hello list, I've been reading a paper recently "B-Format Acoustic Impulse Response Measurement and Analysis in the Forest at Koli National Park, Finland" by Simon Shelley, Damian Murphy and Andrew Chadwick, where a unidirectional loudspeaker is used to collect IRs via a sine-sweep/de-convolution method. To simulate an omnidirectional source, four recordings are made of the sweep with the loudspeaker directed at the mic and at 3 other positions 45 degrees apart.
The IRs for the simulated omnidirectional source are created by: "summing the resulting impulse responses measured for each angle of the loudspeaker. The effect of the summation is to emulate a loudspeaker array of four loudspeakers at the same point in space all pointing in different directions." So, my first question is, is this "summing" really a simple summation of the 4 sets (a b-format mic is used) of IRs with some attenuation of each, or would there be something more complex involved? I'm interested in spatialising the human voice with IRs - and a directional source is probably more appropriate than an omnidirectional one. The above got me thinking though, if it is a simple sum, wouldn't it be possible to simulate the variable orientation of a speaker (human) by doing an equal power crossfade between adjacent angular IRs as this virtual human speaker turns? Am I oversimplifying? Thanks, Iain -- ________ Iain Mott http://reverberant.com http://audiocena.com.br _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
