Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study

Petr Pollak; Michal Borsky

doi:10.1007/978-3-642-35755-8_29

Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study

Petr Pollak, Michal Borsky

Department of Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Citations (Scopus)

Abstract

This paper presents the study of speech recognition accuracy both for small and large vocabulary task with respect to different levels of MP3 compression of processed data. The motivation behind the work was to evaluate the usage of ASR system for off-line automatic transcription of recordings collected from standard present MP3 devices under different levels of background noise and channel distortion. Although MP3 may not be an optimal compression algorithm, the performed experiments have prooved that it does not distort speech signal significantly for higher compression rates. Realized experiments showed also that the accuracy of speech recognition (both small- and large-vocabulary) decreased very slowly for the bit-rate of 24 kbps and higher. However, slightly different setup of speech feature computation is necessary for MP3 speech data, mainly PLP features give significantly better results in comparison to MFCC.

Original language	English
Title of host publication	E-Business and Telecommunications
Subtitle of host publication	International Joint Conference, ICETE 2011 Seville, Spain, July 18-21, 2011 Revised Selected Papers
Editors	Jose L. Sevillano, Joaquim Filipe
Pages	409-419
Number of pages	11
DOIs	https://doi.org/10.1007/978-3-642-35755-8_29
Publication status	Published - 2012
Event	8th International Joint Conference on e-Business and Telecommunications, ICETE 2011 - Seville, Spain Duration: 18 Jul 2011 → 21 Jul 2011

Publication series

Name	Communications in Computer and Information Science
Volume	314
ISSN (Print)	1865-0929

Conference

Conference	8th International Joint Conference on e-Business and Telecommunications, ICETE 2011
Country/Territory	Spain
City	Seville
Period	18/07/11 → 21/07/11

Other keywords

Large vocabulary
LVCSR
MP3
MPEG compression
Noise robustness
Small vocabulary
Speech recognition

Access to Document

10.1007/978-3-642-35755-8_29

Cite this

Pollak, P., & Borsky, M. (2012). Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study. In J. L. Sevillano, & J. Filipe (Eds.), E-Business and Telecommunications: International Joint Conference, ICETE 2011 Seville, Spain, July 18-21, 2011 Revised Selected Papers (pp. 409-419). (Communications in Computer and Information Science; Vol. 314). https://doi.org/10.1007/978-3-642-35755-8_29

Pollak, Petr ; Borsky, Michal. / Small and large vocabulary speech recognition of MP3 data under real-word conditions : Experimental study. E-Business and Telecommunications: International Joint Conference, ICETE 2011 Seville, Spain, July 18-21, 2011 Revised Selected Papers. editor / Jose L. Sevillano ; Joaquim Filipe. 2012. pp. 409-419 (Communications in Computer and Information Science).

@inproceedings{1dc371c519904346bdc8a25557a5f390,

title = "Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study",

abstract = "This paper presents the study of speech recognition accuracy both for small and large vocabulary task with respect to different levels of MP3 compression of processed data. The motivation behind the work was to evaluate the usage of ASR system for off-line automatic transcription of recordings collected from standard present MP3 devices under different levels of background noise and channel distortion. Although MP3 may not be an optimal compression algorithm, the performed experiments have prooved that it does not distort speech signal significantly for higher compression rates. Realized experiments showed also that the accuracy of speech recognition (both small- and large-vocabulary) decreased very slowly for the bit-rate of 24 kbps and higher. However, slightly different setup of speech feature computation is necessary for MP3 speech data, mainly PLP features give significantly better results in comparison to MFCC.",

keywords = "Large vocabulary, LVCSR, MP3, MPEG compression, Noise robustness, Small vocabulary, Speech recognition",

author = "Petr Pollak and Michal Borsky",

year = "2012",

doi = "10.1007/978-3-642-35755-8_29",

language = "English",

isbn = "9783642357541",

series = "Communications in Computer and Information Science",

pages = "409--419",

editor = "Sevillano, {Jose L.} and Joaquim Filipe",

booktitle = "E-Business and Telecommunications",

note = "8th International Joint Conference on e-Business and Telecommunications, ICETE 2011 ; Conference date: 18-07-2011 Through 21-07-2011",

