The earth filter - A simple minded explanation.                                   Remember your mouse!

1. Rocks are elastic. 
2. They will resist movement, but once moved they will rebound.
3. Rebounds will go past original position, causing oscillation. 
4. Inertia will cause phase lags, creating lower dominant frequencies in
    later "legs".
5. Higher onset frequencies will be attenuated with depth.  
<

TUNING, TUNING AND MORE TUNING -

When you add sinusoids, you get sinusoids. Primary events from the top and the bottom of a bed are of opposite polarity. If the bed is too thin, they will cancel each other. If the bed is just the right thickness, a strong event will be recorded (at an offset as seen in the picture). Thickness changes can cause odd apparent dips. 

Take some time to study the red result - The two primary events were drawn using the basic shape shown above, with subsequent lobes at lower dominant frequencies. This distinction was lost in the algebraic addition, and a "Ricker" like shape emerged. The apparent dominant frequency of the result is lower than that of the originals. On real sections one will see higher frequency events interspersed with lower ones. At the very minimum this means the lower ones are tuned phenomena. Of course we are probably not seeing any pure basic wavelet shapes (everything is tuned in most prospects).

So - beware putting too much faith in amplitudes, including common offset ones!

This brings us to the need for inversion.  The plot at the bottom of the picture to the left represents the character stream that would occur from the algebraic mixture of primary reflections from the tops and the bottoms of the two beds. None of the final events have a simple relationship to a single bed, and the maximum energy lobes are really far out. To make it more confusing, any change in bed thickness can cause spurious dips that will confuse structural interpretation.

Obviously we need to remove the effect of the wavelet being leggy to gain resolution, and this is inversion. Most of the PowerPoint shows contain multiple examples of how inversion solves the all important tuning problem! 
 

Now, if you don't see a left panel, click here to go to the right place