Fast Fourier Transform (FFT)

This node performs a forward Fast Fourier Transformation(FFT) on each row of the input table. Generally speaking, it extracts the frequencies of an input signal. The number of columns of the table must be a power of 2, i.e. 2, 4, 8, 16, 32, 64, ... To achieve this you might want to use the Window Slider Node before this node.
There are #(input columns)/2 + 1 output columns, so called frequency bins, e.g. if we have 256 input columns we will get 129 output columns. In order for these frequency bins to be labeled correctly, you need to configure the node with the correct sample rate.
All Frequencies are meant to be in Hz.
The actual output of the FFT is a row of complex numbers, in order to make subsequent processing simpler, this node outputs the absolute value of each of these complex numbers.

This node is part of the DSP Extensions provided by AI.Associates GmbH, Germany. For more information please refer to https://ai.associates/knime or write an email to knime@ai.associates.

Options

Normalization

This option controls the normalization factor of the FFT, in most cases "standard" is a good choice.

Sample Rate

This is important for the accurate labeling of the output columns. The input should be in Hz.

Views

This node has no views

Workflows

• 20220125 Pikairos remove zero value for head and tailKNIME Forum
• AI SignalKNIME Forum
• FFT_DeployKNIME Hub
• forum_19416KNIME Hub
• Fourier Transform for Singal ClassificationKNIME Hub
• Show all 6 workflows

Links

• No links available

Developers