From ftm
Revision as of 23:07, 3 May 2009 by Norbert (talk | contribs)
Jump to: navigation, search
  • gbr.addenv ... Additive synthesis: generate partials with a given envelope
  • gbr.addpartials ... Additive synthesis: generate partials with given frequencies and amplitudes
  • gbr.autox ... Auto correlation and similar
  • gbr.bands ... FFT filter bands
  • gbr.bq ... Constant Q
  • gbr.copy ... Copy vector (fmat) out of a delay line or an fmat or fvec
  • gbr.crossx ... Cross correlation and similar
  • gbr.dct ... Discrete cosine transform
  • gbr.dline~ ... Classical delay line
  • gbr.drain~ ... Forward delay line
  • gbr.fft ... Fast Fourier transform
  • gbr.fire~ ... Gabor timing impulse generator
  • gbr.gen= ... Generate waveform/function
  • gbr.harms ... Estimate harmonics from a given spectrum (or any other vector)
  • gbr.ifft ... Inverse fast Fourier transform
  • gbr.lifter ... Cepstral liftering
  • gbr.lpc ... Linear prediction coefficients
  • gbr.mask ... Partial masking using critical band width
  • gbr.morph ... Partials sets interpolation
  • gbr.ola~ ... Overlap-add
  • gbr.paste ... Paste a grain (fmat or fvec) into a drain
  • gbr.peaks ... Estimate peaks (partials) from a given spectrum (or any other vector)
  • gbr.preemphasis ... Simple first order difference filter
  • gbr.psy~ ... Pysch synchronous (YIN-based) signal slicing
  • gbr.resample ... Resampling
  • gbr.slice~ ... Signal slicing
  • gbr.tapin~ ... Input tap for write delay line
  • gbr.tapout~ ... Output tap for read delay line
  • gbr.timer~ ... Gabor timer
  • gbr.trace ... Trace and index peaks
  • gbr.wind= ... Apply a window to an incoming frame, grain or wave
  • gbr.yin ... Fundamental frequency estimation after de Cheveigne and Kawahara


gbr.addenv Additive synthesis: generate partials with a given envelope
Adds partials with a spectral envolope (given as a vector or list) to an incoming spectrum using the FFT-1 technique. The output is typically connected to gbr.ifft (in real mode) or another gbr.addenv or gbr.addpartials module.
arguments: 1 <num: max> - maximum number of partials [64]
2 <num: size> - spectrum size (0: input size) [0]
attributes: coefs <num: coefs> - set number of FFT-1 spectral bin coefficients [8]
noisy <bool: switch> - enable/disbale noisiness ['off']
zero <bool: switch> - zero spectrum before adding partials ['off']
messages: postdoc - post external doc to console
inlets: 1 <fmat: spectrum> - complex vector (only positive frequencies) to which to add the generated partials
2 <num|fmat|fvec|list: freq(s)> - fundamental frequency (num: for harmonics) or vector of partials frequencies
3 <num|fmat|fvec|list: spectral envelope> - spectral envelope (given values will be linearly interpolated)
4 <num|fmat|fvec|list: phase(s)> - phase (num: for all partials) or vector of phases for the given partials
outlets: 1 <fmat: spectrum> - output spectrum with added partials
gbr.addpartials Additive synthesis: generate partials with given frequencies and amplitudes
Adds partials (given as a vector or list) to an incoming spectrum using the FFT-1 technique. The output is typically connected to gbr.ifft (in real mode) or another gbr.addpartials or gbr.addenv module.
Frequencies and amplitudes can be given by separated vectors with separated inputs (input format 'vec') or as a single matrix to the first inlet. The input format (attribute @format) determines the interpretation of the columns of the incoming matrix: 'fa' requires 2 columns with frequencies and amplitudes, 'ifa' requires 3 columns with partial indices, frequencies and amplitudes.
