Note: Descriptions are shown in the official language in which they were submitted.
TITLE OF THE INVENTION 2 0 3 ~ 5 3 6
MOTION ADAPTIVE LUMINANCE SIGNAL AND
COLOR SIGNAL SEPARATION FILTER
BACKGROUND OF THE INVENTION
Field of the Invention:
The present invention relates to a motion adaptive
luminance signal and color signal separating filter for
separating a luminance signal (hereinafter referred to as "Y
signal" or simply "Y") and a color signal (hereinafter
referred to as "C signal" or simply "C") from a composite
color television signal (hereinafter referred to as "V
signalH~ in which the frequency of the C signal is
multiplexed on the high frequency region of the Y signal.
The motion adaptive YC separating filter is a filter
which locally judges whether a picture is a still or motion
picture and executes YC separation suitable to the pixel
signal in that picture at each of the locations thereof.
Description of the Related Art:
The current NTSC signal system provides a composite
signal comprising a C signal and a Y signal having its
high-frequency region on which the frequency of the C signal
is multiplexed. Therefore, television sets require YC
separation. Imperfect YC separation causes the picture
quality to deteriorate in cross color, dot crawl and so on.
With development of large-capacity digital memories,
there have been proposed various types of signal processing
circuits for improving the quality of picture, for example,
by using a motion adaptive YC separation which utilizes a
delay circuit having a delay time equal to or more than the
2039536
vertical scanning frequency of a television signal.
Figure 10 is a block diagram showing one example of the
conventional motion adaptive YC separating filters. In
Figure 10, the filter receives, at its input terminal 1, a V
signal 101 according to the NTSC system, which signal is
shared to the respective input terminals of infield YC
separation circuit 4, interframe YC separating circuit 5,
Y-signal motion detecting circuit 6 and C-signal motion
detecting circuit 7.
In the infield YC separating circuit 4, the input signal
is infleld separated into a Y signal 102 and a C signal 103
through an infield filter (not shown), the Y and C signals
being then applied respectively to the first inputs of
~-signal mixing circuit 9 and C-signal mixing circuit 10.
In the interframe YC separating circuit 5, the input
signal is interframe separated into a Y signal 104 and a C
signal 105, these Y and C signals being then supplied
respectively to the second inputs of the Y-signal and
C-signal mixing circuits 9 and 10.
On the other hand, a signal 106 indicative of the movement
of Y signal detected by the Y-signal motion detecting circuit
6 is applied to one of the inputs of a synthesizer 8 while a
signal 107 representative of the movement of C signal
detected by the C-signal motion detecting circuit 7 is
supplied to the other input of the synthesizer 8.
The synthesizer 8 forms a motion detection signal 108
which is shared to the respective third inputs of the Y-signal
and C-signal mixing circuits 9 and 10. Thus, the Y-signal
motion detecting circuit 6, C-signal motion detecting circuit
7 and synthesizing circuit 8 define a motion detecting
2039536
circuit 80.
The output 2 of the Y-signal mixing circuit 9 provides a
motion adaptive separated Y signal 109 while the output 3 of
the C-signal mixing circuit 10 provides a motion adaptive
separated C signal 110.
This conventional YC separating circuit will now be
described in operation.
On YC separation of V signal 101, the motion detecting
circuit 80 judges whether the V signal 101 is one indicative
of a still or motion picture, based on the output signal from
the synthesizer 8 in which the outputs of the Y-signal and
C-signal motion detecting circuits 6 and 7 are synthesi~ed.
As shown in Figure 11, the Y-signal motion detecting
circuit 6 may comprise a one-frame delay circuit 53, a
subtracter 54, a low pass filter 55 (hereinafter referred to
as "LPF"), an absolute value circuit 56 and a nonlinear
converting circuit 57. V signal 101 inputted to the Y-signal
motion detecting circuit 6 at its input 51 is delayed by
one frame at the one-frame delay circuit 53. The V signal
101 is also applied directly to the subtracter 54 and then
subtracted from the one-frame delayed signal to determine
one-frame difference therebetween. The one-frame difference
signal is passed through the low pass filter 55 (hereinafter
referred to as "LPF") and then applied to the absolute value
circuit 56 whereat the absolute value thereof is determined.
The determined absolute value is then converted by the
nonlinear converting circuit 57 into a signal 106 indicative
of the amount of movement of the low frequency component in
the Y signal. This signal 106 is outputted from the output 52
of the Y-signal detecting circuit 6. The nonlinear
2039536
converting circuit 57 serves to convert an absolute value
into a data having a magnitude which can be more easily
handled by the system.
