Find the cross-correlation of a signal with itself.
###Input format float* ###Output format float* ###Supported parameters Name: normalize Description: Calculate normalized autocorrelation by dividing each value by lag 0 result (squared signal sum). Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Bandpass filter. Cuts off frequences higher than "frequency_high" and lower than "frequency_low".
###Input format float* ###Output format float* ###Supported parameters Name: frequency_high Description: All frequencies higher will be cut. Default: 24000
Name: frequency_low
Description: All frequencies lower will be cut.
Default: 1
Name: length
Description: Filter size in samples (order).
Default: 256
Name: ripple
Description: Ripple level in dB (used by a subset of filter types).
Default: 1.000000
Name: rolloff
Description: Rolloff level in dB (used by a subset of filter types).
Default: 0.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: IIR filter type (bessel, butterworth, etc.).
Default: chebyshevII
Find the tempo of a musical signal.
###Input format float* ###Output format FixedArray<(unsigned char)2, float>* ###Supported parameters Name: bands Description: The number of bands to sum. Default: 1
Name: debug
Description: Dump the resulting energy vectors.
Default: 0
Name: max_bpm
Description: Maximal number of beats per minute.
Default: 240.000000
Name: min_bpm
Description: Minimal number of beats per minute.
Default: 60.000000
Name: peaks
Description: The number of the most significant peaks to record.
Default: 3
Name: pulses
Description: The number of pulses for the convolution.
Default: 3
Name: resolution1
Description: The bpm resolution during the first pass.
Default: 2.000000
Name: resolution2
Description: The bpm resolution during the second pass.
Default: 0.100000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Converts each complex number to corresponding real numbers.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Window's center of mass in frequency domain calculation.
###Input format float* ###Output format float ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Calculates the magnitude of each complex number, that is, a square root of the sum of squared real and imaginary parts.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Convolve a raw signal with a window function.
###Input format float* ###Output format float* ###Supported parameters Name: length Description: Filter size in samples (order). Default: 256
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: window
Description: Type of the window. E.g. "rectangular" or "hamming".
Default: rectangular
Performs Discrete Cosine Transform on the signal.
###Input format float* ###Output format float*
Discrete Wavelet Packet Transform
###Input format float* ###Output format float* ###Supported parameters Name: order Description: The number of coefficients in the wavelet. Default: 8
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: tree
Description: The wavelet packet binary tree fingerprint.
Default: 3 3 3 4 4 4 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6
Name: type
Description: The type of wavelet to apply. Supported values are daub (Daubechies), coif (Coiflet) and sym (Symlet).
Default: daub
Get the difference between values of adjacent windows.
###Input format float* ###Output format float* ###Supported parameters Name: rlength Description: The linear regression window length. Only odd values greater than 1 are accepted. Default: 5
Name: type
Description: The algorithm to calculate the deltas. Allowed values are "simple" and "regression".
Default: simple
Differentiate the signal.
###Input format float* ###Output format float* ###Supported parameters Name: rectify Description: Perform half-wave rectification (retain only if the difference is positive). Default: 0
Name: swt
Description: Smoothly differentiate using Stationary Wavelet Transform (db1) of the specified level. The level must be greater than or equal to 0. If set to zero, this parameter is ignored.
Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Sound energy calculation.
###Input format float* ###Output format float ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Converts the signal to the selected psychoacoustic scale (default is mel).
###Input format float* ###Output format float* ###Supported parameters Name: debug Description: Dump the precalculated filter bank. Default: 0
Name: frequency_max
Description: Maximal frequency of the filter bank.
Default: 6854.000000
Name: frequency_min
Description: Minimal frequency of the filter bank.
Default: 130.000000
Name: number
Description: The number of triangular filters.
Default: 32
Name: squared
Description: Apply the squared filter bank.
Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: The type of the scale. Supported values are "linear","mel", "bark" and "midi".
Default: mel
Measure of spectral change.
###Input format float* ###Output format float
Clones the windows, increasing their number by a factor of "factor".
