A comparison of t test, F test, and coherence methods of detecting steady-state auditory-evoked potentials, distortion-product otoacoustic emissions, or other sinusoids

Robert A Dobie, Michael J. Wilson

Research output: Contribution to journalArticle

106 Citations (Scopus)

Abstract

Sinusoids in background noise can conveniently be detected using unsegmented power spectra, comparing power at the signal frequency to average power at several neighbor frequencies. In this case, the F test is preferable to t tests based on rms or dB values, because of the skewed distributions of rms and dB when signal-to-noise ratio (SNR)=0. F test performance improves as the number of frequencies increases, to about 15, but can be degraded if the background noise is not white, with a slope exceeding about 10 dB for the range of frequencies sampled. Segment analysis, using magnitude-squared coherence (MSC) or related statistics, has equivalent statistical power; MSC and F each yield unbiased SNR estimates that have identical distributions when SNR=0. Selection of F or MSC for detection of sinusoids will usually be a matter of convenience.

Original languageEnglish (US)
Pages (from-to)2236-2246
Number of pages11
JournalJournal of the Acoustical Society of America
Volume100
Issue number4
DOIs
StatePublished - Oct 1996

Fingerprint

sine waves
signal to noise ratios
background noise
products
performance tests
power spectra
statistics
slopes
Hearing
Evoked Potentials
estimates
Signal-to-noise Ratio

ASJC Scopus subject areas

  • Acoustics and Ultrasonics

Cite this

@article{a83f413c8fb14551b2988d46a3598d6b,
title = "A comparison of t test, F test, and coherence methods of detecting steady-state auditory-evoked potentials, distortion-product otoacoustic emissions, or other sinusoids",
abstract = "Sinusoids in background noise can conveniently be detected using unsegmented power spectra, comparing power at the signal frequency to average power at several neighbor frequencies. In this case, the F test is preferable to t tests based on rms or dB values, because of the skewed distributions of rms and dB when signal-to-noise ratio (SNR)=0. F test performance improves as the number of frequencies increases, to about 15, but can be degraded if the background noise is not white, with a slope exceeding about 10 dB for the range of frequencies sampled. Segment analysis, using magnitude-squared coherence (MSC) or related statistics, has equivalent statistical power; MSC and F each yield unbiased SNR estimates that have identical distributions when SNR=0. Selection of F or MSC for detection of sinusoids will usually be a matter of convenience.",
author = "Dobie, {Robert A} and Wilson, {Michael J.}",
year = "1996",
month = "10",
doi = "10.1121/1.417933",
language = "English (US)",
volume = "100",
pages = "2236--2246",
journal = "Journal of the Acoustical Society of America",
issn = "0001-4966",
publisher = "Acoustical Society of America",
number = "4",

}

TY - JOUR

T1 - A comparison of t test, F test, and coherence methods of detecting steady-state auditory-evoked potentials, distortion-product otoacoustic emissions, or other sinusoids

AU - Dobie, Robert A

AU - Wilson, Michael J.

PY - 1996/10

Y1 - 1996/10

N2 - Sinusoids in background noise can conveniently be detected using unsegmented power spectra, comparing power at the signal frequency to average power at several neighbor frequencies. In this case, the F test is preferable to t tests based on rms or dB values, because of the skewed distributions of rms and dB when signal-to-noise ratio (SNR)=0. F test performance improves as the number of frequencies increases, to about 15, but can be degraded if the background noise is not white, with a slope exceeding about 10 dB for the range of frequencies sampled. Segment analysis, using magnitude-squared coherence (MSC) or related statistics, has equivalent statistical power; MSC and F each yield unbiased SNR estimates that have identical distributions when SNR=0. Selection of F or MSC for detection of sinusoids will usually be a matter of convenience.

AB - Sinusoids in background noise can conveniently be detected using unsegmented power spectra, comparing power at the signal frequency to average power at several neighbor frequencies. In this case, the F test is preferable to t tests based on rms or dB values, because of the skewed distributions of rms and dB when signal-to-noise ratio (SNR)=0. F test performance improves as the number of frequencies increases, to about 15, but can be degraded if the background noise is not white, with a slope exceeding about 10 dB for the range of frequencies sampled. Segment analysis, using magnitude-squared coherence (MSC) or related statistics, has equivalent statistical power; MSC and F each yield unbiased SNR estimates that have identical distributions when SNR=0. Selection of F or MSC for detection of sinusoids will usually be a matter of convenience.

UR - http://www.scopus.com/inward/record.url?scp=0029845757&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029845757&partnerID=8YFLogxK

U2 - 10.1121/1.417933

DO - 10.1121/1.417933

M3 - Article

VL - 100

SP - 2236

EP - 2246

JO - Journal of the Acoustical Society of America

JF - Journal of the Acoustical Society of America

SN - 0001-4966

IS - 4

ER -