As shown in Figure 12, the C-signal motion detecting
circuit 7 may comprise a two-frame delay circuit 81, a
subtracter 82, a band pass filter 83 (hereinafter referred to
as "BPF"), an absolute value circuit 84 and a nonlinear
converting circuit 85. V signal 101 inputted to the C-signal
motion detecting circuit 7 at its input 11 is delayed by
one frame at the two-frame delay circuit 81. The V signal
101 is also applied directly to the subtracter 82 and then
subtracted from the two-frame delayed signal to determine
two-frame difference therebetween. The two-frame difference
signal is passed through the band pass filter 83 and then
applied to the absolute value circuit 84 whereat the absolute
value thereof is determined. The determined absolute value is
then converted by the nonlinear converting circuit 85 into a
signal 107 indicative of the amount of movement in the C
signal. This signal 107 is outputted from the output 89 of
the C-signal detecting circuit 7.
The synthesizing circuit 8 is adapted to select and
output one of the Y-signal and C-signal movement signals 106
and 107 which is larger than the other movement.
Such a judgement is represented by a control signal 108 in
the form Of motion coefficient (0 ~ k ~ 1). If a picture
is judged to be a complete still picture, the motion
coefficient k is equal to zero. If the picture is judged to
be a complete motion picture, the motion coefficient k is
equal to one.
Generally, if a picture is a still picture, the
2039536
interframe correlation is utilized to perform the interframe
YC separation such that Y and C signals are separated from
each other.
As shown in Figure 13, the interframe YC separating
circuit 5 may comprise a one-frame delay circuit 64, an adder
and a subtracter 66. V signal 101 inputted to the
interframe YC separating circuit 5 at its input 61 is delayed
by one frame at the one-frame delay circuit 64 to form a
one-frame delay signal which in turn is added to the V signal
directly inputted to the adder 65. The resultant one-frame
sum provides a YF signal 104 which is outputted from one
output 62 in the interframe YC separating circuit 5. At the
same time, the subtracter 66 subtracts the YF signal 104
from the V signal 101 directly applied from the input 61 to
the subtracter 66 to extract a CF signal 105 which in turn is
outputted from the output 63 of the interframe YC separating
circuit 5.
In general, if a picture is a motion picture, the infield
correlation is utilized to perform the infield YC separation
such that the Y and C signals are separated from each other.
As shown in Figure 14, the infield YC separating circuit
4 may comprise a one-line delay (one horizontal line ... 1 H
delay) circuit 74, an adder 75 and a subtracter 76. V signal
101 inputted to the infield YC separating circuit 4 at its
input 71 is delayed by one line at the one-line delay circuit
74 to form a one-line delay signal which in turn is added to
the V signal directly inputted to the adder 75. The resultant
one-line sum provides a Yf signal 102 which is outputted
from one output 72 in the infield YC separating circuit 5.
At the same time, the subtracter 76 subtracts the Yf signal
2039536
104 from the V signal 101 directly applied from the input 71
to the subtracter 76 to extract a Cf signal 103 which in turn
is outputted from the output 73 of the infield YC separating
circuit 4.
Since the infield and interframe YC separating circuits 4
and 5 are arranged parallel to each other, the motion
adaptive YC separation filter can causes the Y-signal mixing
circuit 9 to calculate the following equation using the
motion coefficient k synthesized by the synthesizer 8:
Y = kYf + (l-k)YF
where Yf is an output Y signal 102 from the infield YC
separation and YF is an output Y signal 104 from the
interframe YC separation. There is thus obtained a motion
adaptive YC separation Y signal 109 which in turn is
outputted from the motion adaptive YC separation filter at
the output Z.
Similarly, the control signal 108 is utilized to cause the
C-signal mixing circuit 10 to calculate the following
equation:
C = kCf + (l-k)CF
where Cf is an output signal 103 from the infield YC
separation and CF is an output signal 105 from the interframe
YC separation. There is thus obtained a motion adaptive YC
separation C signal 110 which in turn is outputted from the
output 3.
The C-signal motion detecting circuit 7 may be arranged
as shown in Figure 15. In this figure, V signal 101 inputted
to the circuit 7 at the input 11 is demodulated by a color
demodulating circuit 86 into two color difference signals R-Y
and B-Y. These color difference signals R-Y and B-Y are then
2039536
applied to a time division multiplexer 87 in which they are
time-division multiplexed at a certain frequency. The output
signal from the time division multiplexer 87 is then
subjected to subtraction from an output signal from a
two-frame delay circuit 81. There is thus obtained a
two-frame difference signal.
