| Résumé |
It has been shown in the past that listeners judge di erently the loudness change of sounds that increase or decrease in levels for the same physical variation. Di erent results were found concerning the direction of this asymmetry. It has been shown that loudness change measurement is strongly dependent on the procedure used: direct or indirect estimations. Direct estimation is mainly inuenced by the global loudness of the sound (see co-submitted work). As a result, loudness change estimated by a direct procedure is greater for a ramped sound and remains whatever is the location in level. For indirect estimations, results are di erent and less obvious. In the present work, indirect estimations consisted in measuring the loudness at the end of the varying-level sounds using magnitude estimation for 1-kHz tones, synthetic vowel sounds and white noises. The loudness change was dened as the ratio between the loudness estimated at the end of the varying sound and the loudness at its start level. The latter originated from the loudness function measured for stationary sounds. Indirect estimation of loudness change is greater for ramped than for damped sounds. Contrary to the result obtained with direct estimation, the di erence between ramped and damped sounds decreased as the location in level of the variation increased; for variations in the more intense levels, the inverse e ect was observed. In most conditions, the loudness at the end of the sounds (ramped or damped) was overestimated compared to the loudness of an equivalent steady state sound. As a result, the ratio (end loudness)/(start loudness) of ramped sounds was, most of time, greater than this ratio for damped sounds. These data showed that the estimation at the end of a sound increasing or decreasing in level explain the asymmetry obtained when measuring loudness changes of ramped and damped sounds indirectly. |
