The influence of alcoholic intoxication on the fundamental frequency of female and male speakers

Abstract

This study investigates long-term features and utterance contours of fundamental frequency (f0) derived from the German Alcohol Language Corpus. The corpus comprises read, spontaneous, and command&control speech uttered by 148 speakers of both genders and various age groups when sober and intoxicated. f0 median, f0 range, and f0 contours are analyzed for intoxication and interactions with gender and age. Contours are compared both directly (root mean squared error, statistical correlation, or the Euclidean distance in the spectral space of the contour) and by parameterization of the contour using discrete cosine transform and the first and second moment of the lower contour spectrum. Results partly confirm earlier findings, i.e., f0 average and range are mostly raised with intoxication, and also suggest that the majority of speakers do not follow a general trend, but show idiosyncratic alterations to f0. f0 contours differ significantly with intoxication, but a more detailed analysis could not assign these changes to specific general form changes like decline or curvature. The results suggest that it is not possible to predict intoxication from f0 in a single model across different speakers. Instead a speaker-dependent model to account for the individual speaker behavior is proposed.

The influence of alcoholic intoxication on the short-time energy function of speech

“…DCT coefficients w x () with lower indices represent low-frequency movements (ripples) in the transformed waveform, while coefficients with higher indices represent ripples of high frequency. The first coefficient of the DCT, w x ( ¼ 1), is the same for all contours due to the preceding normalization and is therefore not considered in the analysis (Baumeister et al, 2012). Varying the number of lower DCT coefficients showed that ripple frequency indices 2 to 7 yielded the best distinction between intoxicated and sober RMS contours,…”

“…Since the average syllable number of the test sentences was 12.3 (which equals a wave length of 0.081 L), the DCT range ¼ 2…51 should roughly cover all RMS movements up to the syllable rate (Baumeister et al, 2012).…”

“…Various distance measures can be calculated between two RMS contours; the measures described below have already been applied to f0 contours in Baumeister et al (2012).…”

“…As a further parameterization of the DCT spectrum the first and second moments of the DCT coefficients were calculated (Baumeister et al, 2012). The first two moments encode basic properties of the DCT spectral shape, i.e., the center of gravity (m 1 ) and the variance across the DCT frequency range (m 2 ).…”

“…Recently new studies based on the publicly available German Alcohol Language Corpus (ALC) (see Sec. III) provide statistically firm findings regarding phonetic standard features (Baumeister et al, 2012;Heinrich and Schiel, 2011;Schiel, 2011;Schiel and Heinrich, 2009;Schiel et al, 2010): in short, the most prominent phonetic features that change under the influence are a decreased speaking rate and raised fundamental frequency (Heinrich and Schiel, 2011;Baumeister et al, 2012). Unfortunately, the same features are also prone to changes under stress (e.g., Hansen and Patil, 2007), by the Lombard effect (e.g., Folk and Schiel, 2011), and when the speaker's emotional state is one of anger, joy, or sadness (e.g., Mathon and de Abreu, 2007;Yildirim et al, 2004).…”