Seismic Data Compression Reference Set
Introduction
This page contains some seismic data. The intention is to make some example data available to
everybody who want to use some real seismic data to test their program or function.
This page should make it possible for us
working with seismic data, compression especially, to use some common data.
Note that we at UiS have done little work in this field this millenum (since year 2000),
and this page has only had cosmetic changes since then.
Seismic data format
SEG-Y is the dominant format for seismic data. Most data here are in another format,
also used by Seismic Unix,
the main difference is that floats are IEEE float in Seismic Unix format
while SEG-Y format uses IBM-floats.
To read these files into Matlab you may use the routines in
SeismicLab
or the slightly modified ones I have used and packed into
readsegy.zip, 4 kB,
(these four files packed: readsegy.m, header.m, segy.mat, count.mat).
Data may then be read into Matlab with the following command:
[X,H] = readsegy('filename','tracl',1,120,1);
Some files may be in Matlab MAT-format and can be loaded into Matlab directly by the load command,
or they may be gzipped Matlab files which must be unzipped before loaded into Matlab.
Pre-stack data
- Common Shot Gather, 1257 kB.
This file is 120 traces marine pre-stack data from the North Sea.
I got this file from Per Tjora, PetroData.
PetroData has given permission that the data may be used for research purposes.
Each trace is a 240 byte header as defined in SEG-Y standard
followed by 2560 samples (IEEE floating point 4 byte). This gives 120*(240+2560*4)=1257600 bytes.
- Trace 112, 20 kB.
This file is the 2560 samples from trace 112 in previous file. File is in MatLab format uncompressed.
Post-stack (or just stacked) data
- stack.mat.gz, 1092 kB.
This file is 512x512 samples of stacked samples from somewhere in North Sea.
I got this file from Tage Røsten,
and he got it from Statoil.
Statoil has given permission that the data may be used for research only.
These data are also used in the JPEG2000 training set.
The format is gzipped MAT-file.
This WWW page is maintained by
Karl Skretting at
University of Stavanger.
Last update: 14.Jan.2015. Comments etc. on this web page may be sent to
karl.skretting@uis.no.