Is the effort dependence of speech loudness explicable on the basis of acoustical cues?

Effort-dependent speech loudness (of sustained vowels) was investigated with the general intention of making a comparison between experimental loudness data and calculated loudness data (based on the Zwicker model). The adjustment method was used in the listening experiment as an alternative to the magnitude-estimation technique used in earlier experiments. The stimuli varied with respect to (1) vocal effort, (2) vowel quality, and (3) initial structure. Intensity, fundamental frequency, and duration were held constant. The speech stimuli, produced by five male speakers, were presented to 25 listeners. The equal loudness data of the listeners indicate that the speech loudness varies significantly with effort and vowel quality, while the initial structure does not have any significant influence. In the second part of the investigation, loudness was calculated on the basis of spectral information. The calculated loudness data were compared with the listener loudness data (phon level). With respect to the relative level, listener loudness and calculated loudness agree to a very large extent in effort-dependent loudness differences and loudness differences dependent on vowel quality. With respect to absolute loudness levels, the listener loudness and the calculated loudness differ almost constantly (by 9 phons). The comparative loudness data suggest that effort-dependent speech loudness changes are due to pure spectral changes, and that a special perceptual mechanism in the loudness perception of speech is questionable. Subject Classification; 70.30, 70.20.