Add options to set the fft size and to normalize the ifft output; calculate the fft by default; and improve the documentation and the tests #3

AngelEzquerra · 2023-11-06T22:16:28Z

This PR does a few small things:

It improves the fft function documentation
It adds a normalize argument to the fft and dft functions. Normalization is disabled by default but if enabled it will normalize the ifft otuput (i.e. it only applies if inverse is true)
It makes inverse=false the default (i.e. the fft function calculates the fft by default)
It adds an n argument to make it possible to choose the FFT size (it still uses the input length by default).
It improves the tests (which could fail if the folds were negative)

AngelEzquerra · 2023-11-07T20:04:13Z

I wonder if there is a more optimum way to implement the "Add n argument to select the FFT size" commit. That is, to only create a copy of the input when needed.

arnetheduck · 2023-12-11T17:17:03Z

src/fftr.nim

+    for i in 0..<result.len:
+      result[i] = result[i] / complexInputLen
+
+func fft*(input: openArray[Complex64], inverse: bool = false, normalize: bool = false, n = 0): seq[Complex64] =


regarding normalization, there are two kinds - 1/n and 1/sqrt(n) depending on what the numbers will be used for - see https://dsp.stackexchange.com/a/63006 and https://docs.scipy.org/doc/scipy/reference/generated/scipy.fft.fft.html#scipy.fft.fft

regarding n, this feels like outside of the domain of the fft function itself - it's a setLen away for anyone that wants to do fft this particular way (though they probably don't)

I knew that scipy's fft had 3 normalization options. If you think that makes sense I can support them all by passing a norm argument that would be an enum.

Regarding n, I think it would be a good idea to support it as well, following scpy's API as much as possible.

I've added norm arguments matching SCIPY's options plus a "disabled" option to completely disable the normalization. I've made the default backwards to mach SCIPY.

arnetheduck · 2023-12-11T17:19:13Z

src/fftr.nim

+    input.toSeq & newSeq[Complex64](size - input.len)
+  elif size < input.len:
+    # Crop the input as much as needed
+    input[0..<size]


if n was pursued, this would introduce an unnecessary copy (that could be avoided with toOpenArray)

You are right of course. I even mentioned in my first comment on this PR :)
I will look into it.

I've improved the code to avoid unnecessary copies.

arnetheduck · 2023-12-11T17:19:33Z

src/fftr.nim

+
+  let input = if size > input.len:
+    # Extend the input with zeros as much as needed
+    input.toSeq & newSeq[Complex64](size - input.len)


if n was pursued, setLen here would involve fewer copies and zeroings

I've improved the code to avoid unnecessary copies.

arnetheduck · 2023-12-11T17:20:32Z

src/fftr.nim

+    # Crop the input as much as needed
+    input[0..<size]
+  else:
+    input.toSeq


and another copy here - adding these copies up would significantly slow down the implementation - you should see this if you run the benchmark before and after

FWIW, I ran the benchmarks and I did not notice any significant differences, but I agree that unnecessary copies should be avoided.

I've improved the code to avoid unnecessary copies.

…egative

The argument name has been chosen to emulate the numpy.fft API.

…ocedure

…ler than the input.

…to match the scpy normalization feature API). The default normalization is set to `backward` to match the scpy behavior. This commit also adds tests for the different normalization modes.

This makes the fft and ifft functions use the same argument order, as well as making this library match scpy's fft argument order.

arnetheduck reviewed Dec 11, 2023

View reviewed changes

AngelEzquerra force-pushed the normalize branch 2 times, most recently from 60a90e4 to c60fee6 Compare December 17, 2023 16:33

AngelEzquerra added 10 commits January 13, 2024 10:42

Add documentation to the main fft function

b0ed5f9

add normalize argument to the fft function

eb08204

add abs to the test checks to avoid false passes when the folds are n…

f2c7039

…egative

Calculate the fft by default (if inverse is not provided)

3b8cbfd

Add n argument to select the FFT size

8565fdc

The argument name has been chosen to emulate the numpy.fft API.

Reduce the amount of copies when using the n argument of the fft pr…

170a972

…ocedure

Add separate tests for FFT with specific sizes larger, equal and smal…

bc0628d

…ler than the input.

Replace the fft normalization parameter with an enum norm parameter (…

898f874

…to match the scpy normalization feature API). The default normalization is set to `backward` to match the scpy behavior. This commit also adds tests for the different normalization modes.

Swap the positions of the norm and n arguments of the fft function

7fe592d

This makes the fft and ifft functions use the same argument order, as well as making this library match scpy's fft argument order.

Add examples of the new arguments to the README

be5daef

AngelEzquerra force-pushed the normalize branch from 8b6e4f0 to be5daef Compare January 13, 2024 09:56

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add options to set the fft size and to normalize the ifft output; calculate the fft by default; and improve the documentation and the tests #3

Add options to set the fft size and to normalize the ifft output; calculate the fft by default; and improve the documentation and the tests #3

AngelEzquerra commented Nov 6, 2023 •

edited

Loading

AngelEzquerra commented Nov 7, 2023 •

edited

Loading

arnetheduck Dec 11, 2023

AngelEzquerra Dec 16, 2023

AngelEzquerra Dec 17, 2023

arnetheduck Dec 11, 2023

AngelEzquerra Dec 17, 2023

AngelEzquerra Dec 17, 2023

arnetheduck Dec 11, 2023

AngelEzquerra Dec 17, 2023

arnetheduck Dec 11, 2023

AngelEzquerra Dec 17, 2023

AngelEzquerra Dec 17, 2023

Add options to set the fft size and to normalize the ifft output; calculate the fft by default; and improve the documentation and the tests #3

Are you sure you want to change the base?

Add options to set the fft size and to normalize the ifft output; calculate the fft by default; and improve the documentation and the tests #3

Conversation

AngelEzquerra commented Nov 6, 2023 • edited Loading

AngelEzquerra commented Nov 7, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

AngelEzquerra commented Nov 6, 2023 •

edited

Loading

AngelEzquerra commented Nov 7, 2023 •

edited

Loading