###Input format float* ###Output format float* ###Supported parameters Name: factor Description: Windows number multiplier value. Default: 4
Cut the frequency bands from sequential windows.
###Input format float* ###Output format float* ###Supported parameters Name: bands Description: Bands configuration. Overrides "number". Default:
Name: filter
Description: IIR filter type to apply.
Default: chebyshevII
Name: lengths
Description: IIR filter orders. "auto" for automatic selection.
Default:
Name: number
Description: The number of bands. It should be the same as the cloning factor of Fork. You may set the bands configuration by hand with "bands" parameter, "number" will be ignored then.
Default: 4
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Highpass filter. All frequences lower than "frequency" shall not pass.
###Input format float* ###Output format float* ###Supported parameters Name: frequency Description: Filter's cutoff frequency. Default: 1
Name: length
Description: Filter size in samples (order).
Default: 256
Name: ripple
Description: Ripple level in dB (used by a subset of filter types).
Default: 1.000000
Name: rolloff
Description: Rolloff level in dB (used by a subset of filter types).
Default: 0.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: IIR filter type (bessel, butterworth, etc.).
Default: chebyshevII
Inverse of "performs Discrete Cosine Transform on the signal.".
###Input format float* ###Output format float*
Inverse of "takes the logarithm from each real value of the signal.".
###Input format float ###Output format float ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Inverse of "takes the logarithm from each real value of the signal.".
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Inverse of "performs Discrete Fourier Transform on the input signal (using real FFT).".
###Input format float* ###Output format float*
Inverse of "squares the signal (window floating point format).".
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Inverse of "splits the raw input signal into numerous windows stepping "step" ms with length "length" ms of type "type".".
###Input format float* ###Output format float* ###Supported parameters Name: count Description: The resulting amount of buffers. Default: 1
Name: interleaved
Description: Interleave windows from sequential input buffers. In inverse mode, treat the windows as interleaved.
Default: 1
Name: length
Description: Window size in samples.
Default: 512
Name: step
Description: Distance between sequential windows in samples.
Default: 205
Inverse of "splits the raw input signal into numerous windows stepping "step" ms with length "length" ms of type "type".".
###Input format short* ###Output format short* ###Supported parameters Name: count Description: The resulting amount of buffers. Default: 1
Name: interleaved
Description: Interleave windows from sequential input buffers. In inverse mode, treat the windows as interleaved.
Default: 1
Name: length
Description: Window size in samples.
Default: 512
Name: step
Description: Distance between sequential windows in samples.
Default: 205
Copy the input to output.
###Input format identity ###Output format identity
Calculate Linear Prediction Coefficients (LPC) using Levinson-Durbin recursion from autocorrelation coefficients.
###Input format float* ###Output format float* ###Supported parameters Name: error Description: Include total estimation error Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Convert Linear Prediction Coefficients (LPC) to Cepstral Coefficients.
###Input format float* ###Output format float* ###Supported parameters Name: size Description: The number of cepstral coefficients. 0 means original LPC length. Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Converts Linear Prediction Coefficients (LPC) to more robust Line Spectral Pairs (LSP) representation.
###Input format float* ###Output format float* ###Supported parameters Name: bisects Description: The number of bisections for the root value refinement. The bigger the number, the more precise the values are and the less probability is to skip a root. Default: 16
Name: intervals
Description: The number of intervals to scan in the range [-1, 1]. The corresponding delta equals to 2 / intervals. The bigger the number, the less probability is to skip a root.
Default: 128
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Takes the logarithm from each real value of the signal.
###Input format float ###Output format float ###Supported parameters Name: add1 Description: Add 1 to values before taking the logarithm. This trick avoids getting NaNs on zeros. Default: 1
Name: base
Description: Logarithm base (2, 10 or e).
Default: e
Name: scale
Description: The number to multiply each value by before taking the logarithm.
Default: 1.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Takes the logarithm from each real value of the signal.
###Input format float* ###Output format float* ###Supported parameters Name: add1 Description: Add 1 to values before taking the logarithm. This trick avoids getting NaNs on zeros. Default: 1
Name: base
Description: Logarithm base (2, 10 or e).
