bad SPS files (Friday afternoon whinge)

Oil and natural gas exploration -- geology and geophysics
Post Reply
Silver Member
Silver Member
Posts: 29
Joined: Fri Nov 24, 2017 4:36 am

bad SPS files (Friday afternoon whinge)

Post by Michael_Seman »

I got a line recorded on an Aram Aries (presumably the original model, as it was in the early noughties) and the XPS file is crap. It has typical records saying channel 1-240 but receiver 101-343, which is inconsistent. Anyway it is explained by some hand-written notes revealing a gap in the spread. Shouldn't they have defined two halves either side of the source??? I wonder did the people at the factory never consider a gap in the spread? I know I have seen lots of stuff where they instead reduced the preamp gain next to the shot.

VIP Member
VIP Member
Posts: 625
Joined: Sat Mar 24, 2012 11:35 pm
Top poster of the Month

Re: bad SPS files (Friday afternoon whinge)

Post by GuyM »


Given the variations and stuff-ups you get in seismic files (SEGD, SEGY) its not a surprise. A fair chunk of software support questions when I was in the seismic processing software business were about bad formats. We had a bunch of tools to help people deal with it but sooner or later you need to get some kind of scripting/coding skills under your belt for this kind of thing. Back in the day I learned awk and sed to do this, but now you'd probably be better off with python.

We could have a whole thing here about standards, but suffice it to say the ones that work in the modern era (and indeed people's expectations) are not for formats that are backwards compatible with reading from magnetic tapes, and that having freely available APIs for standards saves a lot of effort.

Can you imagine how many hours have been wasted by developers writing code to read SEGY (in all of its delicious variants) since the 1980s? And how many hours have been lost digging into stuffed up SEGYs? The mind boggles.

SPS can be even worse mind you; I once had to write a cascading series of scripts to scan 3D SPS files and identify all of the rubbish coding. Those had been created by hand from Observer's Logs at a data library company and never really checked. They loaded fine in some software - no error messages - but then just dropped screeds of data.

So - yeah, not great, but I think it will not go away...

Post Reply
  • Similar Topics
    Last post