cross correlation vs sliding FFT

Hi,

I have a signal with many frequencies appearing at different times. I need to know when a certain frequency appears at this signal. The most obvious thing to do is to have a sliding window FFT and see when the frequency appears. However, what is the difference, or will I see any improvement, if I create a synthetic signal with a desired frequency, and cross correlate this synthetic signal to the original signal? The maximum of the cross-correlation should give me when the desired frequency appears. If there are other frequencies and noise in the signal, it would be difficult to find the frequency in the FFT spectrum, but in the cross correlation it should still be at the maximum position, right?

I feel I am missing some relation between the cross correlation and the FFT that should tell me the possible accuracy I can get. I hope my question is understandable, thanks for any replies.

[Shaurya Garg]

Hi Mortizo! Did you solve the problem of sliding FFT? O have the same problem, could you please help me?