[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: echo in the freq. domain

Hi Dan,
I am trying this in the frequency domain because I have a freq domain "framework" ready: time in > hop size shift > window > dft > SOMETHING > idft > window >overlapp add > time out

I wanted to use this setup to try ANY kind of effect I can just by manipulating with the spectrum bins or whole dft frames. It was pretty easy to get a decent reverb. I also just did simple echo: saved quarter a second worth of dft frames and then added them back (attenuated) to current frames... and voila, not bad echooo choo oo o!


Danijel Domazet

On 13.5.2010 0:02, Dan Stowell wrote:
Hi -

You can do an echo (delay) effect in the frequency domain using convolution, exactly the same as you would for reverb - you would just use a different impulse response (a sparser one, typically, in the time domain).

However, the sparsity means that you can typically implement it efficiently enough in the time domain, so there's not much point doing it frequency-domain. Why do you want to do it?


Danijel Domazet wrote:
Hi auditory,
I'd like to do a simple Echo effect in the frequency domain. Does anyone remember any paper on the subject? There is a nice paper on freq domain reverberation ("Frequency Domain Artificial Reverberation using Spectral Magnitude Decay", Vickers E.), but couldn't find any on the echo....


Danijel Domazet