CN101577851A

CN101577851A - Method for improving stereo crosstalk

Info

Publication number: CN101577851A
Application number: CNA2009100331253A
Authority: CN
Inventors: 张萌; 戴志生; 高星; 吴建辉; 汤佳健; 王声扬; 徐勐; 丁小军; 朱华成; 陆生礼
Original assignee: Southeast University
Current assignee: Southeast University
Priority date: 2009-06-12
Filing date: 2009-06-12
Publication date: 2009-11-11
Anticipated expiration: 2029-06-12
Also published as: CN101577851B

Abstract

The invention proposes a method for improving stereo separation. The method adds a phase compensation module in a stereo decoding circuit, thereby reducing the phase error between the pilot signal and the received stereo signal, and improving the stereo separation. The phase compensation module here is composed of a phase error calculation circuit with a symmetrical internal structure. This circuit calculates the compensation factor required to compensate the phase error, and then uses the compensation factor to compensate the decoded stereo signal, so that the separation of stereo reach the optimum.

Description

A Method of Improving Stereo Separation

技术领域 technical field

本发明涉及通信领域，特别是在FM(调频)接收机立体声解码技术中的一种提高立体声分离度的方法。The invention relates to the communication field, in particular to a method for improving stereo separation in FM (frequency modulation) receiver stereo decoding technology.

背景技术 Background technique

目前，FM立体声接收机被广泛使用在手机、MP3和FM广播等产品中。传统的FM无线电广播以一种标准的立体声方案进行传送。标准的立体声信号频谱图如图1所示。如图所示，立体声信号由左声道和右声道的和信号M＝L+R、左声道和右声道的差信号S＝L-R以及19KHz的导频信号组成。另外，立体声信号中也可能包含被调制在57KHz的RDS(无线电数据系统)信号或者RBDS(无线电广播数据系统)信号。At present, FM stereo receivers are widely used in products such as mobile phones, MP3 and FM broadcasting. Traditional FM radio broadcasts are transmitted in a standard stereo scheme. The spectrum diagram of a standard stereo signal is shown in Figure 1. As shown in the figure, the stereo signal is composed of the sum signal M=L+R of the left and right channels, the difference signal S=L-R of the left and right channels, and a 19KHz pilot signal. In addition, the stereo signal may also contain RDS (Radio Data System) signals or RBDS (Radio Broadcast Data System) signals modulated at 57KHz.

传统立体声解码方法首先需要恢复出19KHz的导频信号，通过这个导频信号以获得38KHz的立体声载波。然后这个载波信号被用来将立体声信号下变频至基带。最后分别得到左声道和右声道信号。The traditional stereo decoding method first needs to recover the 19KHz pilot signal, and obtain the 38KHz stereo carrier through the pilot signal. This carrier signal is then used to downconvert the stereo signal to baseband. Finally, the left and right channel signals are respectively obtained.

与上面立体声解码类似，在发送端，相应地需要一个导频信号用来产生38KHz的立体声载波信号以将基带信号调制到高频。导频信号的使用极大地方便了立体声信号解码操作的硬件实现，但是同时这也带来了一定的问题，比如说，由于发送机内在固有的相位偏移或者由于传输信道的噪声影响，在接收端所恢复出的导频信号和接收的立体声载波信号之间就存在一个相位误差。这个相位误差会降低接收机的立体声分离度。所谓立体声分离度就是指立体声音响系统中左右两个声道之间的分离度，它实际上反应了左右两个声道相互串扰的程度。如果两个声道之间串扰较大，那么声音的立体感将被减弱，这无疑会影响用户的使用体验。Similar to the stereo decoding above, at the sending end, a pilot signal is correspondingly required to generate a 38KHz stereo carrier signal to modulate the baseband signal to a high frequency. The use of the pilot signal greatly facilitates the hardware implementation of the stereo signal decoding operation, but at the same time it also brings certain problems, for example, due to the inherent phase offset of the transmitter or due to the noise of the transmission channel, the receiver There is a phase error between the pilot signal recovered by the terminal and the received stereo carrier signal. This phase error degrades the receiver's stereo separation. The so-called stereo separation refers to the separation between the left and right channels in a stereo system, which actually reflects the degree of crosstalk between the left and right channels. If the crosstalk between the two channels is large, the three-dimensional effect of the sound will be weakened, which will undoubtedly affect the user experience.

