home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!news.claremont.edu!nntp-server.caltech.edu!andrey
- From: andrey@cco.caltech.edu (Andre T. Yew)
- Newsgroups: rec.audio
- Subject: Re: CD Sound Quality
- Date: 30 Dec 1992 02:22:34 GMT
- Organization: California Institute of Technology, Pasadena
- Lines: 57
- Message-ID: <1hr15aINNib7@gap.caltech.edu>
- References: <1992Dec22.090725.11365@leland.Stanford.EDU> <7490273@hpfcso.FC.HP.COM> <vanz.02bd@tragula.equinox.gen.nz> <lk1cpdINNrrh@exodus.Eng.Sun.COM>
- NNTP-Posting-Host: punisher.caltech.edu
-
- chased@rbbb.Eng.Sun.COM (David Chase) writes:
-
- >In article <vanz.02bd@tragula.equinox.gen.nz> vanz@tragula.equinox.gen.nz (Martin Nieuwelaar) writes:
-
- >>FT, apply the box filter, and IFT. Providing you have the
- >>equipment, this seems to me a better way than using (say) an
- >>analog filter that will not have as much slope, and may introduce
- >>all sorts of phase distortions. (Please tell me if I'm wrong).
-
- Well, you're not right. :) I don't know how many people had the
- stomach/patience to sift through my humongous article where I
- tried to prove the sampling theorem, but perhaps I was being
- misleading when I started transforming around like it was necessary.
- It is always the case that you have to analog filter. The Fourier
- transform (FT) is just a way of looking at a signal -- there is
- nothing inherently digital about it. When you lowpass filter,
- you are applying a box filter. You're also interpolating the
- sample points with a sinc (sin(pi*x)/(pi*x), which is 1 at x = 0).
- They are two different sides of the same coin, so to speak. I don't
- know what issues I've addressed, so speak up if I didn't seem
- to answer your question.
-
- >You may be wrong, but it may be my fault.
-
- No, it's mine! :)
-
- >2. In general, the Fourier techniques are more often used as an
- > analysis tool, rather than an actual filter implementation.
- > Naively (or abstractly) speaking, the FT assumes a signal that
- > has infinite duration.
-
- In fact, it assumes nothing like that -- only that its integral
- is bounded. Try FTing a sine wave and see how easy that is. The
- el-cheapo trick I know of doing this is to inverse-FT the delta-
- function and say, "Hey, doesn't that look like ..." I'm sure
- there must be a way of doing the forwards integral.
-
- For those who are curious,
-
- infinity
- /
- |
- FT( f(t) ) = | f(t) exp(- i w t) dt
- |
- /
- - infinity
-
- Where i = sqrt( -1 ). Note that the result of the
- integral is a function of w, frequency. The inverse is similar
- except you have dw and exp(i w t) instead. There are also
- some constants in front and they're inverses of each other in
- the IFT and FT cases, but they're pretty safe to ignore.
-
- --Andre
-
- --
- Andre Yew andrey@cco.caltech.edu (131.215.139.2)
-