NAME
velspec - obtain time and rms velocity semblance of data
set
SYNOPSIS
velspec [ -Nntap ] [ -Ootap ] [ -Ffpik ] [ -reire ] [ -rsirs
] [ -nsns ] [ -nene ] [ -tminstart ] [ -tmaxstop ] [ -vminv-
min ] [ -vmaxvmax ] [ -nvelnvel ] [ -igtigt ] [ tmultmul ] [
dmuldmul ] [ thldthresh ] [ -L ] [ -A ] [ -S ] [ -W ] [ -log
] [ -V ] [ -? ]
DESCRIPTION
A velocity spectrum (linear or hyperbolic) is obtained by
scanning the CMP gather of traces along trajectories for
signal coherence. These scans establish a velocity function
(i.e., velocity versus vertical seismic time) that is used
in calculating the NMO or LNMO corrections prior to the
stacking of the CMP traces. The output is an sis disk (or
pipe) file of the semblance values. velspec takes a set of
i traces and performs a time-trajectory-scan operation to
obtain stacking velocity as a function of stacking time via
a coherence measure (semblance). The method used is the
following. If there are N traces in a CMP gather at dis-
tances r(i), a hyperbolic time function
t(r(i)) = sqrt[t(0)**2 + (r(i)/V)**2],
is swept along the time-axis of t-r plane of traces for a
given set of test velocities, V.
For linear velocity analysis the equation is trivial:
t(r(i)) = r(i)/V
The velocity V that produces the maximum semblance
represents the best-fitting hyperbola and thus produces the
best stack at stacking time t(0). The semblance (Taner and
Koehler, 1969) is defined as
CN(V,t) = [ SUMt [SUMi A(r(i),t) ]**2 ]/[ N SUMt SUMi
[A(r(i),t)]**2 ]
where the SUMi is from i = 1 to N; SUMt is the time-gate sum
from t-T/2 to t+T/2 at time t(0); and A(r(i),t) is the trace
amplitude at distance r and time t.
The procedure is repeated for a collection of zero-offset
(vertical) times t(0), at equally spaced time increments.
The result is called the velocity spectrum at the common
midpoint. The velocity spectrum consists of an sis work file
of semblance values.
A velocity fairway can be input in the form of two xsd pick
segments (made presumably from a typical semblance spectrum)
to limit the semblance calculations. If no fairway is pro-
vided the entire semblance from vmin to vmax is calculated.
velspec gets processing controls from the command line. Rea-
sonable defaults are set up.
Command line arguments
-N ntap
Enter the full path of the file continuing the data
set. If not specified, input is expected on the stan-
dard input. If standard input is not specified, and
there is no input, e.g., program run in background,
expect a very ungracious crash. (default standard
input)
-O otap
Enter the full path of the file to which the output
semblance is to be written. Default is stdout (pipe
out).
-F fpik
(Optional) Enter the full path of the file containing 2
xsd pick segments made on a typical semblance spectrum,
i.e. the user should run a sample record of his data
through velspec and output the entire semblance spec-
trum (no fairway pick file). Then in xsd pick a lower
velocity-time function and an upper velocity-time func-
tion to form a fairway. When this is input in the
major velspec run no semblance calculations will be
made outside this fairway significantly reducing run
time. When saving the picks be sure to fill out the
velocity units and offset for the trace fields (e.g.
for a velspec run using 100 velocities and a vmin=500
and a vmax=6500, i.e. a velocity increment of 60, the
trace units will be 60 and the offset will be 440 [so
the first trace will come out to be 500]).
-rs irs
First record for processing (default = 1)
-re ire
Last record for processing (default = last)
-ns ns
Define the starting trace for analysis (default = 1)
-ne ne
Define the last trace for analysis in each record
(default = ns). The use of the ns and ne pair are use-
ful to be able to skip dead traces or to select just a
portion of the data set for analysis which is loga-
rithmic scale of 2 cycle length.
-L If present linear velocity scans between the minimum
and maximum velocities are performed. Useful for
analyzing linear coherent noise.
-A The semblance is a quantity bounded above by +1. If -A
is invoked the contour plot are scaled so the maximum
is set to 1.0. The effect is to boost the level of the
semblance plot. Otherwise the semblance is an absolute
value lying between 0 and 100
-S If present, calculate the velocity spectra in equal
increments of slowness squared, where
v=sqrt(slowness**2).
-log If present, the velocity (or slowness) scale is loga-
rithmic. This can often result in superior sampling of
the scans.
-V Flag for verbose output on the standard output. This
lists salient trace header information as well as all
command values used for processing, the default ones
given if not overridden.
-tmin start
Starting time of traces in milliseconds (default = 0,
beginning of trace).
-tmax stop
End time in each trace for analysis in milliseconds
(default = end of trace).
-vmin vmin
Minimum velocity value input in ft/sec or m/sec
(default = 1000.0).
-vmax vmax
Maximum velocity value input in ft/sec or m/sec
(default = 10,000.0).
-nvel nvel
Maximum number of velocities to use spaced between vmin
and vmax (default =101)
-igt igt
An integer value (default = 8) to give the time-gate
window T centered at t at time t(0). The half-time
gate T/2 is related to igt by T/2=igt*dt, where dt is
the sample interval in ms. Time gate is small (Robin-
son, 1983). Time gate is usually tens of milliseconds
and generally is approximately equal to the period of
the dominant frequency in the data (Schultz, 1984).
Within this gate the default is to apply a cosine bell
weighting. This has the effect of tightening the sem-
blance wraps over a simple box car (unit) weighting.
-W Include -W on the command line to turn off cosine
weighting within the time gate.
-dmul dmul
Multiply each distance, itr(117), by this factor
(default=1.0). This is useful when the header has dif-
ferent units than expected.
-tmul tmul
Multiply the sampling interval, dt, by this factor
(default = 1.0). This is useful if the sampling inter-
val is in different units than expected.
-thld thresh
Semblance threshold: semblances less than this value
will be zeroed. Default = 0.0 (no thresholding). With
the -A option (max semblances scaled to 1.0) this
represents a relative thresholding.
-? Query mode. With this flag, velspec will give a
description of the command line arguments and stop the
program.
NOTE 1
Plotting the spectra as a function of slowness**2 equalizes
discretization errors in the display for picking. An error
of plus or minus one trace at 1500m/s is 27 times greater
than an error of plus or minus one trace at 4500m/s since
the sensitivity goes as 1./v**3 ! Plotting the spectra as a
function of slowness**2 equalizes this discretization error
to be proportional to the T0 time. SH BUGS It is assumed
that the record(s) to be analyzed should be in a CMP gather.
SEE ALSO
pomega, velin
REFERENCES CITED
Hubral, P. and T. Krey (Larner, K. L., Ed.) (1980). Interval
velocities from seismic reflection time measurements,
Society of Exploration Geophysicists, 203pp.
Robinson, E. A. (1983). Seismic velocity analysis and the
convolution model, IHRDC, 290pp.
Schultz, P. S. (1984). Seismic velocity estimation, in
Proceedings of the IEEE, Oct. 1984, 1330-1339.
Taner, M. T. and F. Koehler (1969). Velocity spectra--Digital
computer derivation and applications of velocity functions,
Geophysics, 34, 859-881.
AUTHOR
B. V. Nguyen and R. B. Herrmann, Saint Louis University,
April 1986.
COPYRIGHT
copyright 2001, Amoco Production Company
All Rights Reserved
an affiliate of BP America Inc.
Man(1) output converted with
man2html