VIDEOSYNC(3dm)VIDEOSYNC(3dm)NAMEvideosync - information about USTs and video timing on SGI systems
SYNOPSIS
#include <sys/videotiming.h>
DESCRIPTION
In various libraries on SGI systems such as the Video Library (VL),
Unadjusted System Times (USTs) are used to indicate the time at which
video fields or frames cross a jack (electrical input or output of the
machine).
Those USTs label a single point in time, yet the video field or frame
occupies a sizeable range of time. Therefore, we must define a point,
within the video waveform, which a video UST labels. This point is known
as the video "synchronization point."
The following rules define the synchronization point for all video signal
formats supported by SGI hardware. Note that the file "videotiming.h"
contains some useful constants related to these rules.
ANALOG VIDEO
These rules define that point for all analog video formats supported by
SGI VL devices and SGI OpenGL devices:
- Every analog video signal contains horizontal and vertical sync pulses.
- If the horizontal and vertical sync pulses are present on one or two
wires that carry no active video information, then we call this
"component sync." In this case, the signal format determines what each
pulse looks like. The choices are "active low" (each pulse consists of a
downward excursion---a high-to-low edge followed by a low-to-high edge)
or "active high" (each pulse consists of an upward excursion---a low-to-
high edge followed by a high-to-low edge). Otherwise, we have "composite
sync" or "embedded sync." The sync pulses always consist of downward
excursions from blanking level (high) to sync level (low). The signal
format also dictates the high and low voltages, which are the same for
all pulses in that signal (signals with tri-level sync may have pulses to
other voltages as well, but we are not interested in those pulses---for
our purposes, those pulses are not sync pulses.).
- The time at which a sync pulse "occurs" is the half-amplitude point of
its leading edge, and the "length" of a sync pulse is the time elapsed
between the half-amplitude points of its leading and trailing edges.
- All the horizontal sync pulses in a video signal are of the same
length. All the vertical sync pulses in a video signal are of the same
length.
Page 1
VIDEOSYNC(3dm)VIDEOSYNC(3dm)
- The vertical sync pulses are always wider than 0.25 times the signal's
line period. Furthermore, the vertical sync pulses are the only sync
pulses present in the signal that are wider than 0.25 times the signal's
line period. These are the the pivotal observations on which our
definition of the video synchronization point is based.
- The vertical sync pulses always occur in contiguous groups of one or
more. The groups are called "vertical sync pulse intervals."
"Contiguous" means that no other sync pulses occur on any wire of the
signal between the time at which the first vertical sync pulse occurs
(inclusive) through the time at which the last vertical sync pulse occurs
(exclusive). This definition even works for signals with separate
horizontal and vertical sync wires, because in those signals there is
only ever one vertical sync pulse in each vertical sync pulse interval.
The time at which a vertical sync pulse interval "occurs" is the time at
which the first vertical sync pulse inside that vertical sync pulse
interval occurs.
- All video signals have fields. There is one vertical sync pulse
interval per field.
- The synchronization point for a video field is the time at which each
vertical sync pulse interval occurs.
- Often it is convenient to group a repeating pattern of one or more
fields of a video signal into frames. The synchronization point for a
video frame is the synchronization point of the first field of that
frame.
EXAMPLE: NTSC VIDEO
The NTSC signal is defined in ANSI/SMPTE 170M-1994 ("SMPTE 170M" below).
Please refer to SMPTE 170M figure 7 for the line and field numbering
scheme we will be using.
The synchronization point for field 1 and field 3 is the half-amplitude
point of the high-to-low transition at the leading edge of line 4.
The synchronization point for field 2 and field 4 is the half-amplitude
point of the high-to-low transition in the center of line 266.
EXAMPLE: PAL VIDEO
The PAL signal is defined in ITU-R BT. 470-3 ("ITU 470" below). Please
refer to ITU 470 figure 5a for the line and field numbering scheme we
will be using.
The synchronization point for field I and field III is the half-amplitude
point of the high-to-low transition at the leading edge of line 1.
The synchronization point for field II and field IV is the half-amplitude
point of the high-to-low transition in the center of line 313.
Page 2
VIDEOSYNC(3dm)VIDEOSYNC(3dm)
DIGITAL VIDEO
ITU-R BT. 601-4 ("ITU 601" below, also known as CCIR 601 or
Recommendation 601) defines a sampling structure for 525- and 625-line
digital video.
We will define the synchronization points for 525- and 625-line digital
video formats in terms of those defined above for the corresponding
analog formats. Note that the line times (1 H) for the analog and
digital signals are the same, but the analog and digital lines and fields
start at different instants, and the analog and digital fields are not
the same length.
525-LINE DIGITAL VIDEO
Please refer to SMPTE 170M figure 7 for the 525-line analog line and
field numbering scheme we will be using. Define point A as the
horizontal reference point (see SMPTE 170M figure 5) leading analog line
1.
ANSI/SMPTE 125M-1992 ("SMPTE 125M" below) defines a 525-line bit-parallel
digital ITU 601 signal. Please refer to SMPTE 125M figure 4 for the
digital line and field numbering scheme we will be using. Define point B
as the half-amplitude point of the low-to-high transition of the clock
(see SMPTE 125M figure 5) signaling the transmission of word 1473 (Y
sample 736) of digital line 1.
