Skip to contents

Sound

Source: R/sound-wrapper.R
Sound.Rd

Represents a digitized acoustic signal (a Praat Sound object).

Arguments

path

Path to an audio file. See the File I/O section for supported formats.

.xptr

Not for direct use. External pointer to the underlying C++ Sound object; set internally when a method returns a new Sound.

Value

A Sound object.

Details

A Sound holds one or more channels of audio sampled at regular intervals. It is the starting point for most acoustic analyses in pladdrr: pitch, formants, intensity, and spectral measures all begin from a Sound.

File I/O

The constructor tries two readers in order:

  1. The native Praat reader, which handles WAV, AIFF, AIFC, NIST, and NeXT/Sun files without extra packages, and is the fastest path.

  2. The av package, used as a fallback for everything else, including FLAC, MP3, and OGG Vorbis. The FLAC and MP3 decoder sources were dropped from the vendored Praat tree in v4.9.5 to keep the CRAN tarball within limits, so those formats need av installed (install.packages("av")). Without it, reading a FLAC or MP3 file raises an error naming the missing package.

Convert to WAV up front if you want to avoid the av dependency entirely.

Usage


# From file
sound <- Sound(path = "audio.wav")

# From numeric data
sound <- Sound$from_values(values, sampling_rate = 44100)

# Synthetic tone
sound <- Sound$create_tone(frequency = 440, duration = 1.0)

Query methods

  • get_duration() - duration in seconds

  • get_sampling_frequency() - sampling rate in Hz

  • get_number_of_samples() - number of samples

  • get_number_of_channels() - number of channels

  • get_value_at_time() - amplitude at a given time

  • get_rms(), get_energy(), get_power() - energy measures

  • get_intensity_db() - intensity in dB

  • get_minimum(), get_maximum(), get_mean() - amplitude statistics

  • get_values(channel) - sample values as a numeric vector

  • get_sample_times() - sample times as a numeric vector

Analysis methods

  • to_pitch() - extract pitch contour (F0)

  • to_formant_burg() - extract formants (F1, F2, F3, ...)

  • to_intensity() - extract intensity contour

  • to_harmonicity_cc() - harmonics-to-noise ratio

  • to_harmonicity_gne() - glottal-to-noise excitation ratio (GNE)

  • extract_electroglottogram(channel, invert) - extract an electroglottogram (EGG) from a channel

  • to_spectrum() - frequency spectrum

  • to_spectrogram() - time-frequency representation

  • to_ltas() - long-term average spectrum

  • to_ltas_pitch_corrected() - pitch-corrected LTAS (voice quality)

  • to_formant_robust() - outlier-resistant formant tracking

  • to_mel_spectrogram() - mel-scale spectrogram

  • to_bark_spectrogram() - Bark-scale spectrogram

  • to_point_process_periodic_cc() - extract glottal pulses

Signal processing

  • lengthen() - time-stretch using overlap-add

  • autocorrelate() - autocorrelation function

  • convolve() - convolve with another sound

  • cross_correlate() - cross-correlate with another sound

  • deepen_band_modulation() - hearing enhancement

  • filter_by_formant() - filter with a Formant object

  • filter_by_formant_noscale() - filter without scaling

Extraction

  • extract_channel() - extract a single channel

  • extract_part(from, to, window_shape, relative_width, preserve_times) - extract a time range, with optional windowing. Supports 12 window shapes (rectangular, triangular, parabolic, hanning, hamming, gaussian1-5, kaiser1-2); see https://www.fon.hum.uva.nl/praat/manual/Sound__Extract_part___.html.

Modification

  • scale_intensity() - scale to a target dB level (in place)

  • scale_peak() - scale peak amplitude (in place)

  • pre_emphasize() - high-pass filter (in place)

  • de_emphasize() - low-pass filter (in place)

  • resample() - resample to a different rate (returns a new object)

  • convert_to_mono() - average channels to mono (returns a new object)

  • concatenate() - append another sound (returns a new object)

  • mix() - mix with another sound (returns a new object)

Export

  • as_matrix() - export as a numeric matrix

  • as_data_frame() - export as a data.frame

  • save() - save to an audio file

See also

Pitch, Formant, Intensity, Spectrum

Examples

# Synthetic tone, no external file needed
sound <- Sound$create_tone(frequency = 440, duration = 1.0)
pitch <- sound$to_pitch()
formants <- sound$to_formant_burg()

cat("Duration:", sound$get_duration(), "s\n")
#> Duration: 1 s
cat("Sample rate:", sound$get_sampling_frequency(), "Hz\n")
#> Sample rate: 44100 Hz

part <- sound$extract_part(0.2, 0.5)