User guide

Upload raw DAS data, review the result, and remove the server files.

HADRON FieldLab is a public technology demonstration for evaluation testing. It is built to accept common DAS and seismic recording formats, return a reviewable result package, and keep the workflow understandable without exposing proprietary implementation details.

1. Before You Upload

Use the site only with data you are authorized to test. The demo server deletes uploaded raw recordings after processing. Result files are kept temporarily so you can download them, then you can delete them from the page.

  • Supported primary input families include HDF5, NetCDF4, TDMS, SEG-Y, MiniSEED, SAC, NPY, NPZ, and CSV/TXT arrays. ZIP archives and TensorFlow record exports are not accepted as primary waveform uploads in the public demo.
  • Optional sidecars can add event windows, geometry, or metadata when you have trusted project context.
  • For commercial use, API access, implementation, publication, redistribution, or operational use of processed outputs, contact JAWD AS first.

2. Run A File

  1. Open hadronfieldlab.com.
  2. Select the raw recording in the upload panel.
  3. Choose the closest dataset preset. If unsure, use the standard format preset that matches the file family.
  4. Add optional event-window, geometry, or metadata sidecars only if they are relevant and accurate.
  5. Press Start run and keep the tab open while the server processes the file.
HADRON FieldLab upload workspace screenshot

3. Read The Result

The result status is intentionally conservative. Enhanced means the automated quality gates passed on that input. Review means there may be useful change, but a domain reviewer should inspect the report before using it. No safe improvement means the run did not find a defensible improvement under the current input and metadata.

The headline metrics compare the original input with the processed output. They are meant to support a scientific review, not replace it.

  • SNR gain: median improvement in event energy relative to measured quiet-window noise.
  • Signal/noise AUC: how well event windows separate from quiet windows. Higher is better; 1.0 is the upper bound.
  • Waveform match: preservation of event-window waveform shape. Values near 1.0 indicate close timing and morphology agreement.
HADRON FieldLab result screenshot

4. Download And Clean Up

A completed run can return enhanced data, a portable HDF5 copy, a PDF report, a preview image, an interactive preview, a machine-readable result summary, and supporting metrics. Download what you need for review, then use Delete to remove the result files from the demo server.

This demo does not grant permission to publish, redistribute, commercialize, or operationally use processed outputs. For those uses, contact JAWD AS at jawdas@jawdas.com.