arguments: 1 <num: max> - maximum number of partials [64]
2 <num: size> - spectrum size (0: input size) [0]
attributes: coefs <num: num> - set number of FFT-1 spectral bin coefficients [8]
noisy <bool: switch> - enable/disbale noisiness ['off']
zero <bool: switch> - zero spectrum before adding partials ['off']
format <'vec'|'fa'|'ifa': format> - set input matrix/vector format ['vec']
messages: postdoc - post external doc to console
inlets: 1 <fmat: spectrum> - complex vector (only positive frequencies) to which to add the generated partials
2 <num|fmat|fvec|list: freq(s)> - set fundamental frequency (num: for harmonics) or vector of partials frequencies
3 <num|fmat|fvec|list: amp(s)> - set amplitude (num: for harmonics) or vector of partials amplitudes
4 <num|fmat|fvec|list: phase(s)> - set phase (num: for all partials) or vector of phases for the given partials
outlets: 1 <fmat: spectrum> - output spectrum with added partials
gbr.autox Auto correlation and similar
Calculates autocorrelation, distance, quadratic distance, sum magnitude difference function and accumulated difference function (yin).
arguments: 1 <num: size> - calculation size [256]
2 <num: width> - window width [256]
attributes: scale <num: factor> - set scaling factor [1]
mode <'corr'|'dist'|'dist2'|'smdf'|'yin': mode> - set calculation mode ['corr']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
width <num: width> - set window width
size <num: size> - set calculation size (maximum output size)
inlets: 1 <fmat|fvec: vector> - input vector
outlets: 1 <fmat: vector> - auto correlation vector
gbr.bands FFT filter bands
Calculate filter bands from FFT spectrum.
arguments: <num: min> <num: max> | <num: insize> <num: outsize> - init bands boundaries or input spectrum size and number of output bands
attributes: bands <'bounds'|'mel'|'htkmel'|'fcmel': mode> - set the bands mode [bounds]
minfreq <num: min> - set the output minimum frequency in hz
maxfreq <num: max> - set the output maximum frequency in hz
scale <num: scale> - set the bands filter scale
integ <'sqrabs'|'abs': spec> - set the spectrum integration type [sqrabs]
domain <num: domain> - set domain of output bands (<= 0 sets to sr/2)
domscale <num: factor> - set scaling factor of output bands (overwrites domain and down) [1]
down <num: down> - set down sampling factor of incoming frames (overwrites domain and domscale)
out <fmat: out> - set output matrix
messages: postdoc - post external doc to console
insize <num: insize> - set the input spectrum size
outsize <num: outsize> - set the number of output bands
bounds <list: boundaries> - set band boundaries
getstate - get the internal weights matrix
inlets: 1 <fmat|fvec: input> - input spectrum
outlets: 1 <fmat: output> - output bands
2 <fmat: weights> - internal weights matrix
gbr.bq Constant Q
Calculates a constant Q transform on an incoming spectrum [J.Brown, M.Puckette 1992]. Its input is typically connected to gbr.fft (in real mode).
arguments: <num: FFT size> <num: min freq> <num: channels per octave> <num: threshold> <num: number of channels> - filter kernel parameters
attributes: channels <num: channels> - set number of channels to calculate [all]
messages: postdoc - post external doc to console
inlets: 1 <fmat: spectrum> - complex vector of (positive frequencies)
outlets: 1 <fmat: coefficients> - filter bands
gbr.copy Copy vector (fmat) out of a delay line or an fmat or fvec
Copies a grain (fmat vector) of a given duration out of a delay line at a given delay time and outputs an fmat reference. If the given delay time is less than the the duration the vector will be shortened
arguments: 1 <delayline|fmat|fvec: source> - source reference
2 <num: duration> - grain duration [100]
attributes: unit <'msec'|'sec'|'samp': unit> - set time unit to msecs, secs or samples ['msec']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
set <delayline|fmat|fvec: source> - set source
bang - copy vector from the beginning of the delay line and output
inlets: 1 <num: delay> - copy and output grain at given delay position
2 <num: duration> - set duration
outlets: 1 <fmat: vector> - copied grain
gbr.crossx Cross correlation and similar
Calculates correlation, distance, quadratic distance and sum magnitude difference function.
arguments: 1 <num: size> - calculation size [256]
2 <num: width> - window width [256]
3 <fmat|fvec: vector> - right operand
attributes: scale <num: factor> - set scaling factor [1]
mode <'corr'|'dist'|'dist2'|'smdf': mode> - set calculation mode
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
size <num: size> - set calculation size (maximum output size)
width <num: width> - set window width
inlets: 1 <fmat|fvec: vector> - left input vector
2 <fmat|fvec: vector> - right input vector
outlets: 1 <fmat: vector> - cross correlation vector
gbr.dct Discrete cosine transform
Calculates a DCT of the incoming vector.
