Using FFT for data compression step is this a lossy method?

 

Using FFT for data compression step is this a lossy method?

Yes, you aren't going to save the entire spectrum, are you? From -∞ to +∞...

 

What type of signal are you trying to compress?

 

Ah, yes, I was assuming a non-trivial signal with many harmonics... if you are recording simple sinusoids, within a restricted range, you shouldn't have any loss.

As crutschow has alluded to, it depends on the types of waveforms you are trying to record....

 

I am planning to use this a pre processing step for satellite data transmission

 

I am planning to use this a pre processing step for satellite data transmission

I assume this is pre-processing of the analog data. What is the nature of that data?

 

Yes ,analog data,after getting the coefficients i will give it to the binary coder.

 

To answer you original question, yes the Fourier transform can be a lossy method of compression, as determined by the number of coefficients. The more coefficients, the more the recovered waveform will match the original. So you trade off accuracy for amount of data transmitted. For example, here's a study that looks at that issue.

 

I have thought of using fft as a pre processing method of the rice codes.

 

I have thought of using fft as a pre processing method of the rice codes.

I'm not that knowledgeable about various coding/compression schemes so can't advise you on that.

 

Sir,do you have any acquaintances to solve my doubts in compression.Please help me and thank you for your reply.

 

Sorry, I know of no one with that expertise.

 

ok..Thank you so much for your reply.