2001 May 30
25
Philips Semiconductors
Preliminary specification
Multistandard video decoder with adaptive
comb filter and component video input
SAA7118
8.1.2.1
Chrominance path
The 9-bit CVBS or chrominance input signal is fed to the
input of a quadrature demodulator, where it is multiplied by
twotime-multiplexedsubcarriersignalsfromthesubcarrier
generation block 1 (0
°
and 90
°
phase relationship to the
demodulator axis). The frequency is dependent on the
chosen colour standard.
The time-multiplexed output signals of the multipliers are
low-pass filtered (low-pass 1). Eight characteristics are
programmable via LCWB3 to LCWB0 to achieve the
desired bandwidth for the colour difference signals (PAL,
NTSC) or the 0
°
and 90
°
FM signals (SECAM).
Thechrominancelow-pass 1characteristicalsoinfluences
the grade of cross-luminance reduction during horizontal
colour transients (large chrominance bandwidth means
strong suppression of cross-luminance). If the Y-comb
filter isdisabled by YCOMB = 0 the filter influences directly
the width of the chrominance notch within the luminance
path (a large chrominance bandwidth means wide
chrominance notch resulting in a lower luminance
bandwidth).
The low-pass filtered signals are fed to the adaptive comb
filter block. The chrominance components are separated
from the luminance via a two line vertical stage (four lines
for PAL standards) and a decision logic between the
filtered and the non-filtered output signals. This block is
bypassed for SECAM signals. The comb filter logic can be
enabled independently for the succeeding luminance and
chrominance processing by YCOMB (subaddress 09H,
bit 6) and/or CCOMB (subaddress 0EH, bit 0). It is always
bypassed during VBI or raw data lines programmable by
the LCRn registers (subaddresses 41H to 57H); see
Section 8.3.
The separated C
B
-C
R
components are further processed
by a second filter stage (low-pass 2) to modify the
chrominance bandwidth without influencing the luminance
path. It’s characteristic is controlled by CHBW
(subaddress 10H, bit 3). For the complete transfer
characteristic of low-passes 1 and 2 see Figs 12 and 13.
The SECAM processing (bypassed for QAM standards)
contains the following blocks:
Baseband ‘bell’ filters to reconstruct the amplitude and
phase equalized 0
°
and 90
°
FM signals
Phase demodulator and differentiator
(FM-demodulation)
De-emphasis filter to compensate the pre-emphasized
input signal, including frequency offset compensation
(DB or DR white carrier values are subtracted from the
signal, controlled by the SECAM switch signal).
The succeeding chrominance gain control block amplifies
or attenuates the C
B
-C
R
signal according to the required
ITU 601/656 levels. It is controlled by the output signal
from the amplitude detection circuit within the burst
processing block.
The burst processing block provides the feedback loop of
the chrominance PLL and contains the following:
Burst gate accumulator
Colour identification and colour killer
Comparisonnominal/actualburstamplitude(PAL/NTSC
standards only)
Loop filter chrominance gain control (PAL/NTSC
standards only)
Loop filter chrominance PLL (only active for PAL/NTSC
standards)
PAL/SECAM sequence detection, H/2-switch
generation.
The increment generation circuit produces the Discrete
Time Oscillator (DTO) increment for both subcarrier
generation blocks. It contains a division by the increment
of the line-locked clock generator to create a stable
phase-locked sine signal under all conditions (e.g. for
non-standard signals).
The PAL delay line block eliminates crosstalk between the
chrominance channels in accordance with the PAL
standard requirements. For NTSC colour standards the
delay line can be used as an additional vertical filter.
If desired, it can be switched off by DCVF = 1. It is always
disabledduring VBIor rawdatalinesprogrammable bythe
LCRn registers (subaddresses 41H to 57H); see
Section 8.3. The embedded line delay is also used for
SECAM recombination (cross-over switches).