| Author |
 Topic  |
|
|
clemon
7 Posts |
 Posted - 07/06/2012 : 8:49:02 PM
|
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
what i am doing atm is creating data in excel and then applying ifft to it. i am trying to see if there is a pattern so that certain fft spectrums sound more desirable - here, harsh.
i have a few questions if that's ok?
1. when i try to export [the result of ifft] as a .wav i am told that the y column does not comtain a sampling interval. can anyone explain what i need to know - esp given the above?
2. what i want to be able to do, is test the fft of sound files [that i have no created] for deviation from a certain form - a curve of the for amp = 1/freq variety. i want to be able to measure when deviation from that function is just "noise" and when e.g. there is more than one genuine peak to the data. or better, how much deviation there is from that function, when noise does not count as deviation and if there is none that returns a high [deviation[ score.
make sense?? i am new to these sorts of questions etc., so if i buy originpro might depennd on how much help i get on the forum - as it's not very googable in general.
cheers :) :) |
|
|
clemon
7 Posts |
Posted - 07/08/2012 : 4:16:09 PM
|
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
copy and pasted from what i said on the cycling 74 forums :)
-------------
so i was talking with someone, and they say fft isn't the way to go with this but: spectrograms.
where do i start? i want to devise a way of measuring the harshness of a recording. perhaps this will be certain frequencies [unpleasantness] perhaps the presence of noise. i dunno yet! but can anyone help me get started with spectrograms - i suppose i'd like to be able to export data so that i can run statistical tests on it??
-----------------
just any advice appreciated! |
 |
|
|
clemon
7 Posts |
Posted - 07/09/2012 : 07:16:04 AM
|
last question - but a better one maybe?
i was trying to run ifft on excel but i then want to manipulate the output [basically - everything above the median amp = 1, everything below = -1. turning into steps]... but this destroys the fft - completely. is there no way around that - i was thinking of using solver and ifft in excel together but that's impossible. so can i do so in originpro and anyway, is this just a completel dead end?
thanks :) |
|
|
|
easwar
USA
1964 Posts |
Posted - 07/09/2012 : 11:49:48 AM
|
Hi clemon,
So the main issue you are facing appears to be that once you do an FFT/IFFT in Origin, you are not able to export the result to a wave file, correct?
This is because currently exporting a column of data to wave file has certain requirements.
Try the following:
1> Go to the result sheet of the FFT/IFFT operation, then right-click on the column that has the "Real" output, and select copy
2> Go to a new worksheet and paste into column 1
3> Double click on the column 1 header to open Properties dialog, and in this dialog set the Plot Designation as Y, and change the Format to Numeric, and set the Data Type to Short(2), then close the dialog
4> Right-click on the column header and from the context menu, select "Set Sampling Interval". In the dialog that opens, set initial x value as 0, and put in some X increment value such as say 1e-5, then close dialog
5> Now try exporting this column as a Wave file, that should work. You can then play with the sampling interval increment number to get a desired play back rate when playing the audio file.
If you still have trouble, send your OPJ with your data and your current results to tech support, using the "Send File to Tech support" link at top right of the page, and please refer to this post in your message.
Easwar
OriginLab
|
|
|
| |
Topic |
|
|
|
some questions for a beginner [sound files]