reduced EG_MIN_RELEASE_TIME to allow to finish kill process of a voice within the same audio fragment, this will be e.g. be mandatory for voice stealing