individual picture memories, one could easily tailor one for badly authored DVDs and one for better discs.

This player has three progressive modes to choose from; Auto, Video 1 and Video 2. We recommend Auto as it does everything correctly. In fact, we are still kind of puzzled as to why Denon included two other modes. Video 1 is a forced video mode that doesn’t do anything that Auto can’t do and Video 2 is pretty much identical to Auto.

For scaling, Denon turned to one of the companies that is a personal favorite of mine; DVDO. I use DVDO’s iScan HD as my reference video processor. I had my Denon DVD-5900 SDI modded and fed the iScan HD for all of my I/P and scaling duties. The scaling solution in the 5910 is not the same as the iScan HD, as this processor is true 10-bit processing while the iScan HD is 8-bit but I could not really tell a difference with the material I played during my tests.

As much as I love the inclusion of the DVDO scaling to this player; it is also the source of my biggest gripe. The 5910 only supports three NTSC resolutions via HDMI or DVI; 480p, 720p, and 1080i. This is a glaring oversight in my opinion and really the Achilles heel of the player. It supports two of the best scaling solutions the market has to offer today, yet it does not take advantage of them to the extent I feel they should. In my opinion this player should have come pre-programmed with most, if not all, of the common resolutions available. I was expecting at least the pre-programmed resolutions one finds with the iScan HD processors. There are way too many odd resolutions out there right now because there are so many display technologies. The drawback of not supporting these resolutions means that the end display has to rescale the image to its native resolution and therefore negates the advantages of having these spectacular resolutions in the player. I also think the 5910 should have supported custom resolutions similar to the Bravo and Momitsu offerings. This would allow anyone to dial in the resolution of the display and achieve 1:1 pixel mapping, which should be the goal of any videophile. This tends to produce the best picture from almost any given display as most if not all of the internal video processing is bypassed and the full benefits of the player’s processing is shown.

At the price point this player comes in at, I think it is a glaring omission by Denon’s engineers. The reality is, I can buy a DVD-5900 or DVD-3910 and feed its digital signal through either via DVI or SDI, to an outboard scaler and basically outperform the 5910, and it would still be cheaper. The de-interlacing and scaling performance would take a bit of a hit but I would still save quite a bit of change. Granted you also wouldn’t reap the other benefits this player may bring to the table such as a huge step up in build quality and what might be an improvement in audio (we have not started our tests on the audio side of this player yet) but this will probably depend more on what you are currently using in your own setup. I tested this idea in my own setup. My reference setup is comprised of a DVD-5900 that has been modified with a SDI output and feeds an iScan HD for video processing. The iScan is sending a 1:1 resolution to my plasma via DVI. The difference between this feed and a standard 480P feed via DVI is startling. This has been the case regardless of the digital display I’ve tested it on at any resolution. The fact is most of the inboard processing in consumer displays does not match the quality you’ll find in outboard processors or high end DVD players. If your display does indeed have a resolution that is different from the standard 5910 modes, going with an outboard scaler would still probably be the preferred route.

On the flipside of this coin, the 5910 is pretty much doing what every other player out there is doing right now though. Other than the previously mentioned Bravo and Momitsu, there aren’t really any custom resolution players available. And the scaling and de-interlacing of this player are unmatched from anything I’ve seen to date. So if your display does use a native resolution of 480p/720p/1080i, the 5910 is by far the de facto standard in this regard. (Just a note, those of you with ED plasmas or 480p projectors be aware that the native resolution may not be true 480p, which is 720x480. Most of these displays use 853x480 to achieve square pixels so the display will have to scale a 480p signal.)

As to the analog side of this player, the core tests went well, but there is some room for improvement. The player does have some pixel cropping that is exaggerated by the player’s output setup. If you leave the player in a default mode, it is clipping a combined 7 pixels from the sides of the image and 1 from the top. The 5910 has an almost perfect white level of 101 IRE and passes a below black pluge signal. The analog video frequency response was ruler flat and there is absolutely zero Y/C delay.

The digital output is slightly different. This player has a choice of either PC (enhanced) or Studio (normal) RBG levels for its DVI or HDMI outputs. It also has the ability to output a pure YCbCr video stream via HDMI. While this player does output a below black signal using Studio RGB levels, it is clipping above white information in the default contrast setting within the user picture settings. This is quite obvious using the reverse gray ramps on Digital Video Essentials. If you set the contrast level to -2, the player passes the above white information properly. We saw something similar to this with the recent Panasonic S-97. I don’t know why Denon overlooked this, as it was not a problem on the 5900, 3910, or 2910 players.

The digital output also has very slight pixel cropping. If you look at the pixel cropping test pattern on AVIA in the default setting, there are about 3 pixels cropped from the sides combined, but these are not affected by the position of the image.

