Gain

Description Apply data dependent (AGC) and data independent gains to a volume.
Module(s) Explorationist (Stack), Image Gather Processing (Gather) Gather Attributes
Requirements Volume
Related Volume Maths, Volume Maths (Simple)
Works with 2D, Stacks, Gathers


The Gain process generates a new volume with a scaling applied to the trace amplitudes. This process can be applied to any volume.

The two main categories of gain function are:

  • Data dependent (AGC)
    Scaling determined by the input amplitudes.
    – Scalar clipping functionality.
    – Can produce a second output volume of scalars.
    – Necessary to later remove AGC (using
    Volume Maths).
  • Data independent
    Scaling determined by time, velocity or other value.
    – Can be applied or removed.

Create new Gain process

Create new Gain process
  1. In the Control Panel, open the Process tab.
  2. Click the Add icon and select New Process.
  3. Double-click Gain.
  4. Type a name for the process and click OK.
  5. In the Details Panel, at Volume, select the volume to apply the gains.

Data-dependent (AGC) gains

Data-dependent (AGC) gains
  1. In the Gain settings section, select Data dependent (AGC).
  2. Select a Gain function to apply:
    • Arithmetic mean within a window
      The average of the window values (see Arithmetic Mean).
    • Median within a window
      The median of the window values (see Median).
    • Geometric mean within a window
      The geometric mean (nth Root of (a * b * c * d... n)) (see Geometric Mean).
    • RMS within a window
      The Root Mean Square / quadratic mean (see Root Mean Square).
    • Median mean within a window divided by the arithmetic mean
      Arithmetic mean of the middle N values in the window, where N is the window trimmed by the specified Percentage trim. Median mean, otherwise known as alpha-trimmed mean, is computed by ranking the values within a window from smallest to largest, removing the smallest and largest %, and averaging the remaining values. For example, a 40% trim would remove the smallest 20% and largest 20% of values, then average the remaining 60%. This is more robust (though more computationally expensive) than straight mean, as it is unaffected by outliers.
  3. The Gain process also comes with a clipping functionality to clip scalars.
  4. Select the Output additional scalars volume check box to produce a volume of scalars as the output, instead of the gained data itself. These can be applied to the data using Volume Maths (data * scalars), and later reversed (seismic/scalars) (see below).
  5. Enter the Window length and Window overlap. For each trace, a single scalar is calculated for each window, blended together in the area of overlap.
  6. As a result of this process, a new volume (with the same name as that given to the process) is available in the Volume tab.

Tip: To quickly open the output volume, right-click the process and select Open Child Volume. Likewise, you can open the process from the Volume tab by right-clicking the volume and selecting Open Parent Process.

Use Volume Maths to remove AGC

Use Volume Maths to remove AGC

Data-independent gains

Data-independent gains
  1. In the Gain settings section, select Data independent.
  2. Select a Gain function to apply (see below for details of each function):
    • Power of t
    • velocity * velocity * time
    • Exponential
    • dB per second (or unit depth)
    • Time/Depth - Gain pairs
  3. Select Direction to Apply or Remove the gain. In all cases, the gain is normalised such that it is 1 at the first data point to which the gain is applied (this may not be the top of the trace due to windowing). The gain functions can be started and stopped at a specified horizon or constant time. In such cases, the gain at the bottom of the selected data range is applied to the remainder of the trace.
  4. Select Windowed or Entire Volume to apply the process within a defined window or on the entire volume.
  5. If Windowed is chosen, select Constant or Horizon to define the windows. If horizons are chosen, you can optionally input a constant to offset the horizon.

As a result of this process, a new volume "Gain" is available in the Volume tab.

Important notes:

  • If you are using a depth-domain input volume, you must have defined a velocity volume (see Converting Time-Depth) in order to use the Power of t, velocity * velocity * time, or Exponential gain functions.
  • For windowed data (where top and bottom horizons are specified), the gain applied is always 1 above the top horizon, and the gain at the bottom horizon is applied to all deeper samples.

Tip: To quickly open the output volume, right-click the process item in the Process tab and click Open Child Volume. Vice versa, you can also open the process from the Volume tab directly. To do this, right-click the volume which was created from a process (indicated in italics) and click Open Parent Process.

Data-independent gain details

Power of 't'

TWT (in seconds) raised to the given exponent (n)

Gain functions involving time differ slightly for windowed vs. entire-volume operation:

  • Windowed: gain(t) = 1 + (t - ttop)n
  • Entire-volume: gain(t) = tn + Constant

velocity * velocity * time (v-squared t)

Velocity squared (in m/s) multiplied by TWT (in seconds)

Gain functions involving time differ slightly for windowed vs. entire-volume operation:

  • Windowed: gain(t) = 1 + velocity * velocity * (t - ttop)
  • Entire-volume: gain(t) = velocity * velocity * t

Exponential

A function of three user-provided constants: scale, power, and bias.

Gain functions involving time differ slightly for windowed vs. entire-volume operation:

  • Windowed: gain(t) = bias + scale * epower * (t - ttop)
  • Entire-volume: gain(t) = bias + scale * epower * t

dB per second (or unit depth)

A function of dB per TWT in seconds (or TVD in km or miles).  Two signals that differ by n dB have an amplitude ratio of 10(n/20)

Gain functions involving time or depth differ slightly for windowed vs. entire-volume operation:

  • Windowed: gain(z) = 10((z - ztop) * dB / 20)
  • Entire-volume: gain(z) = 10(z * dB / 20)

...where z is TWT in seconds (for a time-domain input) or TVD in km or miles (for a depth-domain input, according to the user's preferred units).

Time/Depth - Gain pairs

Click the Add icon to add additional time/gain or depth/gain pairs.

The gain function is a linear interpolation of the specified gain at the specified time/depths.

Example

Example