}

Pollak, P & Borsky, M 2012, Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study. in JL Sevillano & J Filipe (eds), E-Business and Telecommunications: International Joint Conference, ICETE 2011 Seville, Spain, July 18-21, 2011 Revised Selected Papers. Communications in Computer and Information Science, vol. 314, pp. 409-419, 8th International Joint Conference on e-Business and Telecommunications, ICETE 2011, Seville, Spain, 18/07/11. https://doi.org/10.1007/978-3-642-35755-8_29

Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study. / Pollak, Petr; Borsky, Michal.
E-Business and Telecommunications: International Joint Conference, ICETE 2011 Seville, Spain, July 18-21, 2011 Revised Selected Papers. ed. / Jose L. Sevillano; Joaquim Filipe. 2012. p. 409-419 (Communications in Computer and Information Science; Vol. 314).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Small and large vocabulary speech recognition of MP3 data under real-word conditions

T2 - 8th International Joint Conference on e-Business and Telecommunications, ICETE 2011

AU - Pollak, Petr

AU - Borsky, Michal

PY - 2012

Y1 - 2012

N2 - This paper presents the study of speech recognition accuracy both for small and large vocabulary task with respect to different levels of MP3 compression of processed data. The motivation behind the work was to evaluate the usage of ASR system for off-line automatic transcription of recordings collected from standard present MP3 devices under different levels of background noise and channel distortion. Although MP3 may not be an optimal compression algorithm, the performed experiments have prooved that it does not distort speech signal significantly for higher compression rates. Realized experiments showed also that the accuracy of speech recognition (both small- and large-vocabulary) decreased very slowly for the bit-rate of 24 kbps and higher. However, slightly different setup of speech feature computation is necessary for MP3 speech data, mainly PLP features give significantly better results in comparison to MFCC.

AB - This paper presents the study of speech recognition accuracy both for small and large vocabulary task with respect to different levels of MP3 compression of processed data. The motivation behind the work was to evaluate the usage of ASR system for off-line automatic transcription of recordings collected from standard present MP3 devices under different levels of background noise and channel distortion. Although MP3 may not be an optimal compression algorithm, the performed experiments have prooved that it does not distort speech signal significantly for higher compression rates. Realized experiments showed also that the accuracy of speech recognition (both small- and large-vocabulary) decreased very slowly for the bit-rate of 24 kbps and higher. However, slightly different setup of speech feature computation is necessary for MP3 speech data, mainly PLP features give significantly better results in comparison to MFCC.

KW - Large vocabulary

KW - LVCSR

KW - MP3

KW - MPEG compression

KW - Noise robustness

KW - Small vocabulary

KW - Speech recognition

UR - http://www.scopus.com/inward/record.url?scp=84882271145&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-35755-8_29

DO - 10.1007/978-3-642-35755-8_29

M3 - Conference contribution

AN - SCOPUS:84882271145

SN - 9783642357541

T3 - Communications in Computer and Information Science

SP - 409

EP - 419

BT - E-Business and Telecommunications

A2 - Sevillano, Jose L.

A2 - Filipe, Joaquim

Y2 - 18 July 2011 through 21 July 2011

ER -

Pollak P, Borsky M. Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study. In Sevillano JL, Filipe J, editors, E-Business and Telecommunications: International Joint Conference, ICETE 2011 Seville, Spain, July 18-21, 2011 Revised Selected Papers. 2012. p. 409-419. (Communications in Computer and Information Science). doi: 10.1007/978-3-642-35755-8_29

Small and large vocabulary speech recognition of MP3 data under real-word conditions: Experimental study

Abstract

Publication series

Conference

Other keywords

Access to Document

Fingerprint

Cite this