| Package | Description |
|---|---|
| marytts.machinelearning |
Machine learning classes for K-Means clustering, Gaussian Mixture
Models, and manual data generation for testing purposes.
|
| marytts.signalproc |
Signal processing package for Mary consisting of the following
sub-packages:
adaptation: speaker adaptation using various voice conversion
techniques
analysis: feature estimation and analysis
demo: demonstrator with a gui
display: visualization utilities
effects: audio effects for speech and text-to-speech synthesis
output
filter: filter design and filtering utilities
process: various speech signal modification utilities
sinusoidal: sinusoidal speech models and modification
tests: testing tools for signal processing classes
window: various windowing functions
|
| marytts.signalproc.adaptation |
Packages for speaking style and speaker identity adaptation in Mary TTS
supporting various voice conversion algorithms.
|
| marytts.signalproc.adaptation.codebook |
Weighted codebook based voice conversion algorithms.
|
| marytts.signalproc.adaptation.gmm |
Gaussian Mixture Model based voice conversion algorithms.
|
| marytts.signalproc.adaptation.gmm.jointgmm |
Joint source-target Gaussian Mixture Model based voice conversion algorithms.
|
| marytts.signalproc.adaptation.outlier |
Outlier elimination algorithms for voice conversion.
|
| marytts.signalproc.adaptation.prosody |
Prosody transformation algorithms for voice conversion.
A prosody modification framework has been implemented which supports: Mean and standard deviation transformation of f0 Sentence slope transformation Mean and standard deviation transformation is the best method so far. Duration and energy transformation have not yet been implemented. |
| marytts.signalproc.adaptation.smoothing |
Smoothing algorithms for voice conversion.
|
| marytts.signalproc.adaptation.test | |
| marytts.signalproc.analysis |
A collection of analysis algorithms for signal processing.
Important classes are as follows: LpcAnalyser: Linear prediction analysis using autocorrelation appraoch and Durbin recursion LsfAnalyser: Computation of line spectral frequencies (LSFs, or line spectral pairs - LSPs) based on LpcAnalyser EnergyAnalyser: Energy contour estimation with voice activity detection support F0TrackerAutocorrelationHeuristic: An autocorrelation based f0 analysis algorithm extended with heuristic post-processing to reduce voiced/unvoiced errors and f0 doubling/halving problems. |
| marytts.signalproc.analysis.distance |
A collection of popular distance measures in speech processing.
|
| marytts.signalproc.display | |
| marytts.signalproc.filter |
Various classes that support filter and filterbank design and filtering operations.
|
| marytts.signalproc.process | |
| marytts.signalproc.sinusoidal |
Sinusoidal analysis/synthesis framework supporting various approaches:
Conventional sinusoidal analysis/modification/synthesis
Harmonic plus Noise Model (HNM) based analysis/modification/synthesis
Multiresolution sinusoidal analysis/modification/snythesis
Warning: This is a very basic implementation and it does not work properly! Sines+transients+noise based analysis/modification/synthesis |
| marytts.signalproc.sinusoidal.hntm.analysis |
Analysis package for harmonics plus noise speech models.
|
| marytts.signalproc.sinusoidal.hntm.analysis.pitch |
Pitch and voicing analysis package for harmonics plus noise speech models.
|
| marytts.signalproc.sinusoidal.hntm.modification |
PSOLA like prosody modification algorithms for harmonics plus noise models.
|
| marytts.signalproc.sinusoidal.hntm.synthesis |
Synthesis package for harmonics plus noise model consisting of the
following modules:
HarmonicPartLinearPhaseInterpolatorSynthesizer: harmonic part
synthesis with a linear phase interpolator
NoisePartWaveformSynthesizer: Noise part synthesizer when the
noise is kept as original-harmonic waveform
NoisePartLpFilterPostHpfLpcSynthesizer: Noise part synthesizer
using linear prediction forward filter with optional post filtering
with an highpass filter
NoisePartWindowedOverlapAddLpcSynthesizer: Noise part synthesizer using a windowed overlap add approach (supports highpass filtering as well) NoisePartPseudoHarmonicSynthesizer: A pseudo-harmonic approach for noise part generation using parameters obtained by the harmonic part analysis algorithm applied to noise part assuming a fixed f0 TransientPartSynthesizer: A waveform synthesizer for transient parts (performs windowing at transition boundaries) |
| marytts.signalproc.sinusoidal.hntm.synthesis.hybrid | |
| marytts.signalproc.sinusoidal.pitch |
Experimental pitch tracker using the comb filter approach for sinusoidal models.
|
| marytts.signalproc.sinusoidal.test |
A test signal generation package for sinusoidal and harmonics plus noise models.
|
| marytts.signalproc.window | |
| marytts.tools.analysis | |
| marytts.util.data | |
| marytts.util.data.audio |
Various relatively generic utilities for audio input/output.
|
| marytts.util.data.text |
Various relatively generic utilities for text input/output.
|
| marytts.util.display | |
| marytts.util.math |
Various relatively generic utilities for maths.
|
| marytts.util.signal |
Various relatively generic utilities for signal processing.
|
| org.jsresources |
Copyright © 2000–2022 DFKI GmbH. All rights reserved.