Skip to contents

Spectrogram

Source: R/spectrogram-wrapper.R
Spectrogram.Rd

A Spectrogram is a time-frequency representation of sound: a matrix of power values indexed by time (columns) and frequency (rows). Created from a Sound via short-time Fourier transform (STFT).

Value

A Spectrogram object.

Query methods

  • get_start_time(), get_end_time() - time range (s)

  • get_time_step() - time between frames (s)

  • get_number_of_time_bins() - number of time frames

  • get_lowest_frequency(), get_highest_frequency() - frequency range (Hz)

  • get_frequency_step() - frequency resolution (Hz)

  • get_number_of_frequency_bins() - number of frequency bins

Index mapping

  • get_time_from_frame(frame) - time for a frame index

  • get_frame_from_time(time) - frame index for a time

  • get_frequency_from_bin(bin) - frequency for a bin index

  • get_bin_from_frequency(freq) - bin index for a frequency

Power queries

  • get_power_at(time, frequency) - power at a time-frequency point

  • get_power_at_points(times, frequencies) - power at a vector of points (batch)

  • get_frame(time) - full frequency spectrum at one time

  • get_frequency_slice(frequency) - time series at one frequency

  • get_frames(times) - matrix of frames at multiple times

  • get_band_power(fmin, fmax) - integrated power in a frequency band

  • get_spectral_moments_batch(power) - center of gravity, SD, skewness, and kurtosis per frame (single C++ call)

Export

  • as_matrix(include_dimnames) - export as a numeric matrix

  • as_data_frame() - export as a data.frame

Transform

  • to_spectrum(time) - extract a spectrum at one time point

  • to_dtw(reference) - dynamic time warping between spectrograms

See also

Sound, Spectrum, DTW, ComplexSpectrogram

Examples

snd <- Sound$create_tone(duration = 0.5, frequency = 440, sampling_rate =
 44100)
spec <- snd$to_spectrogram(window_length = 0.005, max_frequency = 5000)
power <- spec$get_power_at(time = 0.25, frequency = 440)