arguments: 1 <num: input size> - size of input vector [40]
2 <num: output size> - number of DCT coefficients to be calculated [13]
attributes: mode <'slaney'|'htk'|fc'> - set discrete cosine transform mode ['slaney']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
insize <num: input size> - set size input vector
outsize <num: output size> - set number of DCT coefficients to be calculated
getstate - get the internal weights matrix
inlets: 1 <fmat|fvec: vector> - input vector
outlets: 1 <fmat: vector> - DCT coefficients
2 <fmat: vector> - internal weights matrix
gbr.dline~ Classical delay line
Delay line to be used with gbr.copy and gbr.tapout~.
arguments: <sym: name> <num: size in given unit> - give name and size [none 100]
attributes: scope <'local'|'global': scope> - set delayline name scope ['global']
unit <'msec'|'sec'|'samp': unit> - set time unit to msecs, secs or samples ['msec']
messages: postdoc - post external doc to console
clear - zero delay line
freeze <'0'|'1': freeze>] - enable/disable delay line freeze
inlets: 1 - write signal into delay line
outlets: 1 - thru output (for order-forcing)
gbr.drain~ Forward delay line
Delay line to write with different delays to be used with gbr.paste and gbr.tapin~.
arguments: <sym: name> <num: size in given unit> - give name and size [none 100]
attributes: scope <'local'|'global': scope> - set delayline name scope ['global']
unit <'msec'|'sec'|'samp': unit> - set time unit to msecs, secs or samples ['msec']
messages: postdoc - post external doc to console
clear - clear delay line
inlets: 1 - messages only
outlets: 1 - sum delay line output
gbr.fft Fast Fourier transform
Calculates FFT on incoming vector.
arguments: 1 <num: size> - FFT size (rounded to the next power of 2) [512]
attributes: scale <num: factor> - scaling factor (0 --> 1 / FFT size) [1]
mode <'auto'|'complex'|'real': mode> - FFT mode ['auto']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
inlets: 1 <fmat|fvec: vector> - real or conplex input vector
outlets: 1 <fmat: vector> - real or conplex output vector
gbr.fire~ Gabor timing impulse generator
Periodically outputs a given fmat or a bang within the Gabor scheduling scheme.
arguments: <num: period> [<fmat: vector>] - frequency/period and fmat to fire [0]
attributes: unit <'hz'|'msec'|'sec'|'samp'|'midi'|'midicent': unit> - set frequency/period unit to Hz, msec or samples ['hz']
period <num: period> - set frequency/period [0]
var <num: freq var> - set period variation (0...1) [0]
out <fmat: out> - set output fmat
messages: postdoc - post external doc to console
inlets: 1 <num: freq/period> - fire frequency/period (O is off)
2 <fmat: vector> - set fmat to fire
outlets: 1 - output fmat or bang
gbr.gen= Generate waveform/function
Adds a given (and parametrised) waveform/function to an incomming vector. The user can chose among various waveforms/functions
arguments: <'cosine'|'sine': function> <list: parameters> - generator function and parameters ['cosine']
attributes: none
messages: postdoc - post external doc to console
set <'cosine'|'sine': function> <list: parameters> - set generator function and parameters
inlets: 1 <fmat|fvec: vector> - input vector to which the waveform/function will be added
2 <list: parameters> - set generator parameters
outlets: 1 <fmat|fvec: vector> - output incoming vector with added waveform/function
gbr.harms Estimate harmonics from a given spectrum (or any other vector)
Estimates frequencies (interpolated and scaled indices) and amplitudes of harmonics in an incoming vector. Harmonics are defined as peaks around the multiple of a given value (fundamental frequency) with a given tolerance. The estimation of harmonics in a spectrum works best when a logarthimic amplitude spectrum is provided as input.
