Television – Image signal processing circuitry specific to television – Noise or undesired signal reduction
Reexamination Certificate
2000-10-26
2004-07-20
Lee, Michael H. (Department: 2614)
Television
Image signal processing circuitry specific to television
Noise or undesired signal reduction
C348S726000, C375S235000
Reexamination Certificate
active
06765623
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to a method of signal correction, and to an apparatus for carrying out such correction. More specifically, the invention relates to the correction of nominal in-phase and quadrature received signals, in order to avoid errors that can arise when those received signals are other than exactly orthogonal. In particular, the invention is described with reference to a digital terrestrial television demodulator.
However, the same correction technique could be used in other systems and demodulators for any nominal in-phase and quadrature received signals, provided that, at the transmitter, the information contents on the in-phase and quadrature signals are not correlated, and provided that the signals are pseudo-random.
BACKGROUND OF THE INVENTION
As is well known, the European DVB-T (Digital Video Broadcasting—Terrestrial) standard for digital terrestrial television (DTT) uses Coded Orthogonal Frequency Division Multiplexing (COFDM) of transmitted signals. A DVB-T receiver therefore typically includes an analog tuner for receiving the signals and for down-converting to an intermediate frequency, an analog-digital converter, a digital demodulator, and an MPEG coder for producing output video signals. It will be realized that, although these different stages have been described, there can be overlap between the stages, and some or all functions can advantageously be combined onto a single chip.
Due to signal variations introduced by demodulation of the In-phase (I) and Quadrature (Q) signals in a monolithic (single-chip) integrated circuit (IC), used to mix a complex RF signal down to baseband, the I and Q signals can suffer from the following degradations:
(a) the I and Q signals are not exactly orthogonal (i.e., wherein they would have a 90° phase difference) but, rather, they have a phase difference which is 90°+&PHgr;, where &PHgr; is the phase imbalance in degrees; and
(b) the I and Q signals have different gains, and this difference can be characterized by the ratio R
G
of the gain of Q to the gain of I (R
G
=gain of Q/gain of I).
Typically &PHgr; is on the order of one degree and R
G
is on the order of 0.1 dB.
The following mathematical treatment assumes that the gain imbalance is corrected “up front”, and therefore that the system does not present a gain imbalance.
Recovery of an amplitude modulated signal may be achieved by multiplying the received signal by cos(&ohgr;t) and sin(&ohgr;t), and then low-pass filtering the resulting signals. In particular, if the transmitted signal s(t) is of the form:
s
(
t
)=
a
·cos(&ohgr;
t
)−
b
·sin(&ohgr;
t
), Equation 1
where a and b are the signals of interest that are to be recovered, and cos(&ohgr;t) and sin(&ohgr;t) are the carrier signals.
As is will be appreciated, s(t) can be demodulated to retrieve a and b by multiplying s(t) by cos(&ohgr;t) and sin(&ohgr;t), respectively, and then “low-pass filtering” to remove the double frequency term. For example, function a may be retrieved by multiplying s(t) by cos(&ohgr;t):
r
(
t
)=
s
(
t
)·cos(&ohgr;
t
) Equation 2
Expanding s(t) using Equation 1 yields:
r
(
t
)=
a
·cos
2
(&ohgr;
t
)−
b
·sin(&ohgr;
t
)·cos(&ohgr;
t
)
Using trigonometric identities, this becomes:
r
(
t
)
=
a
2
+
a
2
·
cos
(
2
ω
t
)
-
b
2
·
sin
(
2
ω
t
)
After low-pass filtering to remove the double frequency term, we obtain:
r
(
t
)
=
a
2
,
thus recovering function a.
Function b can be recovered in a similar manner by using sin(&ohgr;t) in place of cos(&ohgr;t) in Equation 2.
If the demodulating signals, cos(&ohgr;t) and sin(&ohgr;t), exhibit a phase imbalance, i.e., they each suffer a phase offset &PHgr;, then they can be written as cos(&ohgr;t+&PHgr;) and sin(&ohgr;t+&PHgr;), respectively.
In the presence of a phase imbalance &PHgr;, Equation 2 then becomes:
r
I
(
t
)=
s
(
t
)·cos(&ohgr;
t
+&PHgr;)
where r
I
(t) represents the in-phase component.
