Synthesizes the harmonic part using the linear phase interpolation and phase unwrapping technique described in: Stylianou, Y., 1996, "Harmonic plus Noise Models for Speech, combined with Statistical Methods, for Speech and Speaker Modification", Ph.D.
Synthesis using harmonics plus noise (and possibly plus transients) model.
A time-domain LP synthesis filter based version of the HNM noise part synthesis algorithm described in: Reference: Stylianou, Y., 1996, "Harmonic plus Noise Models for Speech, combined with Statistical Methods, for Speech and Speaker Modification", Ph.D.
A pseudo-harmonic representation based synthesizer for the noise part.
Synthesizes noise part waveform from non-overlapping chunks of data.
LPC based noise model for HNM using the overlap-add approach, gain normalization according to generated harmonic part gain, and optional triangular energy envelope weighting.
Syntehsizer for the transient part waveform segments.
Copyright © 2000–2018 DFKI GmbH. All rights reserved.