Recently I had the opportunity to test Sony’s flagship DVD player, the DVP-9100ES. While I was impressed with the build, features, and video processing, the player needed work in its core performance. This has become more and more of an issue it seems in the last year or so. As video processing has been getting better and better, attention to detail in some of the most critical areas has been getting worse. The unfortunate side of this is the core performance has more to do with the manufacturer then anything else. Does this mean they are getting lazy, or that they just don’t care? It is a hard question to ask but an even harder one to answer. Most video processing chips do just fine in their default state nowadays, so implementing them is not hard. Most user interfaces are rehashes of older designs and are generic amongst whole product lines most of the time. But tweaking the full performance of the core video properties takes time and attention to detail. It means you have to look for the issues before hand and work with the chipmakers to get it right. I’ve seen some companies invest serious amounts of time in this, and others just push out the same problems over and over again, year after year.
For this review I took a crack at one of Sony’s entry offerings, the DVP-NS90V. This player is much different than the DVP-9100ES in every way. The build quality is in an entirely different league and its video processing isn’t at the same level. But here is a player that actually performed better on the core side. The DVP-NS90V uses a different MPEG decoder; that much was apparent right from the start. So this could mean that the chip’s default performance was better from the start and didn’t require much tweaking. But it really makes me wonder why a player that costs less than half of another would actually have these important details worked out.
I was unable to figure out what MPEG decoder is being used for the DVP-NS90V. Most Sony DVD players have the information on the top of the chip scraped off, blacked out, or stamped with a generic Sony stamp. I have some suspicions that it may be a Mediatek offering, but I can’t prove it. Its responsiveness and chroma performance suggest it though. Sony is not using a two-chip solution here, so all MPEG decoding and de-interlacing/scaling are being done with this single chip. The video processing performance is also inline with what I would expect from a solution like this.
The NS90V is more of an entry DVD player by price point, but it does have some attractive features that have been making their way down the price ladder. This includes support for SACD (Sony’s high-resolution audio format), HDMI, and scaling to high definition resolutions. The player is average in build for its price point, and has a clean design as well as a reasonably friendly user interface. The HDMI output supports standard PCM and Dolby Digital and DTS soundtracks, but not SACD or DVD-A digital out. The HDMI video output supports 480p, 720p, and 1080i. Support for 480i is not offered, making this player a bad choice as a straight digital transport. That is unfortunate as its core performance on the HDMI side is pretty good.
The player is very good from a usability standpoint. Menu navigation and chapter skips were very fast, making operability smooth. The initial load times were a bit on the sluggish side though, which seems to be a reoccurring theme with the Sony players. The player’s layer change was very fast and clocked in at less than a second.
I started my tests with the digital side of the house first. I evaluated all of the HDMI output resolutions but conducted the majority of the benchmark tests with 720p to match the native resolution of the projector I had at the time. The core performance of the HDMI output was quite good and better than the analog component output. You can choose from two output color spaces: YCbCr and RGB. The YCbCr output is upsampled to 4:4:4, which is another indicator that this is indeed a Mediatek-based player (the new Mediatek chips are having chroma issues with 4:2:2 YCbCr so they don’t include this output option). The luma scale is correct and uses video RGB levels with black correctly set at digital 16 and white at digital 235. Above white and below black information are retained.
The HDMI output did extremely well with our chroma upsampling error (CUE) tests, failing only the expected 4:2:0 ICP test. It actually did quite poor in this test though with very pronounced streaking in moving objects, in fact, far more than any other Mediatek player I’ve looked at in the past. The DVP-NS90V had no signs of Y/C delay via HDMI, and pixel cropping was limited to just 3 pixels on the left side.
The component outputs had somewhat different results. While they performed the same with our CUE tests, the analog outputs suffered from over a full pixel of Y/C delay. All NTSC video is basically two complete frames of video overlaid onto each other. One contains the luma (black and white) data, and the other has the chroma (color) data. This information is combined to give you the full color image. This is why grayscale is adjusted when a display is calibrated. If the grayscale (the luma information from black to white)is incorrect, it can skew the color information. If there is Y/C delay, the images do not line up properly on the display causing what looks like ghosting or ringing on one side of the picture. With the NS90V, the chroma information is delayed to the left side so the right side of an object will appear to have some ringing. Images also tend to be softer from this problem similar to what you see when convergence is off in an analog or 3-chip digital display solution.
The white level of the player is just a tad low but still passed our criteria. Using a gray ramp and a digital oscilloscope, we measured the expected 100 IRE output at 99 IRE. This should have a negligible effect on contrast. The overall video frequency response of the analog video was quite good and near flat. There was just some mild increase in the upper frequency range, but I did not see any ringing as a result. The component output does not crop any pixels and retains the full active image.
Moving on to our de-interlacing tests, the Sony did about average for its price point. This is the drawback of a one-chip solution. At this time there are quite a few other players on the market at this price point using the Genesis FLI-2310 solution which offers far better de-interlacing performance, but even that solution has its caveats. The now infamous “macroblocking” bug manifests with a lot of displays out there, making the image near un-watchable for some consumers. At this time, there are just no other low-cost video processing solutions being used, so the lower end market tends to be below average in overall video prowess.
The NS90V does do true inverse telecine de-interlacing with both 3-2 and 2-2 based cadences, which covers the majority of the software on the market today. This player also features motion adaptive de-interlacing for video based material and does a fairly decent job with diagonal line processing. While not nearly as good as Faroudja’s DCDi solution or some of the other newer chips out there, it is not nearly as bad as most current single chip DVD players.
The Sony did fail the majority of our “hiccup” tests. These include chapter breaks and bad flagging. Recovery times are good, but you will notice either combing or softening of the image if the DVDs have mastering issues. This is where higher end de-interlacing solutions outperform these budget players and offer the best playback experience.
Conclusions
Overall, the Sony NS90V is quite a decent player. While it is by no means the best budget player I’ve reviewed, it only has a few drawbacks and will perform fine for the majority of DVDs out there. It offers some nice features for the money, and its video performance is about average for this part of the market. If you are looking for a decent low budget player, then the Sony DVP-NS90V might be one player to consider.
