Remarks on the Lax conjecture for hyperbolic polynomials