The two-frame difference signal is passed through LPF 88
wherein a Y-signal component is removed therefrom. The output
signal of the LPF 88 is then applied to an absolute value
circuit 84 to extract an absolute value therefrom. The
absolute value is then applied to a nonlinear converter 85
wherein it is nonlinearly converted into a C-signal motion
detection signal 107 which in turn is outputted from the
output 89 of the C-signal motion detecting circuit 7.
It will be apparent from the foregoing that Yf and Cf
signals from the infield YC separating circuit 4 and YF and
CF signals from the interframe YC separating circuit 5 are
respectively mixed with each other, based on the amount of
movement which is obtained by synthesizing the motion signals
from the respective Y-signal and C-signal motion detecting
circuits 6 and 7.
Therefore, the filter characteristics for the still
picture will be completely different from that for the motion
picture. If a picture is switched from a still to a motion
picture or vice versa, the resolution is subjected to severe
change such that the quality of picture will be remarkably
degraded on processing the motion picture.
SUMMARY OF THE INVENTION
In order to overcome the above problem in the prior art,
2039536
it is therefore an object of the present invention to provide
a motion adaptive YC separation filter which can reproduce
even such a multi-switched picture as described above with an
increased resolution and with a reduced degradation of image
quality.
To this end, the present invention provides a motion
adaptive YC separation filter comprising YC separation
in three fields circuit means which can provide Y and C
signals from the YC separation in three fields by locally
detecting the correlation between frames when a motion
picture is detected by a motion detecting circuit, the
detected results being used to perform the adaptive selection
of plural interfield processing operations including
calculations in three fields.
If a motion picture is detected by the motion detecting
circuit, the motion adaptive YC separating filter of the
present invention detects the correlation between the frames.
The motion adaptive YC separating filter includes three
YC separating in three fields circuits, one of which,
depending on the magnitude of the detected correlation, is
selected to provide Y and C signals from the YC separation in
three field.
In accordance with the present invention, the motion
adaptive YC separating filter comprises YC separation in
three fields circuit means which includes YC separating
in three fields filters performing luminance signal band
limitations from three different three-interfield
calculations by detecting the local interframe correlation
when a motion picture is detected by the motion detecting
circuit or four different YC separating in three fields
2039536
filters including luminance signal band limitations from
infield calculations, which filters are adaptively selected
by detecting the local interframe correlation when a motion
picture i8 detected by the motion detecti~g circult.
Therefore, the optimum YC separation can be carried out by
utilizing the correlations in the motion picture to process
it without degradation of resolution.
In accordance with the present invention, moreover, the
motion adaptive YC separating filter comprises YC separation
in three fields circuit means which includes YC separating
in three fields filters performing color signal band
limitations from three different calculations in three fields
by detecting the local interframe correlation when a
motion picture is detected by the motion detectlng circuit or
four different YC separating in three fields filters including
color signal band limitat$ons from infield calculations,which
filters are adaptively selected by detecting the local
interframe correlation when a motion picture $s detected by
the motion detecting circuit. Therefore, the optimum YC
separation can be carried out by utilizing the correlation in
the motion picture to process it without degradation
of resolution.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is a block diagram of one embodlment of a motion
adaptive YC separating filter constructed in accordance with
the present invention.
Figure 2 is a block diagram of the details of a YC
separating in three fields circuit used in the embodiment of
the present invention shown in Figure 1.
A
2039536
Figure 3 is a block diagram of the detail of another
YC separating in three fields circuit usable in the embodiment
of the present invention shown in Figure 1.
Figure 4 is a block diagram of the detail of still
another YC separating in three fields circuit usable in the
embodiment of the present invention shown in Figure 1.
Figure 5A is a plan view illustrating the arrangement of a
V signal represented by the use of t-axis and y-axis, the V
signal being digitized at a frequency four times as high as
that of a chrominance subcarrier in a three-dimensional time
space.
Figures 5B and 5C are plan views illustrating the
arrangement of the same V signal represented by the use of
x-axis and y-axis.
Figure 6A is an oblique view of the spectrum
distribution of the V signal in a three-dimensional frequency
space.
Figure 6B is a view of the spectrum distribution of
Figure 6A as viewed along the f-axis from the negative
side.
Figure 6C is a view of the spectrum distribution of
Figure 6A as viewed along the ~ -axis from the positive
side.
Figure 7A is an oblique view of the spectrum
distribution of Y and C signals obtained from the first
YC separation in three fields according to the present
invention in a three-dimensional frequency space.