This player also has an “Auto” selection for its output resolution via HDMI or DVI. This is for displays that will communicate what their optimal resolution is, but this is rarely correct, so be cautious when using it. When selected, it will indicate the end resolution, so there is no concern of whether it “might have” selected the proper output resolution. We tested this with the new Samsung DLP front projector, and the player automatically selected 480p, despite the projector having a native resolution of 720p. This confirms the fact that the feature may not always work as designed.

On a usability scale, the 5910 gets big points. This player has a faster user interface than my DVD-5900, but is still not quite to the level of the DVD-2900. Menu navigation is fast for the most part, but there were some times that the player staggered a bit once an entry was made. All of the menus are easy to navigate, and the player still features five different user setup memories for advanced picture settings.

From a build standpoint, the 5910 is an absolute tank. Denon went away from the standard connectors on the back panel and added some beefy ones that are well spaced, reminding me of the DVD-9000. The new remote is completely backlit and even offers full operation of the DVI/HDMI outputs and SACD modes.

Without a doubt this is the best DVD player I’ve used from Denon so far, but it has the potential to be so much more. Hopefully, Denon will provide a software update that offers the flexibility in video processing that the 5910 deserves. At $3,500 consumers will expect, and want, that kind of performance.

Toshiba - SD-V592

MPEG Maker: Mediatek MPEG Model: MT1389FE Deinterlacer Maker: Genesis Deinterlacer Model: FLI-2310     MSRP: $249.99 Website: http://www.toshiba.com

The Toshiba SD-V592 is the very first DVD/VCR combo we’ve ever had in the Benchmark. Usually, combo players leave little to be desired, but this was a rare treat that really caught me by surprise.

The player is based on the Mediatek MPEG decoder, which is quickly becoming one of my favorites. The chip provides lightning fast access and seamless layer changes, as well as some of the best chroma performance on the market.

If you are using the component video outputs, the Mediatek chip does all of the video processing, including the de-interlacing duties. This chip varies in performance, and its ultimate levels depend greatly on the manufacturer’s implementation. This is one of the better ones. Thankfully, Toshiba has ensured that this player is motion adaptive, which is a rarity with this chip. The de-interlacing scores were quite good for not having a standalone solution. The player did fail our bad edit and 2-2 cadence tests, but this is pretty common, and even the big dogs like Faroudja and Silicon Image have trouble with these from time to time.

For our core tests, the analog side of the player did excellent. The only chroma issue I saw was the 4:2:0 ICP, but it was very slight. The Mediatek is one of the best chips out there in this regard. The video frequency response was extremely flat, with only a tiny bump down in the upper end. Y/C delay was perfect, and the player did pass a below black pluge. The biggest drawback I found via component was the white level of the player. It is quite low and measured in at 94 IRE.

Navigation was extremely fast and almost instantaneous. There are times that once the selection was made, the player would hang a tiny bit, but it was rare. Again, layer changes are absolutely seamless.

One of the bonus features of this player is the inclusion of an HDMI output. This output allows for resolutions of 480i, 480p, 720p, and 1080i, making it one of the few players that actually allows for 480i via a digital output. We have had reports though that some people’s players won’t output 480i. Realize that this is not a viable option if you are converting to DVI before you go into your display. Video processing via the HDMI output is handled by the Genesis FLI-2310 and is HDCP compliant. The HDMI output has some output problems that we are hoping Toshiba may be able to remedy with a firmware update. The Faroudja chip has several different video processing capabilities and one of these is a fine detail enhancer. Unfortunately this setting is maxed out on the chip and this results in blotching in colors. This can be easily fixed with software but until it does, the problem is quite obvious in solid colors. This only affects the HDMI output though.

Another issue is the white level of the player. This player is clipping some of the upper white information, but is retaining above white information. This is very odd and is the first time we’ve seen a player do this. It is quite apparent using the test patterns for contrast on AVIA. This player also exhibits the macroblocking problem that has become a staple of the Genesis FLI-23xx series chips. As we have stated many times now, this is always display dependant so be sure to test the player out on your own display before passing judgment.

This is the first player that has used the Faroudja chip and managed a seamless layer change, a feat that even Denon has not been able to pull off despite it being one of their trademark features. This is no doubt due to the Mediatek MPEG decoder which has been extremely fast in all of its implementations.

The player passes a below-black pluge via its HDMI output and conforms to the Studio RGB black levels (black = 16/ white = 235). I was not able to connect it to a DVI display input, so I don’t know if it correctly maps the RGB levels when converted to DVI. Since the HDMI output uses an entirely different de-interlacing solution, the results were a bit different. The Faroudja chip was employed well, and the only test that tripped up the player was the 2-2 mastering. Toshiba did not include a mode to support it, even though the Genesis chip can support that cadence.

This player really has the potential to be a great one. A simple firmware fix could put this player way up there in terms of performance via HDMI. As it stands now, this is a solid offering for supplying 480i digitally to an outboard scaler. I will be following up with Toshiba on this player’s drawbacks and update the site with any news on this player as we get it.

