ANSI SCTE 27:2016 pdf free download – Subtitling Methods for Broadcast Cable
4.5 Display Standards
Subtitles are authored with a specific display standard in mind (NTSC vs. PAL, HDTV vs. SDTV, etc.). The display standard used influences choices regarding text placement on the display grid. The targeted display standard is indicated in each subtitle message, given as the size of the display grid horizontally and vertically. Note that in some instances it may be possible to process a subtitle stream intended for one display standard on video of another. For example, it may be practical to define a service containing both an HDTV component and an SDTV component (dual carriage of video), but with just one subtitle stream. The stream would be formatted for NTSC display, to cover both cases. In this example, the NTSC version represents the least common denominator between the two formats. It is the responsibility of the HDTV Decoder to scale the vertical position based on vertical resolution, and adjust positions horizontally, as appropriate, in order to accommodate the foreign-format subtitle. For horizontal placement of bitmap or text objects, the 720 NTSC pixels would be centered within the HDTV display region. The actual sizes of bitmaps and frames shall not be required to be scaled except by a factor of two (line doubling). Specification of specific requirements for support of foreign-format subtitles are outside the scope of this Standard.
MPEG-2 Transport Stream packets come into the model at the left, and are filtered by PID. Packets whose PID matches the target value flow into the 512 byte transport buffer. These buffered packets are removed at a rate of 32 kbps (kilobits/second) for depacketization and desegmentation, and the result is stored in the input buffer as the data input to the subtitling application. In the case of segmented subtitle messages (see descriptions in Section 4.6), only completely reassembled messages should be kept for subsequent processing. Incomplete messages shall be discarded. Subsequent processing of each reassembled message includes decompression, rendering, and queuing for display at the specified in-cue time. In order to define the input buffer and display queue sizes, maximum data sizes for uncompressed and rendered subtitles are defined. Although derived from the NTSC display standard, the same maxima are also assumed for other display standards. The Title Safe Area on a standard NTSC screen is 80% of the screen. The normal subtitle region can be up to 120 lines within the title safe area. The 120 lines may be divided into groups of contiguous lines corresponding to disjoint sub-regions. Therefore, for monochrome bitmaps a maximum sized subtitle region is 576 x 120 = 69120 bits = 8640 bytes. The size of the input buffer is 16 Kbytes, which can hold at least three largest compressed subtitles if the compression ratio is no worse than 2. Data stored in the input buffer is removed and processed at a rate of 128 kbps. This Standard assumes that each displayed pixel is a 4-bit colored pixel. Consequently, the size of a rendered subtitle is up to 4 times the displayed screen size and the size of a subtitle being processed can be as large as 5 (4 for the rendered subtitle plus 1 for the original decompressed bitmap) times the screen size, a total of at least 8640 x 9 bytes. In addition to the queuing function, the display queue can also be used as a work space for data decompression and subtitle building.