|
<< Click to Display Table of Contents >> Navigation: Migration > Depth velocity updater |
The Depth velocity updater (second variant) is an interactive interval velocity analysis tool designed to iteratively update a depth-domain velocity model using pre-stack depth migrated (PSDM) common image gathers (CIGs). It operates through a layer-stripping workflow: the user selects a layer defined by input horizons, computes depth-moveout (DM) semblance over a range of velocity perturbations for that layer, picks the best-fitting velocity correction, and applies it to update the model one layer at a time from the shallowest to the deepest.
For each selected bin the module migrates raw shot data on the fly using the current velocity model, produces inline and crossline stacks and CIG gathers in both time and depth domains, and computes DM semblance (a horizon-based inversion filter) that shows the residual velocity error as a function of velocity perturbation and horizon. The user inspects the semblance panel, places picks manually or runs the automatic horizontal picking action, and then triggers the model update. The module supports super-gather stacking to boost signal-to-noise ratio when individual bins are sparsely populated, and optional mute picking to suppress noise or multiples before semblance computation.
This module differs from the first Depth velocity updater variant by using a horizon-based inversion scheme (hodograph-driven semblance) rather than a direct image-space approach. It is best suited for 3D datasets where clear reflectors are identifiable on depth-domain stacks and a layered velocity model can be defined via horizon picks.
Connect the pre-stack SEG-Y data item that contains the raw shot or CMP gathers to be migrated. This handle gives the module random-access to individual traces, which it reads on demand during the on-the-fly Kirchhoff depth migration performed for each selected bin. The data must be in the same coordinate system as the depth velocity model.
Connect the trace header vector associated with the input SEG-Y data. The headers supply the source and receiver coordinates, offsets, and bin assignments used to sort and select traces for each inline and crossline stack computation. This item must correspond exactly to the data handle connected above.
The current interval velocity model in the depth domain, expressed as a velocity gather (traces versus depth samples in m/s). This is the model to be updated. It is used both as the migration velocity field for on-the-fly PSDM and as the starting point for velocity perturbation during semblance analysis. After each update cycle the output Updated depth velocity should be connected back here for the next iteration.
A horizon picking item defining the layer boundaries in the velocity model. Each horizon corresponds to an interface between velocity layers. The module uses these horizons to isolate one layer at a time during the layer-stripping workflow and to constrain the hodograph-based semblance computation to the depth extent of the selected layer. At least one horizon must be present; the horizons are listed by name in the Current layer dropdown.
An optional mute picking item that defines offset-time or offset-depth mute curves. When Use mute is enabled, these curves are applied to each CIG gather before semblance computation to suppress direct arrivals, multiples, or noise at near or far offsets. Use the Clear mute picking, Load mute picking, and Save mute picking custom actions to manage the mute interactively. The mute can also be painted directly on the depth gather view using the interactive mute picking overlay.
When enabled, the module treats the input SEG-Y data as being in the time domain and performs depth migration internally. When disabled (default), the module expects data that have already been depth-migrated (i.e., pre-stack depth-migrated CIGs). In most interactive velocity update workflows the input is raw time-domain shot gathers and this option should be left enabled, allowing the module to run on-the-fly Kirchhoff PSDM for each selected bin.
Default: Off (false).
The sample interval of the time-domain stack output, in seconds. This controls the temporal resolution of the inline and crossline time stacks displayed in the vista panels. Finer values improve resolution but increase memory usage. The value must be greater than 0.00001 s.
Default: 0.004 s (4 ms).
The total recording length of the time-domain output stacks, in seconds. Set this to cover the full two-way travel time of the interval of interest. Traces are truncated or zero-padded to this length. The value must be at least 0.004 s.
Default: 4 s.
The replacement (datum) velocity used when applying static datum corrections to the time-domain stacks, in m/s. This value is used to shift traces to and from the floating datum defined by the Datum parameter. Set V0 to a value representative of the near-surface velocity in the survey area. The minimum allowed value is 100 m/s.
Default: 4000 m/s.
The reference datum depth in meters. Time-domain and depth-domain stacks are corrected so that depth zero corresponds to this datum level. Use this to align the output imagery with the geological datum used in your project. A value of 0 means sea level or the acquisition surface.
Default: 0 m.
When enabled, the mute curves from the connected Mute picking item are applied to CIG gathers before stacking and semblance computation. Muting noisy or contaminated offset ranges improves semblance quality and the reliability of velocity picks. Disable this option if no mute item is connected or if the data quality is uniformly good across all offsets.
Default: On (true).
