SAA7118 Datasheet, PDF (52/169 Pages) NXP Semiconductors – Multistandard video decoder with adaptive comb filter and component video input

Philips Semiconductors
Multistandard video decoder with adaptive
comb filter and component video input
Preliminary specification
SAA7118
Remark: Due to bandwidth considerations XPSC[5:0] and
XACL[5:0] can be chosen different to the previously
mentioned equations or Table 12, as the H-phase scaling
is able to scale in the range from zooming up by factor 3 to
downscale by a factor of ¹⁰²⁴⁄₈₁₉₁.
Figs 33 and 34 show some resulting frequency
characteristics of the prescaler.
Table 12 shows the recommended prescaler
programming. Other programmings, other than given in
Table 12, may result in better alias suppression, but the
resulting DC gain amplification needs to be compensated
by the BCS control, according to the equation:
CONT[7:0] = SATN[7:0] = lower integer of _D-----C2----^X--g-^D--a-^C---i^G-n---^[-²-_×--^:-⁰--6^]---4--
Where:
2^XDCG[2:0] ≥ DC gain
DC gain = (XC2_1 + 1) × XACL[5:0] + (1 − XC2_1).
For example, if XACL[5:0] = 5, XC2_1 = 1, then the
DC gain = 10 and the required XDCG[2:0] = 4.
The horizontal source acquisition timing and the
prescaling ratio is identical for both the luminance path and
chrominance path, but the FIR filter settings can be
defined differently in the two channels.
Fade-in and fade-out of the filters is achieved by copying
an original source sample each as first and last pixel after
prescaling.
Figs 31 and 32 show the frequency characteristics of the
selectable FIR filters.
Table 11 FIR prefilter functions
PFUV[1:0] A2H[7:6]
PFY[1:0] A2H[5:4]
00
01
10
11
LUMINANCE FILTER COEFFICIENTS
bypassed
121
−1 1 1.75 4.5 1.75 1 −1
12221
CHROMINANCE COEFFICIENTS
bypassed
121
3 8 10 8 3
12221
2001 May 30
52