发明内容 Contents of the invention

本发明提出一种提高立体声分离度方法，该方法通过在立体声解码电路中增加一个相位补偿模块，从而降低导频信号和接收到的立体声信号之间的相位误差，提高了立体声的分离度。The invention proposes a method for improving stereo separation. The method adds a phase compensation module in a stereo decoding circuit, thereby reducing the phase error between a pilot signal and a received stereo signal, and improving the stereo separation.

本发明方法是这样实现的：一种提高立体声分离度的方法，其特征是在立体声解码电路中增加一个相位补偿模块以降低导频信号和接收到的立体声信号之间的相位误差，相位补偿模块位于低通滤波器之后，立体声混合器之前，是一个内部结构对称的相位误差计算电路，通过该电路计算出要补偿相位误差所需要的补偿因子，然后利用该补偿因子来补偿解码后的立体声信号；相位误差计算电路含乘法器、加法器、除法器、平方和的开方器组成，低通滤波器的输出先经过乘法器，再依次经过加法器，除法器，平方和的开方器之后再经过乘法器，最后输出到立体声混合器，使得立体声的分离度达到最优。The method of the present invention is achieved in that a method for improving stereo separation is characterized in that a phase compensation module is added in the stereo decoding circuit to reduce the phase error between the pilot signal and the received stereo signal, and the phase compensation module After the low-pass filter and before the stereo mixer, there is a phase error calculation circuit with a symmetrical internal structure, through which the compensation factor required to compensate the phase error is calculated, and then the compensation factor is used to compensate the decoded stereo signal ; The phase error calculation circuit consists of a multiplier, an adder, a divider, and a square root of the sum of squares. The output of the low-pass filter first passes through the multiplier, and then passes through the adder, divider, and square root of the sum of squares in turn. After passing through the multiplier, it is finally output to the stereo mixer, so that the stereo separation can be optimized.

具体步骤如下：Specific steps are as follows:

(1)立体声信号经过低通滤波器得到左声道和右声道的和信号M＝L+R；(1) Stereo signal obtains the sum signal M=L+R of left sound channel and right sound channel through low-pass filter;

(2)立体声信号与信号cos(2π×38×10³t)，sin(2π×38×10³t)相乘经过两个低通滤波器输出到相位误差计算电路，相位误差计算电路由乘法器、加法器、除法器和平方和的开方器组成，其输出即为左声道和右声道的差信号S＝L-R。将和信号M和差信号S输入到立体声混合器后经过去加重器即得到左声道和右声道。(2) Stereo signal and signal cos(2π×38×10 ³ t), sin(2π×38×10 ³ t) are multiplied and output to the phase error calculation circuit through two low-pass filters, and the phase error calculation circuit is composed of multiplication It is composed of an adder, an adder, a divider and a square root of the sum of squares, and its output is the difference signal S=LR between the left and right channels. The sum signal M and the difference signal S are input to the stereo mixer and passed through the de-emphasis device to obtain the left channel and the right channel.

(3)立体声信号与信号cos(2π×20×10³t)相乘经过一个低通滤波器输出到数字锁相环和立体声检测器，完成跟踪导频信号和立体声检测。(3) Stereo signal is multiplied by signal cos(2π×20×10 ³ t) and output to digital phase-locked loop and stereo detector through a low-pass filter to complete tracking pilot signal and stereo detection.

