The problem is compounded by the virtual impossibility of achieving a common frame rate.
Because the ATSC committee could not agree on supporting a single scanning format, North America is faced with a choice of 18 formats if we choose a 60Hz-related frame rate, or 36 formats if we take into consideration the legacy NTSC-related 59.94Hz frame rate. The ATSC document does not describe analog, digital and other signal characteristics that led to a flurry of standard developing activities in the SMPTE to fill the vacuum. The activity is still merrily going on as existing standards are updated and new standards are developed and issued. On the international scale the problem is compounded by the virtual impossibility of achieving a common frame rate due to the entrenched 60Hz and 50Hz power-line related scanning formats.
In April 2000 the International Telecommunications Union issued the fourth revision of the ITU-R BT.709 Standard, Parameter Values for the HDTV Standards for production and International Program Exchange. This version of the standard reaches out to encompass the 50Hz- and 60Hz-based 1920 × 1080 scanning formats in both their progressive and interlaced aspects. In addition it introduces the recently developed progressive sequential frame (PsF) concept as applied to the 24-, 25- and 30Hz frame systems. This standard makes no mention of the 1280 × 720 scanning format that is not recognized as an HDTV format.
This article outlines several characteristics of the 1920 × 1080 HDTV picture formats.
The 1920 × 1080 picture formats
There are two distinct categories:
- 60Hz-related: 60p, 60i, 30PsF and 30p with a choice of nominal or NTSC-friendly vertical scanning, numbered 1 through 8. All formats in this category feature 2200 Y samples per total line and 1100 each Cb and Cr samples each per total line. The H and V scanning rates and the related Y, Cb and Cr sampling frequencies depend on the scanning system. The highest quality format, 60p, has a Y digital sampling frequency of 148.5MHz and a Cb, Cr digital sampling frequency of 74.25MHz each. At 10 bits per sample this results in a Y, Cb, Cr time-division-multiplexed serial distribution bit rate of 2.97Gb/s. The other formats in this category have an equivalent serial bit rate of 1.485Gb/s.
- 50Hz-related: 50p, 50i, 25PsF and 25p numbered 9 through 12. All formats in this category feature 2640 Y samples per total line and 1320 each Cb and Cr samples per total line. The H and V scanning rates and the related Y, Cb and Cr sampling frequencies depend on the scanning system. The highest quality format, 50p, has a Y digital sampling frequency of 148.5MHz and a Cb Cr digital sampling frequency of 74.25MHz each. At 10 bits per sample, this results in a Y, Cb, Cr time-division multiplexed serial distribution bit-rate of 2.97GB/s. The other formats in this category have an equivalent serial bit rate of 1.485Gb/s.
- 24Hz-related: 24PsF and 24p with a choice of nominal or NTSC-friendly vertical scanning, numbered 13 through 16. All formats in this category feature 2750 Y samples per total line and 1375 each Cb and Cr samples per total line. The H and V scanning rates and the related Y, Cb and Cr sampling frequencies are the same for all formats. All formats have a Y sampling frequency of 74.25MHz and a Cb, Cr sampling frequency of 37.125MHz each.The various 1920 × 1080 categories all belong to a “Common Image Format,” which features a constant number of active lines (1080) and samples per active line (1920) independent of the picture rate.The digital coding is based on one luminance E'y, and two color-difference signals E'cb and E'cr or on the use of three primary signals E'g, E'b and E'r. The most common digital representation assumes a dual-channel parallel data stream consisting of:
- A digital data stream conveying digitized luminance signal (Y) sampled at one of the frequencies listed in Table 1 with 10-bit accuracy.
- A digital data stream conveying time-division-multiplexed signals Cb and Cr.Each data stream carries the active video information, the timing reference signal (TRS) information and the ancillary data if present.A similar situation exists in the Cb and Cr scanning line not shown here. All timing figures are expressed in multiples of T, where T is the reciprocal of the digital sampling frequency fs of the luminance (Y) component signal. Note the peculiar tri-level analog sync waveform. In all likelihood you may never see this analog waveform because you will be mostly dealing with digital signals that do not carry the analog sync.The 0h datum represents the analog line start. A digital sampling instant in a line is defined by a number from 0 (the first active digital sample) through a number representing the total number of samples in a line less one.The digital line starts with sample “o,” which assumes the value 0 in all systems. The last active video sample “p” assumes the value 1919 in all scanning systems. This indicates that all scanning systems have the same number (1920) of active samples per line. Samples “a,” “b,” “c” and “d” identify the four EAV (End of Active Video) words and assume identical values (1920, 1921, 1922 and 1923) in all scanning systems. Sample “e,” the first sample in the horizontal digital blanking interval, assumes the value 1924 in all scanning systems. The samples “f” through and including “n” assume different values depending on the scanning system. As a result symbol “B,” representing the duration of the horizontal blanking interval, assumes a value of 272T for systems 1, 2, 3, 4, 5, 6, 7 and 8. For systems 9, 10, 11 and 12, “B” assumes a value of 822T. This results in an equal number of 1920 active samples per line in all scanning standards, a requirement of the HDTV common image format (CIF) concept.The duration of EAV and SAV (Start of Active Video) signals are equal in all scanning standards. They each consist of a four-word sequence:
- The three synchronizing words with hexadecimal values of, respectively, 3FF, 000 and 000.
- The XYZ word that carries the V bit, the F bit and the H bit, which define the vertical and horizontal blanking. In addition, bits P0, P1, P2 and P3, which assume values depending on the status of the V, F and H bits, provide a limited error correction (single errors) and detection (two errors) of these bits.
Michael Robin, former engineer with the Canadian Broadcasting Corporation's engineering headquarters, is an independent broadcast consultant in Montreal Canada. He is the co-author of Digital Television Fundamentals, published by McGraw-Hill.
Send questions and comments to: firstname.lastname@example.org