Simultaneous Sourcing without Compromise

We show that iterative methods can improve the separation and remove noise for seismograms recorded with multiple vibrators operating simultaneously with a single sweep. Interference noise arises because multiple seismograms need to be recovered or separated from a smaller number of vibroseis field records. This noise represents a compromise made for high efficiency acquisition compared to methods, such as HFVS (High Fidelity Vibratory Seismic), in which there are as many sweeps as there are vibrators. The iterative methods start with an initial separation to as many seismograms as there are vibrators. Then updated seismograms are produced by adjusting them to match the survey data and to satisfy characteristics of noise-free seismograms. The iterations can be repeated until the measure of error is satisfactory. The match to survey data involves reconvolving parts of the initial seismograms with an estimate of the vibrator signatures and stacking. The characteristics of noise-free seismograms can include the fact that signal decreases with time but noise does not. We show improved separation and reduced noise for both pseudo-random sweeping and for long segmented sweeps (Continuous-HFVS). With these iterative methods, the full efficiency of simultaneous sourcing can be obtained without a noise compromise.