Namespace AudioAnalysisTools.DSP
Classes
Clipping
TODO: This class should be Unit tested on a variety of clipped recordings. TODO: The calculations employed in this class to estimate clipping need to be revisted. Not clear what to do due to resampling. Estimates of clipping are complicated by the fact that down sampling greatly reduces the degree of clipping in a recording. Therefore it is difficult to know how much of the original recording was clipped after it has been downsampled. The assumption in the current calculations is that we want to know that a recording was clipped before it was subsequently processed.
DSP_Frames
Digital signal processing methods.
DSP_Frames.EnvelopeAndFft
DSP_IIRFilter
digital signal processing FILTERS methods
"Finite impulse response" (FIR) filters use only the input signals, while an "infinite impulse response" filter (IIR) uses both the input signal and previous samples of the output signal. FIR filters are always stable, while IIR filters may be unstable.
DspFilters
digital signal processing FILTERS methods.
FeatureExtraction
This class is designed to extract clustering features for target input recordings.
FeatureLearning
This class is designed to learn bases (cluster centroids) through feature learning process.
FeatureLearningSettings
FFT2D
Performs two dimensional FFT on a matrix of values. IMPORTANT: The matrix passed to this class for performing of 2D FFT need not necessarily have width equal to height but both width and height MUST be a power of two.
FrequencyScale
KmeansClustering
KmeansClustering.Output
LocalContrastNormalisation
Performs local contrast normalisation on a matrix of values where the matrix is assumed to be derived from an image.
MFCCStuff
NoiseProfile
contains info re noise profile of an entire spectrogram.
NoiseRemoval_Briggs
NoiseRemovalModal
OctaveFreqScale
PatchSampling
PcaWhitening
PcaWhitening.Output
Outputting the Projection Matrix, whitened matrix, eigen vectors, and the number of PCA components that is used to to transform the data into the new feature subspace. in Accord.net, this matrix is called "ComponentVectors", which its columns contain the principle components, known as Eigenvectors.
PowerSpectralDensity
SNR
SNR.BackgroundNoise
SNR.SnrStatistics
used to store info about the SNR in a signal using db units.
WavInfo
Wav Info.
WavWriter
Enums
FreqScaleType
All the below octave scale options are designed for a final freq scale having 256 bins. Scale name indicates its structure. You cannot vary the structure.
NoiseReductionType
PatchSampling.SamplingMethod
sample a set of patches ("sequential" or "random" or "overlapped random") from a spectrogram in "sequential" mode, it generates non-overlapping patches from the whole input matrix, and in this case the "numOfPatches" can be simply set to zero. However, in "random" mode, the method requires an input for the "numOfPatches" parameter.