Effects of spatial and temporal integration of a single early reflection on speech intelligibilitya)
Spatial configurations of the sources used in the different experiments. The direct sound of the speech material was always presented frontally (S 0); noise was also presented frontally (N 0), laterally (N 135), or diffusely (ND , not shown). Black speakers indicate diotic reflections (R 0 and R 180), dark gray speakers indicate a reflection from the same side as the lateral noise source (R 45, R 90, and R 135), and light gray speakers indicate a reflection from the opposite direction (R 225, R 270, and R 315). The azimuth of the single reflection varied in the experiments in steps of 45°.
Mean SRTs and interindividual standard deviations measured with frontally presented direct speech, noise, and reflection of the same amplitude as a direct sound (AR = 1) and of reduced amplitude (AR = 0.3). Right: Mean SRTs and corresponding standard deviation for diotic speech condition (S 0 R 0) obtained in noise with (N 0, R ) and without (N 0) speech reflection.
Mean SRTs and interindividual standard deviations as a function of delay of a frontal reflection of the speech signal in diotic noise (circles), laterally located noise (squares), and diffuse noise (triangles). Direct soundwas presented frontally. Bottom: Mean binaural intelligibility level differences for laterally located (squares) and diffuse noise (triangles) as a function of delay of the reflection.
Mean SRTs and standard deviations as a function of delay of a reflection of the speech signal at an azimuth of 45° in laterally located noise (squares connected with a solid line) and diffuse noise (triangles connected with a solid line). The direct sound was presented frontally. Dashed, gray curves represent data measured with a frontal reflection (Fig. 2 ). Bottom: Differences in SRT between a frontal and a lateral reflection for lateral (squares) and diffuse noise (triangles) as a function of delay of the reflection.
SRTs measured in Expt. VI (frontal speech, diffuse noise, different reflection azimuths). Top panel: SRTs as a function of reflection delay. Black symbols connected with a solid line indicate a diotic speech signal, dark gray symbols connected with dotted lines indicate a reflection azimuth at the right side of the listeners' head, and light gray symbols connected with dashed lines indicate a reflection from the left side. For clarity, curves are slightly shifted horizontally. Bottom panel: SRTs as a function of azimuthal angle. Circles connected with a solid, black line indicate data for a reflection delay of 10 ms, squares connected with a dark gray line show threshold for a delay of 50 ms, and triangles connected with a light gray solid line indicate data for a delay of 200 ms.
Same data representation as in Fig. 5 , but for lateral noise instead of diffuse noise. The arrow at an azimuth of 135° indicates the direction of the noise source.
Experimental parameters used in the different experiments of this study: speech azimuth (ϕS ), reflection amplitude relative to the direct sound (Ar ), reflection azimuth (ϕR ), noise azimuth (ϕN ), and delay between direct sound and reflection (Δt). D denotes diffuse characteristics.
Article metrics loading...
Full text loading...