NaN values from misfit using Real data

[raulleoncz]

Hello Dr. Daniel,

My name is Raul Leon. I have been using DENISE for some synthetic acoustic inversion problems and now I am trying to run an acoustic FWI using real data. Since DENISE creates the seismograms using an old version of seismic unix, I tried to run the inversion using ascii data-format but I got the following error:

Calculate residuals

Message from PE 1
R U N - T I M E E R R O R:
Field data file could not be opened !
...now exiting to system.
Message from PE 0
R U N - T I M E E R R O R:
Field data file could not be opened !
...now exiting to system.

Based on that, it seems that FWI workflow using ascii data-format is not available yet. So, in order to run the inversion using the seismic unix data format, which version of SU would you recommend me to use?

I used the 44R24 version but I've been getting the following error:

Calculate residuals

PE 0 is writing 648 seismograms of pressure to
su/DENISE_L0970_Tiber_p.su.shot1.it1

==================================================================================

***** Starting simulation (adjoint wavefield) **********

==================================================================================

Time step: 121; pvy: nan
Message from PE 0
R U N - T I M E E R R O R:
Simulation is unstable !
...now exiting to system.

From the output file I know that my simulation should be stable so that makes me believe the error could be related to the seismic unix version that I am using to convert from segy to su. Could you give me any recommendation to avoid these types of problems?

I really appreciate your help.

[raulleoncz]

Update:

Seismic unix converts segy to su using the xdr format. This format is different from the one needed in DENISE.
I solved my issue using 'suswapbytes' command.