question on data padding

[Stan Voynick]

While I have generally assumed that all remaining bytes in the file should
be zero, regardless of the data format (equivalent to your option (a)), I
wanted to point out that your option (b) is not a feasible interpretation
anyway, as I often use BSCALE/BZERO scalings for our 16-bit images in
which no integer value even comes close to mapping to "zero intensity"
after being scaled by BSCALE and BZERO. 

                                  - Stan Voynick -
                                <svoynick at netcom.com>


In article <tdame.1185583563A at cfanews.harvard.edu>,
Thomas Dame <tdame at cfa.harvard.edu> wrote:
>According to the FITS standard, the data array must be 
>an integer multiple of 2880 bytes in length... [.. snip ..]
>
>When the data is stored as 16-bit integers (BITPIX=16), it
>is not clear to me whether the padding should be:
>
>a) integer zero (all bits zero)
>
>b) the integer representing zero intensity in the image 
>  (i.e., the integer I such that BZERO + I*BSCALE = 0).
>
>c) the integer BLANK