上面步骤(2)的实现流程如下：The implementation process of the above step (2) is as follows:

立体声信号stereo signal

y(t)＝(L+R)+Acos(2π×19×10³×t+γ)+(L-R)cos(2π×38×10³×t+η)与信号cos(2π×38×10³t)，sin(2π×38×10³t)相乘后得y(t)=(L+R)+Acos(2π×19×10 ³ ×t+γ)+(LR)cos(2π×38×10 ³ ×t+η) and signal cos(2π×38×10 ³ t), after multiplying sin(2π×38×10 ³ t), we get

y(t)·cos(2π×38×10³×t)＝(L+R)cos(2π×38×10³×t)+y(t)·cos(2π×38×10 ³ ×t)＝(L+R)cos(2π×38×10 ³ ×t)+

Acos(2π×19×10³×t+γ)cos(2π×38×10³×t)+Acos(2π×19×10 ³ ×t+γ)cos(2π×38×10 ³ ×t)+

(L-R)cos(2π×38×10³×t+η)cos(2π×38×10³×t)(LR)cos(2π×38×10 ³ ×t+η)cos(2π×38×10 ³ ×t)

y(t)·sin(2π×38×10³×t)＝(L+R)sin(2π×38×10³×t)+y(t) sin(2π×38×10 ³ ×t)＝(L+R)sin(2π×38×10 ³ ×t)+

Acos(2π×19×10³×t+γ)sin(2π×38×10³×t)+Acos(2π×19×10 ³ ×t+γ)sin(2π×38×10 ³ ×t)+

(L-R)cos(2π×38×10³×t+η)sin(2π×38×10³×t)(LR)cos(2π×38×10 ³ ×t+η)sin(2π×38×10 ³ ×t)

经过低通滤波器后得After low pass filter

LPF_out cos＝2×LPF{y(t)·cos(2π×38×10³×t)}＝(L-R)cosηLPF _{out cos} ＝2×LPF{y(t)·cos(2π×38×10 ³ ×t)}＝(LR)cosη

LPF_out sin＝2×LPF{y(t)·sin(2π×38×10³×t)}＝(L-R)sinηLPF _{out sin} ＝2×LPF{y(t)·sin(2π×38×10 ³ ×t)}＝(LR)sinη

LPF_out cos和LPF_out sin并行地经过一对乘法器与cosθ，-sinθ相乘后再经过一个加法器得LPF _{out cos} and LPF _{out sin} pass through a pair of multipliers in parallel with cosθ, -sinθ and then pass through an adder to get

S_um430＝(L-R)cosηcosθ+(L-R)sinηsinθ＝(L-R)cos(θ-η)S _um430 ＝(LR)cosηcosθ+(LR)sinηsinθ＝(LR)cos(θ-η)

同时，LPF_out cos和LPF_out sin并行地经过另一对乘法器与-cosθ，-sinθ相乘后再经过另-个加法器得At the same time, LPF _{out cos} and LPF _{out sin} pass through another pair of multipliers in parallel with -cosθ, -sinθ and then pass through another adder to get

S_ubtract440＝(L-R)cosηsinθ-(L-R)sinηcosθ＝(L-R)sin(θ-η) _Subtract440 = (LR) cos η sin θ - (LR) sin η cos θ = (LR) sin (θ - η)

S_um430与S_ubtract440经过除法器得 _Sum430 and _Subtract440 get through the divider

$Q Q = = \frac{{S S}_{ubtract extract 440440}}{{S S}_{um um 430430}} = = tan the tan ((θ θ - - η η))$

Q再经过平方和的开方器得Q is then obtained by the square extractor of the sum of squares

$\sqrt{11 + + {Q Q}^{22}} = = \sqrt{11 + + {tan the tan}^{22} ((θ θ - - η η))} = = \frac{11}{cos cos ((θ θ - - η η))}$

补偿操作通过再一个乘法器来实现，即The compensation operation is realized by another multiplier, namely

${S S}_{compensated compensated} = = {S S}_{um um 430430} \cdot &Center Dot; \sqrt{11 + + {Q Q}^{22}} = = ((L L - - R R)) cos cos ((θ θ - - η η)) \frac{11}{cos cos ((θ θ - - η η))} = = ((L L - - R R))$

这样，即得左声道和右声道的差信号S＝S_compensated＝L-R。In this way, the difference signal S=S _compensated =LR of the left and right channels is obtained.