When enabled, traces from neighboring bins within the Super gather aperture radius are included when building the CIG and stack for the selected bin. This increases fold and improves signal-to-noise ratio, which is especially beneficial in areas of sparse acquisition. The trade-off is a slight loss of lateral resolution. Disable this option when the dataset has high fold and good per-bin signal quality.
Default: On (true).
The spatial radius in meters within which neighboring bins are gathered into the super gather for the selected bin. All bins whose bin center falls within this distance are included. Larger apertures improve fold and signal-to-noise ratio but reduce lateral resolution and slow computation. Active only when Use super gather is enabled.
Default: 250 m. Minimum: 0 m.
Controls the geometric spreading compensation applied during on-the-fly migration. Options are:
none — no normalization; raw migration amplitudes are preserved.
2D — applies 2D geometric spreading correction, suitable for 2D line data or when a 2D approximation is acceptable.
3D — applies full 3D geometric spreading correction; use this for 3D datasets to obtain amplitude-balanced CIGs.
Default: 2D.
The order of the Eikonal solver used to compute travel-time tables for on-the-fly depth migration. Higher accuracy corresponds to a higher-order finite-difference approximation of the Eikonal equation, yielding more precise travel times in complex velocity models but requiring more computation time. Options are Low, Medium, and High.
Use Low during initial coarse velocity iterations for speed. Switch to Medium or High for final updates when the velocity model is close to correct and accurate imaging is required.
Default: Medium.
The number of traces read at once from disk when building the inline and crossline stacks. Larger values speed up I/O by reducing the number of disk reads but increase memory consumption. Reduce this value if you encounter out-of-memory errors on machines with limited RAM. Minimum value: 1.
Default: 10000 traces.
This parameter group contains the settings controlling the layer-stripping velocity update workflow: which layer to update, the hodograph geometry used for semblance computation, the spatial calculation and display grids, and optional horizon and residual smoothing.
Selects which velocity layer to analyze and update. The dropdown is populated from the horizon names in the connected Input horizons item. In the layer-stripping workflow, process layers from shallowest to deepest. The semblance panel, hodographs, and DM picks all correspond to the selected layer. After picking and applying an update, move to the next deeper layer.
The spacing in meters between modeled hodograph offsets used to compute the DM semblance. Smaller step values sample the offset axis more finely, producing smoother semblance curves but requiring more computation. A step of 100–200 m is typically adequate for most surveys. Minimum: 1 m.
Default: 200 m.
The maximum source-receiver offset in meters used when computing hodographs for DM semblance. Offsets beyond this value are excluded from the semblance calculation. Set this to match the maximum usable offset in your dataset — typically the offset at which reflections from the target layer remain coherent. Range: 0 to 100000 m.
Default: 5000 m.
The spatial step in the inline (X) direction in meters at which DM semblance is computed across the survey area. The module calculates semblance at a regular grid of analysis points spaced by this value. Larger steps speed up computation but reduce the spatial detail of the velocity update. Minimum: 1 m.
Default: 500 m.
The spatial step in the crossline (Y) direction in meters at which DM semblance is computed. Works in combination with Calculate grid step X to define the analysis grid over the survey area. Minimum: 1 m.
Default: 500 m.
The display grid step in the inline direction in meters for the map view overlays (such as the Layers V average point cloud). This controls how densely the velocity correction map is displayed in the map vista, without affecting the actual computation grid. Minimum: 5 m.
Default: 50 m.
The display grid step in the crossline direction in meters for the map view overlays. Works analogously to Map step X. Minimum: 5 m.
Default: 50 m.
When enabled, the selected horizon is spatially smoothed before being used in the DM semblance computation. This can help suppress picking noise on rough or noisy horizon surfaces, leading to cleaner semblance panels. When enabled, the Smooth horizon window X and Smooth horizon window Y parameters become active.
Default: Off (false).
The half-width of the spatial smoothing operator applied to the selected horizon in the inline direction, in meters. Larger values produce a smoother horizon but may blur genuine structural dip. Active only when Smooth selected horizon is enabled. Minimum: 0 m.
Default: 500 m.
The half-width of the spatial smoothing operator applied to the selected horizon in the crossline direction, in meters. Works analogously to Smooth horizon window X. Active only when Smooth selected horizon is enabled. Minimum: 0 m.
Default: 500 m.
When enabled, the picked velocity residuals (DM corrections) are spatially smoothed along horizon surfaces before being applied to the velocity model. This prevents abrupt lateral velocity changes and produces geologically plausible, laterally consistent updates. When enabled, the Smooth residuals window X and Smooth residuals window Y parameters become active.
Default: Off (false).