arguments: 1 <num: max harms> - maximum number of harmonics to be estimated [16]
2 <num: freq> - fundamental frequency in Hz [10]
3 <num: factor> - allowed deviation factor from theoretic harmonic frequency (linear factor of f0) [1.0]
attributes: none
messages: postdoc - post external doc to console
max <num: max harms> - set maximum number of harmonics to be estimated
freq <num: freq> - set fundamental frequency in Hz
delta <num: factor> - set allowed deviation factor from theoretic harmonic frequency (linear factor of f0)
interval <num: cent> - set allowed deviation factor from theoretic harmonic frequency (in cent)
height <num: amp> - set minimum height for harmonic peaks
width <num: freq> - set maximum width for harmonic peaks
inlets: 1 <fmat|fvec: vector> - input vector>
outlets: 1 <fmat: vector> - vector of harmonics
gbr.ifft Inverse fast Fourier transform
Calculates inverse FFT on incoming vector.
arguments: 1 <num: size> - FFT size (rounded to the next power of 2) [512]
attributes: scale <num: factor> - scaling factor (0 --> 1 / FFT size) [1]
mode <'auto'|'complex'|'real': mode> - FFT mode ['auto']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
inlets: 1 - messages only
outlets: 1 <fmat> - no description
gbr.lifter Cepstral liftering
Cepstral liftering (HTK and Auditory Toolbox styles)
arguments: 1 <num: size> - size of input vector [13]
2 <num: factor> - filtering factor [0]
attributes: inv <bool: switch> - enable/disable the inverse liftering mode [0]
mode <'exp'|'sin': mode> - set liftering type: exponential (Auditory Toolbox-like) or sinusoidal (HTK-like) ['exp']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
insize <num: size> - size of input vector
factor <num: factor> - set filtering factor
getstate - get internal weights matrix
inlets: 1 <fmat|fvec: vector> - cepstrum vector
outlets: 1 <fmat: vector> - liftered cepstrum
2 <fmat: weights> - internal weights matrix
gbr.lpc Linear prediction coefficients
Calculates LPC coefficients from incoming sinal frame.
arguments: 1 <num: order> - LPC order [12]
attributes: errasfloat <bool: switch> - enable/disable float number output [off]
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
order <num: order> - set LPC order
inlets: 1 <fmat|fvec: vector> - input vector
outlets: 1 <fmat: vector> - LPC coefficients
2 <num|fmat: error> - prediciton error
3 <fmat: vector> - autocorrelation
4 <fmat: vector> - internal values
gbr.mask Partial masking using critical band width
Calculates and applies masking to incoming vector of partials. The input format (attribute @format) determines the interpretation of the columns of the incoming matrix: 'fa' requires 2 columns with frequencies and amplitudes, 'ifa' requires 3 columns with partial indices, frequencies and amplitudes.
arguments: none
attributes: format - 'fa'|'ifa': input format> - set input matrix format ['fa']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
slope <num: slope> [<num: slope>] - set (left and right) masking slope
calcpeaks <num: (left) peaks> [<num: right peaks>] - set number of peaks on the left and right to be taken into account in calculation
inlets: 1 <fmat: partials> - vector of partials (in given format)
outlets: 1 <fmat: partials> - vector of partials (in given format)
gbr.morph Partials sets interpolation
Partials sets interpolation using indexes, frequencies and amplitudes
arguments: 1 <fmat: partials> - vector of paritals corresponding to the interpolation factor 0
2 <fmat: partials> - vector of paritals corresponding to the interpolation factor 1
attributes: format - 'fa'|'ifa'|'plain': input format> - set input matrix format ['fa']
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
inlets: 1 <num: factor> - morphing factor (0...1)
2 <fmat: partials> - vector of paritals corresponding to the interpolation factor 0
3 <fmat: partials> - vector of paritals corresponding to the interpolation factor 1
outlets: 1 <fmat: partials> - vector of morphed paritals
gbr.ola~ Overlap-add
Performs the overlap-add of incoming vectors into a forward delayline. The vector will be shortened at the end of the delayline.
arguments: 1 <num: channels> - number of output channels [1]
2 <num: size> - buffer size [100]
3 <num: delay> - delay position [0]
4 <num: channel> - output channel (0 switches off) [1]
attributes: interp <bool: switch> - enable/disable interpolation ['off']
unit <'msec'|'sec'|'samp': unit> - set time unit to msecs, secs or samples ['msec']
messages: postdoc - post external doc to console
clear - clear delay line
inlets: 1 <fmat|fvec: vector> - overlap-add vector at given delay position
2 <num: delay> - set delay
3 <num: channel> - set output channel (0 switches off)
outlets: 1 - sum delay line output
gbr.paste Paste a grain (fmat or fvec) into a drain
Copies a vector into a drain with a given delay. The vector will be shortened at the end of the drain.