左声道和右声道的差信号S与和信号M经过立体声混合器后再经过去加重器即得左声道信号L和右声道信号R。The difference signal S and the sum signal M of the left channel and the right channel pass through the stereo mixer and then pass through the de-emphasis device to obtain the left channel signal L and the right channel signal R.

本发明的优点及显著效果：(1)即使在很差的工作条件下，立体声分离度也能得到有效的提高；(2)本FM立体声解码器可以全数字实现，因此众多的软件方法可以利用，而且系统的参数是可配置的；(3)由于对导频恢复质量的要求降低了，所以DPLL(数字锁相环)的设计得到简化。Advantages and remarkable effects of the present invention: (1) even under very poor working conditions, stereo separation can be effectively improved; (2) this FM stereo decoder can be fully digitally realized, so numerous software methods can utilize , and the parameters of the system are configurable; (3) The design of DPLL (Digital Phase Locked Loop) is simplified because the requirements for the quality of pilot recovery are reduced.

附图说明 Description of drawings

图1是FM解调后一个典型的立体声信号的频谱分布图；Figure 1 is a spectrum distribution diagram of a typical stereo signal after FM demodulation;

图2是传统立体声解码电路的方框图；Fig. 2 is the block diagram of traditional stereo decoding circuit;

图3是根据本发明实施的立体声解码器的方框图。Figure 3 is a block diagram of a stereo decoder implemented in accordance with the present invention.

具体实施方式 Detailed ways

首先阐述本发明方法的原理：First set forth the principle of the inventive method:

解调后的FM信号包含三种不同的分量，其频谱图如图1所示。具体地说，解调后的FM信号包括：代表立体声左声道与右声道的和信号M＝L+R，它占据从0到15KHz频率范围；如FCC(联邦通信委员会)规则所要求的那样，位于19KHz频率处，有一个19KHz的导频信号；以38KHz频率即导频信号频率的两倍为中心，左声道减去右声道的信息S＝L-R被调制于抑制载波的DSB-SC(双边带调制)信号中，其频带占据从23KHz到53KHz的30KHz范围。The demodulated FM signal contains three different components, and its frequency spectrum is shown in Figure 1. Specifically, the demodulated FM signal includes: the sum signal M=L+R representing the stereo left and right channels, which occupies the frequency range from 0 to 15KHz; as required by FCC (Federal Communications Commission) regulations In that way, at 19KHz frequency, there is a 19KHz pilot signal; centered at 38KHz frequency, which is twice the frequency of the pilot signal, the information S=L-R of the left channel minus the right channel is modulated on the DSB- of the suppressed carrier In the SC (double sideband modulation) signal, its frequency band occupies a 30KHz range from 23KHz to 53KHz.

传统的立体声解码器通过导频信号来恢复出左声道和右声道的和以及差信号，最后将这两个信号混合以得到左右声道信号。图2显示了一个传统的立体声解码器200的结构框图。在该图中，输入信号为FM解调之后的MPX(立体声)信号201。首先MPX信号经过DPLL 206跟踪导频信号，恢复处的导频信号207分为两路：一路进入立体声检测模块以判断该MPX信号中是否含有立体声分量；另外一路经过倍频器得到38KHz的立体声调制载波。该载波被用来与输入端的MPX信号相乘，再经过一个低通滤波器204以产生差信号L-R信号。同时，MPX 201直接经过另外一个低通滤波器202以产生和信号L+R。最后和信号和差信号同时进入立体声混合电路218来分别产生左声道和右声道信号。A traditional stereo decoder restores the sum and difference signals of the left and right channels through the pilot signal, and finally mixes these two signals to obtain the left and right channel signals. FIG. 2 shows a block diagram of a conventional stereo decoder 200 . In this figure, the input signal is an MPX (stereo) signal 201 after FM demodulation. First, the MPX signal passes through the DPLL 206 to track the pilot signal, and the pilot signal 207 at the recovery point is divided into two paths: one path enters the stereo detection module to determine whether the MPX signal contains a stereo component; the other path passes through a frequency multiplier to obtain 38KHz stereo modulation carrier. The carrier is multiplied with the input MPX signal and passed through a low pass filter 204 to generate the difference signal L-R signal. Meanwhile, MPX 201 directly passes through another low-pass filter 202 to generate sum signal L+R. Finally, the sum signal and the difference signal enter the stereo mixing circuit 218 to generate left and right channel signals respectively.

