Assessment of the starting point of the Lombard effect

Pasquale Bottalico, Ivano Ipsaro Passione, Simone Graetzer, Eric J. Hunter

Research output: Contribution to journalArticlepeer-review


The Lombard Effect is an involuntary tendency to raise the level of the voice with an increase in the background noise level. The aim of this study was to determine whether there is a starting point of the Lombard Effect in terms of background noise level (Ln).Twenty subjects were instructed to read a passage with the goal of being understood by a listener at 2m. A loudspeaker, also located at 2m, emitted pink noise. Ln varied in level from 25 dBA to 70 dBA in 5 dB increments. A regression model was fit with segmented relationships to the data, estimating the slopes and the change point(s) in the relationship between the response variable, within-subject normalised SPL, and the explanatory variable, Ln. One change point was identified at Ln 48.3 dBA. The rate of change in speech level below 48.3 dBA of noise was about 0.24 dB per dBA of noise while the rate of change above 48.3 dBA was nearly 0.5 dB per dBA of noise. Hence, a change point was identified in the speaker’s adaptation to the background noise level, which could be said to mark the starting point of the Lombard Effect, as typically described.
Original languageEnglish (US)
Pages (from-to)2015-2015
JournalThe Journal of the Acoustical Society of America
Issue number4
StatePublished - Apr 1 2016
Externally publishedYes


Dive into the research topics of 'Assessment of the starting point of the Lombard effect'. Together they form a unique fingerprint.

Cite this