A spreadsheet that can do spectral magnitude and phase

[gpapag]

Hi all

The concept, the push and the main mathematical formulas are all from SY. (*see references)
I put the bricks together.

4096 is the max data points that excel can perform Fourier analysis on.
The free program I use to convert wav. Files to .txt (“Watex_1”) can do max 4096. (Download Watex 1.0 Free - Wave to txt-file converter - Softpedia )

I import the .txt data on column C (C2:C4097).
Then, I select:
[Tools],
[Data Analysis],
[Fourier Analysis], [Input Range (C2:C4097), Output Range (E2:E4097)].

Then, I input in shell “W3” the number in seconds (exact) of the duration of the wave file that was converted to text.

In shell “W6” I input the number of samples that I used in the wav.to .txt converter program.

After inputing all this data, the diagrams take their shape.

You can have a look on the spreadsheet ( it is unlocked, no macros), test it by importing your wav. to txt file and tell me if you find anything wrong.

Regards
George

*references:
John Curl's Blowtorch preamplifier part II
John Curl's Blowtorch preamplifier part II
John Curl's Blowtorch preamplifier part II
John Curl's Blowtorch preamplifier part II
John Curl's Blowtorch preamplifier part II
John Curl's Blowtorch preamplifier part II
John Curl's Blowtorch preamplifier part II

[wintermute]

Hi George, I think that amount of effort at least deserves a response!!!

I assume that this was done as a learning exercise, as I know you know other (much easier) ways to get fft data from wav files

I stumbled across this post because I was looking at your thread history to find the Y out thread, as I was feeling the need to solder something (now that I have completed my crossovers), and there is a packet from Greece that needs assembly!

Tony.

[gpapag]

Hi Tony
Thanks for visiting and commenting.
Yes, you are right. There are less masochistic programs to do FFT.
But I had to pay a closer look at the suggestion of SY about the phase (in addition to amplitude) diagram of an FFT. The other FFT programs to which I have access to, don’t do this.
Alas, 4096 data points is really restricting. The remedy was to reduce the time slice. So with a a very short sample (small fraction of a second), I can go high in frequency.
The diagram 3 is very helpful, as it shows the waveform of the sampled input after the sampling. There I can easily watch for any inappropriate sampling, as in this case, the waveform will look distorted. (apart from the resulting spiky FFT).
All in all, it was a learning exercise and this, remains a learning tool for me.
Regards
George

[gpapag]

There are less masochistic programs to do FFT

He he. I can't help it, but I still remember what a tedious process was to make the Sinclair ZX Spectrum (1kB RAM, audio cassette as mass storage and BASICA software) print a primitively simple diagram!

[wintermute]

ah yes, I remember being utterly disapointed that our comodore 1526 printer (with our commodore 64) could not print graphics. It was a text only dot matrix printer.

But that didn't deter me. The printer had the ability to print user defined characters, so I wrote a program in assembler that would take an image and convert it into (I think it was 8X8) programable characters and output it on the printer. 6502 assembler is horrid!! But it actually worked. The only problem was the images were tiny

Tony.

[gpapag]

Graphics printer back then?
Were you reading science fiction that extensively?

[wintermute]

Haha I think my first science fiction was in year nine at school (1981) Robert Heinlein "Assignment in Eternity"

Even earlier than that my Dad brought home some pictures that had been printed on a mainframe printer. Very clever, it overlaid different characters to increase the density to achive from black to grey. you had to look at it from a distance of course.

Tony.

[gpapag]

So, no one can blame you. It was the environment you grew-up in