应用本发明方法后的立体声解码器的模块框图如图3所示。The module block diagram of the stereo decoder after applying the method of the present invention is shown in FIG. 3 .

该FM立体声解码器一共包括了4个低通滤波器402、406、414和462，8个乘法器404、408、412、416、420、422、432、460、2个加法器430和440、1个除法器442、1个平方和的开方器444、1个立体声混合器450、两个去加重器470和480、1个数字锁相环DPLL 464和1个立体声检测器466，其中5个乘法器408、416、420、422、432、2个加法器430、440、1个除法器442、1个平方和的开方器444组成的模块即为本发明的相位误差计算电路，如图3中的虚线部分所示。The FM stereo decoder includes 4 low-pass filters 402, 406, 414 and 462, 8 multipliers 404, 408, 412, 416, 420, 422, 432, 460, 2 adders 430 and 440, 1 divider 442, 1 square root 444, 1 stereo mixer 450, two de-emphasis 470 and 480, 1 digital phase-locked loop DPLL 464 and 1 stereo detector 466, wherein 5 A module composed of a multiplier 408,416,420,422,432, 2 adders 430,440, 1 divider 442, 1 square root 444 is the phase error calculation circuit of the present invention, as Shown by the dotted line in Figure 3.

该立体声解码器的工作机理是，首先立体声信号201经过截止频率为15KHz的低通滤波器402，15KHz频带外的信号被滤除，只通过L+R信号。所以滤波器402的输出是L与R的和信号。同时，立体声信号201也连接到两个乘法器404和412。乘法器404的另外一个输入是cos(2π×38×10³t)，其中t代表时间，乘法器412的另外一个输入是sin(2π×38×10³t)。这两个乘法器的输出信号分别经过低通滤波器406和414以滤出不必要的带外信号(15KHz频带之外的信号)。然后这两个低通滤波器的输出进乘法器408和416分别与信号cosθ和-sinθ相乘，这里的θ代表恢复出的导频相位。这两个乘法器的输出由一个加法器430相加。另外，滤波器406和414同时进入两个对称的乘法器与-sinθ和-cosθ信号相乘，这两个积的差信号有减法器440获得。和信号与差信号然后分别进入补偿因子计算单元以补偿相位误差带来的影响。最后，补偿后的L-R和L+R信号经立体声混合器得到左声道和右声道信号，然后经去加重电路去加重后输出。另外，MPX信号201同时还会进入乘法器460与cos(2π×20×10³t)相乘，它们的乘积经过截止频率1.8KHz的低通滤波器462后分为两路：一路进入立体声检测器466以检测MPX信号是否为立体声信号；另一路进入数字锁相环464以跟踪导频信号获得相位信息cosθ和sinθ。The working mechanism of the stereo decoder is that first the stereo signal 201 passes through the low-pass filter 402 with a cut-off frequency of 15KHz, and the signals outside the 15KHz frequency band are filtered out, and only the L+R signal passes through. So the output of filter 402 is the sum signal of L and R. At the same time, the stereo signal 201 is also connected to two multipliers 404 and 412 . Another input of the multiplier 404 is cos(2π×38×10 ³ t), where t represents time, and another input of the multiplier 412 is sin(2π×38×10 ³ t). The output signals of these two multipliers respectively pass through low-pass filters 406 and 414 to filter out unnecessary out-of-band signals (signals outside the 15KHz frequency band). Then the outputs of these two low-pass filters are multiplied by multipliers 408 and 416 respectively with signals cosθ and -sinθ, where θ represents the recovered pilot phase. The outputs of these two multipliers are summed by an adder 430 . In addition, the filters 406 and 414 simultaneously enter two symmetrical multipliers to multiply the -sinθ and -cosθ signals, and the difference signal of these two products is obtained by the subtractor 440 . The sum signal and the difference signal then respectively enter the compensation factor calculation unit to compensate the influence brought by the phase error. Finally, the compensated LR and L+R signals are passed through a stereo mixer to obtain left and right channel signals, and then output after being de-emphasized by a de-emphasis circuit. In addition, the MPX signal 201 also enters the multiplier 460 to be multiplied by cos(2π×20×10 ³ t), and their product is divided into two paths after passing through the low-pass filter 462 with a cutoff frequency of 1.8KHz: one path enters the stereo detection The device 466 is used to detect whether the MPX signal is a stereo signal; the other way enters the digital phase-locked loop 464 to track the pilot signal to obtain phase information cosθ and sinθ.