Denon - DVD-2910

MPEG Maker: ESS MPEG Model: ESS Vibrato Deinterlacer Maker: Genesis Deinterlacer Model: FLI-2310     MSRP: $699.00 Website: http://www.usa.denon.com

Video Frequency Response
Denon

Samsung - DVD HD-1000

MPEG Maker: LSI MPEG Model: Ziva5 Deinterlacer Maker: Genesis Deinterlacer Model: FLI-2310     MSRP: $1500.00 Website: http://www.samsungusa.com

Video Frequency Response
Samsung

Sony - DVP-NS975V

MPEG Maker: Sony MPEG Model: CDX9304 Deinterlacer Maker: Sony Deinterlacer Model: CDX9304     MSRP: $349.99 Website: http://www.sony.com

Previously, I reviewed the Sony 555ES changer. I was quite impressed with it compared to the other Sony players and looked forward to seeing where their future offerings would go. Unfortunately Sony has been extremely quiet in the land of mid and high range players. It has been a few years now since they introduced the last ES player, and they didn’t have really anything new to show at CEDIA 2004.

The DVP-N975V is part of their entry line of players, but it has more in common with their ES line then the lower line. The player is based on their own MPEG decoder, the CDX9304, which also doubles as a de-interlacer and scaler for their HDMI output. This is the first time I’ve seen Sony use a proprietary solution for their MPEG decoder duties or de-interlacing, and it has paid off.

The 975 passed all of our chroma tests with the exception of the 4:2:0 ICP test. This was expected since this test requires special chroma filters normally found only on outboard processors or the Faroudja video processing solutions. It did an excellent job with alternating 3-2 material though, with no obvious flickering or banding. However,the MPEG decoder could not properly synch its subtitle output with the progressive frames, and combing was quite evident with our reference test using “The Abyss”.

The 975 passed a below black pluge via its component output, and the white level measured in a tad hot at 102 IRE, but was still within our test criteria. Pixel cropping was excessive on both the component and HDMI output, with 17 pixels cropped on the left and right side combined and 7 pixels on the upper part of the image.

The HDMI output offers several output resolutions including 480i, 480p, 720p, and 1080i. This makes another player on the market to offer 480i via HDMI. Unfortunately, the HDMI output does not pass a below black pluge and didn’t have any selections for RGB or YCbCr colorspace. Since HDMI supports several different output formats, including 10 and 12 bit YCbCr, I would like to see more manufacturers support its potential.

Navigating the player’s setup menus was a breeze, and the overall response time of the player was excellent. Menu navigation with our test discs was fast, and chapter skips had little hesitation. The layer change clocked in at a respectable 0.75 seconds.

The most impressive part of this player was its de-interlacing performance. The player offers two progressive modes that are easily accessed via the front panel progressive button; Progressive Auto and Progressive Video. The Auto mode did a superb job in our testing and reminded me more of a high line solution than a one chip proprietary solution. It passed all of our de-interlacing tests except our Bad Edit test, which is a common failure even for higher line de-interlacers. Recovery time during our cadence tests was a tad sluggish at about 15 frames, but otherwise this is the best performance I’ve seen from Sony yet.

Progressive Video is a forced video mode and should only be used with known video material. Setting the player up in this mode for normal film based material will result in an extremely soft image with obvious artifacts.

Another addition to this player is support for Sony’s SACD format. It is nice to see more support for this format from Sony in their lower priced players. The HDMI output of this player is the new generation and would support high resolution DVD-A, but as of yet Sony has not supported the competing format. We hope that Sony will eventually allow for full SACD support on HDMI making for a more streamlined hookup to future products.

I was really impressed with the performance of this DVD player, and Sony continues on the upward trend. I am hopeful that Sony will announce a new player for their ES line at CES and will approve HDMI as a transmission method for their popular SACD format. The biggest downside that I could find here was the lack of support for below black information via HDMI. This could be fixed via firmware, and hopefully Sony will address it. The de-interlacing performance was excellent though, and as a component based player, this is a great option.

Video Frequency Response
Sony

Marantz - DV-9500

MPEG Maker: unknown MPEG Model: unknown Deinterlacer Maker: unknown Deinterlacer Model: unknown     MSRP: $2099.00 Website:

I have always been a fan of Marantz’s DVD player offerings, but more for their audio implementation than their video. Their past lines of players have been modeled after the Pioneer Elite line and inherited a lot of the limitations that we were finding with the Elites. With this new model, the DV-9500, Marantz has separated themselves from that trend and developed an entirely new platform that doesn’t seem to share the same attributes that plagued the early Pioneer models.

The DV-9500 looks to be based on a Mitsubishi MPEG decoder, but I couldn’t tell since the chip was mainly blacked out. The user interface reminded me a lot of the new Samsung line, but based on our tests, this player is not a Zoran based player as the Samsungs are. That is a good thing. The 9500 passed our main CUE tests but still came up short with 4:2:0 material and material mastered with an alternating 3-2 cadence. This is extremely common among MPEG decoders today, and these drawbacks rarely show themselves. Thankfully, this player does not have any of the flicker effect that I’ve seen with other players when it comes to alternating 3-2 material. That really draws attention to the issue compared to slight banding.