Figure 7B is a view of the spectrum distribution of
Figure 7A as viewed along the f-axis from the negative side.
Figure 7C is a view of the spectrum distribution of
1 0
2039536
Figure 7A as viewed along the ~ -axis from the positive side.
Figure 8A is an oblique view of the spectrum
distribution of Y and C signals obtained from the second
YC separation in three fields according to the present
invention in a three-dimensional frequency space.
Figure 8B is a view of the spectrum distribution of
Figure 8A as viewed along the f-axis from the negative side.
Figure 8C is a view of the spectrum distribution of
Figure 8A as viewed along the ~ -axis from the positive side.
Figure 9A is an oblique view of the spectrum
distribution of Y and C signals obtained from the third
YC separation in three fields according to the present
invention in a three-dimensional frequency space.
Figure 9B is a view of the spectrum distribution of
Figure 9A as viewed along the f-axis from the negative side.
Figure 9C is a view of the spectrum distribution of
Figure 9A as viewed along the ~ -axis from the positive side.
Figure 10 is a block diagram of a conventional motion
adaptive YC separation filter.
Figure 11 is a block diagram of the details of a Y-signal
motion detecting circuit in the conventional motion adaptive
YC separation filter shown in Figure 10.
Figure 12 is a block diagram of the details of a C-signal
motion detecting circuit in the conventional motion adaptive
YC separation filter shown in Figure 10.
Figure 13 is a block diagram of the details of an
interframe YC separating circuit in the conventional motion
adaptive YC separation filter shown in Figure 10.
Figure 14 is a block diagram of the details of an
infield YC separating circuit in the conventional motion
203~536
adaptive YC separation filter shown in Figure 10.
Figure 15 is a block diagram illustrating another example
of the conventional C-signal motion detecting circuits.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
The present invention will now be described by way of
example with reference to the drawings.
Figure 1 shows a block diagram of one embodiment of a
motion adaptive luminance and color signal separating filter
constructed in accordance with the present invention. In the
arrangement shown in Figure 1 is distinguished from that of
Figure 10 only in that the infield YC separating circuit 4 is
replaced by a YC separating in three fields circuit 50.
Therefore, the remaining common parts will not be further
described herein.
The details of the YC separating in three fields circuit
50 shown in Figure 1 is illustrated in Figure 2 by a block
diagram.
Referring now to Figure 2, the filter receives, at its
input from one input terminal 11, a V signal 101 which in turn
is applied to the inputs of a 262-line (262H) delay circuit
14, one-line (lH) delay circuit 28 and two-pixel (2D~ delay
circuit 31 and the first input of a subtracter 16.
The V signal is delayed by 262 lines at the 262H delay
circuit 14 and then provided to the input of a one-line (lH)
delay circuit 15. After delayed by one line, the V signal is
applied to the inputs of a 262-line (262H) delay circuit 17
and two-pixel (2D) delay circuit 20 and the second input of
the subtracter 16. After delayed by 262 lines at the 262H
delay circuit 17, the V signal is provided to the inputs of a
2039536
one-line (lH) delay circuit 18 and four-pixel (4D) delay
circuit 30 and the first input of a subtracter 32. After
delayed by one line at the lH delay circuit 18, the V signal
is applied to the input of a two-pixel (2D) delay circuit 21.
The output of the subtracter 16 is applied to the input
of one-line (lH) delay circuit 19, the output of which in
turn is provided to the input of four-pixel (4D) delay
circuit 22 and the first input of an adder 24. The output of
the 4D delay circuit 22 is applied to the first input of an
adder 23.
After delayed by two pixels at the 2D delay circuit 20,
the V signal is provided to the respective first inputs of
subtracter 25, adder 27 and subtracter 41. The output of the
2D delay circuit 21 is applied to the second inputs of the
subtracter 25 and adder 27 and the first input of a
subtracter 34. The output of the subtracter 25 is provided
to the second inputs of adders 23 and 24 and the input of an
LPF 26. The output of the adder 27 is applied to the first
input of an adder 40.
The output signal of the adder 23 is applied to the first
input of a signal selection circuit 39. The output signal of
the adder 24 is provided to the second input of the signal
selection circuit 39. The output signal of the LPF 26 is
supplied to the third input of the signal selection circuit
39.
After delayed by one line at the lH delay circuit 28, the
V signal is applied to the input of a four-pixel (4D) delay
circuit 29 and the first input of a subtracter 33. After
delayed by four pixels at the 4D delay circuit 29, the V
signal is provided to the second input of the subtracter 32.
