In this paper we revisit the problem of calibrating stochastic volatility

models. By finding smart initial parameters, we improve robustness of

Levenberg-Marquardt. Applying this technique to the SABR and Heston