下面从数学的角度来阐述根据本发明方法的工作原理。In the following, the working principle of the method according to the invention will be explained from a mathematical point of view.

立体声MPX信号仍然表示为：Stereo MPX signals are still represented as:

y(t)＝(L+R)+Acos(2π×19×10³×t+γ)+(L-R)cos(2π×38×10³×t+η)y(t)=(L+R)+Acos(2π×19×10 ³ ×t+γ)+(LR)cos(2π×38×10 ³ ×t+η)

根据上面的分析，经过滤波器402之后的信号为L+R。According to the above analysis, the signal after the filter 402 is L+R.

首先考虑与乘法器460相连的这路信号，经乘法操作之后，输出信号为：First consider the signal connected to the multiplier 460. After the multiplication operation, the output signal is:

$y the y ((t t)) \cdot \cdot cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t)) = = ((L L + + R R)) cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t)) + +$

$A A cos cos ((22 π π \times \times 1919 \times \times 1010^{33} \times \times t t + + γ γ)) cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t)) + +$

$((L L - - R R)) cos cos ((22 π π \times \times 3838 \times \times 1010^{33} \times \times t t + + η η)) cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t))$

$= = ((L L + + R R)) cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t)) + +$

$((L L - - R R)) cos cos ((22 π π \times \times 3838 \times \times 1010^{33} \times \times t t + + η η)) cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t)) + +$

$\frac{11}{22} A A cos cos ((22 π π \times \times 11 \times \times 1010^{33} \times \times t t - - γ γ)) + + \frac{11}{22} A A cos cos ((22 π π \times \times 3939 \times \times 1010^{33} \times \times t t - - γ γ))$

该信号为位于1KHz处的中频信号，然后通过一个截止频率为1.8KHz的低通滤波器462以滤除1.8KHz频带外的信号。因此滤波器462的输出为：The signal is an intermediate frequency signal located at 1KHz, and then passes through a low-pass filter 462 with a cutoff frequency of 1.8KHz to filter out signals outside the 1.8KHz frequency band. The output of filter 462 is thus:

$LPF LPF {{y the y ((t t)) \cdot \cdot cos cos ((22 π π \times \times 2020 \times \times 1010^{33} \times \times t t))}} = = \frac{11}{22} A A cos cos ((22 π π \times \times 11 \times \times 1010^{33} \times \times t t - - γ γ))$

由于混频之后信号的频率较低，因此在实际物理实现上可以考虑降低采样率，这会极大地降低功耗以及电路面积。然后，立体声检测电路检测输入信号中是否存在立体声信号。另外数字锁相环DPLL 464用来跟踪导频信号，输出导频相位信息cosθ和sinθ，这里的θ等于γ。Since the frequency of the signal after frequency mixing is low, the sampling rate can be considered to be reduced in actual physical implementation, which will greatly reduce power consumption and circuit area. Then, the stereo detection circuit detects whether there is a stereo signal in the input signal. In addition, the digital phase-locked loop DPLL 464 is used to track the pilot signal and output the pilot phase information cosθ and sinθ, where θ is equal to γ.

