In model testing for ocean waves, the data recording duration is usually longer the better especially for irregular waves. The recording length shall be long enough to achieve sufficiently small bandwidth in the spectral analysis. There must be a limit according to bandwidth of the wave spectra. But is it true that the time recording length is inversely proportional to the frequency bandwidth?

In model testing for ocean waves, the data recording duration is usually longer the better especially for irregular waves. The recording length shall be long enough to achieve sufficiently small bandwidth in the spectral analysis.

In ideal wave spectrum, the bound waves can also present, so the bandwidth can be large, but does it mean the recording duration can be shorter even?

On the other end, for a regular wave (zero bandwidth), then only the one period length of recording is needed, right?

So is confused completely. The time recording length is neither inversely proportional to the frequency bandwidth nor otherwise.

There must be a way to follow the bandwidth of the wave spectra. It might be a stupid question, please help with some basics.