Comment on "Loss-error compensation in quantum-state measurements"