Quality Improvements of Zero-Concatenation-Cost Chain Based Unit Selection

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F14%3A43923340" target="_blank" >RIV/49777513:23520/14:43923340 - isvavai.cz</a>

Výsledek na webu

<a href="http://link.springer.com/chapter/10.1007%2F978-3-319-11581-8_47" target="_blank" >http://link.springer.com/chapter/10.1007%2F978-3-319-11581-8_47</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-319-11581-8_47" target="_blank" >10.1007/978-3-319-11581-8_47</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Quality Improvements of Zero-Concatenation-Cost Chain Based Unit Selection

Popis výsledku v původním jazyce

In our previous work, we introduced a zero-concatenation-cost (ZCC) chain based framework of unit-selection speech synthesis. This framework proved to be very fast as it reduced the computational load of a unit-selection system up to hundreds of time. Since the ZCC chain based algorithm principally prefers to select longer segments of speech, an increased number of audible artifacts were expected to occur at concatenation points of longer ZCC chains. Indeed, listening tests revealed a number of artifacts present in synthetic speech; however, the artifacts occurred in a similar extent in synthetic speech produced by both ZCC chain based and standard Viterbi search algorithms. In this paper, we focus on the sources of the artifacts and we propose improvements of the synthetic speech quality within the ZCC algorithm. The quality and computational demands of the improved ZCC algorithm are compared to the unit-selection algorithm based on the standard Viterbi search.

Název v anglickém jazyce

Quality Improvements of Zero-Concatenation-Cost Chain Based Unit Selection

Popis výsledku anglicky

In our previous work, we introduced a zero-concatenation-cost (ZCC) chain based framework of unit-selection speech synthesis. This framework proved to be very fast as it reduced the computational load of a unit-selection system up to hundreds of time. Since the ZCC chain based algorithm principally prefers to select longer segments of speech, an increased number of audible artifacts were expected to occur at concatenation points of longer ZCC chains. Indeed, listening tests revealed a number of artifacts present in synthetic speech; however, the artifacts occurred in a similar extent in synthetic speech produced by both ZCC chain based and standard Viterbi search algorithms. In this paper, we focus on the sources of the artifacts and we propose improvements of the synthetic speech quality within the ZCC algorithm. The quality and computational demands of the improved ZCC algorithm are compared to the unit-selection algorithm based on the standard Viterbi search.

Druh

D - Stať ve sborníku

CEP obor

JD - Využití počítačů, robotika a její aplikace

OECD FORD obor

—

Projekt

<a href="/cs/project/TA01030476" target="_blank" >TA01030476: Inteligentní technologie pro zvýšení bezpečnosti letového provozu</a><br>

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2014

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Speech and Computer, 16th International Conference, SPECOM 2014, Novi Sad, Serbia, October 5-9, 2014, Proceedings

ISBN

978-3-319-11580-1

ISSN

0302-9743

e-ISSN

—

Počet stran výsledku

10

Strana od-do

376-385

Název nakladatele

Springer

Místo vydání

Heidelberg

Místo konání akce

Novi Sad, Serbia

Datum konání akce

5. 10. 2014

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—