marytts.signalproc.sinusoidal.hntm.synthesis
Class NoisePartWaveformSynthesizer
java.lang.Object
marytts.signalproc.sinusoidal.hntm.synthesis.NoisePartWaveformSynthesizer
public class NoisePartWaveformSynthesizer
- extends java.lang.Object
Synthesizes noise part waveform from non-overlapping chunks of data.
This model is the most natural one since it involves no noise models.
- Author:
- oytun.turk
|
Method Summary |
static void |
processFrame(HntmSpeechFrame prevFrame,
HntmSpeechFrame currentFrame,
HntmSpeechFrame nextFrame,
int samplingRateInHz,
boolean isFirstSynthesisFrame,
boolean isLastSynthesisFrame,
double[] noisePartWaveform,
double[] wgts,
boolean existsLeftContexts,
HntmSpeechFrame currentLeftContext,
boolean existsRightContexts,
HntmSpeechFrame currentRightContext)
|
static double[] |
synthesize(HntmSpeechSignal hnmSignal,
HntmSpeechFrame[] leftContexts,
HntmSpeechFrame[] rightContexts,
HntmAnalyzerParams analysisParams)
|
| Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
NoisePartWaveformSynthesizer
public NoisePartWaveformSynthesizer()
synthesize
public static double[] synthesize(HntmSpeechSignal hnmSignal,
HntmSpeechFrame[] leftContexts,
HntmSpeechFrame[] rightContexts,
HntmAnalyzerParams analysisParams)
processFrame
public static void processFrame(HntmSpeechFrame prevFrame,
HntmSpeechFrame currentFrame,
HntmSpeechFrame nextFrame,
int samplingRateInHz,
boolean isFirstSynthesisFrame,
boolean isLastSynthesisFrame,
double[] noisePartWaveform,
double[] wgts,
boolean existsLeftContexts,
HntmSpeechFrame currentLeftContext,
boolean existsRightContexts,
HntmSpeechFrame currentRightContext)