|
<< Click to Display Table of Contents >> Navigation: Velocity > DVA |
DVA (Dip Velocity Analysis) is a gather-attribute computation module that measures local energy focusing quality at each sample in a seismic gather. For every trace-time sample, the module computes an entropy-based statistic over a sliding 2D window: it squares all amplitude values in the window, normalizes them by the local mean squared amplitude, and evaluates the quantity q·ln(q) for each normalized value. The average of these values over the window is output as the DVA attribute.
A highly focused (well-corrected) gather produces a peaked, non-uniform amplitude distribution within the window, resulting in a large negative entropy value. A poorly focused gather — for example one with residual moveout or excessive noise — has a flat amplitude distribution, producing an entropy value near zero. The DVA output gather can therefore serve as a focusing quality indicator to guide velocity updates, residual moveout corrections, or gather conditioning workflows.
Note: This module is deprecated. For current velocity analysis and residual moveout workflows, consider using the Residual DVA module, which extends this analysis across multiple velocity-factor panels to produce a spatially interpolated optimal velocity field.
The input seismic gather on which the DVA entropy attribute will be computed. This should be a stack-format gather (one trace per offset or CMP bin). The gather is expected to have residual moveout already applied or partially corrected, so that the focusing quality measurement is meaningful. The module processes one gather at a time and supports multi-threaded execution.
The half-width of the analysis window in the trace (offset or lateral) direction, measured in number of traces. For a given trace, the module uses traces from (current trace − Trace window) to (current trace + Trace window) inclusive when computing the local entropy. The default value is 10 traces, giving a total window spanning 21 traces. Increase this value to smooth the DVA attribute more heavily across offsets, which stabilizes results in noisy data but reduces lateral resolution. Decrease it to obtain a more trace-local focusing measure. At the gather edges the window is automatically clipped to the available traces.
The half-width of the analysis window in the time direction, specified in seconds. For a given time sample, the module uses samples from (current time − Sample window) to (current time + Sample window) inclusive. The default value is 0.01 s (10 ms). The window length in samples is computed automatically from this value and the input gather sample interval, so you can specify it in physical time units regardless of the data sampling rate. Larger values integrate the entropy statistic over a longer time interval, which is useful for broadband data or noisy records; smaller values produce a higher-resolution attribute that responds to individual reflectors. At the gather boundaries the window is automatically clipped to the available samples.
The DVA entropy attribute gather. It has the same dimensions (number of traces and number of samples) as the input gather. Each sample value represents the local entropy of the squared amplitude distribution within the analysis window centered on that trace-time position. Values are negative or zero: larger negative values indicate stronger amplitude focusing (higher energy concentration in fewer samples), while values close to zero indicate poor focusing or uniformly distributed energy. Samples where the local mean squared amplitude is zero are assigned a value of zero.