2039536
After delayed by four pixels at a four-pixel (4D) delay
circuit 30, the V signal is provided to the second input of a
subtracter 33. After delayed by two pixels at the 2D delay
circuit 31, the V signal is applied to the second input of a
subtracter 34.
The output of the subtracter 32 is applied to the input
of an absolute value (ABS) circuit 35; the output of the
subtracter 33 to the input of an ABS circuit 36; and the
output of the subtracter 34 to the input of an ABS circuit 37.
The output of the A~S circuit 35 is provided to the first
input of a minimum value selection circuit 38; the output of
the ABS circuit 36 to the second input of the minimum value
selection circuit 38; and the output of the ABS circuit 37 to
the third input of the minimum value selection circuit 38.
The output of the minimum value selection circuit 38 is
applied to the fourth input of the signal selection circuit
39, thereby selecting and controlling the first to third
inputs of the same circuit 39.
The output of the signal selection circuit 39 is applied
to the second input of an adder 40, the output of which is
provided to the second input of a subtracter 41 and also
outputted through an output terminal 12 as a Y signal 112
from the YC separation in three fields.
The operation will be described below:
Assuming that a scene includes a horizontal x-axis, a
vertical y-axis extending perpendicular to the x-axis in the
same scene and a time t-axis extending perpendicular to a
plane defined by the x- and y-axes, it can be believed that a
space defined by the three x-, y- and t-axes is a
three-dimensional time space.
1 4
203q536
Figure 5 shows such a three-dimensional time space.
Figure SA shows a plane defined by the t- and y-axes while
Figures 5B and 5B show a plane defined by the x- and y-axes.
Figure 5A also shows interlacing scan lines with a broken
line illustrating one field. Solid lines depict that
chrominance subcarriers are in phase.
In Figure 5B, solid and broken lines represent scan lines
in n and (n-l) fields, respectively. Four marks ~o n,
"~ " and "~ " on each scan line represent sample points at
which chrominance subcarriers are in phase when the V signal
is digitized with a sampling ~requency four times as high as
the frequency fsc (=3.58 MHz) of the chrominance subcarrier.
In Figure 5C, solid and broken lines represent scan lines
in (n+l) and _ fields, respectively. Four marks "O ", "- ",
~ n and ~ H on each scan line are similar to those of
Figure B. The sample points "O , "~ ", "- ",and " " have
chrominance subcarriers which are out of phase by each 90
in such an order as described.
If it is assumed that an objective sample point is repre-
sented by a mark "~ ", the chrominance subcarriers are out of
phase by 180at four points a, b, c and d, which are at the
respective second sample points measured forward and backward
from the objective sample point ll~ l and on the respective
first scan line spaced vertically away from the scan line of
the objective sample point in samefield.
Therefore, there can be constructed a comb line filter
comprising a digital circuit, an adapted YC separation filter
as disclosed in Japanese Patent Laid-Open No. 58-242367 and
so on.
Since the chrominance subcarriers are out of phase by
1 5
2039536
180 at the identical sample points spaced away from each
other by one frame as shown in Figure 5A, the present
invention can provide an inframe YC separating filter.
As can be seen from Figure 5B, the phase of the
chrominance subcarrier is reversed in (n-1) field spaced by
one field forwardly apart from an objective sample point at a
sample point on a line immediately above the scan line on
which the objective sample point is located or at two sample
points on a line immediately below that scan line. Therefore,
interfield YC separation can be made from an arithmetic
operation between any one of these three points e, f and g
and the objective point X .
As can be seen from Figure 5C, the phase of the
chrominance subcarrier is reversed in (n+1) field spaced by
one field backwardly apart from an objective sample point at
a sample point h on a line immediately above the scan line on
which the objective sample point is located or at two sample
points i and i on a line immediately below that scan line.
Therefore, interfield YC separation can be made from an
arithmetic operation between any one of these three points _,
i and ~ and the objective point X .
If it is assumed that a horizontal frequency axis
corresponding to the x-axis is ~ -axis, a vertical frequency
axis corresponding to the y-axis is v -axis and a time
frequency axis corresponding to the t-axis is f-axis, it can
be believed that there is a three-dimensional frequency space
defined by these ~ -, v - and f-axes perpendicular to each
other.
Figure 6 depicts such a three-dimensional frequency space
in projection. Figure 6A is an oblique view of the
1 6
2039536
three-dimensional frequency space; Figure 6B is a view of the
three-dimensional frequency space as viewed along the f-axis
from the negative side; and Figure 6C is a view of the
three-dimensional frequency space as viewed along the ~ -axis
from the positive side.