Expanding this equation using cos(A+B)=cos Acos B-sin Asin B, yields:
r
I
(
t
)=(
a
·cos(&ohgr;
t
)−
b
·sin(&ohgr;
t
))(cos(&ohgr;
t
)·cos&PHgr;−sin (&ohgr;
t
)·sin&PHgr;)
then
r
I
(
t
)=
a
·cos
2
(&ohgr;
t
)·cos&PHgr;−
a
·cos(&ohgr;
t
)·sin (&ohgr;
t
)·sin&PHgr;−
b
·sin(&ohgr;
t
)·cos(&ohgr;
t
)·cos&PHgr;+
b
·sin
2
(&ohgr;
t
)·sin&PHgr;.
Using the following trigonometric identities:
cos
2
x
=
1
2
·
(
cos
(
2
x
)
+
1
)
;
sin
2
x
=
1
2
·
(
1
-
cos
(
2
x
)
)
;
sin
(
x
)
·
cos
(
x
)
=
1
2
·
sin
(
2
x
)
;
we obtain:
r
I
(
t
)
=
a
2
·
(
cos
(
2
ω
t
)
+
1
)
·
cos
Φ
-
a
2
·
sin
(
2
ω
t
)
·
sin
Φ
-
b
·
sin
(
2
ω
t
)
·
cos
Φ
+
b
2
·
(
1
-
cos
(
2
ω
t
)
)
·
sin
Φ
;
Low-pass filtering to remove the 2&ohgr;t terms (double-frequency) finally yields the In-phase component:
I
(
t
)
=
a
2
·
cos
Φ
+
b
2
·
sin
Φ
Equation
3
When recovering function b, we have:
r
Q
(
t
)=
s
(
t
)·sin(&ohgr;
t
+&PHgr;)
where r
Q
(t) represents the Quadrature component.
Expanding this equation using sin(A+B)=sin Acos B+sin Bcos A, gives:
r
Q
(
t
)=(
a
·cos(&ohgr;
t
)−
b
·sin(&ohgr;
t
))·
(sin(&ohgr;
t
)·cos&PHgr;+sin &PHgr;·cos(&ohgr;
t
)
r
Q
(
t
)=
a
·cos(&ohgr;
t
)·sin(&ohgr;
t
)·cos&PHgr;
+
a
·cos
2
(&ohgr;
t
)·sin&PHgr;−
b
·sin
2
(&ohgr;
t
)·cos &PHgr;−
b
·sin(&ohgr;
t
)·cos(&ohgr;
t
)·sin&PHgr;;
r
Q
(
t
)
=
a
2
·
sin
(
2
ω
t
)
·
cos
Φ
+
a
2
·
(
1
+
cos
(
2
ω
t
)
)
·
sin
Φ
-
b
2
·
(
1
-
cos
(
2
ω
t
)
)
·
cos
Φ
-
b
2
·
sin
(
2
ω
t
)
·
sin
Φ
.
Low-pass filtering to remove the 2&ohgr;t terms (double-frequency) yields the quadrature component:
Q
(
t
)
=
a
2
·
sin
Φ
-
b
2
·
cos
Φ
Equation
4
Equations 3 and 4 give the signals that will be retrieved in the I and Q outputs of a tuner with phase imbalance. If an estimate of &PHgr; can be obtained, then Equations 3 and 4 form a system of two equations with two unknowns, that can readily be solved, providing perfect correction for the phase imbalance.
SUMMARY OF THE INVENTION
If an analog tuner is used to demodulate the signal into separate in-phase (I) and quadrature (Q) signals, then, typically, the generated I and Q signals are not always reliably separated by exactly 90° of phase, as they should be. When those signals are then provided as input to a demodulator, any phase imbalance will cause additional noise, due to cross-talk between the I and Q signals. This invention relates to a technique for correcting for any such imbalances.
In particular, the invention relies on an understanding that, if the phase difference between the received nominal in-phase and quadrature signals is anything other than exactly 90°, there will be a degree of correlation between the received signals. Moreover, the degree of correlation can be used as a measure of the amount of phase imbalance, allowing an appropriate correction to be applied so that the signals used for further processing are exactly orthogonal.
According to the invention, there are therefore provided a signal processor, and a receiver, which apply a correction to a received signal, based on a phase imbalance estimated from
Conexant Systems Inc.
Lee Michael H.
Natnael Paulos
The Eclipse Group
LandOfFree
Method and apparatus for correcting phase imbalance in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for correcting phase imbalance in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for correcting phase imbalance in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3224530