DECRIPTION_CONFIG_FILE

Decription of the configuration parameters

% Signal Parameters
SAMPLE_RATE             % Sampling Rate
FDLP_LEN                % FDLP frame length in seconds (1 second)
FLAG_FULL_SIG           % Using the entire signal in one frame length (useful for short speech files of 2-3s)

% Type of Feature
FEAT_TYPE               % Type of features (fdlps=1/fdlpm=2/fdlp_plp2=3)
AXIS                    % Frequency Axis   (bark=1/mel=2/linear=3)
MEL_WARP                % Applies only for converting linear to mel-scale
DCT_LOW                 % Begin of Frequency Axis in DCT domain
DCT_HIGH                % End of Frequency Axis in DCT domain
SKIP_BANDS              % Skip the initial bands in feature computation

% Feature configuration
FLAG_GAIN_NORM          % Gain normalization
MODEL_ORDER             % Model order per sub-band per second
FLAG_LP_TYPE            % Type of linear prediction (autocorr=1/ls=2)
FLAG_WIENER             % Apply Wiener filtering in feature

% Modulation feature configuration
TEMP_CEPS               % Number of modulation components per band
FLAG_ADAPT              % Including static and adaptive compression
                        % Requires mex file adapt_m

% Spectral feature parameters
SPEC_CEPST              % Number of cepstral components
FLAG_C0                 % Flag set to 1 for using C1-C13
SPEC_FR_LEN             % Frame length for spectral frame (ms)
SPEC_FR_SHIFT           % Frame shift for spectral fram (ms)
FLAG_DELTA              % Apply Delta and Acceleration