Why damping amplitude the amplitude shifting position of y azis

I damped the amplitude signal using by multiplied the freq with 0.5 after fft result, and returning the signal to time domain using ifft, but the returning signal that i get is shifting in Y axis as shown below. Should I multiplied y to some number to ifft result. Why it is happened? Is it effect of fft? But it does't happend if i use sinewave starting with 0 in y axis.

%DAMPLING AMPLITUDE USING REAL DATA
SMR=load ("sig_fft.txt");
SMR_data=SMR(:,2);
srateSMR=1/100;
SMR_Damp=real(ifft(fft(SMR_data)*0.5));
tSMR=(1:length(SMR_data));
figure(2)
plot(tSMR,SMR_data,'r',tSMR,SMR_Damp,'b');

Answer 1

[Star Strider]

It would help to have the file to demonstrate with it, however that may not be absolutely necessary.

The signal in the file (red curve) has an obvious constnt offset (D-C offset) value that looks to be about -1.2. Muttiplying the fft of that signal by 0.5 produces a result (blue curve) that is offset by about -0.6. The sine curve used to test it has a 0 offset, so there is no similar shift.

[nirwana]

@Star Strider

I see, so it just because amplidue is "minimize" from the bigining. May I extend the question, so if i want to get my resut (blue curve) in the same position of red curve, i just need to multiply in plot part

[Star Strider]

The easiest way to do that would be not to multiuply it, so do this:

SMR_Damp=real(ifft(fft(SMR_data)));

instead.

Then, they should both plot at approximately the same locations, unless other transformations (for example, filtering) are involved.

[nirwana]

but by write it mean that the damping amplitude is zero, what i want is damping amp stil perform but plot the signal with overlap (in the same start of y axis)

[Star Strider]

I am not certain that I understand.

If you want to decrease the signal amplitude without affecting the D-C offset, do something like this —

Fs = 10000;
Tlen = 1000;
t = linspace(0, Tlen*Fs, Tlen*Fs+1).'/Fs;
SMR_data = sin(2*pi*0.015*t) .* sin(2*pi*0.001*t+pi)-1.2;
SMR_Damp = (real(ifft(fft(SMR_data)))-mean(SMR_data))*0.5 + mean(SMR_data);

figure
plot(t, SMR_data,'r', t, SMR_Damp,'b')
legend('SMR\_data','SMR\_Damp', 'Location','best')

The procedure is to subtract the mean (D-C offset), do the multiplication, then add the mean value back.

.