Transformations of frequency domain information
This module is a collection of functions to work with spectrum of a signal
Before we do anything in the field of spectral modeling, we must be able to competently compute the spectrum of a signal. The spectrum is given by the Fourier transform of a signal, but in virtually all cases, the result from the DFT has to be converted into polar coordinates in order to permit the desired modifications in an appropriate way as magnitudes and phases
This module contains function to work with the result of a DFT (or FFT), the signal in the frequency domain.
This is part of dsp-kit
References
- Polar notation: http://www.dspguide.com/ch8/8.htm
- Source:
Example
const dsp = require('dsp-kit')
dsp.spectrum(dft.fft(signal))Methods
(static) bandFrequency(index, size, sampleRate) → {Number}
Calculates the center frequency of an DFT band (or bin)
Parameters:
| Name | Type | Description |
|---|---|---|
index |
Integer | The index of the FFT band. |
size |
Integer | the DFT (or FFT) buffer size |
sampleRate |
Integer | the sample rate of the original signal |
- Source:
Returns:
-
the center frequency of the DFT band
- Type
- Number
-
The middle frequency in Hz.
(static) bandWidth(size, sampleRate) → {Number}
Get band width of a result of a fourier transformation
It calculates the size of each bin of the spectrum in Hertzs.
Parameters:
| Name | Type | Description |
|---|---|---|
size |
Integer | the DFT (or FFT) buffer size |
sampleRate |
Integer | the sample rate of the original signal |
- Source:
Returns:
the frequency width of each bin
- Type
- Number
(static) polar(freqDomain, output) → {Array}
Convert a signal in frequency domain from rectangular { real, imag } to
polar form { magnitudes, phases }
It returns an object with two arrays: { magnitures: <Array>, phases: <Array> }
If not provided, the magnitudes and phases lengths will be the same as the
real and imaginary parts (you can remove calculations by providing arrays
of length = (N / 2) + 1 in real signals because the symetric properties)
Parameters:
| Name | Type | Description |
|---|---|---|
freqDomain |
Object | the frequency domain data in rectangular notation |
output |
Object | (Optional) the buffers to store the data in polar form if you want to reuse buffers for performance reasons |
- Source:
Returns:
the frequency domain data in polar notation: an object
with the form: { magnitudes: <Array>, phases: <Array> }
- Type
- Array
Example
const dsp = require('dsp-kit')
dsp.polar(dsp.fft(signal)).magnitudes(static) rectangular()
Given a spectrum in rectangular form ({ magnitudes, phases }) convert
into a spectrum in polar form ({ real, imag }).
This is the inverse operation of polar function
- Source:
- See:
-
- polar
(static) unwrap(data, output) → {Array}
Perfroms a phase-unwrapping of a phase data
Parameters:
| Name | Type | Description |
|---|---|---|
data |
Array | phase data |
output |
Array | (Optional) the output array to store the unrapped phase data (or a new array will be created if not provided) |
- Source:
Returns:
the unrapped phase data
- Type
- Array
Example
// get the spectrum of a 1024 size signal fragment
const spectrum = dsp.spectrum(dsp.fft(1024, signal))
// unwrap the phases
const unwrapped = dsp.unwrap(spectrum.phases)