No profile info found for toddwelti

Re: Constructing vector for IFFT

Sounds right.  In my case I was interpolating if f domain though.  I expected to get zero padding in time.Unfortunately my figures that i inserted in one of my...

Re: Constructing vector for IFFT

Can you explain what you mean? If you are interpolating complex numbers in f domain, you can do linear (default in Matlab) or nearest neighbor or other like spline...

Re: Constructing vector for IFFT

Yes, but only if your freq bin size didn't change. I made some plots in Matlab.  Starting with a short section of sine wave:Zero padding in f domain gives the expected...

Re: Constructing vector for IFFT

Got it, thanks.  Interpolating in one domain is zero padding in the other.  I haven't thought about this stuff for a ling time!

Re: Constructing vector for IFFT

OK, so if you zero pad, then Fs is increased proportionately, and, yes your IR has more points, but higher Fs = same length.  But, if you add more points in the...

Re: Constructing vector for IFFT

I was maybe being a little loose with terms.  I really mean a DFT, Im just used to calling it  FFT.  You could calculate it in several different ways, FFT being...

Re: Constructing vector for IFFT

Yes, I ws trying to figure out how to do it without needing to decimate, but in the end it is simpler to just do it that way!

Re: Constructing vector for IFFT

I may not have been clear as to when I was talking about frequency values or indices.  In my zero padded version (yes, not required except I wanted to get more...

Re: Constructing vector for IFFT

I've done this a few times over the years (I used to do it more often), and it always seems like there ONE bin off somewhere that screws it up.  For years, I just...

Constructing vector for IFFT

I have some frequency domain data and I want to try to construct the proper vector for IFFT'ing to get an impulse response.  In the current case, I have a starting...

Use this form to contact toddwelti

Before you can contact a member of the *Related Sites:

• You must be logged in (register here)
• You must confirm you email address