Figures 6A, 6B and 6C represent the spectrum distribution
of a V signal in the three-dimensional frequency space. As
seen from these figures, the spectrum of Y signal extends
around the origin of the three-dimensional frequency space.
C signal has four spectrums located in the three-dimensional
frequency space at four points as shown in Figures 6A to 6C
since I and Q signals are modulated into two orthogonal
phases at the frequency fsc of the chrominance subcarrier.
If the V signal is observed on the ~ -axis as shown in
Figure 6C, however, the spectrums of the C signal will be
only on the second and fourth quadrants.
This corresponds to the fact that solid lines
representing the in-phase state of the chrominance subcarrier
extend upwardly with the passage of time as shown in Figure
5A.
The conventional motion adaptive YC separating filters
performed YC separation by the use of infield correlation
when a motion picture was detected. Although the conventional
filters could carry out the band limitations in the directions
of ~ -axis and v -axis, they could not take the band
limitation in the direction of f-axis. This will cause a
fre~uency space originally including Y signal to be separated
as C signal, so that the band of Y signal in the motion
picture will be decreased.
If the YC separation is made according to the
2039536
aforementioned interfield processing operation, the band of Y
signal in the motion picture can be increased.
Referring again to Figure 5B, the (n-l) field includes
three sample points " ~ e, f and g which are near the
objective sample point "~ " X and have chrominance
subcarriers angularly spaced away from one another by 180 .
Referring again to Figure 5C, the (n+l) field includes three
sample points n Il _, i and ~ which are near the objective
sample point "~ " X and have chrominance subcarriers
angularly spaced away from one another by 180 . Calculation
for any one of the six sample points permits the
three-interfield YC separation.
First of all, a low-frequency component in the
three-dimensional frequency space which is part of Y signal
can be taken out from the sum of the objective sample point
~lo n X with the sample point "- ~ e in Figure SB.
Moreover, a high-frequency component in the three-dimensional
frequency space which includes C signals can be taken out
from the difference between the objective sample point X
and the sample point e. The C signals may be removed from
the high-frequency component when the latter is passed
through the LPF. The addition of these outputs provides a Y
signal. The subtraction of the Y signal from the V signal
provides C signals. This is referred to as "the first
YC separation in three field".
Figures 7A, 7B and 7C are respectively similar to Figures
6A, 6B and 6C and illustrate a three-dimensional frequency
space including the Y and C signals which have been obtained
from the first YC separation in three fields.
A sample point "O " k is now considered which is in the
1 8
2039536
same positional relation as that of the objective sample
point X with the sample point e relative to the sample point
i in Figures 5B and 5C. Secondly, when the difference
between the sample points i and k is added to the difference
between the objective sample point X and the sample point
e, C signals can be removed from the high-frequency component
in the three-dimensional frequency space. When this
high-frequency component is summed with the low-frequency
component in the three-dimensional frequency space which is
part of the Y signal obtained from the sum of the objective
sample point X with the sample point e, Y signal can be
obtained. The subtraction of the Y signal from the V signal
provides C signals. This is referred to as "the second
YC separation in three fields".
Figures 8A, 8B and 8C similarly illustrate a three-
dimensional frequency space including Y and C signals which
have been obtained from the second YC separation in three
fields. It appears from these figures that the C signals are
partially included within the separated Y signal. However,
there is an extremely little possibility that the C signals
are contained in the Y signal, since the great correlation
exists between the Y and C signals.
A sample point "O " 1 is now considered which is in the
same positional relation as that of the objective sample
point X with the sample point e relative to the sample point
~ in Figures 5B and 5C. Thirdly, when the difference between
the sample points ~ and 1 is added to the difference between
the objective sample pointX and the sample point e, C
signals can be removed from the high-frequency component in
the three-dimensional frequency space. When this
1 9
2039536 - - -
high-frequency component is summed with the low-frequency
component in the three-dimensional frequency space which is
part of the Y signal obtained from the sum of the objective
sample point X with the sample point e, Y signal can be
obtained. The subtraction of the Y signal from the V signal
provides C signals. This is referred to as "the third
YC separation in three fields".
Figures 9A, 9B and 9C similarly illustrate a three-
dimensional frequency space including Y and C signals which
have been obtained from the third YC separation in three
fields. It appears from these figures that the C signals are
partially included within the separated Y signal. However,
there is an extremely little possibility that the C signals
are contained in the Y signal, for the same reason as in
Figure 8.
