[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20030604: IMG* use of site-developed MSG calibration module

>From:  "=?ISO-8859-1?Q?Marianne=20K=F6nig?=" <address@hidden>
>Organization:  EUMETSAT
>Keywords:  200306040725.h547PDLd003216 McIDAS-X calibration

Hi Marianne,

>I ran into a mcidas problem (what else??), and before I dig into that I
>thought maybe you'd immediately know what the problem is:

Ready...

>Using my MSG server I did an IMGCOPY to an AREA file. Seemed ok (file
>size, AREA directory entries all look fine and as expected). When I
>display the AREA I just get a black screen, an IMGPROBE on some
>location writes out the RAW value (correctly, so that actually is in
>the AREA file), but temperature, radiance and BRIT of 0 (that's why it
>is black). I suspect it has to do with my cal module that I wrote for
>MSG but cannot really see anything wrong with it. Called from within
>the MSG server it works fine. In case you instantly know what the
>proble is or could be, please let me know -

Did you add your calibration module to the list of modules that
will be recognized by McIDAS?  What I mean by this is that you need
to add your calibration module to the list of calibration modules
in the McIDAS makefile:

######################################################################
#-------
# Fortran kbx & nvx dynamic load modules for libmcidas.
#-------

CalDlm_Srcs             = \
                        kbxaaa.dlm      \
                        kbxamsu.dlm     \
                        kbxavh3.dlm     \
                        kbxavhr.dlm     \
                        kbxeas.dlm      \
                        kbxfy1.dlm      \
                        kbxgms.dlm      \
 ...

In the same section of the make file, there are three places that your
code will need to be referenced:

- in the CalDlm_Srcs section

- in the Dlm3_Gens section

- and in the section of dependencies:

kb1aaa.f        kb2aaa.f        kb3aaa.f        : $(convdlm) kbxaaa.dlm
                                                @ $(CONVCMD) kbxaaa.dlm
kb1amsu.f       kb2amsu.f       kb3amsu.f       : $(convdlm) kbxamsu.dlm
                                                @ $(CONVCMD) kbxamsu.dlm
kb1avh3.f       kb2avh3.f       kb3avh3.f       : $(convdlm) kbxavh3.dlm
                                                @ $(CONVCMD) kbxavh3.dlm
 ...

You will then need to rebuild McIDAS so that your calibration module
will get added to libmcidas.a and so that kbprep.for will get remade
and compiled to have a reference to your calibration module.  Finally,
in the rebuild the IMG* routines will get relinked and your calibration
module will be available for use.

WARNING:

If you use the SSEC method of building McIDAS (i.e., running their script
that unpacks the distribution and runs the compilations), your modifications
to 'makefile' will be lost.  I suggest that you add the needed lines to
'makefile' and then rebuild/relink the distribution "by hand":

<login as 'mcidas'>
cd mcidas2002/src               <- or wherever directory your current McIDAS
                                   distribution is found
<edit makefile and add the needed lines for your cal module>

make all
make install.bin

>if not, I'll look into
>this. I see nothing obviously wrong with the AREA file - e.g. it
>contains the correct cal section. What I am not sure is this conversion
>to BRIT - should my calmodule also take care of that?

Your cal module will only be called if the IMG* routines have had it
linked in.  This will only be done if you add your cal module to the
makefile as I outlined above.

>What are your plans for the MUG meeting?

That is _so_ far off that I havn't really thought about it seriously.
I will probably arrive early and stay at the Edgewater.

>I'll arrive on 11 October. Don't have a room yet in the Edgewater for
>the first night.

So, you have already made reservations for the other nights at the
Edgewater?  You are incredibly organized to get this done so far in
advance!

Cheers,

Tom