MultitaperTfr

Namespace: RTPROCESSINGLIB  ·  Library: DSP Library

Python equivalent

mne.time_frequency.tfr_multitaper in MNE-Python.

#include <dsp/multitaper_tfr.h>

class UTILSLIB::MultitaperTfr

Sliding-window multitaper time-frequency representation.

Slides a fixed-length analysis window across the data and computes a multitaper PSD at each position, yielding a time-frequency power map per channel.

src/examples/ex_dsp_spectral/main.cpp

const MultitaperTfrResult tfr = MultitaperTfr::compute(signal, kSFreq, 128, 32); // 128-sample windows, 32-sample hop

---

Static Methods

compute(matData, sfreq, windowSize, stepSize, halfBandwidth, nTapers)

Compute sliding-window multitaper TFR for every channel of a data matrix.

Parameters:

• matData : const Eigen::MatrixXd & Data matrix (n_channels × n_times).

• sfreq : double Sampling frequency in Hz.

• windowSize : int Analysis window length in samples (default 256).

• stepSize : int Step size in samples; -1 → windowSize / 2.

• halfBandwidth : double Half-bandwidth parameter (NW); default 4.0.

• nTapers : int Number of DPSS tapers; -1 → floor(2*halfBandwidth - 1).

Returns:

• MultitaperTfrResult — MultitaperTfrResult with tfrData, vecFreqs, vecTimes.

---

computeArray(matData, sfreq, vecFreqs, nCycles, timeBandwidth, zeroMean)

Compute multitaper power at every sample with frequency-specific DPSS wavelets, as mne.time_frequency.tfr_array_multitaper(..., output="power") does for one epoch.

Parameters:

• matData : const Eigen::MatrixXd & Data matrix (n_channels × n_times).

• sfreq : double Sampling frequency in Hz.

• vecFreqs : const Eigen::RowVectorXd & Frequencies of interest in Hz.

• nCycles : double Number of cycles in each wavelet (default 7).

• timeBandwidth : double Time-bandwidth product; floor(timeBandwidth - 1) tapers (default 4).

• zeroMean : bool Make each wavelet zero mean (default true).

Returns:

• MultitaperTfrResult — MultitaperTfrResult: per channel n_freqs × n_times, vecTimes = sample times.

---

Example

• Full source: src/examples/ex_dsp_spectral/main.cpp
• Build target: ex_dsp_spectral (runs under CTest)

Used in examples

ex_dsp_spectral

Spectral and filtering tools of the dsp library on signals with known content.

src/examples/ex_dsp_spectral/main.cpp, line 145

const MultitaperTfrResult tfr = MultitaperTfr::compute(signal, kSFreq, 128, 32); // 128-sample windows, 32-sample hop
const RowVectorXd tfrMean = tfr.tfrData[0].rowwise().mean().transpose();
ok &= expect(std::fabs(frequencyOfMax(tfrMean, tfr.vecFreqs) - 20.0) < 2.0 && tfr.vecTimes.size() == tfr.tfrData[0].cols(),
             QString("multitaper TFR: %1 windows, strongest band near 20 Hz").arg(tfr.vecTimes.size()));

Authors of this file

• Christoph Dinh <christoph.dinh@mne-cpp.org>