Samsung engineers, it turned out, had forced an internal video noise-reduction circuit to a full-on setting in the first BD-P1000 units to hit the market. Not only did that soften the picture, it also gave the inherent grain in film images "a somewhat mottled look," accounting for the observed "crawlies."
But a second sample BD-P1000 with the noise-reduction circuit disabled by Samsung engineers produced a picture with "no obvious distinction," at first, from the original sample's not-all-that-impressive picture. Hours of further testing revealed that the second sample did offer a number of very subtle picture improvements, however: more detail, more "pop." Many of the remaining picture-quality issues may have been inherent to the first crop of Blu-ray titles to become available, since, often, picture flaws came and went on a scene-by-scene basis.
At any rate, Samsung has promised that BD-P1000s made starting in August 2006 would either disable the noise-reduction circuit outright or allow the user to opt to turn it off or on. Owners of earlier units will be provided with updated firmware to accomplish the same thing. It will have to be provided on a disc, since the BD-P1000 has no Ethernet port.
Another disappointment: the BD-P1000 outputs 1080p, but only after converting the 1080p/24 images on the disc, first to 1080i/60, and then to 1080p/60. A review in The Perfect Vision issue for September 2006 — which is not available online, except as a download of the entire issue here — says:
Movies are stored on the disc at 1080p/24 (24 frames per second); when the disc is played, it is converted to 1080i/60 (60 fields per second) using 3:2 pulldown. Then, it is deinterlaced to 1080p/60 for the HDMI output. This process introduces all the familiar 3:2 artifacts that must be compensated for. To be fair, HD DVD’s 1080i/60 output has the same problem when deinterlaced by the display.3:2 pulldown is the method used to change the 24 frames per second encoded on a Blu-ray disc — the frame rate of motion picture film — to 1080i/60's sixty interlaced fields per second. Each 1080i frame is composed of two fields, with one field containing the odd-numbered pixel rows of the two-field frame, the other with the even-numbered lines.
These artifacts could be completely avoided by simply sending 1080p/24 (or, better yet, 1080p/48 by sending each frame twice) to the display. Unfortunately, there are very few displays that can accept 1080p/24 or 48. All of this leads me to implore manufacturers of displays and high-def disc players to implement 1080p/48 or 72 to avoid all the headaches associated with 3:2 pulldown.
So the BD-P1000 converts 24 progressive frames on the disc to 60 interlaced fields per second and then deinterlaces those fields to make 60 progressive frames per second, which it sends to the TV in digital form over HDMI. On the other hand, if the BD-P1000 is told just to send the 60 interlaced fields to the TV in digital form, the TV itself will deinterlace them, giving what the Sound & Vision reviewer says is an identical result.
Any time deinterlacing of 3:2-pulldown material is done, visible artifacts can mar the picture. For instance, vertical or diagonal edges of moving objects can develop vibrating, serrated/jagged edges; this is called "flagging." Flagging can happen if the de-interlacer doesn't catch on to, or loses track of, the 3:2 cadence of film-based material.
The root of the problem is that two of every five interlaced video frames have mismatched fields, one from a particular film frame and the other from the next film frame. Static objects in the picture are unaffected, but two different versions of moving objects will be picked up in a single video frame. If the deinterlacer doesn't properly compensate for this, you'll see flagging.
There is thus no real advantage to using the 1080p/60 output which the BD-P1000 player can provide — and the first HD DVD players from Toshiba cannot — unless, of course, your TV does a poorer job at "reverse 3:2 pulldown" than the BD-P1000 does.
We are accordingly never going to get the best possible picture from Blu-ray discs until (a) the players can output 1080p at 48 or 72 frames per second, by simply doubling or tripling the 24-fps frame rate on the disc, with no interlacing or 3:2 pulldown, and (b) 1080p TVs can accept digital inputs at those rates.
Yet another disappointment with the BD-P1000 is that it cannot utilize any of the advanced audio compression modes which can optionally be included on Blu-ray discs: Dolby Digital Plus, Dolby Digital TrueHD, DTS HD, or DTS HD Master Audio. It can utilize the uncompressed 6-channel linear PCM audio tracks found on most of the Blu-ray discs released to date and send it out over analog connections to a suitably equipped A/V receiver. (It can also decode the familiar Dolby Digital 5.1 tracks also found on many of the first Blu-ray releases.)
One reason why the advanced audio compression modes have been short-shrifted is that they can't be output in digital form on HDMI, to be decoded by a receiver. That won't even be an option until HDMI version 1.3 arrives sometime in 2007. When receivers start becoming HDMI 1.3-capable, presumably future Blu-ray player models will be designed to read and output the advanced-compression-mode tracks — and/or decode them internally.
Finally, the BD-P1000 disappoints because it doesn't implement BD Java, the computer language which is supposed to allow Blu-ray players of the future to marry disc play with interactive Internet content. Perhaps this is why Samsung omitted the Ethernet networking capability so nicely (but as yet uselessly) built into the first HD DVD players from Toshiba.
When you combine all the disappointments of the first Samsung Blu-ray player with the failure of the Blu-ray camp to utilize the two advanced, highly efficient video compression methods that have been promised for the future, AVC and VC-1, or to be able to manufacture dual-layer discs as yet — the two delays together possibly explain the dodgy video quality of the first Blu-ray releases — you get the impression of a home entertainment format that has been rushed too soon to market, and is not yet ready for prime time.
The Blu-ray camp has long said its format beats rival HD DVD because it holds more data. More recently, Blu-ray has boasted of its true 1080p output, where HD DVD currently maxes out at 1080i. And the extra Blu-ray data space has been touted as perfect for fostering interactive applications, with possible Internet ties.
So what do we get with the first Blu-ray player and Blu-ray discs to be released?
- We get less room for data than on present dual-layer HD DVD media, where what room there is on Blu-ray is taken up by inefficient MPEG-2 video encoding and uncompressed linear PCM audio encoding, the latter being of exceptionally high quality but possibly indistinguishably better than, say, losslessly compressed Dolby Digital TrueHD or DTS HD Master Audio.
- We get 1080p output, yes, but it has been compromised by passing it through intermediate 3:2 pulldown/deinterlacing stages in the player. 1080p/48 and /72 output, which are what is really needed, await TVs that can take advantage of them.
- We get digital audio output over HDMI, yes, but not of any of the missing advanced audio codecs, and not of uncompressed linear PCM, which can only be output in 6-channel analog form. Later Blu-ray players will mate with compatible audio/video receivers over HDMI 1.3 to rectify this oversight, it is true, but that won't be a reality until 2007.
- We get a player that can't connect to a network over Ethernet, can't go online, and can't use BD Java (which isn't implemented on any of the first disc releases anyway.)
- We get discs that don't, from beginning to end, have the eye-popping video quality uniformly found on initial HD DVD releases — discs whose insufficiencies are possibly aggravated by a player that may or may not be doing unwanted video noise reduction.