Paper Info
Title
Revisiting scenarios and methods for variable frame rate analysis in automatic speech recognition
Abstract
In this paper we present a revision and evaluation of some of the main methods used in variable frame rate (VFR) analysis, applied to speech recognition systems. The work found in the literature in this area usually deals with restricted conditions and scenarios and we have revisited the main algorithmic alter- natives and evaluated them under the same experimental frame- work, so that we have been able to establish objective consider- ations for each of them, selecting the most adequate strategy. We also show till what extent VFR analysis is useful in its three main application scenarios, namely "reduction of com- putational load", "improve acoustic modelling" and "handling additive noise conditions in the time domain". From our evalu- ation on a difficult telephone large vocabulary task, we establish that VFR analysis does not significantly improve the results ob- tained using the traditional fixed frame rate analysis (FFR), ex- cept when additive noise is present in the database and specially for low SNRs.
Year
Venue
Keywords
2003
INTERSPEECH
speech recognition,automatic speech recognition,time domain
Field
DocType
Citations 
Time domain,Computer science,Speech recognition,Frame rate,Artificial intelligence,Variable frame rate,Vocabulary,Machine learning
Conference
10
PageRank 
References 
Authors
0.95
4
6
Name
Order
Citations
PageRank
Javier Macías Guarasa113825.19
J. Ordonez2121.81
Juan Manuel Montero321831.51
Javier Ferreiros4100.95
Ricardo De Córdoba514225.58
Luis Fernando D'Haro618125.97