Re: My problem or yours?

[Andrew Bennett <andrew.bennett@ns.sympatico.ca>]

On Tue, 23 Oct 2001 21:45:13 -0700, Chris Albertson
<chrisalbertson90278@yahoo.com> wrote:

>
>Tom,
>
>As you say the FITS header will take care of it.

Ha!!

>
>In the header are (well should be) both the dimensions of
>the raw image array AND the location within that array of
>the actual image pixels.  So you can fix the PROM code when
>you can get around to it and then fix the header to match.
>In theory you would not have to tell anyone as FITS files
>are self descriptive.
>
>There are two keywords IMAGESEC and DATASEC that are synonyms
>for the area that contains the "real image data" and the FITS file
>should contain lines line this:
>
>   IMAGESEC   [10:2042,5:2037]
>   DATASEC    [10:2042,5:2037]

My image actually has
DATASEC = '[ 17:2064, 1:2048]'
which is not right. It is flagged in the FITS header as being
a "place keeper."

Anybody blindly using packaged FITS routines is getting
silly answers.

> ... Your program ... will "see" only the 2032^^2 image pixels.

Not quite. Data is being lost. As I said in my post
to Tom, this leaves the image non-square and means that
Bennett is going to have to try one more time to sort out
if Axis1 is X or Y. Yeccchhh!

By the way, Tom is wrong in thinking that the end of
the data overwrites the beginning. 

I am not too sure what is there but I am sure it is
not the end data. I think it is one row or is that column
of garbage, one row of covered pixels and then the data with
a horrid ramp (1400 adu or so) superposed. This has to be
the start of the data recorded in more or les its proper
place?

>  Those headers save the software
>and the programmer a lot of work.

again: Ha!!

Andrew Bennett, Avondale Vineyard