Component: Fast Fourier Transform (FFT) (DSP Operators)

Author	Matrix Ltd
Version	1.2
Category	DSP Operators

Fast Fourier Transform (FFT) component

Provides a way of converting a buffer full of time domain data into frequency domain data. The output of the FFT is a set of frequency bins which correspond to the frequencies present in the signal. The number of frequency bins is equal to half the input buffer size with each bin being responsible for a portion of the frequency up to 1/2 the nyquist.

Component Source Code

Please click here to download the component source project: FC_Comp_Source_DSP_FFT.fcfx

Please click here to view the component source code (Beta): FC_Comp_Source_DSP_FFT.fcfx

Detailed description

No detailed description exists yet for this component

Examples

Example takes two adjustable frequency sine wave inputs and combines them and then passes the collection of 256 samples to the FFT function which produces 128 individual frequency banks which are then plotted from left (lowest frequency) to right (highest frequency) using an XY chart component.

FFT Example

Frequency output for a set of data

Frequency banks can be worked out by using the sample frequency and buffer size. The number of frequency banks is equal to the number of samples divided by 2 minus 1. We subtract the one because bank 0 is always empty and so can be discounted.

FrequencyBanks = (SampleCount / 2) - 1

The frequency ranges present inside a bank can be worked out as follows.

FreqMin = ((SampleRate / 2) / (SampleCount / 2)) * Bank

FreqMax = ((SampleRate / 2) / (SampleCount / 2)) * (Bank + 1)

For example if we sample at 10KHz and collect 256 samples then frequency bank 10 would represent any frequencies in the range of 390.6Hz to 429.7Hz.

Here is a spreadsheet document that will calculate the frequencies present in each bank. The highlighted green fields are editable and will update the rest of the values.

File:FFT Frequency Bank Calculator.xls

Macro reference

FFT

	FFT
Performs the data conversion from time based data to frequency based data. Takes a full buffer input and converts it into real and imaginary frequency data ready for you to read the individual frequency bins.
- VOID	Return

QueueData

	QueueData
Takes a single value from the buffer input and stores it, once there are enough values queued up it returns a 1.
- BOOL	ProcessWhenFull
0 = Do Nothing when Buffer is full, 1 = Auto call FFT macro when Buffer is full
- BOOL	Return

ReadFrequencyBank

	ReadFrequencyBank
Reads the value from one of the frequency banks after the FFT conversion has been performed. Valid frequency banks range from 1 to the allocated buffer size / 2 assuming the buffers are sized to be equal to a power of 2.
- UINT	BankIdx
- UINT	Return

Property reference

	Properties
	Connect To
DSP component with output buffer to collect our data from.
	Buffer Size
Number of individual elements the buffer can store, must be a power of 2, default 256.
	Buffer Type
Sets the buffer data type.
	Filter Input Data
Applies a hamming window filter to the time domain sample data to provide better frequency detection.
	Calculations
	Sample Rate
Sample rate used to perform simple calculations like nyquist and bin size.
	Nyquist Frequency
The maximum frequency that can theoretically be detected by the FFT component assuming the sample rate is setup correctly.
	Frequency Bin Count
The number of frequency bins that the component will provide.
	Frequency Bin Size (Hz)
The frequency range covered by each frequency bin.