Default: e
Name: scale
Description: The number to multiply each value by before taking the logarithm.
Default: 1.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Lowpass filter. All frequences higher than "frequency" shall not pass.
###Input format float* ###Output format float* ###Supported parameters Name: frequency Description: Filter's cutoff frequency. Default: 1
Name: length
Description: Filter size in samples (order).
Default: 256
Name: ripple
Description: Ripple level in dB (used by a subset of filter types).
Default: 1.000000
Name: rolloff
Description: Rolloff level in dB (used by a subset of filter types).
Default: 0.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: IIR filter type (bessel, butterworth, etc.).
Default: chebyshevII
Window means calculation.
###Input format float* ###Output format sound_feature_extraction::formats::FixedArray<(unsigned char)2, float> ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Name: types
Description: Mean types to calculate (names separated with spaces).
Default: arithmetic
Merge all single-s to one solid array.
###Input format float ###Output format float*
Mix stereo audio channels together.
###Input format short* ###Output format short* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Peak analysis. This is a modified version of the algorithm from "Automatic Musical Genre Classification Of Audio Signals".
###Input format FixedArray<(unsigned char)2, float>* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Accurate average estimation using dynamic programming approach.
###Input format FixedArray<(unsigned char)2, float>* ###Output format float ###Supported parameters Name: mind_values Description: Whether to favor points with higher values. Default: 0
Peak detection, that is, extraction of maximums and/or minimums from the signal. The first item is the normalized position, the second is the peak value.
###Input format float* ###Output format FixedArray<(unsigned char)2, float>* ###Supported parameters Name: max_pos Description: The real maximal value of the position of each peak. Default: 1.000000
Name: min_pos
Description: The real minimal value of the position of each peak.
Default: 0.000000
Name: number
Description: The maximal number of peaks to extract. The rest of the slots will be set to (min_pos, 0).
Default: 3
Name: sort
Description: The order in which to sort peaks (position, value, or both).
Default: 3
Name: swt_level
Description: SWT level, that is, the number of smoothing passes."0" means do not do SWT.
Default: 0
Name: swt_order
Description: The number of coefficients in the wavelet.
Default: 4
Name: swt_type
Description: Type of Stationary Wavelet Transform which will be done before taking the peaks. Supported values are daub (Daubechies), coif (Coiflet) and sym (Symlet). Please note that SWT introduces a phase shift.
Default: daub
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: The type of the peaks: max, min or all.
Default: 1
Filter the signal with a first-order high-pass filter y[n] = 1 - k * x[n - 1].
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Name: value
Description: The filter coefficient from range (0..1]. The higher, the more emphasis occurs.
Default: 0.900000
Converts each real number to complex number (imaginary part is set to zero).
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Perform RASTA filtering.
###Input format float* ###Output format float* ###Supported parameters Name: length Description: Filter size in samples (order). Default: 256
Name: pole
Description: Pole value. It normally lies within (1, 0.9].
Default: 0.940000
Name: ripple
Description: Ripple level in dB (used by a subset of filter types).
Default: 1.000000
Name: rolloff
Description: Rolloff level in dB (used by a subset of filter types).
Default: 0.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: IIR filter type (bessel, butterworth, etc.).
Default: chebyshevII
Performs Discrete Fourier Transform on the input signal (using real FFT).
###Input format float* ###Output format float*
Wave rectification to decrease high-frequency content.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Reorders the values in the array according to the specified algorithm.
###Input format float* ###Output format float* ###Supported parameters Name: algorithm Description: The way to reorder the values. Default: chroma
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Measure of spectral shape.
###Input format float* ###Output format float ###Supported parameters Name: ratio Description: The ratio between the partial sum and the whole sum. Default: 0.850000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Rotate the buffers, so that all elements with the same index get to the same buffer.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Spectral Flatness Measure calculation.
###Input format sound_feature_extraction::formats::FixedArray<(unsigned char)2, float> ###Output format float ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Apply spectral harmonics correlation.
