A (probably very straightforeward) question for you guys,when calibrating a Gaussian short rate model (vasicek/g2++), one would normally search for the parameters that make the model prices as close as possible to the observed market prices. In the literature (Brigo, for example), calibration is done on cap volatilities or the swaption volatility cube, so you have to transform your model prices to volatilities (using Black's formula) and then match them with the market volatilities.My question: is this equivalent with matching model prices with market prices (so not using Black's formula to obtain the volatilities)? If so, why would one bother to transform the prices to volatilities?Thanks in advance

Prob just market convention. BTW- Don't you have to strip the cap vols into caplet vols?