The core performance of this player is just about perfect. There is absolutely no pixel cropping, which is an extreme rarity these days, and there is no Y/C delay at all. The analog video frequency response is outstanding and one of the best we’ve seen at any price point. The player passes a below black pluge via component or HDMI out, and the HDMI output can be configured for either PC or Studio RGB levels to accommodate your display. The white level via component was a tad hot at 102 IRE but was still within our passing criteria.

For progressive processing, the Marantz offers no less then six different modes: Auto 1-3 and Video 1-3. Each of these modes was only slightly different then the other, and we recommend Auto 2 for all viewing habits. This mode did well with all of our tests, with the exception of 2-2 based material. None of the de-interlacing modes would pass this test though. The player is cadence based and motion adaptive but had a somewhat slower time recovering from a cadence break, similar to what you see with a Faroudja solution.

From a user standpoint, this player performed very well. Setup menus were very concise and easy to navigate. The only setup I didn’t care for was the HDMI modes, which is a bit cumbersome to access. I wish more manufacturers would allow for a single resolution select button on the remote instead of burying it in their setup menus. Navigation of disc menus and chapters was very fast and never an issue. The player’s layer change clocked in at a respectable one second.

From a build standpoint, the DV-9500 continues Marantz’s established reputation of phenomenal styling and detail. All of the interior boards are laid out carefully, and the power supply is excellent, resulting in a very clean video output. The audio connections are extremely robust and laid out with plenty of space between each other, which is a problem with the majority of nicer DVD players I’ve seen. This player is a universal transport that supports DVD-A and SACD. SACD is converted to PCM though if time alignment is used.

This player is also the first one I’ve seen that boasts DTS 96/24 support internally via its multi-channel outputs. This is a nice feature for those who want this support without having to buy a new receiver or processor. The only thing I found really lacking in the audio department was Firewire or “I-link” support, which is becoming extremely popular in high end crowds. This player does support the new version of HDMI though and can pass the high resolution PCM audio of DVD-A. SACD is not supported via HDMI at this time.

This is by far the best player I’ve yet seen from Marantz. A Firewire output would have been the icing on the cake in my opinion, but the support of the new HDMI standard pretty much makes up for it. I am glad that Marantz took this player’s video side in their own hands, and whatever de-interlacer they’re using seems to be doing a great job. I look forward to seeing what they have in store down the road.

Samsung - DVD-HD941

MPEG Maker: Zoran MPEG Model: Vaddis 778 Deinterlacer Maker: Genesis Deinterlacer Model: FLI-2310     MSRP: $299.99 Website: http://www.samsungusa.com

I have really not been impressed with the last few Samsung players I’ve tested. The HD-1000 was a step in the right direction, but it was on the market for only a very short time. The HD-931 and HD-841 had some things going for them, but overall, implementation was severely lacking. I know a lot of people were hoping that the HD941 would be the saving grace, but unfortunately it suffers the same fate and ultimately comes up short.

For a long time it seemed that this player would never make it to the market. It had been announced a long time ago but never seemed to materialize on the market. I was hoping that this delay was based on feedback on the previous upscaling models and Samsung was doing some fine tweaking before putting it out on the market, but that doesn’t seem to be the case.

The Samsung HD941 is based on the Zoran Vaddis 778 chip, which is one of the less capable Zoran chips. I would have preferred the Vaddis 5 or 6 which don’t exhibit the same issues the 778 does. Thankfully, the Faroudja processing makes up for a lot of the shortcomings, especially with the chroma tests. In progressive mode, this player passes all of our chroma tests. But this MPEG decoder does not properly synch subtitles to the progressive frames, and this causes obvious combing artifacts in the displayed subtitles.

The player passes a below black pluge via both its component and HDMI output, and the HDMI output is properly configured for Studio RGB levels. But there is a catch, it only works at 480p or 1080i. If you set the output to 720p, it remaps to PC levels and clips the below black and above white levels. I have no idea why they did this. This is an improvement over the previous 841 and 931 players, but still a glaring oversight. Also, if you convert the HDMI signal to DVI, the RGB levels remap to PC levels and truncate the below black and above white information, so I don't recommend this player in that application. The white level is a bit hot via the component output and measured in at 103 IRE. This will artificially enhance contrast levels and tend to cause whites to look overdriven at times.

The core performance was below average compared to most players at this price point. The 941 has obvious Y/C delay that measured out greater than a full pixel between the luminance and chroma channels. There was no Pb/Pr delay though. The HDMI output did not show the same Y/C delay as the component output and is the preferred output for this player. The component outputs overall frequency response was exaggerated in the mid band and then rolled off a bit in the upper end. This could lead to artificial ringing in the image and a loss of fine detail.

