A user had problems with some SEG-Y files from the
Knudsen 320 Echosounder. The seafloor changed very rapidly
and while the realtime display followed the data, the SEG-Y
data files did NOT - data were dropped from the SEG-Y file.
(Click here for a description and a way to prevent it.)
The Melville 2000 example has some additional information.
(plot1 and plot2).
These are some of the steps I used to examine the problem. Two
bad spots were day 232 1008z to 1012z and day 232 2326z to 2332z.
ls *sgy
L021_001.sgy L023_003.sgy L024_006.sgy L024_014.sgy L025_004.sgy L027_004.sgy
L021_002.sgy L023_004.sgy L024_007.sgy L024_015.sgy L025_005.sgy L028_001.sgy
L021_003.sgy L023_005.sgy L024_008.sgy L024_016.sgy L026_001.sgy L028_002.sgy
L022_001.sgy L024_001.sgy L024_009.sgy L024_017.sgy L026_002.sgy L029_001.sgy
L022_002.sgy L024_002.sgy L024_010.sgy L024_018.sgy L026_003.sgy L1A_001.sgy
L022_003.sgy L024_003.sgy L024_011.sgy L025_001.sgy L027_001.sgy L1B_001.sgy
L023_001.sgy L024_004.sgy L024_012.sgy L025_002.sgy L027_002.sgy L1C_001.sgy
L023_002.sgy L024_005.sgy L024_013.sgy L025_003.sgy L027_003.sgy L1D_001.sgy
The file naming convention is:
format: Lnnn_rrr.sgy
where: L = low frequency channel indicator (H for high)
nnn = line number
rrr = "reel" number
There are two interpretations to the SEG-Y revision 0 standard.
Knudsen's interpretation is that an SEG-Y file MUST have uniform
trace sampling. Whenever the number of trace samples changes a
new SEG-Y file is opened.
lsd L025_001.sgy 21
SHOT TR RP TR ID RANGE DELAY NSAMPS SI YR DAY HR MIN SEC
17560 1 0 0 1 0 26 6995 67 2001 232 23 26 56
17581 1 0 0 1 0 26 6995 67 2001 232 23 27 57
17602 1 0 0 1 0 26 6995 67 2001 232 23 28 57
.
.
.
18925 1 0 0 1 0 26 6995 67 2001 233 0 32 0
.
.
.
19093 1 0 0 1 0 26 6995 67 2001 233 0 40 0
A generalized script to create the plots is:
sioseis << eof
procs diskin t2f f2t gains diskoa end
diskin
fno 2758 lno 3050 allno no ipath L021_001.sgy end
! fno 17560 lno 18000 allno no ipath L025_001.sgy end
end
gains
type 7 end ! complex modulus
end
t2f
end
end
f2t
type analytic end
end
plot
wiggle 0 ann gmtint anninc 1
trpin 150 clip .005 def .0015 tlines .01
nibs 2859 vscale 24.728 srpath sunfil nsecs .6 end
end
end
eof
xloadimage -r 90 sunfil &
The conclusion of this analysis is that the Knudsen 320
SEG-Y files are truncated under certain situations. As of
December 2001, there is no easy way to predict when the SEG-Y
data are trunacted. For the detailed explanation click here.
Return to SIOSEIS examples. Go to the list of seismic processes. Go to SIOSEIS introduction.