Topic:
How  is seismic data stored in IES and IESX

Description:
        

Solution:
IES:

8 bit storage.

Loading 32 bit data to IES 8 bit storage (such as SEGY tape loading into IES
system). The data is processed trace by trace. Each 32 bit floating point
input trace is scanned for a maximum absolute value (MAXVAL). A scalar
(SCALAR) is then calculated, where SCALAR = MAXVAL / 128, which is stored
with the 8 bit IES output trace. On an 8 bit IES output trace, there are 256
possible unique values (VAL) (ranging from -128 to +127). The amplitude of
each floating point sample of the input trace (FLOATAMP) is divided by
SCALAR, such that VAL = FLOATAMP / SCALAR, and a VAL is stored for each
sample of the output 8 bit IES trace. When an IES trace is read back in an
interpretation session, it is converted back into a 32 format by multiplying
each sample VAL by the stored SCALAR for that trace, to form a
"reconstituted" 32 bit value. After this process is complete for each trace,
the trace will contain sample values across the same range of floating point
numbers as the original 32 bit SEGY data, but limited to a maximum of 256
unique values.  

16 bit storage.

Loading 32 bit data to IES 16 bit storage (such as SEGY tape loading into
IES system). The data is processed trace by trace. Each 32 bit floating
point input trace is scanned for a maximum absolute value (MAXVAL). A scalar
(SCALAR) is then calculated, where SCALAR = MAXVAL / 32768, which is stored
with the 8 bit IES output trace. On an 8 bit IES output trace, there are
65536 possible unique values (VAL) (ranging from -32768 to +32767). The
amplitude of each floating point sample of the input trace (FLOATAMP) is
divided by  SCALAR, such that VAL = FLOATAMP / SCALAR, and a VAL is stored
for each sample of the output 16 bit IES trace. When an IES trace is read
back in an interpretation session, it is converted back into a 32 format by
multiplying each sample VAL by the stored SCALAR for that trace, to form a
"reconstituted" 32 bit value. After this process is complete for each trace,
the trace will contain sample values across the same range of floating point
numbers as the original 32 bit SEGY data, but limited to a maximum of 65536
unique values.

IESX:

Both 8 bit and 16 bit storage in IESX work similarly to IES, with one
difference. Instead of scanning purely for a trace absolute maximum to
calculate the scalar, the minimum and maximum actual values on the trace are
determined, and the data compression is done only between those two values.
To illustrate this difference, assume a trace with original values that
range between -25,000 and +40,000. IES would have found the absolute maximum
value (40,000) and would have calculated a scalar to fit the signed range of
numbers -40,000 to +40,000 (of which the value range -40,000 <= FLOATAMP <
-25,000 would not be used). IESX records the amplitude range of -25,000 to
+40,000 and scales the data precisely within that range. 

The IESX method is therefore an overall improvement on the IES method.

In general, the 8 bit storage compression method retains absolutely the
trace to trace amplitude variations along a seismic line. If a seismic line
has had short window time variant post stack scaling, or AGC, applied, then
8 bit storage in IES/IESX should provide adequate dynamic resolution for
most interpretation needs. If the aim of the interpretation includes
amplitude studies, and no windowed scaling is applied to the seismic data
for that reason, then 8 bit, and in some cases even 16 bit, storage in
IES/IESX will probably not allow adequate amplitude resolution. A general
rule for the interpreation and analysis of unscaled seismic data would be to
use 32 bit storage in IES/IESX.




Last Modified on: 14-SEP-98