Pixel cropping was a bit on the severe via both HDMI and component outputs. The left and right side crop over 20 pixels together, and the top and bottom combine for 7 missing pixels.

Overall, navigation was quite sluggish with some menus taking a LONG time to load or respond to inputs. Chapter skips were a bit better though. The player’s setup menus were easy to navigate, and the HDMI setup was quite easy. The user can select from three different output resolutions via HDMI; 480p, 720p, and 1080i. I was disappointed that 480i wasn’t supported, as this could bypass the Faroudja chip and remedy the macroblocking issue, yet still provide a digital interface for the monitor or an outboard scaler. Our layer change test clocked in at a very slow 2.5 seconds, which is too long for a new generation DVD player.

The player did as expected with our de-interlacing tests. But, for some reason, the Faroudja processing was extremely slow at picking up the proper cadence. This was quite evident with even the most basic 3-2 cadence tests. Once the cadence was locked down though, the image was quite stable. This player does not have any pre-selectable progressive modes, so what you see is what you get. The player passed all of our tests with the exception of the 2-2 cadence test. This was expected since the player didn’t provide any alternate progressive modes.

The HD941 is a universal transport and supports both the DVD-A and SACD formats. It does not use the newer HDMI spec though, so it will not pass DVD-A material over its HDMI connection.

Overall, this is an improvement over the HD-931 and HD-841 but still falls short of the other players with the same features and general price point. I really hope that Samsung will put a bit more effort into their future offerings, especially since this is their third generation of upscaling player.

Samsung - DVD-HD1000

MPEG Maker: LSI MPEG Model: Ziva5 Deinterlacer Maker: Genesis Deinterlacer Model: FLI2310     MSRP: $1500.00 Website: http://www.samsungusa.com

The Samsung HD-1000 was a limited run player that did not get a big reception here in the United States. Pity, it is a beautiful beast of a player that rivals almost anything I’ve seen in terms of aesthetic appeal.

The player features full DVD Audio support and HDCD as well, but not SACD. The biggest draw for the video enthusiast is full scaling support over the component outputs to 720p or 1080i using the Genesis/Faroudja FLI-2310 chip. The only other player I know of that does this is the discontinued Zenith DVB-318. But this is a far better player in almost every regard.

The platform is based around the LSI Ziva 5 MPEG Decoder which is not one of the better designs out there. It does have some chroma issues associated with it, but the Genesis chip’s chroma filters mask them when using progressive playback. The chip also doesn’t synch its subtitle overlays to the video fields which creates some noticeable combing.

The manual for the Samsung says that the up-scaling feature only works with non-copyrighted material, but I tried various DVDs and all seemed to work fine at any resolution I chose. Selecting the output resolution is a bit tedious though as it is accomplished via a small switch on the back of the player. This is inconvenient for those who plan on putting this on a rack since the switch won’t be very accessible.

As I mentioned earlier, this player features the new Genesis FLI-2310 chip with full DCDi support. This chip has had its ups and downs over the last year. While it does support great de-interlacing and scaling features, the high bit processing has had some pretty obvious incompatibility issues with displays that require further processing of the image. This results in digital artifacts that can vary in intensity from display to display. Since I am always using the same display, it is easy for me to ascertain the level that each player shows the problem. Unfortunately, this is one of the worse I have seen so far. Blocking was extremely evident and showed up throughout the entire testing process. Most of the newer players have only shown the problem using specific test material, but that's not the case here.

In terms of de-interlacing, the HD-1000 did very well but did not have a mode to support material with a 2-2 cadence. This is pretty common though for cadence based models.

Since this player doesn’t feature any digital video outputs, the importance of a solid analog section goes without question. Peeking under the hood revealed an extremely robust power supply section and that translated to a very crisp image on screen. The overall video frequency response of the player was quite good, regardless of the output resolution. There was some slight Y/C delay and it varied in intensity depending on the resolution selected. 720p had the least amount and was less then a half pixel. The biggest flaw I could find came from the pixel cropping. For 720p and 1080i, there were more then 19 pixels cropped on the right side of the image alone!! 480p was the best in this regard, but it suffered from the worst Y/C delay with a full pixel delay.

While I didn’t do a full evaluation of the audio side, I was inclined to see what type of setup options would be included. Unfortunately, this was a bit of a letdown. I didn’t see any options for time alignment between speakers or even channel level controls for that matter. This is unfortunate, since these are necessities for a great multi-channel experience, and with a player built to this magnitude, I expected more.

Overall, this is the best scaling player I’ve seen from Samsung so far, but since it is hard to find and I think discontinued, we might not see one like it again. Like so many other companies, Samsung has gotten more in the habit of releasing smaller players with cheaper designs. A huge thanks goes out to Eric Eash for lending us this player for the benchmark, we really appreciated the opportunity!!!