The half-width of the spatial smoothing operator applied to the picked velocity residuals in the inline direction, in meters. A larger window produces a smoother, more laterally continuous velocity update. Use values consistent with the expected lateral scale of velocity variations in the target layer. Active only when Smooth residuals along horizons is enabled. Minimum: 0 m.
Default: 500 m.
The half-width of the spatial smoothing operator applied to the picked velocity residuals in the crossline direction, in meters. Works analogously to Smooth residuals window X. Active only when Smooth residuals along horizons is enabled. Minimum: 0 m.
Default: 500 m.
This parameter group contains the settings that define the velocity perturbation scan range and the shape of the residual moveout model used during DM semblance computation, as well as optional semblance smoothing.
The length of the time window in seconds used when computing DM semblance at each offset-depth sample. A longer window averages semblance over more of the reflection wavelet, producing a smoother panel but reducing temporal resolution. Typical values are 0.02–0.08 s. Range: 0 to 1 s.
Default: 0.04 s (40 ms).
The minimum velocity perturbation in m/s included in the semblance scan. The DM semblance is computed for a range of velocity corrections from Min delta V to Max delta V at intervals of Step delta V. Set Min delta V to a sufficiently negative value to capture the case where the current model is too fast. Range: -10000 to 10000 m/s.
Default: -1000 m/s.
The maximum velocity perturbation in m/s included in the semblance scan. Set Max delta V to a sufficiently positive value to capture the case where the current model is too slow. Together with Min delta V and Step delta V, this defines the full search range displayed on the semblance panel. Range: -10000 to 10000 m/s.
Default: 1000 m/s.
The velocity step in m/s between successive semblance panels in the scan. Smaller steps produce a finer velocity grid on the semblance display, allowing more precise picking, but increase computation time proportionally. A step of 50–200 m/s is typical. Minimum: 0.001 m/s.
Default: 100 m/s.
The functional form assumed for the velocity perturbation when computing modeled hodographs for each semblance trial. Options are:
Constant — the velocity correction is uniform throughout the layer. Use this for simple, laterally homogeneous layers.
Gradient — the correction varies linearly with depth inside the layer, scaled by the Gradient value parameter. Use this when the layer has a significant velocity gradient.
Parabolic — the correction follows a parabolic function of offset, accommodating anisotropy or higher-order moveout effects.
Default: Constant.
When enabled, the computed DM semblance volume is smoothed spatially (X and Y) and in the velocity axis (Z) before being displayed and picked. Smoothing suppresses random noise in the semblance and makes the peak easier to identify visually and for automatic picking. When enabled, the Smooth semblance window X, Y, and Z parameters become editable.
Default: Off (false).
The spatial smoothing half-width applied to the DM semblance in the inline direction, in meters. Larger values produce a smoother semblance panel at the cost of reduced lateral resolution in the velocity estimate. Active only when Smooth semblance is enabled. Minimum: 0 m.
Default: 500 m.
The spatial smoothing half-width applied to the DM semblance in the crossline direction, in meters. Works analogously to Smooth semblance window X. Active only when Smooth semblance is enabled. Minimum: 0 m.
Default: 500 m.
The smoothing half-width applied to the DM semblance in the velocity axis (delta-V axis), in m/s. Smoothing in velocity helps merge closely spaced semblance peaks caused by noise into a single coherent maximum, making automatic picking more reliable. Active only when Smooth semblance is enabled. Minimum: 0 m/s.
Default: 500 m/s.
This group contains fine-tuning parameters for the layer boundary handling during the velocity update. These control how the boundary samples are treated at the top and bottom of the selected layer when the model is modified.
The number of depth samples by which the top boundary of the current layer is shifted inward (into the layer) before applying the velocity update. A positive value trims samples from the top of the layer to avoid contaminating the updated interval with the velocity of the overlying layer. Range: -20 to 20 samples.
Default: 5 samples.
The number of depth samples by which the bottom boundary of the current layer is shifted inward before applying the velocity update. A positive value trims samples from the bottom of the layer to prevent the correction from bleeding into the next layer below. Range: -20 to 20 samples.
Default: 2 samples.
This group contains the spatial interpolation settings used when spreading the picked velocity corrections from the sparse analysis grid onto the full velocity model grid. The interpolation parameters (method, covariance type, search radius, and number of neighboring points) are inherited from the base interpolator component and control how smoothly the corrections are extrapolated between analysis locations. Kriging or triangulation can be selected depending on the lateral complexity of the expected velocity variations.
The module produces numerous output gathers displayed interactively in the vista panels. The key outputs for the velocity update workflow are the Updated depth velocity and Depth delta gathers. The stack and CIG outputs are used for quality control.