In order to adaptively control switching one of the
three, first, second and third interfield YC separations, it
is required that correlations in the picture is detected in
the directions of connection between the objective sample
point "~ " X and the respective one of the sample points
"- " e, f, and g. The correlations of the picture in the
respective directions may be detected by calculating the
sample points "- " e, f and g in the (n-l) field and the
sample points "- " _, i and ~ in the (n+l) field, the
objective sample point "~ " X being located between the
(n-l) and (n+1) fields. In such a manner, control signals can
be obtained.
The inframe YC separation circuit shown in Figure 2 will
be described in operation below:
The present invention is characterized by that when a
2 0
2039536
motion picture is detected by the motion detecting circuit
80, the motion picture is processed by the optimum selected
one of the inframe YC separations including the aforementioned
first, second and third calculations in three fields, in place
of the infield YC separation.
Referring now to Figure 2, it is assumed herein that V
signal 101 provided through the input 11 is one in the (n+l)
field. The V signal 101 is delayed by 262 lines (one field)
at the 262H delay circuit 14 and further delayed by one line
at the lH delay circuit 15 from which a _-field signal is
outputted. This n-field signal is delayed by 262 lines at
the 262H delay circuit 17 and further delayed by one line at
the lH delay circuit 18 from which a (n-l) field signal i5
outputted.
The _-field V signal, which is the output signal of the
lH delay circuit 15, is further delayed by two pixels at the
2D delay circuit 20 to which a signal at the objective sample
point "~ " X is outputted. At this point, the V signal
delayed by two pixels at the 2D delay circuit 21 is a signal
at the (n-l) field sample point ..~ n . These signals are
then subjected to subtraction at the subtracter 25 to provide
a difference between the objective sample point X and the
sample point e. This difference is passed through the LPF 26
whereat C signals for the first YC separation in three fields
are removed to provide a high-frequency component in the
three-dimensional frequency space.
The V signal 101 inputted through the input 11 is
subtracted from the output of the lH delay circuit 15 at the
subtracter 16 and then delayed by one line at the lH delay
circuit 19 to provide a difference between the sample points
2039536
i and k. This difference is added to the output of the
subtracter 25 at the adder 24. As a result, a high-frequency
component in the three-dimensional frequency space is
obtained from which C signals for the second YC separation in
three fields are removed.
The output of the lH delay circuit 19 is delayed by four
pixels at the 4D delay circuit 22 to provide a difference
between the sample points i and 1. This difference is then
added to the output of the subtracter 25 at the adder 23 to
provide a high-frequency component in the three-dimensional
frequency space from which C signals for the third
YC separation in three fields are removed~
The three calculations between three fields are inputted
to the signal selection circuit 39 and selected by the output
of the minimum value selection circuit 38 as will be
described.
The outputs of the 262H and 4D delay circuit 17, 29 are
subjected to subtraction from each other at the subtracter 32,
with the result being converted into an absolute value by the
absolute value circuit 35, thereby detecting a correlation
between the sample points g and ~ shown in Figures 5B and 5C.
The outputs of the 4D and lH delay circuits 30, 28 are
subjected to subtraction from each other at the subtracter
33, with the result being converted into an absolute value
by the absolute value circuit 36, thereby detecting a
correlation between the sample points f and i shown in
Figures 5B and 5C. The outputs of the 2D delay circuits 21,
31 are subjected to subtraction from each other at the
subtracter 34, with the result being converted into an
absolute value by the absolute value circuit 37, thereby
2 2
203~536
detecting a correlation between the sample points e and _
shown in Figures 5B and 5C.
The minimum value selecting circuit 38 selects the
minimum one of the aforementioned three absolute value
outputs, which is maximum in the detection of correlation.
The minimum value is then used to control the signal
selection circuit 39.
More particularly, the signal selection circuit 39
selects the output of the adder 23 if the output of the ABS
circuit 35 is minimum; the output of the adder 24 if the
output of the ABS circuit 36 is minimum; and the output of
the LPF 26 if the output of the ABS circuit 37 is minimum,
respectively.
Moreover, the output of the signal selection circuit 39
is added at the adder 40 to the low-frequency component in the
three- dimensional frequency space which is the output of the
adder 27, thereby providing a Y signal 112 from the
YC separation in three fields.
The subtracter 41 subtracts the Y signal 112 of
YC separation in three fields from the V signal which is the
output of the 2D delay circuit 20, so as to provide C signals
113 of YC separation in three fields.
Figure 3 illustrates a block diagram of the second
embodiment of a YC separation in three fields circuit 50 which
is shown in Figure 1 and constructed in accordance with the
present invention.