Video Frequency Response
Samsung

Cary Audio - DVD-6

MPEG Maker: ESS MPEG Model: ESS Videodrive ES6028F Deinterlacer Maker: nDSP Deinterlacer Model: NV221     MSRP: $2000.00 Website: www.caryaudio.com

Cary Audio is a small independent company based out of North Carolina that specializes in high end audio equipment including processors and amplifiers. We have reviewed quite a bit of their audio components but this was the first time I had the chance to look at one of their DVD players. For the benchmark they sent out the DVD-6, but they informed me that they have some other players currently in the works.

The DVD-6 caught me completely by surprise and is probably the one of the biggest highlights of this benchmark. This thing is one of the best built DVD players I’ve seen regardless of price. This is not just some re-badge of a mass market machine; this was designed from scratch from the ground up. Looking under the hood revealed one of the best layouts and startling designs I’ve ever seen. The audio board alone was a thing of beauty and the power section featured one of the largest transformers I’ve ever seen in a DVD player. I cannot wait to have the opportunity to test out the audio side of their players in the future.

This player is built upon a MPEG decoder from ESS, though it is not the typical Vibratto model. This player uses more commercial based video processing. This had its ups and downs overall, but mainly ups.

The MPEG decoder didn’t pass the two most common CUE errors, 4:2:0 ICP and 3-2 alternating, but few do without chroma filtering. It doesn’t have any issues with video or normal 3-2 based material though. There was some combing with subtitles but it was light. This transport and decoder is extremely fast. Menu navigation and chapter skips were almost instantaneous and reminded me a lot of a Mediatek solution or the popular Denon DVD-2900. All of the setup menus were very straight forward and easy to navigate making this one of the most user friendly designs I’ve encountered.

For de-interlacing the DVD-6 uses the NV221 chip from nDSP. The only other player I know of that uses this chip is the Primare DVD player which I haven’t had the chance to review yet. This is a commercial design that is rarely used in consumer DVD player but it performs very similar to the Silicon Image chips. In fact, up until the engineer told me what solution they were actually using, I thought it was probably the SIL-504 chip. The player features three different modes; Auto, Film and Video. These modes seem to be identical though as they all did exactly the same thing regardless of the material I put on. The player is motion adaptive and it passed the majority of our tests with flying colors with the exception of the Bad Edit and 2-2 cadence tests, which are common hiccups for most de-interlacers including the Silicon Image chip. For any end users I would just recommend keeping the player in Auto.

The DVD-6 did extremely well in our core testing. Frequency response was excellent and just about ruler flat across the board. There was absolutely no Y/C delay either. The biggest flaw I could find was pixel cropping which was quite excessive. The player is clipping more then 20 active pixels from the right side alone and the top and bottom of the image are combining for another 15. Hopefully this is something they can resolve with future models or a firmware update.

I personally can’t wait to see Cary Audio’s follow ups to this player as it is one of the best specialty market players I’ve seen to date. Judging from the inside I would bet this player sings too. Look for more reviews of their upcoming players in the next year with full reviews of their audio side as well.

Video Frequency Response
Cary Audio

Arcam - FMJ DV-29

MPEG Maker: Zoran MPEG Model: Vaddis 5 Deinterlacer Maker: Zoran Deinterlacer Model: Vaddis 5     MSRP: $2999.00 Website: http://www.arcam.co.uk

Arcam returns with a new flagship offering that replaces the excellent FMJ DV-27A. I had the opportunity to review that player a few months ago and it is easily one of my favorites to date. This new platform continues Arcam’s reputation of phenomenal build quality and audio prowess.

Like its predecessors, the DV-29 is based around the Zoran Vaddis 5 MPEG decoder. But something has changed this time around. In previous reviews the chroma problems associated with this chip were minimal and with the DV-27A, I could barely detect the problem at all. But this time around it was quite obvious using the same test material as before. I don’t know if this was a filter issue or not but banding was easily detected using our tests.

Another drawback in comparison to the 27A is the lack of the Silicon Image de-interlacer, which continues to be one of the best in the industry. Arcam dropped that solution from this design and relied on the Vaddis 5 for its I/P conversion duties. This was a mistake in my opinion. Flagship designs deserve flagship performance and the Vaddis doesn’t quite deliver that. It is still a good motion adaptive solution, but it tends to comb a bit too often in comparison to the better solutions out there. Now this player’s de-interlacing performance is more on line with the DV-79 which we previously reviewed but didn’t like as much as the DV-27A.

The plus side here is the HDMI output that was lacking on the 27A. And since Arcam was good enough to allow for a 480i output, you will eventually be able to feed an outboard video processor an interlaced signal that is completely digital for de-interlacing and scaling. I really wish more manufacturers’ that are including an HDMI output would include this feature. On top of that, this is one of the first HDMI players that meet the new high resolution audio specification allowing full DVD-Audio high resolution audio streams via HDMI. This makes the player fully compliant with the new processors and receivers that will feature HDMI in the upcoming year.

