A MFCC-Based CELP Speech Coder for Server-Based Speech Recognition in Network Environments

Jae Sam YOON; Gil Ho LEE; Hong Kook KIM

doi:10.1093/ietfec/e90-a.3.626

A MFCC-Based CELP Speech Coder for Server-Based Speech Recognition in Network Environments

Jae Sam YOON, Gil Ho LEE, Hong Kook KIM

Full Text Views

0

Share
Cite this

Summary :

Existing standard speech coders can provide high quality speech communication. However, they tend to degrade the performance of automatic speech recognition (ASR) systems that use the reconstructed speech. The main cause of the degradation is in that the linear predictive coefficients (LPCs), which are typical spectral envelope parameters in speech coding, are optimized to speech quality rather than to the performance of speech recognition. In this paper, we propose a speech coder using mel-frequency cepstral coefficients (MFCCs) instead of LPCs to improve the performance of a server-based speech recognition system in network environments. To develop the proposed speech coder with a low-bit rate, we first explore the interframe correlation of MFCCs, which results in the predictive quantization of MFCC. Second, a safety-net scheme is proposed to make the MFCC-based speech coder robust to channel errors. As a result, we propose an 8.7 kbps MFCC-based CELP coder. It is shown that the proposed speech coder has a comparable speech quality to 8 kbps G.729 and the ASR system using the proposed speech coder gives the relative word error rate reduction by 6.8% as compared to the ASR system using G.729 on a large vocabulary task (AURORA4).

Publication: IEICE TRANSACTIONS on Fundamentals Vol.E90-A No.3 pp.626-632

Publication Date: 2007/03/01

Publicized

Online ISSN: 1745-1337

DOI: 10.1093/ietfec/e90-a.3.626

Type of Manuscript: Special Section PAPER (Special Section on Multimedia and Mobile Signal Processing)

Category: Speech/Audio Processing

Cite this

Copy

Jae Sam YOON, Gil Ho LEE, Hong Kook KIM, "A MFCC-Based CELP Speech Coder for Server-Based Speech Recognition in Network Environments" in IEICE TRANSACTIONS on Fundamentals, vol. E90-A, no. 3, pp. 626-632, March 2007, doi: 10.1093/ietfec/e90-a.3.626.
Abstract: Existing standard speech coders can provide high quality speech communication. However, they tend to degrade the performance of automatic speech recognition (ASR) systems that use the reconstructed speech. The main cause of the degradation is in that the linear predictive coefficients (LPCs), which are typical spectral envelope parameters in speech coding, are optimized to speech quality rather than to the performance of speech recognition. In this paper, we propose a speech coder using mel-frequency cepstral coefficients (MFCCs) instead of LPCs to improve the performance of a server-based speech recognition system in network environments. To develop the proposed speech coder with a low-bit rate, we first explore the interframe correlation of MFCCs, which results in the predictive quantization of MFCC. Second, a safety-net scheme is proposed to make the MFCC-based speech coder robust to channel errors. As a result, we propose an 8.7 kbps MFCC-based CELP coder. It is shown that the proposed speech coder has a comparable speech quality to 8 kbps G.729 and the ASR system using the proposed speech coder gives the relative word error rate reduction by 6.8% as compared to the ASR system using G.729 on a large vocabulary task (AURORA4).
URL: https://globals.ieice.org/en_transactions/fundamentals/10.1093/ietfec/e90-a.3.626/_p

Copy

@ARTICLE{e90-a_3_626,
author={Jae Sam YOON, Gil Ho LEE, Hong Kook KIM, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={A MFCC-Based CELP Speech Coder for Server-Based Speech Recognition in Network Environments},
year={2007},
volume={E90-A},
number={3},
pages={626-632},
abstract={Existing standard speech coders can provide high quality speech communication. However, they tend to degrade the performance of automatic speech recognition (ASR) systems that use the reconstructed speech. The main cause of the degradation is in that the linear predictive coefficients (LPCs), which are typical spectral envelope parameters in speech coding, are optimized to speech quality rather than to the performance of speech recognition. In this paper, we propose a speech coder using mel-frequency cepstral coefficients (MFCCs) instead of LPCs to improve the performance of a server-based speech recognition system in network environments. To develop the proposed speech coder with a low-bit rate, we first explore the interframe correlation of MFCCs, which results in the predictive quantization of MFCC. Second, a safety-net scheme is proposed to make the MFCC-based speech coder robust to channel errors. As a result, we propose an 8.7 kbps MFCC-based CELP coder. It is shown that the proposed speech coder has a comparable speech quality to 8 kbps G.729 and the ASR system using the proposed speech coder gives the relative word error rate reduction by 6.8% as compared to the ASR system using G.729 on a large vocabulary task (AURORA4).},
keywords={},
doi={10.1093/ietfec/e90-a.3.626},
ISSN={1745-1337},
month={March},}

Copy

TY - JOUR
TI - A MFCC-Based CELP Speech Coder for Server-Based Speech Recognition in Network Environments
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 626
EP - 632
AU - Jae Sam YOON
AU - Gil Ho LEE
AU - Hong Kook KIM
PY - 2007
DO - 10.1093/ietfec/e90-a.3.626
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E90-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2007
AB - Existing standard speech coders can provide high quality speech communication. However, they tend to degrade the performance of automatic speech recognition (ASR) systems that use the reconstructed speech. The main cause of the degradation is in that the linear predictive coefficients (LPCs), which are typical spectral envelope parameters in speech coding, are optimized to speech quality rather than to the performance of speech recognition. In this paper, we propose a speech coder using mel-frequency cepstral coefficients (MFCCs) instead of LPCs to improve the performance of a server-based speech recognition system in network environments. To develop the proposed speech coder with a low-bit rate, we first explore the interframe correlation of MFCCs, which results in the predictive quantization of MFCC. Second, a safety-net scheme is proposed to make the MFCC-based speech coder robust to channel errors. As a result, we propose an 8.7 kbps MFCC-based CELP coder. It is shown that the proposed speech coder has a comparable speech quality to 8 kbps G.729 and the ASR system using the proposed speech coder gives the relative word error rate reduction by 6.8% as compared to the ASR system using G.729 on a large vocabulary task (AURORA4).
ER -