At least in the paper "N. Ma et al. - A machine-hearing system exploiting head movements for binaural sound localisation in reverberant conditions, ICASSP 2015" we tried different head rotation strategies and could show that they have an impact on the localisation results.
At the moment we have only a fixed head rotation strategy implemented in our model (and if I remember correctly not the one with the best results). It would be nice, if we could add different strategies and the possibility to switch between them. Another nice option would be to add your own strategy or real head movement data from a listening experiment.
@ningma97 replied already that this should be not to hard, but he has to migrate his old code to the new blackboard first.
At least in the paper "N. Ma et al. - A machine-hearing system exploiting head movements for binaural sound localisation in reverberant conditions, ICASSP 2015" we tried different head rotation strategies and could show that they have an impact on the localisation results.
At the moment we have only a fixed head rotation strategy implemented in our model (and if I remember correctly not the one with the best results). It would be nice, if we could add different strategies and the possibility to switch between them. Another nice option would be to add your own strategy or real head movement data from a listening experiment.
@ningma97 replied already that this should be not to hard, but he has to migrate his old code to the new blackboard first.
Note, this is a continuation of https://dev.qu.tu-berlin.de/issues/2733