Author Topic: some experiments  (Read 10605 times)

ygfperson

  • Founders
  • Posts: 601
  • Karma: +10/-1
    • Last.fm
some experiments
« on: June 30, 2005, 02:29:26 PM »
wire - the commercial

in messing with the fast fourier transform stuff, I forgot to tell it to read the other half of the file.

how half a song sounds like on the frequency domain

ober

  • Ashton Shagger
  • Ass Wipe
  • Posts: 14305
  • Karma: +73/-790
  • mini-ober is taking over
    • Windy Hill Web Solutions
some experiments
« Reply #1 on: June 30, 2005, 03:46:01 PM »
Umm.... err... interesting?

Rob

  • New improved. Now with added something...
  • Jackass In Charge
  • Posts: 5959
  • Karma: +86/-149
  • Approaching 60 from the wrong damn direction...
some experiments
« Reply #2 on: June 30, 2005, 05:58:15 PM »
Soooo. You did some FFT stuff on a file, but only read one channel, and the second file is the result? How'm I doing so far?

ygfperson

  • Founders
  • Posts: 601
  • Karma: +10/-1
    • Last.fm
some experiments
« Reply #3 on: June 30, 2005, 06:28:28 PM »
sort of... the operation goes like this:
data + data(i) -> fft = newdata + newdata(i)
newdata + newdata(i) -> fft = data + data(i)

the data array is the song basically. data(i) is exactly the same
newdata is only half as big, but newdata(i) is also different data that's half as big, making it practically the same size. i wanted to see what would happen if i replaced newdata(i) with all 0's.

i just thought it was interesting to see a different perception of what it means to miss a half of something. normally you think of missing a stereo channel, or cutting a cassette tape in half (to get two halves of the song), or reducing the bits used per sample to get something of crappier quality. with this you can hear the song in the background, but some of the information saying where some frequencies belong in time are missing.

Rob

  • New improved. Now with added something...
  • Jackass In Charge
  • Posts: 5959
  • Karma: +86/-149
  • Approaching 60 from the wrong damn direction...
some experiments
« Reply #4 on: June 30, 2005, 06:38:41 PM »
M'kay, sounds interesting (pun intended). I really need to get into DSP stuff, but as soon as I hear the words "Fourier Transform", I kind of switch off. It would be cool to understand all the time / frequency domain stuff though.

ygfperson

  • Founders
  • Posts: 601
  • Karma: +10/-1
    • Last.fm
some experiments
« Reply #5 on: June 30, 2005, 10:57:23 PM »
Quote
It would be cool to understand all the time / frequency domain stuff though.

hehe... ditto.