同时，与乘法器402和412相乘之后的信号分别经过截止频率为15KHz的低通滤波器LPF 406和414。这里的LPF拥有放大系数2。输出为：At the same time, the signals multiplied by the multipliers 402 and 412 respectively pass through low-pass filters LPF 406 and 414 with a cutoff frequency of 15KHz. The LPF here has an amplification factor of 2. The output is:

LPF_out cos＝2×LPF{y(t)·cos(2π×38×10³×t)}LPF _{out cos} ＝2×LPF{y(t)·cos(2π×38×10 ³ ×t)}

＝2×LPF{(L+R)cos(2π×38×10³×t)+＝2×LPF{(L+R)cos(2π×38×10 ³ ×t)+

(L-R)cos(2π×38×10³×t+η)cos(2π×38×10³×t)}(LR)cos(2π×38×10 ³ ×t+η)cos(2π×38×10 ³ ×t)}

＝(L-R)cosη＝(L-R)cosη

LPF_out sin＝2×LPF{y(t)·sin(2π×38×10³×t)}LPF _{out sin} ＝2×LPF{y(t) sin(2π×38×10 ³ ×t)}

＝2×LPF{(L+R)sin(2π×38×10³×t)+＝2×LPF{(L+R)sin(2π×38×10 ³ ×t)+

(L-R)cos(2π×38×10³×t+η)sin(2π×38×10³×t)}(LR)cos(2π×38×10 ³ ×t+η)sin(2π×38×10 ³ ×t)}

＝(L-R)sinη＝(L-R)sinη

经过低通滤波之后的信号LPF_outcos和LPF_outsin将会和cosθ或者sinθ信号以不同的组合分别相乘，这种组合如图3所示，然后这四个乘积之间分别相加或者相减，所得到的和信号和差信号分别为：The low-pass filtered signals LPF _outcos and LPF _outsin will be multiplied with cosθ or sinθ signals in different combinations. This combination is shown in Figure 3, and then the four products are added or subtracted respectively. The resulting sum and difference signals are:

S_um430＝(L-R)cosηcosθ+(L-R)sinηsinθS _um430 ＝(LR)cosηcosθ+(LR)sinηsinθ

＝(L-R)cos(θ-η)＝(L-R)cos(θ-η)

S_ubtract440＝(L-R)cosηsinθ-(L-R)sinηcosθ _Subtract440 ＝(LR)cosηsinθ-(LR)sinηcosθ

＝(L-R)sin(θ-η)＝(L-R)sin(θ-η)

这两个信号相除，得到商：These two signals are divided to get the quotient:

从Q可以很容易确定函数cos(θ-η)。这可以利用泰勒级数展开的方法来实现。为了补偿(L-R)cos(θ-η)，只需要得到cos(θ-η)的倒数即可。由公式：The function cos(θ-η) can be easily determined from Q. This can be achieved using the method of Taylor series expansion. In order to compensate for (L-R)cos(θ-η), it is only necessary to obtain the reciprocal of cos(θ-η). By the formula:

$\frac{11}{cos cos ((θ θ - - η η))} = = \sqrt{11 + + {tan the tan}^{22} ((θ θ - - η η))} = = \sqrt{11 + + {Q Q}^{22}}$

最后，补偿操作通过一个乘法器432来实现：Finally, the compensation operation is implemented by a multiplier 432:

${S S}_{compensated compensated} = = ((L L - - R R)) cos cos ((θ θ - - η η)) \frac{11}{cos cos ((θ θ - - η η))} = = ((L L - - R R))$

因此，本发明提出的方法能有效地补偿由于导频相位和立体声载波之间的差异而导致的立体声分离度下降。Therefore, the method proposed by the present invention can effectively compensate for the decrease in stereo separation caused by the difference between the pilot phase and the stereo carrier.