Although the arrangement of Figure 2 adaptively selects
filters performing Y-signal band limitations due to three
three-interfield calculations, the arrangement shown in
Figure 3 adaptively selects filters performing C-signal band
2 3
2039536
llmitations due to three calculations in three fields. Thus,
only a part of the YC separation in three fields circuit shown
in Figure 3, which makes the C-signal band limitations
dlfferent from these of Flgure 2, wlll bo describ~d below. In
Figure 3, parts similar to those of Figure 2 are denoted by
similar reference numerals.
The output of the subtracter 25 is passed through the BPF
44 to provide C signals for the YC separation in three fileds.
The subtracter 43 subtracts the difference between sample
points i and k which is the output signals of the lH delay
circuit 19 from the output of the subtracter 25 to provide a
difference signal which in turn provides C signal for the
second YC separation in three fields.
The subtracter 42 subtracts the difference between sample
points i and 1 which is the output signal of the 4D delay
circuit 22 from the output of the subtracter 25 to provide a
difference signal which in turn provides C signal for the
third YC separation in three flelds.
The outp~t of the signal selection circuit 39 is used to
select C signal from any one of the three calculations in
three fields to provide a C signal 113 for the YC separation
in three fields.
The subtracter 45 subtracts this C signal 113 from the V
signal which is the output of the 2D delay circuit 20, so as
to provide a Y signal 112 for the YC 8eparation ln three
fields.
Figure 4 is a block diagram of the third embodiment of
the YC separation in three fields circuit S0 shown in Figure
1.
The arrangement of Figure 4 is di~tinguished from that of
2 4
A
2039536
Figure 2 only in that there is an infield YC separation circuit
using infield calculation in addition to the YC separation in
three fields circuits including three different calculations in
three fields. The optimum one of the above four calcualtions
is selected and utilized. There will be described only an
interframe correlation detecting circuit in the Y~
separating in three fields circuits of Figure 4, which is
different from those of Figure 2. Parts similar to those of
Figure 2 are designated by similar reference numerals.
The output of the adder 23 is applied to the first input
of the signal selection circuit 49. The output of the adder
24 is provided to the second input of the signal selection
circuit 49. The output of the LPF 26 is applied to the third
input of the signal selection circuit 49. The output of the
2D delay circuit 20 is provided to the first inputs of the
subtracter 25, subtracter 41 and adder 27 and-further applied
to the infield YC separation circuit 46. The infield YC
separation circuit 46 is defined only by infield calculations
as in the conventional infield YC separation circuit 4 shown
in Figure 10. The output of the infield YC separation
circuit 46 is applied to the fourth input of the signal
selection circuit 49.
The output of the ABS circuit 35 is applied to the first
inputs of the maximum and minimum value selection circuits 47
and 38, respectively. The output of the ABS circuit 36 is
provided to the second inputs of the maximum and minimum
value selection circuits 47 and 38, respectively. The output
of the ABS circuit 37 is provided to the third inputs of the
maximum and minimum value selection circuits 47 and 38,
respectively. The output of the maximum value selection
2 5
2039536
circuit 47 is provided to the first input of a threshold
discriminating circuit 48. The output of the minimum value
selection circuit 38 is provided to the second input of a
threshold discriminating circuit 4B and also to the fifth
input of the signal selection circuit 49. The output of the
threshold discriminating circuit 48 is provided to the sixth
input of the signal selection circuit 49. If the maximum one
of the three interframe correlations is smaller than a first
threshold a or if the minimum one of the three interframe
correlations is larger than a second threshold ~ , the
threshold discriminating circuit 48 controls the signal
selection circuit 49 to select the output of the infield YC
separation circuit 46. On the other hand, if the maximum one
of the three interframe correlations is larger than the first
threshold a or if the minimum one of the three interframe
correlations is smaller than the second threshold ~ , the
threshold discriminating circuit 48 causes the output of the
minimum value selection circuit 38 to control the signal
selection circuit 49 such that the latter will select the
output of the adder 23 if the output of the ABS circuit 35 is
minimum; the output of the adder 24 if the output of the ABS
circuit 36 is minimum; and the output of the LPF 26 if the
output of the ABS circuit 37 is minimum. As in the
embodiment of Figure 2, this will permit the adaptive
YC separation in three fields including three calculations in
three fields. However, there must be a < ~ .
As in the embodiment of Figure 4, the embodiment of
Figure 3 can also adaptively control switching the YC
separations utilizing only the infield band limitations and
the three YC separations in three fields by using the infield
2 6
2039536
YC separating circuit 46, maximum value selection circuit 47,
threshold discriminating circuit 48 and signal selection
circuit 49.
2 7
