Detail publikace

Smooth Pitch Tracker Based on Harmonic and Noise Model

SZŐKE, I.

Originální název

Smooth Pitch Tracker Based on Harmonic and Noise Model

Typ

článek ve sborníku mimo WoS a Scopus

Jazyk

angličtina

Originální abstrakt

The paper deals with an approach to a smooth and precise fundamental frequency estimator with voiced/unvoiced decision. Our method is based on a parametric Harmonic and Noise Model used especially in speech modification and synthesis. Errors between speech signal and resynthesized ones with different estimations of fundamental frequencies are computed. Dynamic programing is used to retrieve the best path in an "error map". It uses a constant for setting smoothness of the path. Reliable voiced/unvoiced segmentation can be done using thresholding of the error.

Klíčová slova

Pitch, tracking, harmonic and noise model, speech, fundamental frequency

Autoři

SZŐKE, I.

Rok RIV

2005

Vydáno

16. 4. 2005

Nakladatel

Faculty of Information Technology BUT

Místo

Brno

ISBN

80-214-2890-2

Kniha

STUDENT EEICT 2005

Strany od

673

Strany do

677

Strany počet

5

URL

BibTex

@inproceedings{BUT18910,
  author="Igor {Szőke}",
  title="Smooth Pitch Tracker Based on Harmonic and Noise Model",
  booktitle="STUDENT EEICT 2005",
  year="2005",
  pages="673--677",
  publisher="Faculty of Information Technology BUT",
  address="Brno",
  isbn="80-214-2890-2",
  url="https://www.fit.vutbr.cz/~szoke/papers/eeict_2005.pdf"
}