Claims

1, a kind of method that improves stereo crosstalk, it is characterized in that in the stereo decoding circuit, increasing a phase compensation block with the phase error between the stereophonic signal that reduces pilot signal and receive, phase compensation block is positioned at after the low pass filter, before the stereo mix device, it is the phase error computation circuit of an internal structure symmetry, calculate by this circuit and to want the needed compensating factor of compensation of phase error, utilize this compensating factor to compensate decoded stereophonic signal then; The phase error computation circuit contains the square root extractor of multiplier, adder, divider, quadratic sum, the output of low pass filter through multiplier, is passed through adder, divider earlier more successively, pass through multiplier again after the square root extractor of quadratic sum, output to the stereo mix device at last.

2, raising stereo crosstalk method according to claim 1, the concrete steps of its feature are as follows:

(1) stereophonic signal obtain L channel and R channel through low pass filter with signal M=L+R;

(2) stereophonic signal and signal cos (2 π * 38 * 10 ³T), sin (2 π * 38 * 10 ³T) multiply each other and output to the phase error computation circuit through two low pass filters, its output is the difference signal S=L-R of L channel and R channel.Promptly obtain L channel and R channel through deac after will being input to the stereo mix device with signal M and difference signal S.

(3) stereophonic signal and signal cos (2 π * 20 * 10 ³T) multiply each other and output to digital phase-locked loop and stereo detector, finish TPS Tracking Pilot Signal and stereo detection through a low pass filter.

3,, it is characterized in that the flow process of step (2) is as follows according to the method for the described raising stereo crosstalk of claim 2:

Stereophonic signal

Y (t)=(L+R)+Acos (2 π * 19 * 10 ³* t+ γ)+(L-R) cos (2 π * 38 * 10 ³* t+ η) with signal cos (2 π * 38 * 10 ³T), sin (2 π * 38 * 10 ³T) after multiplying each other

y(t)·cos(2π×38×10 ³×t)＝(L+R)cos(2π×38×10 ³×t)+

Acos(2π×19×10 ³×t+γ)cos(2π×38×10 ³×t)+

(L-R)cos(2π×38×10 ³×t+η)cos(2π×38×10 ³×t)

y(t)·sin(2π×38×10 ³×t)＝(L+R)sin(2π×38×10 ³×t)+

Acos(2π×19×10 ³×t+γ)sin(2π×38×10 ³×t)+

(L-R) cos (2 π * 38 * 10 ³* t+ η) sin (2 π * 38 * 10 ³* t) through getting behind the low pass filter

LPF _out?cos＝2×LPF{y(t)·cos(2π×38×10 ³×t)}＝(L-R)cosη

LPF _out?sin＝2×LPF{y(t)·sin(2π×38×10 ³×t)}＝(L-R)sinη

LPF _{Out cos}And LPF _{Out sin}Concurrently through a pair of multiplier respectively with signal cos θ and-get through an adder again after sin θ multiplies each other

S _um430＝(L-R)cosηcosθ+(L-R)sinηsinθ＝(L-R)cos(θ-η)

Simultaneously, LPF _OutcosAnd LPF _OutsinConcurrently through another to multiplier respectively with signal-cos θ and-get through another adder again after sin θ multiplies each other

S _Ubtract440=(L-R) sin of the sin η cos θ of cos η sin θ-(L-R)=(the L-R) (S of θ-η) _Um430With S _Ubtract440Get through divider

Q = \frac{S_{ubtract 440}}{S_{um 430}} = \tan (θ - η)

Q gets through the square root extractor of quadratic sum again

\sqrt{1 + Q^{2}} = \sqrt{1 + \tan^{2} (θ - η)} = \frac{1}{\cos (θ - η)}

Compensating operation realizes by another multiplier, promptly

S_{compensated} = S_{um 430} \cdot \sqrt{1 + Q^{2}} = (L - R) \cos (θ - η) \frac{1}{\cos (θ - η)} = (L - R)

Like this, promptly get the difference signal S=S of L channel and R channel _Compensated=L-R.

The difference signal S of L channel and R channel with and signal M through promptly getting left channel signals L and right-channel signals R through deac again behind the stereo mix device.