SENSITIVITY TO CHANGES IN LEVEL AND ENVELOPE PATTERNS ACROSS FREQUENCY

In the first experiment, two measurements were compared-sensitivity to across-frequency changes in level and sensitivity to across-frequency changes in the modulation phase of SAM tones. For the level task, mul ti-tone stimuli composed of 2-80 tones ranging in frequency from 200 t o 5000 Hz were used. For the phase task, the same frequency range was used, and 2-80 SAM tones were tested. For the level task, observers di scriminated between a multi-tone, equal-amplitude standard and one of two signals-a one-step or an up-down signal, The one-step signal had h igher levels at low frequencies and lower levels at high frequencies. The up-down signal had components with levels that varied high-low-hig h-low. For the phase task, the standard was the sum of SAM tones with identical modulator phases across frequency. The one-step signal had a common modulator phase at low frequencies and a different common modu lator phase at high frequencies. The up-down signal had modulator phas es that varied lag-lead-lag-lead. The results suggest that sensitivity to across-frequency changes in level and modulation phase reflect sim ilar initial processing stages. In a second experiment, SAM tones were used, and psychometric functions were measured for the level task, th e phase task, and a condition in which changes in level and modulator phase were both present. The standard was ''flat,'' and an up-down sig nal was to be detected. For one observer, the data suggest that level and phase information an independently represented. For the other two observers, interactions between the two features of the stimuli are ap parent. A multiple-looks model was moderately successful in accounting for the data. (C) 1998 Acoustical Society of America. [S0001-4966(98) 0111-4].