Finding line spectral frequencies using the fast fourier transform

Tom Bäckström; Christian Fischer Pedersen; Johannes Fischer; Grzegorz Pietrzyk

doi:10.1109/ICASSP.2015.7178947

Finding line spectral frequencies using the fast fourier transform

Tom Bäckström, Christian Fischer Pedersen, Johannes Fischer, Grzegorz Pietrzyk

Not published at Aalto University

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

1 Citation (Scopus)

Abstract

Main-stream speech codecs are based on modelling the speech source by a linear predictor. An efficient domain for quantization and coding of this linear predictor is the line spectral frequency representation, where the predictor is encoded into an ordered set of frequencies that correspond to the roots of the corresponding line spectral polynomials. While this representation is robust in terms of quantization, methods available for finding the line spectral frequencies are computationally complex. In this work, we present a method for finding these frequencies using the FFT, including methods for limiting numerical range in fixed-point implementations. Our experiments show that, in comparison to a zero-crossing search in the Chebyshev domain, the proposed method reduces complexity and improves robustness, while retaining accuracy.

Original language	English
Title of host publication	2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings
Publisher	IEEE
Pages	5122-5126
Number of pages	5
Volume	2015-August
ISBN (Electronic)	9781467369978
DOIs	https://doi.org/10.1109/ICASSP.2015.7178947
Publication status	Published - 1 Jan 2015
MoE publication type	A4 Article in a conference publication
Event	IEEE International Conference on Acoustics, Speech, and Signal Processing - Brisbane, Australia Duration: 19 Apr 2015 → 24 Apr 2015 Conference number: 40

Conference

Conference	IEEE International Conference on Acoustics, Speech, and Signal Processing
Abbreviated title	ICASSP
Country	Australia
City	Brisbane
Period	19/04/2015 → 24/04/2015

Keywords

line spectral frequencies
linear prediction
root finding
speech coding

Access to Document

10.1109/ICASSP.2015.7178947

Link to publication in Scopus

Fingerprint Dive into the research topics of 'Finding line spectral frequencies using the fast fourier transform'. Together they form a unique fingerprint.

Cite this

@inproceedings{627278e9e6434390b2de807e19c999b2,

title = "Finding line spectral frequencies using the fast fourier transform",

abstract = "Main-stream speech codecs are based on modelling the speech source by a linear predictor. An efficient domain for quantization and coding of this linear predictor is the line spectral frequency representation, where the predictor is encoded into an ordered set of frequencies that correspond to the roots of the corresponding line spectral polynomials. While this representation is robust in terms of quantization, methods available for finding the line spectral frequencies are computationally complex. In this work, we present a method for finding these frequencies using the FFT, including methods for limiting numerical range in fixed-point implementations. Our experiments show that, in comparison to a zero-crossing search in the Chebyshev domain, the proposed method reduces complexity and improves robustness, while retaining accuracy.",

keywords = "line spectral frequencies, linear prediction, root finding, speech coding",

author = "Tom B{\"a}ckstr{\"o}m and Pedersen, {Christian Fischer} and Johannes Fischer and Grzegorz Pietrzyk",

year = "2015",

month = jan,

day = "1",

doi = "10.1109/ICASSP.2015.7178947",

language = "English",

volume = "2015-August",

pages = "5122--5126",

booktitle = "2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings",

publisher = "IEEE",

address = "United States",

note = "IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP ; Conference date: 19-04-2015 Through 24-04-2015",

}

Bäckström, T, Pedersen, CF, Fischer, J & Pietrzyk, G 2015, Finding line spectral frequencies using the fast fourier transform. in 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings. vol. 2015-August, 7178947, IEEE, pp. 5122-5126, IEEE International Conference on Acoustics, Speech, and Signal Processing, Brisbane, Australia, 19/04/2015. https://doi.org/10.1109/ICASSP.2015.7178947

Finding line spectral frequencies using the fast fourier transform. / Bäckström, Tom; Pedersen, Christian Fischer; Fischer, Johannes; Pietrzyk, Grzegorz.

2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings. Vol. 2015-August IEEE, 2015. p. 5122-5126 7178947.

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

TY - GEN

T1 - Finding line spectral frequencies using the fast fourier transform

AU - Bäckström, Tom

AU - Pedersen, Christian Fischer

AU - Fischer, Johannes

AU - Pietrzyk, Grzegorz

N1 - Conference code: 40

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Main-stream speech codecs are based on modelling the speech source by a linear predictor. An efficient domain for quantization and coding of this linear predictor is the line spectral frequency representation, where the predictor is encoded into an ordered set of frequencies that correspond to the roots of the corresponding line spectral polynomials. While this representation is robust in terms of quantization, methods available for finding the line spectral frequencies are computationally complex. In this work, we present a method for finding these frequencies using the FFT, including methods for limiting numerical range in fixed-point implementations. Our experiments show that, in comparison to a zero-crossing search in the Chebyshev domain, the proposed method reduces complexity and improves robustness, while retaining accuracy.

AB - Main-stream speech codecs are based on modelling the speech source by a linear predictor. An efficient domain for quantization and coding of this linear predictor is the line spectral frequency representation, where the predictor is encoded into an ordered set of frequencies that correspond to the roots of the corresponding line spectral polynomials. While this representation is robust in terms of quantization, methods available for finding the line spectral frequencies are computationally complex. In this work, we present a method for finding these frequencies using the FFT, including methods for limiting numerical range in fixed-point implementations. Our experiments show that, in comparison to a zero-crossing search in the Chebyshev domain, the proposed method reduces complexity and improves robustness, while retaining accuracy.

KW - line spectral frequencies

KW - linear prediction

KW - root finding

KW - speech coding

UR - http://www.scopus.com/inward/record.url?scp=84946015977&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2015.7178947

DO - 10.1109/ICASSP.2015.7178947

M3 - Conference contribution

AN - SCOPUS:84946015977

VL - 2015-August

SP - 5122

EP - 5126

BT - 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings

PB - IEEE

T2 - IEEE International Conference on Acoustics, Speech, and Signal Processing

Y2 - 19 April 2015 through 24 April 2015

ER -