Thread: Re: Having problems creating ALPHA input files

Started: 2005-11-11 22:02:29
Last activity: 2005-11-11 22:02:29
Topics: SAC Help
Robert Casey
2005-11-11 22:02:29

Hi Carlos (cpena<at>email.arizona.edu)-


Did you read ac_new[1].dat using the RA command in a linux version of
SAC?.

No, I performed this operation on a Mac.

Before I tried your scheme, I just tried to read the file you created
using RA
and SAC responds with "WARNING: Trouble reading line. Skipping ...
(list of 5
column data points)"; one message per row of data points, and after it
finishes
with the warning messages it states "Segmentation fault" and proceeds
to kick me
out of SAC. The same thing happended to me when I tried to read one
column data.

Not sure why this happens on the Linux machine. The text file has New
Line characters for line terminators and I would think it would work
fine on a Linux box.


Moreover, when I convert a SAC format data file to an ALPHA format
data file,
the resulting data file is all "screwed up". SAC changes all the header
information and considers only a fraction of the data points. You can
try this
by doing FG SEISMOGRAM and converting the resulting data file from SAC
to
ALPHA.

Not sure of the reasons for this. Might be worth trying
space-separated data fields instead of tabs. I was taking a first
guess at a solution.


Could there be a bug in the linux version of SAC?; other people I've
talked to
tried it in a MAC version, using RA on one column data and no header
info and
it worked OK.

I have no idea if there is a bug with the Linux version. We don't yet
have good tabulation of known issues for SAC here at IRIS (as far as I
am aware). I think it may be that the Mac-created text file


By the way, since you got it to plot, can you send me the raw
acceleration time
series plot, followed by a filtered acceleration, velocity and
displacement
time series plots?. Consider a 4th degree Butterworth low pass filter
with a
lowcut at 4 Hz to filter raw acceleration points and a 4th degree
Butterworth
high pass filter with a high cut at 2 Hz for velocity and displacement
time
series. Please normalize acceleration time series with respect to the
mean
BEFORE you apply filter and also normalize with respect to mean
velocity and
displacement time series AFTER you apply the filter.

This is something I don't have experience with or the time to work
with. Perhaps a more SAC-inclined user can step in and help you with
your processing issue. (anyone?) I was trying to demonstrate that
through a UNIX shell it is not difficult to convert a single column of
text data to five columns of data. Beyond this, it's getting out of my
depth to manage in a time-effective manner.

-Rob


08:53:04 v.01697673