Stack inline time / Stack crossline time — the CMP stack along the selected inline and crossline in the time domain, after datum correction. Used to verify that the current velocity model produces well-focused, flat reflectors.
Stack inline time after time processing / Stack crossline depth after time processing — the same stacks after additional time-domain sub-processing (e.g., filtering applied via the sub-procedure chain). Used for comparison with the raw stack to evaluate the effect of post-stack processing.
Stack inline depth / Stack crossline depth — the depth-domain CMP stacks along the selected inline and crossline, produced by depth migration of the raw data with the current velocity model. Flat events in these panels indicate a correct velocity model for the corresponding layer.
Inline V depth / Crossline V depth — the interval velocity model displayed along the selected inline and crossline in depth, before any well-tie correction.
Tied inline V depth / Tied crossline V depth — the velocity model after well-tie calibration, displayed along the inline and crossline sections.
Vertical time tables inline / Vertical time tables crossline — the vertical travel-time table (one-way time as a function of depth) computed from the current interval velocity model, displayed for the selected inline and crossline. Useful for QC of time-depth conversion consistency.
VRMS from interval cube / VRMS from interval datum inline / VRMS from interval datum crossline — the RMS velocity derived from the interval velocity model, displayed in the time domain. Use these to cross-check against independently picked Vrms curves.
Selected CIG time / Selected CIG depth / Selected CIG depth muted — the common image gather for the selected bin, displayed in time domain, depth domain, and depth domain with mute applied. These are the primary QC displays for assessing gather flatness and the quality of the velocity correction for the current bin.
Updated depth velocity — the full depth-domain interval velocity model after the velocity corrections from the current layer update have been applied. Connect this output back to the Depth velocity input to proceed with the next layer or the next iteration.
Depth delta — the spatial map of velocity corrections (delta-V) applied to the current layer, stored as a depth-domain gather. Inspect this output to verify that the corrections are geologically reasonable and spatially consistent before accepting the update.
Expected CIG time — the modeled CIG hodograph computed from the picked velocity picks, shown in the time domain. Overlay this with the observed CIG to assess how well the picked correction explains the residual moveout.
Removes all mute picks currently stored in memory, resetting the mute to a fully open state (no muting applied). Use this to start fresh when the existing mute definition is no longer appropriate for the current dataset or bin selection.
Loads a previously saved mute definition from a file and applies it as the current mute picking. Use this to restore a mute that was saved in an earlier session or to apply a mute prepared for a different module to this workflow.
Saves the current mute picking definition to a file for later reuse. Run this action after interactively adjusting the mute in the depth gather vista to preserve the mute before closing or reconfiguring the module.
Computes and saves the complete inline and crossline depth stacks covering all bins in the dataset, using the current velocity model. This action is more computationally intensive than a single-bin display update and writes the stacks to disk for later inspection or export. Run this after finalizing a layer update to produce QC images of the full section.
Computes and saves the full set of common image gathers for all bins using the current (updated) velocity model. The saved CIGs can be used for AVO analysis, further velocity updating in subsequent iterations, or export to an external workflow.
Initializes the horizon-based inversion filter for the currently selected layer. This action must be run before computing the DM semblance for a new layer or after changing the Current layer selection. It reads the input horizons, builds the internal hodograph geometry, and prepares the semblance calculation engine. If the semblance panel is empty or incorrect, run Init filter first.
Computes the DM semblance for the selected layer across the full scan range defined by Min delta V, Max delta V, and Step delta V. This is the main analysis step of the velocity update workflow. After running Init filter, click this action to populate the DM semblance panels for inline and crossline directions. The result is displayed as a semblance-versus-delta-V panel from which the user picks the optimal velocity correction.
Automatically picks the semblance peak at each analysis location for the selected layer and stores the picks. The auto-picker finds the delta-V value with maximum semblance at each grid point. Run this action after Calculate layer semblance to get an initial set of picks that can then be edited manually in the DM semblance vista panels before applying the update.
Applies the current DM semblance picks to the velocity model, updating the interval velocity of the selected layer. The picked delta-V corrections are spatially interpolated (and optionally smoothed) across the analysis grid and then added to the existing velocity values within the layer boundaries. After running this action, the Updated depth velocity output reflects the corrected model. Verify the result using the depth stacks and CIG outputs before proceeding to the next layer.
Exports the current DM semblance picks (the picked delta-V corrections at each analysis location) to a plain-text ASCII file. This allows the picks to be archived, shared with other processing applications, or reloaded in a future session via Import residual picking from ASCII file.
Loads DM semblance picks from a previously exported ASCII file and restores them as the current active picks. Use this to resume a velocity update session that was interrupted, or to apply picks produced by an external quality-control step without re-running the semblance computation.