###Input format float* ###Output format float* ###Supported parameters Name: harmonics Description: The number of harmonics. Default: 3
Name: max
Description: The maximal frequency to scan.
Default: 600
Name: min
Description: The minimal frequency to scan.
Default: 50
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: window
Description: The spectral window length, in Hz.
Default: 40
Calculate short-time mean and scale normalized values, that is,$stmsn_n[i] = rac{w_n[i] - \sum_{k=n-L/2}^{n+L/2}{w_k[i]}}{\max_{k=n-L/2}^{n+L/2}{w_k[i]} -\min_{k=n-L/2}^{n+L/2}{w_k[i]}}$.
###Input format float* ###Output format float* ###Supported parameters Name: length Description: The amount of local values to average. Default: 300
Selects the specified part of input.
###Input format float* ###Output format float* ###Supported parameters Name: from Description: The anchor of the selection. Can be either "left" or "right". Default: left
Name: length
Description: The resulting vector length. 0 means the length of the input.
Default: 0
Name: select
Description: The length of picked up input values, The rest will be set to zero. 0 means the length of the output.
Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Calculates the squared magnitude of each complex number, that is, the sum of squared real and imaginary parts.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Squares the signal (window floating point format).
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Calculate statistical measures, such as skew or kurtosis.
###Input format float* ###Output format float* ###Supported parameters Name: interval Description: "Texture" interval, must be bigger than 1. Zero means all buffers. Default: 0
Name: overlap
Description: Overlap value, cannot be greater than (interval - 1).
Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: types
Description: Stats types to calculate (names separated with spaces, "all" for all).
Default: average stddev skewness kurtosis
Stretch the signal by an integer factor.
###Input format float* ###Output format float* ###Supported parameters Name: center Description: Center the intervals, extending the corners (useful for stretching after splitting into windows.). Default: 0
Name: factor
Description: Stretch the signal by duplicating its values.
Default: 1.000000
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Calculates the subband energies (subbands are defined witha binary tree fingerprint identical to used in DWPT).
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Name: tree
Description: The subbands binary tree fingerprint.
Default: 3 3 3 4 4 4 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6
Subsample the signal by an integer factor.
###Input format float* ###Output format float* ###Supported parameters Name: factor Description: Calculate normalized subsampling by dividing each value by lag 0 result (squared signal sum). Default: 1
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Unpacks the result after applying RDFT.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Splits the raw input signal into numerous windows stepping "step" ms with length "length" ms of type "type".
###Input format float* ###Output format float* ###Supported parameters Name: interleaved Description: Interleave windows from sequential input buffers. In inverse mode, treat the windows as interleaved. Default: 1
Name: length
Description: Window size in samples.
Default: 512
Name: step
Description: Distance between sequential windows in samples.
Default: 205
Name: type
Description: Type of the window. E.g. "rectangular" or "hamming".
Default: hamming
Splits the raw input signal into numerous windows stepping "step" ms with length "length" ms of type "type".
###Input format short* ###Output format short* ###Supported parameters Name: interleaved Description: Interleave windows from sequential input buffers. In inverse mode, treat the windows as interleaved. Default: 1
Name: length
Description: Window size in samples.
Default: 512
Name: step
Description: Distance between sequential windows in samples.
Default: 205
Name: type
Description: Type of the window. E.g. "rectangular" or "hamming".
Default: hamming
Applies a window function to each window.
###Input format float* ###Output format float* ###Supported parameters Name: predft Description: Apply Discrete Fourier Transform to window function. Default: 0
Name: threads_number
Description: The maximal number of OpenMP threads.
Default: 8
Name: type
Description: Type of the window. E.g. "rectangular" or "hamming".
Default: hamming
Number of time domain zero crossings of the signal.
###Input format float* ###Output format int ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Number of time domain zero crossings of the signal.
###Input format short* ###Output format int ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8
Pads signal with zeros to make it's length a power of 2.
###Input format float* ###Output format float* ###Supported parameters Name: threads_number Description: The maximal number of OpenMP threads. Default: 8