[SAC-HELP] SAC Issue

Chad Trabant chad at iris.washington.edu
Tue Nov 3 17:40:52 PST 2009


Hi Stuart,

SAC is now a community developed and supported program, the best place  
to find help is the sac-help mailing list (CC'd), list info is here: http://www.iris.washington.edu/mailman/listinfo/sac-help 
.

What you describe smells like an endian issue, search through the  
email archive I found this message:
http://www.iris.washington.edu/pipermail/sac-help/2009-January/000484.html

Using sac 101.2 I was able to successfully read back data written on  
my little-endian Mac by changing the sample type ("datatype") in  
the .wfdisc table-file from 't4' (big-endian) to 'f4' (little- 
endian).  So there is a bug in 101.2 regarding the correct sample type  
designation, my guess is that it always writes host-endian floats and  
always writes 't4' as the datatype which is causing problems on little- 
endian machines.

Unfortunately the current release of SAC, 101.3b, seems to have  
another bug in that it wouldn't even read the input css:

  SEISMIC ANALYSIS CODE [08/25/2009 (Version 101.3b)]
  Copyright 1995 Regents of the University of California

SAC> fg seismograph
SAC> writecss lala
-99999.00 truncated to (-99999.) in order to fit format requirements.
-99999.00 truncated to (-99999.) in order to fit format requirements.
-99999.00 truncated to (-99999.) in order to fit format requirements.
SAC> readcss lala.wfdisc
No wfdisc file(s) found in the specified directory
  ERROR 1301: No data files read in.

Hopefully someone on this list will have time to look into that.

regards,
Chad


>> From: Stuart Weinstein <stuart at ptwc.noaa.gov>
>> Date: November 2, 2009 2:40:23 PM HST
>> To: Rhett Butler <rhett at iris.edu>
>> Subject: SAC Issue
>>
>> Rhett,
>>
>> Can you forward this email to someone at IRIS who can help us  
>> understand and perhaps solve
>> the issue?
>>
>> We are using SAC version 101.2, which was emailed to myself and  
>> Chip by IRIS a few months
>> back. This is the LINUX version. We have discovered some odd  
>> behavior.
>>
>> CSS files written by writecss under SAC are not readable by  
>> readcss. Oddly enough,
>> css files written on our SUNS (Big Endian) are readable by readcss  
>> on LINUX (little Endian)!
>>
>> Is there a switch or something that we are missing? The css files  
>> written by SAC V101.2
>> for LINUX are not understandable by SAC on LINUX. We have confirmed  
>> this behavior
>> numerous times by reading a SAC file, and writing it in CSS format  
>> and the trying to read
>> it with readcss. We get nonsense. This procedure always works on  
>> Solaris, but never on LINUX.
>>
>> Thanks,
>>
>> Stuart






More information about the sac-help mailing list