Kyma Forum
  Tips & Techniques
  Imaginary vs real fft

Post New Topic  Post A Reply
profile | register | preferences | faq | search

next newest topic | next oldest topic
Author Topic:   Imaginary vs real fft
KX
Member
posted 05 September 2005 10:08         Edit/Delete Message   Reply w/Quote
What is the "imaginary" part of a fft?
I've seen this term many times and I don't have any clue about whta it is...

IP: Logged

pete
Member
posted 05 September 2005 13:21         Edit/Delete Message   Reply w/Quote
If you think of an audio signal as a bunch of slowly changing sine waves all mixed together. Then we could think of an FFT as a way of showing the levels of each of those sine waves in a Chunk of timecalled a frame. In theory we alt to be able to do the reverse and get back the original signal if we had only those levels to work with. But we cannot because we are missing a vital piece of information. The levels tell us how loud the sine waves were but not what phase the were in.

So if we had instead keeped two sets of levels (one that represented the level of sine and one that represented the level of cos) then we could reconstruct the original signal. One set of levels is called the real and the other set is called the imaginary.

But phase is a relative thing, so relative to what, I hear you say. Normaly it's the begining of the frame that the FFT is analising. If we have a sin wave and a cos wave (a cos wave is a sine wave 90 degs out of phase ) of the same frequency we can make a sine/cos wave of the same frequency of any phase and level we choose , just by adjusting those two signals levels and mixing the results. Don't forget that the levels could be positive or negative . Yes a sine wave (0 degs) and a cos wave (90Deg) mixed together make a single wave 45degs out of phase but with exactly the same shape. BTW there is no other wavefrom that has this feature as any other waveform would produce something entirely different.

So in an FFT, if we analised just a single sine wave of level one that was at its peek at the time the frame started , our FFT would give a value of 1 in the cos represening levels (real), and a value of zero in the sine representing levels (imaginary). If the wave form was passing through zero at the time the frame started the FFT would give a value of zero in the cos represening levels (real) and a value of one in the sine representing levels (imaginary). If it started the frame at negative peek then the cos value would be minus one and the sine value would be zero.

If you start playing with FFTs and inverse FFTs ( the ones we stick on the output to get your original signal back) you find that they are almost identical. So you find that the an FFTs input also has two space for two signals in the same way as the output, but in most cases the second input signal is silent. If we are analising real input signals, then it gets feed into the FFTs real input, In this case the other input is purely imaginary.

I hope this makes some form of sence

Pete

[This message has been edited by pete (edited 05 September 2005).]

IP: Logged

KX
Member
posted 05 September 2005 18:48         Edit/Delete Message   Reply w/Quote
Thanks! That make sense.

IP: Logged

All times are CT (US)

next newest topic | next oldest topic

Administrative Options: Close Topic | Archive/Move | Delete Topic
Post New Topic  Post A Reply

Contact Us | Symbolic Sound Home

This forum is provided solely for the support and edification of the customers of Symbolic Sound Corporation.


Ultimate Bulletin Board 5.45c