Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : C-Cube -- Ignore unavailable to you. Want to Upgrade?

To: let who wrote (35144)	8/12/1998 4:39:00 PM
From: John Rieman	Respond to of 50808

The trouble with NTSC.................................. tvbroadcast.com More than Meets the Eye? The issues related to transcoding between the "mezzanine level" compression formats, which Sony and other manufacturers are developing for studio applications, and MPEG-2 emission formats could be significant, potentially having an impact on delivered image quality. The reluctance of the ATSC to permit this small change, to a table which the FCC has already eliminated from the standard, appears to be based on fears that a change to the standard this close to the introduction of DTV broadcast services could delay the introduction of new DTV receivers, or worse, cause some receivers to "go black" on broadcasts with 720 samples per line. Jukka Hamalainen, vice president and director of the Matsushita Applied Research Laboratory, told Television Broadcast that Panasonic voted against changing Table 3 because many manufacturers have already committed to the designs of the first generation decoder silicon that will be used in the DTV receivers to be introduced this fall. He went on to say, however, that "the 704/720 issue is not a significant concern, as our decoder implementation will support both." Apparently, virtually all current MPEG-2 decoder implementations are capable of decoding either 704 or 720 pixels per line, along with a variety of other variations permitted by the MPEG profiles that are used in virtually all DTV and related digital video standards worldwide--including DSS and DVD-V. The adoption of 704 in the ATSC standard appears to be rooted in MPEG numerology and to concerns about coding areas of the picture that may contain processing artifacts on the vertical edges of the active picture area. These edges of the picture are typically lost to overscan on CRT receivers, though they may be displayed on systems that present video within a window. MPEG-1 was based on source image format (SIF) with a resolution of 352 samples per line; 704 doubles this resolution, and is still used in many MPEG-2 encoders. MPEG-2 MP@ML also permits the use of 544 samples per line, which is a better match for the resolution of NTSC and most existing television receivers. For example, most broadcasts from DirecTV are encoded with 544 samples per line. Many broadcasts on the European DVB system--which specifies 720 samples per line--are encoded with 704 samples. Receivers are expected to scale any of the line lengths, available in MPEG profiles, to fill the width of the display. This type of scaling is a prerequisite for ATSC compliant decoders, as they must deal with multiple formats that vary both in vertical and horizontal resolution. Other video encoding standards and industry practices crop some of the 720 samples in ITU-R 601 source material. When 601 video is converted to NTSC using the standard 13.5 MHz pixel clock, approximately 712 of the 720 pixels in each video line are preserved. Thus, pictures transmitted digitally with a horizontal resolution of 704 pixels would suffer a loss of approximately 1 percent of their active picture area, relative to the NTSC nominal active picture width. It should be noted that SMPTE Recommended Practice 187 defines the width of the "clean aperture" of Rec. 601 video as 708 pixels. Relative to this benchmark, about 0.5 percent of the picture area would be lost in cropping to 704 pixel wide lines for transmission. All this might lead one to decide that the controversy surrounding this issue is without significant merit...i.e. "much ado about nothing." But the more recent development of mezzanine compression standards for studio and contribution applications could cause a problem. MPEG is based on the coding of 8 x 8 blocks of pixels using the discrete cosine transform (DCT). An array of four DCT blocks (16 x 16 pixels, called a macroblock) is used for the motion compensated prediction that gives MPEG an additional compression boost, removing redundancy between video frames. When images are cropped by removing eight pixels from each side of the image, the macroblock alignment is destroyed (see figure 1). Thus the image must be converted back to baseband and re-encoded, with the potential that some image quality could be lost due to concatenation errors. The "mole" technique preserves the information about macroblock alignment and motion vectors when an image is decoded. This information can then be used when an image is re-encoded, after, for example, the insertion of an emergency message or a station's logo. If the mole encoded image is cropped all of the encoding information is rendered useless.