arguments: 1 <delayline|fmat|fvec: destination> - destination (write delay line or fmat)
2 <num: delay> - delay position [0]
attributes: interp <bool: switch> - enable/disable interpolation ['off']
unit <'msec'|'sec'|'samp': unit> - set time unit to msecs, secs or samples ['msec']
messages: postdoc - post external doc to console
set <delayline|fmat|fvec: destination> - set destination (write delay line, fmat or fvec)
inlets: 1 <fmat|fvec: vector> - paste vector at given delay position
2 <num: delay> - set delay position
outlets: none
gbr.peaks Estimate peaks (partials) from a given spectrum (or any other vector)
Estimates frequencies (interpolated and scaled indices) and amplitudes of peaks in an incoming vector. The estimation of partials in a spectrum works best when a logarthimic amplitude spectrum (positive frequencies) is provided as input.
arguments: 1 <num: max peaks> - maximum number of peaks to be estimated [16]
attributes: domain <num: domain> - set domain of output peaks (<= 0 sets to sr/2) [sr/2]
domscale <num: factor> - set scaling factor of output peaks (overwrites domain and down)
down <num: down> - set down sampling factor of incoming frames (overwrites domain and domscale) [1]
keep <'lowest'|'strongest': mode> - keep first or strongest peaks ['lowest']
range <min: boundary> <max: boundary> - band where to search for peaks
messages: postdoc - post external doc to console
max <num: max peaks> - set maximum number of peaks to be estimated
height <num: amp> - set minimum height for peaks
width <num: freq> - set maximum width for peaks (indicates sinuso�dality)
dev <num: value> - set maximum deviation from mean value
inlets: 1 <fmat|fvec: vector> - input vector (spectrum)>
outlets: 1 <fmat: vector> - vector of peaks
gbr.preemphasis Simple first order difference filter
Simple first order difference filter
arguments: 1 <num: factor> - filtering factor [0]
attributes: out <fmat: out> - set output vector
messages: postdoc - post external doc to console
factor <num: factor> - set filtering factor
getstate - get the previous sample
clear - clear any previous sample
inlets: 1 <fmat|fvec: vector> - input signal
outlets: 1 <fmat: vector> - output preemphasised frame
2 <fmat: vector> - previous frame
gbr.psy~ Pysch synchronous (YIN-based) signal slicing
Cuts incoming signal into elementary waveforms. Outputs vectors corresponding to two periods of the estimated frequency or fixed duration (256 points) when unvoiced.
arguments: none
attributes: active <bool: switch> - activate/deactivate calculation and output ['on']
threshold <num: pitch> [<num: noise>] - set pitch and noise quality treshold [0.6838 0.4523]
messages: postdoc - post external doc to console
inlets: 1 - input signal to be analysed and cut into elemetary waveforms
outlets: 1 <fmat: vector> - elementary waveforms (2 periods)
2 <num: freq> - estimated frequency in Hz (0, when unvoiced)
3 <num: periodicity> - yin periodicity/quality factor
4 <num: factor> (linear) energy
gbr.resample Resampling
Resamples incoming vector in different modes: 'cubic' (cubic interpolation), 'downmean' (downsampling by calculating the mean of a given number of values), 'downremove' (picks nearest value)
arguments: 1 <num: order> - resampling order/increment [1]
attributes: mode <'cubic'|'downmean'|'downremove': mode> - set resampling mode
out <fmat: output> - set output vector
messages: postdoc - post external doc to console
inlets: 1 <fmat: vector> - input vector
2 <num: order> - set resampling order/increment
outlets: 1 <fmat: vector> - output vector
gbr.slice~ Signal slicing
Cuts incoming signal into frames of given size with given period (hop size).