ANSI/SMPTE 259M-1993 ("SMPTE 259M" below) defines a 525-line bit-serial
digital ITU 601 signal. It uses the same line and field numbering scheme
as SMPTE 125M. Define s[n] as the n'th scrambled bit that is transmitted
over a SMPTE 259M connection. A SMPTE 259M receiver must descramble the
bits s[n] into an unscrambled sequence u[n] as per SMPTE 259M 5.2.
Therefore, all SMPTE 259M receivers compute u[n] using this math:
u[n] = s[n] xor s[n-1] xor
s[n-5] xor s[n-6] xor
s[n-9] xor s[n-10]
Define M so that u[M] is the least significant bit of word 1473 (Y sample
736) of digital line 1. Point C lies midway between the leading and
trailing edges of the bit s[M] which is used to compute u[M] using the
formula above.
Consider a SMPTE 170M analog signal, a SMPTE 125M digital signal, and a
SMPTE 259M digital signal which are lined up so that point A, B, and C
coincide.
The synchronization point of field 1 of the digital signals is the
synchronization point of field 1 (or 3) of the analog signal. The
synchronization point of field 2 of the digital signals is the
synchronization point of field 2 (or 4) of the analog signal.
For convenience, we work the values out using the information above.
Page 3
VIDEOSYNC(3dm)VIDEOSYNC(3dm)
The synchronization point for field 1 of the bit-parallel signal is the
half-amplitude point of the low-to-high transition of the clock signaling
the transmission of word 1473 (Y sample 736) of digital line 4.
The synchronization point for field 2 of the bit-parallel signal is the
half-amplitude point of the low-to-high transition of the clock signaling
the transmission of word 615 (Y sample 307) of digital line 266.
The synchronization point for field 1 of the bit-serial signal lies
midway between the leading and trailing edges of the bit s[F1], such that
a receiver computes u[F1] using the formula shown above, and u[F1] is
word 1473 (Y sample 736) of digital line 4.
The synchronization point for field 1 of the bit-serial signal lies
midway between the leading and trailing edges of the bit s[F2], such that
a receiver computes u[F2] using the formula shown above, and u[F2] is
word 615 (Y sample 307) of digital line 266.
Note: As shown in SMPTE 125M figure 2b and 4, a digital line starts with
EAV at word 1440 and wraps from word 1715 to word 0 after SAV.
625-LINE DIGITAL VIDEO
Please refer to ITU 470 figure 5a for the analog line and field numbering
scheme we will be using. Define point A as the Oh instant (see ITU 470
figure 1a) leading analog line 1.
ITU-R BT. 656-2 ("ITU 656" below) defines 525- and 625-line versions of a
bit-serial and bit-parallel digital ITU 601 signal. We will not refer to
the 525-line digital specification in ITU 656, as it is for use with M-
PAL and not NTSC. Please refer to ITU 656 part 1 table 1 for the digital
line and field numbering scheme we will be using.
For the bit-parallel digital signal, define point B as the half-amplitude
point of the low-to-high transition of the clock (see ITU 656 part 2
figure 2) signaling the transmission of Y sample 732 of digital line 1.
For the bit-serial digital signal, as per ITU 656 part 3 section 2, the
bits of each sample are scrambled using the same math as described for
SMPTE 259M signals above. Using the same definitions of u[n] and s[n] as
seen above, define M so that u[M] is the least significant bit of Y
sample 732 of digital line 1. Point C lies midway between the leading
and trailing edges of the bit s[M] which is used to compute u[M] using
the formula above.
Consider an ITU 470 analog signal, a bit-parallel ITU 656 digital signal,
and a bit-serial ITU 656 digital signal which are lined up so that point
A, B, and C coincide.
The synchronization point of field 1 of the digital signals is the
synchronization point of field 1 (or 3) of the analog signal. The
synchronization point of field 2 of the digital signals is the
Page 4
VIDEOSYNC(3dm)VIDEOSYNC(3dm)
synchronization point of field 2 (or 4) of the analog signal.
For convenience, we work the values out using the information above.
The synchronization point for field 1 of the bit-parallel signal is the
half-amplitude point of the low-to-high transition of the clock signaling
the transmission of Y sample 732 of digital line 1.
The synchronization point for field 2 of the bit-parallel signal is the
half-amplitude point of the low-to-high transition of the clock signaling
the transmission of Y sample 300 of digital line 313.
The synchronization point for field 1 of the bit-serial signal lies
midway between the leading and trailing edges of the bit s[F1], such that
a receiver computes u[F1] using the formula shown above, and u[F1] is Y
sample 732 of digital line 1.
The synchronization point for field 1 of the bit-serial signal lies
midway between the leading and trailing edges of the bit s[F2], such that
a receiver computes u[F2] using the formula shown above, and u[F2] is Y
sample 300 of digital line 313.
Note: As per ITU 656 part 1 table 1 note 1 and 2, a digital line starts
with EAV (Cb sample 360 then Y sample 720) and wraps from (Cr sample 431
then Y sample 863) to (Cb sample 0 then Y sample 0) after SAV.
SEE ALSOdmGetUST(3dm), vlGetFrontierMSC(3dm), vlGetUSTMSCPair(3dm),
vlGetUSTPerMSC(3dm)
Page 5
