A remark on the asymmetry of convolution operators