Skip to content
TensorBundle
TensorBundle
Get in touch
EN TR
SeismicBundle

Seismic processing and interpretation software

SeismicBundle turns advances in AI into tools for seismic processing and interpretation. Its tools are built for field data, helping geophysicists automate routine work and tackle more demanding analysis.

Request access See the modules
SeismicBundle / North Basin 3D
A shot gather after noise attenuation Shot 412
Offset (m) · Time (ms) Amplitude − +

Modules

  1. A marine shot gather before and after swell noise attenuation Raw shot After processing

    Processing

    Prepare prestack data for velocity analysis by attenuating noise, removing multiples and correcting for ghosts. You can also separate blended shots and reconstruct missing traces. Automatic first-break picking reduces the number of picks you need to make by hand.

  2. A migrated section before and after image enhancement Conventional SeismicBundle

    Velocity and imaging

    Velocity picking, model building and migration are part of the same workflow. Update the velocity model with tomography, then use Kirchhoff or reverse time migration to image the subsurface. The module supports both time and depth imaging.

  3. A velocity model before and after full waveform inversion Starting model After FWI

    Full waveform inversion

    Start with a model built from tomography and well data, then refine it with full waveform inversion. For field recordings with limited low-frequency coverage, the module estimates the missing content and controls cycle skipping as the inversion progresses to higher frequencies.

  4. An inverted property section with three wells

    Seismic inversion

    Our inversion models estimate impedance and other rock properties from seismic and well data. Each prediction includes an uncertainty estimate, so you can consider the range of possible values when interpreting the subsurface.

  5. A seismic section with two faults and three horizons

    Interpretation

    Automatic fault detection and horizon tracking reduce repetitive work across 3D seismic volumes. The same module extracts salt bodies, channels and other geobodies. You can examine these structures in individual sections or across the full volume.

Our seismic research

Our seismic research focuses on three constraints: sparse well coverage, missing low frequencies and limited interpretation labels.

See TensorBundle's research areas

  • Learning from unlabeled seismic

    Our seismic foundation models learn geological patterns from unlabeled surveys. With a small set of expert picks, you can adapt them to a new area's geology.

  • Inversion with geological constraints

    Our probabilistic inversion models combine seismic, well logs and geological constraints to estimate rock properties. They return plausible property ranges alongside the preferred model.

  • FWI with limited low frequencies

    Learned low-frequency extrapolation gives FWI a starting point when field recordings lack low frequencies. Adaptive inversion objectives reduce sensitivity to cycle skipping.

  • Interpretation in three dimensions

    Our interpretation models follow geological structures across neighboring inlines and crosslines. Reading these sections together helps distinguish faults and horizons that are ambiguous in a single section.

Evaluate SeismicBundle on your survey

Start with a problem in your own survey. We'll evaluate the relevant modules on a sample of your data and review the output alongside your existing results.

Request an evaluation

Request access

Tell us about your survey and what you'd like to test. We'll work out which modules to use, what data we'll need and how to compare the outputs with your current results.

Or email us at [email protected]

Modules you're interested in