Although the SEG standard for SEGY does not apply to stacked data, stacks have been stored on tape media for quite some time in a format roughly following the SEG publication. For tape media, the EBCDIC, the binary file header, and individual traces (single trace blocks) are put on tape, each separated by an inter-record gap and then followed by a single EOF. Several lines may be stored on tape in the above sequence with one EOF between lines and two EOFs after the last line. Subsequent tapes in multi-tape volumes should have EBCDIC and binary headers at the start of each tape.

Stacked data has also been stored as files on disk for use by workstations and as a method of exchange, primarily on CDs. Many different versions of “SEGY” files for stacked data now exist because of the absence of a SEG standard and because several companies want their own customized version.

This article outlines a GENERIC SEGY format for stacked data, providing essential information and workstation data formats while maintaining the original SEGY specifications as much as possible.

The stacked data for each line may be stored in a single disk file consisting of the 3200 byte comment header, the 400 byte binary header, and the data traces, each trace having a 240 byte binary header.

It is recommended that the file and trace binary headers contain only 2-byte or 4-byte integers (in IBM mainframe byte order) and that the 4-byte integers be aligned on 4-byte boundaries. Alphanumeric or float values are not recommended, since neither their format nor their locations are specifically defined anywhere in published form. The use of integers only conforms with the SEG standard. Shotpoint numbers are multiplied by 1000 so that float values are not necessary.

To accommodate ease of use on workstations and PCs, the comment header may be in EBCDIC or ASCII (detection is easily accomplished by examining the space characters), and more data formats are allowed:

6 for SUN IEEE 32-bit floating point
11 for PC or DEC IEEE 32-bit floating point

The only difference between formats 6 and 11 is the byte order of the trace data. Format 11 is introduced here because of the dominance of the PC in the world today and the increasing use of that platform for workstations. However, it is recommended that format 6 be used, even by PC workstations, rather than 11 so that data files remain compatible with current UNIX workstations. The SEGY format 1 bug is addressed here by the use of format 6 for IEEE 32-bit floating point, format 1 being used only for the standard SEGY IBM 32-bit floating point.

The binary file and trace headers maintain IBM (or SUN) byte ordering since these headers contain values of variable byte sizes. Only the comment header and the trace data are allowed to have formats not specified in the SEG publication since their conversion involves a single do loop of all elements. Also, since the unassigned portions of the binary file header and binary trace headers are not specifically documented, these values should be all 32 bit integers.

This is a very minimal specification and workstations should be able to access this data with little or no programming changes.

This outline closely follows the suggestion of the 1994 CSEG committee report except for some important items which were omitted or misplaced (indicated by a *). These include the treatment of 3D stacked data similar to that of 2D shot data: several inline records containing crossline traces. This allows the cdp number for 2D and 3D data to occupy byte locations 21-24 in the trace header as originally specified in the SEG standard.

3200 byte Comment header, 2D and 3D data

* Client name
Line name
Area name
* Acquisition parameters
Processing center and processing date
Datum elevation used for structure statics
Time to first sample
Processing history as detailed as possible
Survey datum (NAD 27, NAD 83)
UTM Zone or central meridian
Header byte locations for items not mentioned here.

Additional Information for 3D:
* NE corner: Trace count, XUTM, YUTM
* NW corner: Trace count, XUTM, YUTM
* SE corner: Trace count, XUTM, YUTM
* SW corner: Trace count, XUTM, YUTM

400 byte binary file header, 2D and 3D data

Bytes Format Value
* 13-14 16 bit integer Number of data traces/record: 1 for 2D stacks number of traces/inline for 3D stacks
17-18 16 bit integer Sample rate of this data in microseconds
21-22 16 bit integer Number of samples for this data
25-26 16 bit integer Data Sample format code:
1 IBM 32 bit floating point
2 32 bit integer
3 16 bit integer
6 SUN IEEE 32 bit floating point
8 8 bit integer
11 PC IEEE 32 bit floating point
* 27-28 16 bit integer CDP fold
* 29-30 16 bit integer Trace sorting code: 4=stacked
55-56 16 bit integer Measurement system: 1=metres
The following locations in the binary file header may be used for optional additional information and were chosen to minimize any conflicts with the existing SEIS-X locations.
* 97-100 32 bit integer The time of the first sample in ms. May be negative, zero, or positive.
Information for 2D line extents:
* 133-136 32 bit integer (station_number*1000) @ start of line
* 137-140 32 bit integer (station_number*1000) @ end of line
Information for possible data base searching: (Latitudes and Longitudes are in seconds of arc)
* 141-144 32 bit integer Latitude minimum
* 145-148 32 bit integer Longitude minimum
* 149-152 32 bit integer Latitude maximum
* 153-156 32 bit integer Longitude maximum
* 157-160 32 bit integer Central meridian of the UTM coordinates
Information for 3D corner locations: (This may facilitate more automatic workstation loading)
* 161-164 32 bit integer UTM zone
* 165-168 32 bit integer Coordinate scaler:
1 for UTMs in metres
-10 for UTMs in decimetres
* 169-172 32 bit integer NE corner UTM X
* 173-176 32 bit integer NE corner UTM Y
* 177-180 32 bit integer NW corner UTM X
* 181-184 32 bit integer NW corner UTM Y
* 185-188 32 bit integer SE corner UTM X
* 189-192 32 bit integer SE corner UTM Y 
* 193-196 32 bit integer SW corner UTM X
* 197-200 32 bit integer SW corner UTM Y

240 byte trace headers, 2D and 3D data

Bytes Format Value
1-4 32 bit integer Trace count within line
5-8 32 bit integer Trace count within volume
* 9-12 32 bit integer 3D Inline number
* 13-16 32 bit integer 3D Trace number within the inline This is also the crossline number
* 17-20 32 bit integer 2D Interpolated station number above this CDP multiplied by 1000
* 21-24 32 bit integer CDP number, 2D or 3D
29-30 16 bit integer Trace ID code: 1=live trace, 2=dead trace
* 41-44 32 bit integer Surface elevation above cdp if available.
* 69-70 16 bit integer Elevation scalar:
1 for elevations in metres
-10 for elevations in decimetres
71-72 16 bit integer Coordinate scalar:
1 for UTMs in metres
-10 for UTMs in decimetres
81-84 32 bit integer  X coordinate of CDP
85-88 32 bit integer Y coordinate of CDP
89-90 16 bit integer Coordinate units: 1 for distance

All traces, even zero fold dead traces should contain proper header information.

Bytes 115-118 (the number of samples and the sample interval) may be set but should not be required since this information must be in the binary file header, bytes 17-22, and these values should be the same for all traces within the file.

The SEG has not endorsed any changes to it’s original SEGY specification of 1975 or any changes recommended by groups such as the CSEG. Because of the improvements in hardware, media, processing and related information, a new exchange standard for processed seismic data is needed, but this is not likely to occur in the near future. Until then, a simple workstation-independent specification that deviates as little as possible from the SEGY stan-dard seems to be the most appropriate.

Comments from the readers are welcome and should be addressed to the Editor.

References

“Recommended Standards for Digital Tape Formats”, (SEGY) Barry et al. Digital Tape Standards, Society of Exploration Geophysicists, 1980

“CSEG Geophysical Workstation SEGY Standards”, Bath et al., CSEG Recorder, June 1994