Extract cepstral features from audio segment
Audio Toolbox / Measurements
The Cepstral Feature Extractor block extracts cepstral features from an audio segment. Cepstral features are commonly used to characterize speech and music signals.
Port_1
— Audio input to cepstral feature extractorAudio input to the cepstral feature extractor, specified as a column vector or a matrix. If specified as a matrix, the columns are treated as independent audio channels.
Data Types: single
 double
coeffs
— Cepstral coefficientsCepstral coefficients, returned as a column vector or a matrix. If the coefficients matrix is an NbyM matrix, N is determined by the values you specify in the Number of coefficients to return and Log energy usage parameters. M equals the number of input audio channels.
When the Log energy usage parameter is set to:
Append
–– The block prepends
the log energy value to the coefficients vector. The length
of the coefficients vector is 1 +
NumCoeffs, where
NumCoeffs is the value specified in
the Number of coefficients to return
parameter.
Replace
–– The block replaces
the first coefficient with the log energy of the signal. The
length of the coefficients vector is
NumCoeffs.
Ignore
–– The block does not
calculate or return the log energy.
This port is unnamed until you select Output delta parameter, the Output deltadelta parameter, or both.
Data Types: single
 double
delta
— Change in coefficientsChange in coefficients over consecutive calls to the algorithm, returned as a column vector or a matrix. The delta array is of the same size and data type as the coeffs array.
To enable this port, select the Output delta parameter.
Data Types: single
 double
deltaDelta
— Change in delta valuesChange in delta values over consecutive calls to the algorithm, returned as a column vector or a matrix. The deltaDelta array is the same size and data type as the coeffs and delta arrays.
To enable this port, select the Output deltadelta parameter.
Data Types: single
 double
If a parameter is listed as tunable, then you can change its value during simulation.
Filter bank type
— Type of filter bankMel
(default)  Gammatone
Type of filter bank, specified as either Mel
or
Gammatone
:
Mel
–– The block computes the mel
frequency cepstral coefficients (MFCC).
Gammatone
–– The block computes the
gammatone cepstral coefficients (GTCC).
Tunable: No
Domain of the input signal
— Input signal domainTime
(default)  Frequency
Input signal domain, specified as either Time
or Frequency
.
Tunable: No
Number of coefficients to return
— Number of coefficients to return13
(default)  positive integerNumber of coefficients to return, specified as an integer in the range [2, v], where v is the number of valid passbands. The number of valid passbands depends on the type of filter bank:
Mel
–– The number of valid
passbands is defined as sum(κ <=
floor(fs/2))2
, where κ
is the
number of band edges in the mel filter bank and
fs
is the sample rate.
Gammatone
–– The number of valid
passbands is defined as ceil(
,
where hz2erb
(R(2))hz2erb
(R(1)))R
is the frequency range of the
gammatone filter bank.
Tunable: No
Data Types: single
 double
Nonlinear rectification
— Type of nonlinear rectificationLog
(default)  CubicRoot
Type of nonlinear rectification applied prior to the discrete cosine transform.
Tunable: No
Inherit FFT length from input dimensions
— Inherit FFT length from inputon
(default)  off
When you select this parameter, the FFT length is equal to the number of rows in the input signal.
Tunable: No
To enable this parameter, set Domain of the input
signal to Time
.
FFTLength
— FFT length[]
(default)  positive integerFFT length, specified as a positive integer. The default,
[]
, means that the FFT length is equal to the number
of rows in the input signal.
Tunable: No
To enable this parameter, set Domain of the input
signal to Time
and select the
Inherit FFT length from input dimensions
parameter.
Data Types: single
 double
 int8
 int16
 int32
 int64
 uint8
 uint16
 uint32
 uint64
Log energy usage
— Specify how the log energy is shownAppend
(default)  Replace
 Ignore
Specify how the log energy is shown in the coefficients vector output, specified as:
Append
–– The block prepends the
log energy to the coefficients vector. The length of the
coefficients vector is 1 + NumCoeffs, where
NumCoeffs is the value specified in the
Number of coefficients to return
parameter.
Replace
–– The block replaces the
first coefficient with the log energy of the signal. The length
of the coefficients vector is
NumCoeffs.
Ignore
–– The block does not
calculate or return the log energy.
Tunable: No
Output delta
— Output delta valuesoff
(default)  on
When you select this parameter, an additional output port, delta, is added to the block. This port outputs the change in coefficients over consecutive calls to the algorithm.
Tunable: No
Output deltadelta
— Output deltadelta valuesoff
(default)  on
When you select this parameter, an additional output port, deltaDelta, is added to the block. This port outputs the change in delta values over consecutive calls to the algorithm.
Tunable: No
Inherit sample rate from input
— Specify source of input sample rateoff
(default)  on
When you select this parameter, the block inherits its sample rate from the input signal. When you clear this parameter, you specify the sample rate in Input sample rate (Hz) parameter.
Tunable: No
Input sample rate (Hz)
— Sample rate of input16000
(default)  positive scalarInput sample rate in Hz, specified as a real positive scalar.
To enable this parameter, clear the Inherit sample rate from input parameter.
Simulate using
— Specify type of simulation to runCode generation
(default)  Interpreted execution
Code generation
–– Simulate model using
generated C code. The first time you run a simulation, Simulink^{®} generates C code for the block. The C code is reused
for subsequent simulations, as long as the model does not change.
This option requires additional startup time, but the speed of the
subsequent simulations is comparable to Interpreted
execution
.
Interpreted execution
–– Simulate model
using the MATLAB^{®} interpreter. This option shortens startup time but has
a slower simulation speed than Code
generation
. In this mode, you can debug the source
code of the block.
Tunable: No
Gammatone frequency range (Hz)
— Frequency range of gammatone filter bank (Hz)[50 8000]
(default)  twoelement row vectorFrequency range of the gammatone filter bank in Hz, specified as a positive, monotonically increasing twoelement row vector. The maximum frequency range can be any finite number. The center frequencies of the filter bank are equally spaced across the frequency range on the ERB scale.
Tunable: No
To enable this parameter, set Filter bank type to
Gammatone
.
Band edges of Mel filter bank (Hz)
— Band edges of mel filter bankBand edges of the filter bank in Hz, specified as a nonnegative monotonically increasing row vector in the range [0, ∞). The maximum bandedge frequency can be any finite number. The number of bandedges must be in the range [4, 80].
The default band edges are spaced linearly for the first ten and then logarithmically thereafter. The default band edges are set as recommended by [1].
Tunable: No
To enable this parameter, set Filter bank type to
Mel
.
Domain for Mel filter bank design
— Mel filter bank design domainHz
(default)  Bin
Mel filter bank design domain, specified as either
Hz
or Bin
. The
filterbank is designed as overlapped triangles with band edges specified by
the Band edges of filter bank (Hz) parameter.
The band edges are specified in Hz. When you set the design domain to:
Tunable: No
To enable this parameter, set Filter bank type to
Mel
.
Filter bank normalization
— Normalize filter bankBandwidth
(default)  Area
 None
Normalization technique used to normalize the weights of the filter bank, specified as:
Bandwidth
–– The weights of each
bandpass filter are normalized by the corresponding bandwidth of
the filter.
Area
–– The weights of each
bandpass filter are normalized by the corresponding area of the
bandpass filter.
None
–– The weights of the filter
are not normalized.
Tunable: No
Data Types 