arguments: 1 <num: size> - frame size [512]
2 <num: size> - hop size [256]
attributes: size <num: size> - set frame size
period <num: size> - set hop size
active <bool: switch> - activate/deactivate calculation and output ['on']
unit <'samp'|'msec'|'sec': unit> - set time unit to msecs, secs or samples ['samp']
messages: postdoc - post external doc to console
inlets: 1 - input signal to be sliced into frames
outlets: 1 <fmat: vector> - signal frames
2 <undefined> - no description
gbr.tapin~ Input tap for write delay line
Simple input tap for write delay line defined by gbr.dline~.
arguments: 1 <delayline: write> - delay line (defined by gbr.drain~)
2 <num: delay> - delay time (in given unit) [0]
attributes: unit <'msec'|'sec'|'samp': unit> - set delay unit to msecs, secs or samples ['msec']
messages: postdoc - post external doc to console
set <delayline: write> - set delay line (defined by gbr.drain~)
inlets: 1 - input signal written to delay line
2 - set delay time
outlets: 1 - signal outputundefined
gbr.tapout~ Output tap for read delay line
Simple output tap for read delay line defined by gbr.drain~.
arguments: 1 <delayline: read> - delay line (defined by gbr.dline~)
2 - delay time (in given unit) [0]
attributes: interp - 0|1|'off'|'on|'cubic'|'linear': mode> - interpolation mode ['off']
unit <'msec'|'sec'|'samp': unit> - set delay unit to msecs, secs or samples ['msec']
messages: postdoc - post external doc to console
set <delayline: read> - set delay line (defined by gbr.dline~)
inlets: 1 - messages only
2 <sig|num: delay> - delay time
outlets: 1 - delayed signal
gbr.timer~ Gabor timer
stop watch in Gabor scheduling scheme
arguments: none
attributes: unit <'msec'|'sec'|'samp'|'hz'|: unit> - set timer unit to msecs, secs, samples or Hz ['msec']
messages: postdoc - post external doc to console
bang - report time since last bang
inlets: 1 - messages only
outlets: 1 <num: time> - time in given unit
gbr.trace Trace and index peaks
Trace peaks from frame to frame to associate indexes to the peaks
arguments: 1 <num: max> - maximum number of peaks [200]
attributes: max <num: max> - maximum number of partials [200]
relfreq <num: var> - allowed relative frequency variation for a peak to keep its index in cent > 0. (default: 20.)
absfreq <num: var> - allowed frequency variation for a peak to keep its index in Hz > 0. (default: 50.)
absamp <num: var> - allowed linear amplitude variation for a peak to keep its index > 0. (default: 0.5)
maxpasses <num: max> - maximum number of connection passes for the algorithm (default 4)
messages: postdoc - post external doc to console
clear - reset (indexes start from 0, drops previous peaks set)
inlets: 1 <fmat: partials> - vector of partials
outlets: 1 <fmat: partials> - vector of traced partials (with index)
gbr.wind= Apply a window to an incoming frame, grain or wave
Applies a chosen (and parametrized) window to the incomming fmat (column by column). The user can chose among various window types (see help patch).
arguments: <sym: function> <list: parameters> - window function and parameters ['hann']
attributes: none
messages: postdoc - post external doc to console
set <sym: function> <list: parameters> - set window function and parameters
inlets: 1 <fmat|fvec: vector> - input vector to be windowed
2 <list: parameters> - set window parameters
outlets: 1 <fmat|fvec: vector> - output incoming vector with applied window
gbr.yin Fundamental frequency estimation after de Cheveigne and Kawahara
Estimates fundamental frequency and outputs energy, periodicity factor, and auto correlation coefficients.
arguments: 1 <num: min freq> - lowest estimated frequency in Hz [50.]
2 - quality/periodicity threshold [0.68]
attributes: minfreq <num: min freq> - lowest estimated frequency in Hz
threshold <num: threshold> - quality/periodicity threshold
down <num: factor> - down sampling factor of incoming frames [1]
messages: postdoc - post external doc to console
inlets: 1 - signal frame (fmat or fvec
outlets: 1 <num: freq> - estimated frequency in Hz
2 <num: energy> - energy factor
3 <num: perodicity> - quality/periodicity factor
4 <num: ac1> - 2nd autocorrelation coefficient (ac1)
5 <fmat: acf> - vector of autocorrelation coefficients
Retrieved from "https://ftm2.ircam.fr/index.php?title=Gabor_Modules&oldid=2604"