In this paper we extend the approach of directly using all the available tick-by-tick data to the realized correlation estimation. As for the realized volatility, the presence of market microstructure can induce signiflcant bias in standard realized covariance measure computed with artiflcially regularly spaced returns. Contrary to these standard approaches we analyse a very simple and unbiased realized covariance estimator which does not resort on the construction of a regular grid, but directly and e‐ciently employs the raw tick-by-tick returns of the two series. Montecarlo simulations calibrated on realistic market microstructure conditions show that this simple tick-by-tick covariance posses no bias and the smallest dispersion, among the covariance estimators considered in the study. Combining this ralized covariance measure together with tickby-tick volatility estimator we obtain a realized correlation measure where both the volatilities and the covariances are computed from tick-by-tick data. In the empirical analysis performed on S&P 500 and US bond data we apply the HAR model to the tick-by-tick realized correlation obtaining remarkably good out of sample forecasting performance.