TM1100 Preliminary Data Book
Philips Semiconductors
6-6
PRELIMINARY INFORMATION
File: vin.fm5, modified 7/24/99
FIELD2: Indicates whether the eld currently being
received is a eld1 or 2. This ag gets updated based
on the F eld of every received SAV code. Note that
eld1 is the ‘top’ eld, i.e. the eld containing the top-
most visible line. Field1 contains lines 1,3,5 etc.
Field2 contains lines 2,4,6,8 etc.
Table 6-3 illustrates common digital camera standards
and the number of active pixels per line, lines per field
and fields per second. Note that any source is accept-
able to VI, as long as the maximum VI_CLK rate is not
exceeded.
Figure 6-9 shows the details of an incoming field and the
captured image. The incoming field consists of N hori-
zontal lines, each line having M pixels labeled 0 through
M–1. Lines are numbered from 0 through N–1. The cap-
tured image is a subset of the incoming image. It is de-
fined by the capture parameters (START_X, START_Y,
WIDTH, HEIGHT) held in the VI_CAP_START and
START_X: Denes the starting pixel number or (X-
coordinate of the starting pixel). START_X must be
even, and greater or equal zero.
START_Y: Denes the starting line number or (Y-
coordinate of the starting pixel). START_Y must be
greater or equal zero.
WIDTH: Denes the width of the captured image in
pixels. WIDTH must be even.
HEIGHT: Denes the height of the captured image in
lines.
Image capture starts after the following conditions are
met:
VI_CTL.CAPTURE ENABLE is asserted.
VI_STATUS.CAPTURE COMPLETE is de-asserted,
indicating that any previously captured image has
been acknowledged.
CUR_Y = START_Y occurs.
Once image capture is started, HEIGHT ‘lines’ are cap-
tured. Each ‘line’ capture starts if:
The previous line capture, if any, is completed.
CUR_X = START_X
Once line capture starts, it continues for 2*WIDTH pixel
clocks1 in which VI_DVALID is asserted, irrespective of
the presence of 1 or more EAV codes.
Note that capture continues regardless of any horizontal
or vertical retrace and associated CUR_Y or CUR_X re-
set. This provides special applications with the ability to
capture information embedded inside the horizontal or
vertical blanking interval. If it is desirable to capture ‘pix-
els’ in the horizontal blanking interval, a minimum time
separation of 1
s is required between the last pixel cap-
tured on line
y and the first pixel captured on line y+1. An
exception to this rule is allowed if and only if the storage
parameters below are chosen such that the last and first
pixel end up in adjacent memory locations. Note that
blanking information capture only makes sense in fullres
mode, with co-sited sampling. All other modes apply fil-
tering, which will distort the numeric sample values.
The captured image is stored in SDRAM at a location de-
fined by the storage parameters in MMIO registers
(Y_BASE_ADR, Y_DELTA, U_BASE_ADR, U_DELTA,
V_BASE_ADR, V_DELTA). Note that the base-address
registers force alignment to 64-byte boundaries (six
LSBs are always zero). The default memory packing is
big-endian although little-endian packing is also support-
ed by setting the LITTLE_ENDIAN bit in the VI_CTL reg-
ister.
Y_BASE_ADR: The desired starting (byte) address
in SDRAM memory where the rst Y (Luminance)
sample of the captured image will be stored. This
address is forced to be 64-byte aligned (six LSBs
always zero).
Y_DELTA: The desired address difference between
the last sample of a line and the address of the rst
sample on the next line. Note that the value of
Y_DELTA must be chosen so that all line-start
addresses are 64-byte aligned.
U_BASE_ADR,
U_DELTA,
V_BASE_ADR,
V_DELTA: Same functions and alignment restrictions
as above, but for chrominance-component samples.
Horizontally-adjacent samples are stored at successive
byte addresses, resulting in a packed form (four 8-bit
samples are packed into one 32-bit word). Upon horizon-
tal retrace, pixel storage addresses are incremented by
the corresponding DELTA to compute the starting byte
address for the next line. Note that DELTA is a 16-bit un-
signed quantity. This process continues until HEIGHT
lines of WIDTH samples have been stored in memory for
luminance (Y). For chrominance, HEIGHT lines of half
Modifications to Y_BASE_ADR, U_BASE_ADR and
V_BASE_ADR have no effect until the start of next cap-
ture, i.e. VI hardware maintains a separate pointer to
track the current address. Modifications to Y_DELTA,
Table 6-3. Common Video Source Parameters.
Video Source
M
(# active pixels)
N
(# active lines)
Field
Rate
(Hz)
CCIR601
50 Hz/625 lines
720
288
50
CCIR601
60 Hz/525 lines
720
240
60
square pixel
50 Hz/625 lines
768
288
50
square pixel
60 Hz/525 lines
640
240
60
1.
Four clocks for each Cb,Y,Cr,Y group representing two
luminance pixels
2.
Note that consecutive pixel components of each line
are stored in consecutive memory addresses but con-
secutive lines need not be in consecutive memory ad-
dresses