Seismic data processing · Interpretation · Software development

From field records
to final image.

SDPC is a geophysical services company. We process and interpret seismic data, and we build the software that does it — grounded in physics, validated on benchmark models, engineered without shortcuts.

Fig. 1 — Variable-area wiggle display. Synthetic reflectivity convolved with a 28 Hz Ricker wavelet, rendered live.

01

Services

Three disciplines, one team. Every project runs through transparent, reproducible workflows with full dataset lineage — you see exactly what was done to your data, and why.

Seismic Data Processing

Complete land processing sequences from raw field data through final gathers and stacks — with particular depth in near-surface model building and signal preservation.

  • First-arrival traveltime tomography, GLI & refraction statics
  • Surface-consistent amplitude, deconvolution & residual statics
  • Noise attenuation: FX/FXY, wavelet, Radon, FK/FKK & 30+ methods
  • 5D interpolation & regularization (MWNI, POCS, MSSA)
  • Velocity analysis, NMO/RMO, trim statics
  • Kirchhoff PSTM / PSDM and RTM imaging

Interpretation & Quantitative Analysis

From structural imaging to rock properties — attribute-driven analysis that connects seismic response to geology.

  • 40+ seismic attributes: spectral, geometric, texture (GLCM)
  • Coherence, curvature & discontinuity mapping
  • AVO intercept–gradient & fluid analysis
  • Rock-physics modelling from well logs
  • Multi-attribute classification with machine learning
  • Integrated well-log (LAS) workflows

Geophysical Software Development

We take algorithms from the primary literature to production code — for ourselves and for clients who need capabilities no commercial package offers.

  • Custom processing algorithms & plugins
  • GPU compute kernels (Apple Metal, native C++)
  • Numerical solvers: FD modelling, eikonal, ADMM inversion
  • Data infrastructure: SEG-Y / SEG-D readers, Zarr stores
  • Clean architecture, protocol-driven design
  • Automated testing & validation suites
02

The SDPS Platform

Our flagship in-house system: a complete seismic data processing platform with a visual flow editor, plugin architecture, and GPU-accelerated kernels. Every service we offer runs on it — and every line of it is ours.

25 processing plugins
80 flow-node algorithms
3,400+ automated tests
360k lines of scientific Python

Imaging & Modelling

  • Kirchhoff PSTM — VTI anisotropy, curved-ray, true-amplitude weighting
  • Kirchhoff PSDM with factored-eikonal traveltimes
  • 2D/3D acoustic RTM — PML boundaries, adjoint checkpointing
  • Finite-difference forward modelling
  • 3D ray-tracing illumination & survey design

Near-surface & Statics

  • First-arrival traveltime tomography (FATT)
  • Multiscale TV-regularized tomography — ADMM, fast-sweeping adjoint
  • GLI & refraction / delay-time statics
  • Residual & trim statics (windowed cross-correlation)
  • Interactive first-break picking

Signal Processing

  • Surface-consistent amplitude, decon & statics
  • Spiking & predictive deconvolution, spectral balancing
  • Radon transforms — linear, parabolic, hyperbolic
  • FK / FKK / dip filtering, radial trace transform
  • 30+ denoise methods: FX-decon, wavelet, SST, VMD, BM3D, NLM…

Velocity & Regularization

  • Interactive semblance velocity analysis with live stack
  • Velocity model building & editing tools
  • NMO / RMO correction, RMO picking
  • 3D–5D interpolation: MWNI (beyond-alias), POCS, MSSA

Attributes, AVO & QI

  • 40+ attributes: instantaneous, spectral decomposition, sweetness
  • Eigenstructure coherence, curvature suite, GLCM texture
  • AVO — Shuey intercept/gradient, angle gathers & mutes
  • Rock-property modelling from well logs
  • Multi-attribute classification & clustering

Data & Platform

  • SEG-Y import/export with full header mapping
  • SEG-D rev 3 — including STRYDE nodal data
  • SPS geometry, LAS well logs
  • Zarr + Parquet internal store, dataset lineage
  • Visual flow (DAG) editor, 2D/3D viewers, PPTX reporting
  • Native GPU kernels for Apple Silicon (Metal), CPU fallbacks
03

A Scientific Approach

Geophysics rewards rigor. Our working principles are simple to state and expensive to keep — which is exactly why we keep them.

§1

Physics first

Algorithms are implemented from the primary literature and verified against analytic solutions and synthetic references before they touch real data.

§2

Benchmark validation

Imaging and inversion codes are validated on standard models — Marmousi, controlled synthetics, and field-data comparisons — with documented results.

§3

Fail fast, never silently

No silent fallbacks, no graceful degradation, no placeholder results. If something is wrong with the data or the parameters, you hear about it immediately.

§4

Reproducibility by design

Every dataset carries its full processing lineage. Flows are deterministic and re-runnable; 3,400+ automated tests guard every release.

Fig. 2 — 28 Hz Ricker wavelet (left) and its amplitude spectrum (right), as used in Fig. 1.
04

Work with us

Whether you have a survey to process, a dataset to understand, or an algorithm that needs to become production software — tell us about it.

Email

ohnordavind@gmail.com

Focus areas

land seismic · near-surface tomography · depth imaging · AVO & QI · nodal (STRYDE) data · custom geophysical software