最近在做频谱分析，发现使用hanning窗函数+fft算法，对matlab生成信号：

0.7*sin(2*pi*50*t) + sin(2*pi*120*t)进行频谱分析，发现结果中信号幅值衰减，代码如下，如何对幅值修正，才能使fft结果中的幅值不衰减？

Fs = 9000; % Sampling frequency

Ts = 1/Fs; % Sampling period

L = 17000; % Length of signal

t = (0:L-1)*Ts; % Time vector

Signal = 0.7*sin(2*pi*50*t) + sin(2*pi*120*t);

Signal = Signal';

%%

fsU = 1/Ts;

%% frequency window

WinCoeff = 2;

HannWin = hann(L);

UPhaCurrWin = Signal .* HannWin;

%% fft

%fft

FreSpcU = fft(UPhaCurrWin);

%amplitude

AmpU = abs(FreSpcU/L);

AmpUtemp = AmpU(1:(fix(L/2)));% remove the direct current component,because the 0Hz component has not been divide into two parts.

AmpUtemp(2:end) = 2*AmpUtemp(2:end);%above 0Hz components, the Amp need been multiplicated 2, because the result of fft is symmetric spectrum there are two

AmpUtemp = AmpUtemp *WinCoeff;

%Fre

f = Fs/L *(0:(fix(L/2)-1));

%Plot

plot(f,AmpUtemp);

title('Frequency Spectrum')

xlabel('Fre/Hz')

ylabel('Amp/A')