The analog side of the player was just about perfect. Y/C delay was perfect, the frequency response was superb and the overall usability of the player is excellent. Menu navigation and chapter skips are speedy but the layer change was too slow at 2 seconds. This was probably the biggest drawback to the player from a usability standpoint.

The audio side has been revamped and seems to retain the same glowing attributes the DV-27A had including full Dolby Pro Logic II support inboard.

I really hope that Arcam will return to a standalone de-interlacing solution like the Silicon Image or even something newer. Relying on their MPEG decoder at this price is not the way to go, especially to stay in line with the overall performance their product line is known for. Otherwise this is still one of the best built DVD players out there.

Toshiba - SD-4960

MPEG Maker: Zoran MPEG Model: Vaddis 778 Deinterlacer Maker: Zoran Deinterlacer Model: Vaddis 778     MSRP: $129.99 Website: http://www.toshiba.com

Toshiba ventures into the Universal player market with their first single disc universal offering, the SD-4960. At only $129.99, this is one of the lowest priced universal offerings I’ve seen to date.

The 4960 is based on the Zoran Vaddis 778 MPEG decoder and uses this chip for all its video processing duties. There are plenty of players out there that use the Zoran chips, but we have found that their quality varies substantially with each different model. The 778 chip didn’t do near as well as the Vaddis 5 does, which is the preferred Zoran chip.

This player is not motion adaptive, which at this point in the game is a must for a progressive scan DVD player. This player will drops into video mode with any flag issue and combing is quite evident. If the DVD is flagged correctly (which most aren’t), artifacts will be few and far between and the image should look quite good.

This chip does have an advantage over the Vaddis 5 though, better chroma filtering. This player did a superb job with almost all of our chroma tests and even passed the 4:2:0 ICP test. With normal 3-2 based material, there was a slight amount of jaggedness around the edges, so I gave it a borderline score for that test. The MPEG decoder did an excellent job synching the subtitles to the video frames and didn’t show any signs of combing.

The analog stage of this player was pretty good. There was no Y/C delay at all using Video Essentials Bowtie pattern and the overall frequency response of the player was decent, with only a very slight roll off in the upper frequencies.

Navigating menus was quick and painless and chapter skips were decent. I clocked the layer change in at about 1 second, which is respectable.

This player does pass a below-black pluge pattern but its white level is set a tad high at 103 IRE. This gives the image a rather exaggerated look in contrast levels.

This player does support DVD Audio and SA-CD and offers a single crossover point and level adjustments for each channel. It does not offer time alignment though, which is crucial for proper multi-channel audio playback.

Overall this is a decent entry for Toshiba, especially given the price point. I wish they would have gone with a slightly better MPEG decoder and video processor though.

Video Frequency Response
Toshiba

Crystalio - VPS-2300

MPEG Maker: N/A MPEG Model: N/A Deinterlacer Maker: Genesis Deinterlacer Model: FLI-2300     MSRP: $3499.00 Website: www.crystalio.com

The Crystalio VPS-2300 is a new stand alone video processor that incorporates the latest video processing chip from Genesis, the FLI-2300. This chip was designed by Faroudja laboratories and features both DCDi and TrueLife processing. Crystalio has also added some of their own features including 1:1 pixel mapping and a slew of digital video enhancement tools.

The FLI-2300 chip is used to its full capabilities here and Crystalio has included a huge amount of user tweakability to dial in the performance of the chip. This includes full use of the TrueLife processing set. TrueLife processing allows for noise reduction and digital enhancement without the ringing that is normally associated with these functions.

For our de-interlacing tests, the Genesis did the same as its other implementations, which is great. In the Auto mode, this processor passed all of the tests with the exception of 2-2 based de-interlacing. You can use the pre-set 2:2 Odd or NTSC 30 de-interlacing modes for this and the processor will lock in just fine.

I used my Denon DVD-5900 that has been SDI modded to test the analog frequency response of this processor. The player has a slight roll-off in the upper end that contributes to a slightly soft look. Y/C delay was fine and can be adjusted in the processors menus for tweaks that need it.

The new firmware for this processor fixes the video output level via component out. Previously the white level was too low and now it measure in at almost a perfect 100 IRE.

Like every other product that uses the FLI-23xx series chips, this processor suffers from the digital artifacting commonly know as “Macroblocking”. Using my reference, Chapter 22 of "A Bug’s Life", digital artifacts were clearly present in backgrounds and varied from scene to scene. While this wasn’t the worst case of it I’ve seen, it was still quite apparent. Those of you with plasma displays or fixed pixel monitors should be wary of this and make sure you can return the unit if you have the compatibility issue.

Look for a full review of this processor very shortly from our staff writer Brian Weatherhead.

Video Frequency Response
Crystalio

Lumagen - Vision HDP

MPEG Maker: N/A MPEG Model: N/A Deinterlacer Maker: Silicon Image Deinterlacer Model: SIL-504     MSRP: $1499.00 Website: http://www.lumagen.com

The Lumagen Vision HDP is almost identical to the Vision DVI. The case, inputs, outputs and menu structure are all identical. The differences you will see come from the internal processing and capabilities of the unit. This processor is based more on the flagship VisionProHDP then the VisionDVI.

