Trim statics - Simple

<< Click to Display Table of Contents >>

Navigation:  Statics >

Trim statics - Simple

 

Description

Trim statics - Simple computes and applies trace-by-trace residual static corrections directly within the processing sequence, without requiring a separate Calculate and Apply step. For each CDP gather the module cross-correlates every receiver trace against a model trace (optionally derived from a super gather) and solves for the set of time shifts that maximises inter-trace coherence within the specified time window. The resulting shifts are applied to the output traces before passing them to the next module.

This module is suited for fast trim static corrections in a batch sequence. Use the more comprehensive Trim statics - Calculate / Edit / Apply workflow when you need interactive quality control, smoothing between adjacent gathers, or separate editing of the shift field. Both constant-time-window and time-variant window modes are supported.

Input data

Input DataItem

The NMO-corrected or otherwise pre-processed seismic dataset. Connect from the preceding module in the processing sequence.

Input SEG-Y data handle

SEG-Y file handle passed through the sequence for downstream random-access reading.

Input trace headers

Trace header table passed through the sequence.

Input gather

The current CDP gather provided by the sequence iterator. This is the gather on which trim statics are computed and applied during each processing iteration.

Input stack line

Stack section passed through the sequence unchanged.

Input crooked line

Crooked-line geometry descriptor passed through the sequence.

Input bin grid

Bin grid passed through the sequence.

Input sorted headers

Sorted gather index vector passed through the sequence for downstream random access.

Model

Container grouping the model gather inputs. The model trace used for cross-correlation can be the stack of the current gather itself, or a super gather assembled from neighbouring CDPs.

Super gather (optional)

Optional super gather assembled from multiple adjacent CDP gathers, typically produced by the Trim statics - Calculate module running in super gather mode. When connected, this higher-fold gather is used as the cross-correlation model instead of the single-bin stack, improving the stability of the shift estimates on low-fold data.

Partial xcor summing for each bin

When enabled (default), the super gather model is recomputed for each bin by updating incrementally rather than assembling from scratch, which reduces memory usage and speeds processing. Disable only if the incremental computation produces artefacts on data with highly irregular bin geometry.

Parameters

Selection type { Constant time, Time variant }

Selects how the cross-correlation window is defined. Constant time (default) uses a single fixed window defined by From time and To time. Time variant uses a table of multiple windows, each with its own time range and maximum shift, enabling different trim parameters at different depths. Selecting Constant time activates the From time, To time, and Max time shift parameters; selecting Time variant activates the Time variant table.

Time variant

Table of time-variant cross-correlation windows. Each row defines: Start time (s), End time (s), Analysis windows step (s), and Max shift (s) for that interval. Add a row for each depth zone requiring different static parameters — for example, a shallow window with a small maximum shift and a deeper window with a larger maximum shift. Active only when Selection type is set to Time variant.

From time

Start of the cross-correlation analysis window (s). Default: 0.0 s. Set to the time of the shallowest reflector of interest. Active when Selection type is Constant time.

To time

End of the cross-correlation analysis window (s). Default: 6.0 s. Set to the time of the deepest reflector of interest. Active when Selection type is Constant time.

Max time shift

Maximum allowed trim static shift (s) per trace. Default: 0.01 s. Estimated shifts larger than this value are clipped. Set to the maximum expected residual after long-wavelength statics correction; typical values are 0.008–0.032 s. Active when Selection type is Constant time.

Overlapping

Overlap (s) between adjacent analysis sub-windows used during time-variant shift estimation. Default: 0.01 s. A small overlap ensures smooth transitions between depth zones; use values between 0 and the sub-window length.

Massive cross-correlation approach

When enabled, cross-correlations are computed between all pairs of traces in the gather rather than only between each trace and the model stack. Default: off. The massive approach can improve shift accuracy on gathers with strong amplitude variations between traces, at the cost of significantly higher computational load. Use on key gathers for QC or when standard cross-correlation gives unsatisfactory results.

Solving method { Lsqr, Gauss-Seidel }

Linear solver used to decompose the measured inter-trace cross-correlation shifts into per-trace static corrections. Lsqr (default) uses a least-squares conjugate gradient solver that is fast and robust on well-conditioned gathers. Gauss-Seidel is an iterative relaxation solver that may converge more reliably on gathers with many dead or noisy traces.

Remove deviations from trend detection

When enabled (default), traces whose estimated shift deviates significantly from the offset-based trend are detected as outliers and their shifts are replaced by the trend value. This prevents isolated noisy traces from being over-corrected. Disable on very high fold data where all traces have reliable cross-correlation estimates.

Settings

Auto-connection

When enabled, data items are connected automatically when the module is added to the sequence. Disable to configure connections manually.

Bad data values option { Fix, Notify, Continue }

Controls what happens when NaN or Inf values are detected in the input data. Fix replaces bad values with zero silently. Notify replaces bad values and logs a warning. Continue passes bad values through unchanged.

Execute on { CPU, GPU }

Selects whether the cross-correlation computation runs on the CPU or a supported GPU. GPU execution can significantly reduce processing time on large surveys.

Distributed execution

Configuration for distributed (cluster) execution. These settings control node assignment and job scheduling when the sequence runs on a compute cluster.

Bulk size

Minimum number of gathers sent to each compute node in a single chunk during distributed execution.

Limit number of threads on nodes

When enabled, limits CPU threads per cluster node to the value set in Number of threads.

Job suffix

Optional text appended to the distributed job name. Use to distinguish simultaneous jobs of the same type on the cluster.

Set custom affinity

When enabled, sets the CPU core affinity for this module explicitly using the Affinity mask.

Affinity

CPU affinity mask applied when Set custom affinity is enabled.

Number of threads

Maximum number of CPU threads used during execution. Higher values increase throughput on multi-core workstations.

Skip

When enabled, this module is bypassed during sequence execution and data passes through unchanged.

Output data

Output DataItem

The dataset with trim static corrections applied to every trace. Pass to the next processing module.

Output SEG-Y data handle

Updated SEG-Y file handle for downstream random-access reading.

Output trace headers

Updated trace headers with shift values recorded in the appropriate header fields.

Output gather

The trim-static-corrected CDP gather for the current iteration, available for display or downstream processing.

Output stack line

Stack line passed through unchanged.

Output crooked line

Crooked-line geometry descriptor passed through unchanged.

Output bin grid

Bin grid passed through unchanged.

Output sorted headers

Sorted gather index vector passed through for downstream use.

Information

Graphics

Custom actions