Direct Feedthrough 

Multidimensional Signals 

VariableSize Signals 

ZeroCrossing Detection 

Auditory cepstrum coefficients are popular features extracted from speech signals for use in recognition tasks. In the sourcefilter model of speech, cepstral coefficients are understood to represent the filter (vocal tract). The vocal tract frequency response is relatively smooth, whereas the source of voiced speech can be modeled as an impulse train. As a result, the vocal tract can be estimated by the spectral envelope of a speech segment.
The motivating idea of cepstral coefficients is to compress information about the vocal tract (smoothed spectrum) into a small number of coefficients based on an understanding of the cochlea. Although there is no hard standard for calculating the coefficients, the basic steps are outlined by the diagram.
Two popular implementations of the filter bank are the mel filter bank and the gammatone filter bank.
The default mel filter bank linearly spaces the first 10 triangular filters and logarithmically spaces the remaining filters.
The default gammatone filter bank is composed of gammatone filters spaced linearly
on the ERB scale between 50 and 8000 Hz. The filter bank is designed by gammatoneFilterBank
.
If the input (x) is a timedomain signal, the log energy is computed using the following equation:
$$\mathrm{log}E=\mathrm{log}(\text{sum}({x}^{2}))$$
If the input (x) is a frequencydomain signal, the log energy is computed using the following equation:
$$\mathrm{log}E=\mathrm{log}\left(\text{sum}\left({\leftx\right}^{2}\right)/FFTLength\right)$$
[1] Auditory Toolbox. https://engineering.purdue.edu/~malcolm/interval/1998010/AuditoryToolboxTechReport.pdf
[2] ETSI ES 201 108 V1.1.3 (200309). https://www.etsi.org/deliver/etsi_es/201100_201199/201108/01.01.03_60/es_201108v010103p.pdf
Voice Activity
Detector  cepstralFeatureExtractor
 mfcc
 pitch
 voiceActivityDetector
You clicked a link that corresponds to this MATLAB command:
Run the command by entering it in the MATLAB Command Window. Web browsers do not support MATLAB commands.
Choose a web site to get translated content where available and see local events and offers. Based on your location, we recommend that you select: .
Select web siteYou can also select a web site from the following list:
Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.