Some of the new features you’ll find with the HDP include full support of 1080P via component or DVI outputs. The VisionHDP will also de-interlace 1080i, but the processing is not motion adaptive. This is similar to most processors out there at this point. Lumagen is working on a motion adaptive solution and plans to support this feature with a future software upgrade.

Another cool feature is the ability to transcode RGB and DVI to component video. This is a huge plus for those using analog based rear projection televisions or older projectors. Being able to feed the processor digital information from your DVD player and skipping the D to A/A to D steps results in a much cleaner image in the end. Like the VisionDVI, this unit will accept DVI signals encoded with HDCP and do full video processing to it then output again with the HDCP encryption.

The processing is done at 10 bits like the VisionDVI but this processor does oversampling which allows for a slightly cleaner output. This is a nice addition if you are using the analog video outputs of the processor.

Like the rest of the Lumagen line, the VisionHDP uses the Silicon Image de-interlacing solution and their proprietary scaling algorithums. The processor will let you output any resolution you want via any output. The scaling performance was excellent all around and didn’t add any noticeable artifacts regardless of the test material I threw at it. The only shortcoming was with 2-2 based material and Lumagen is currently working on a software upgrade for this.

Like the VisionDVI my chief complaint is a lack of audio delay support, which I have become accustomed to with the DVDO Iscan HD. I would have also liked some type of user interface on the front panel, although the menus were easy to navigate from the on-screen display. Be aware that the flagship VisionProHDP unit does have a front panel display and interface for those that demand it. The manual for the unit steps you through all of the setup very easily and includes the key strokes needed for anything you might want to do.

Again given its price point and features the Lumagen processors are really a bargain. Keep a look out for a full review of the Lumagen processors in the months to come.

Video Frequency Response
Lumagen

Lumagen - Vision DVI

Lumagen was nice enough to lend us two of their newest video processors for this benchmark. The VisionDVI is the entry level offering at an attractive pricepoint of $999. This small unit is exactly the same size and shape as the Vision HDP, with just a few features missing from the other sibling.

Like the previous Lumagen processors, the DVI uses the Silicon Image processor for its de-interlacing duties and a proprietary solution for its scaling. The de-interlacing tests were a breeze for this unit with the only failure coming from the 2-2 based material. Lumagen is aware of this issue and is working on a software upgrade that deals with the issue. This is a pretty common problem for the Silicon Image chip and it usually needs to be forced into a film bias mode to pass this test.

The Vision DVI that I requested was outfitted with optional SDI inputs. This allows me to measure the frequency response, Y/C delay, and white levels that the processor puts out over its component outputs. The frequency response of the DVI was just about ruler flat and didn’t have a single deviation greater than .5db. Y/C delay was dead on and the processor allows for adjustments to compensate for your input source our monitors issues. White level was also spot on at 100 IRE.

There are a host of great features with this unit. Each input can have up to 4 independent memories for picture adjustments and component video and RGB inputs can be transcoded to DVI. This unit will also accept any resolution via its DVI input and output any resolution you want, with the exception of 1080i to 1080P, even if the incoming signal is HDCP encrypted. The DVI output will adhere to the HDCP encrypting though. The unit will support any aspect ratio you want regardless of input resolution and comes pre-programmed with several choices for the output aspect ratio.

Some of the other great features included in this processor are color space conversion, frame rate conversion and 1:1 pixel mapping via any output. The color space conversion is a huge treat. There are a lot of monitors that lock into a HD color matrix when fed a higher resolution signal, even if the source was originally 480P. Since DVDs are mastered in the SMPTE C color space, and HD is not, it is important that the monitor is fed the proper color space or that the processor can convert it before hand. 1 to 1 pixel mapping is the biggest benefit a processor can give you. With compliant displays, this allows all of the internal video processing of the monitor to be bypassed which usually results in a far cleaner image that isn’t limited to the low bit processing most displays use.

The Vision DVI does have simple chroma filtering to remedy chroma upsampling issues. This was apparent when I hooked up my DVD-5900 via SDI. Our test patterns revealed the 4:2:0 ICP issues associated with the 5900 were remedied. This also works for 3-2 Film and Video based cadences.

One of the features I would have liked to have seen with the Lumagen processors was audio delays to compensate for the video processing. This is a feature that DVDO included on their new processor and it makes a big difference since most receivers and processors don’t allow for global delays. Overall though this is an outstanding processor given its price point. The scaling looked excellent regardless of the test material I presented it with and the DVI and SDI support is a huge plus. This unit will also re-scale any HD input and will output via any connector you want. No other processor I know of will do this at this price point. This unit also supports 480i via DVI, which is a first to my knowledge.

Overall this is an outstanding value with top notch processing. Definitely one to put on the short list of video processors.

Video Frequency Response
Lumagen