On Thu, 28 Jun 2012, Tim Jenness wrote:
> On Thu, Jun 28, 2012 at 4:24 AM, David Berry <[log in to unmask]> wrote:
>> Hi Andy,
>> I think Tim is right that the behaviour you are seeing
>> is what is intended. I suppose there could be a case for an option
Ah, OK - well it was only speculation on my part :)
>> However, I think Tim may be wrong about the TRIM=FALSE TRIMWCS=TRUE
>> issue. If TRIM is FALSE, then TRIMWCS is not used - in effect a value
>> of FALSE is always used for TRIMWCS if TRIM is FALSE. So I think you
>> must be loosing the WCS axis somewhere else.
I checked and the SCUBA-2 pipeline never explicitly sets TRIM=FALSE in
any call to ndfcopy. The offending ndfcopy call is actually at the
point when the pipeline is trying to find the peak in the map. This
codes makes the assumption that the input file has 3 dimensions when
it only has 2 in this case. Seeing as ndftrace is called immediately
before to get the pixel scale, I can clearly get the actual number of
dimensions rather than always assuming 3.
>> setbound gs20120627_6_850_foc1(,,1:1)
>>
>> is used, thus re-introducing a 3rd pixel axis, but no 3rd wcs axis.
>
> Aha. So it's deliberate. The NDF fix sounds like it's all that is needed.
Yes - looking at the code and thinking back to when I added it (it was
based on code Brad wrote for ACSIS) I added a 3rd pixel axis to each
plane to be added to the cube, and only later did I add the 3rd WCS
axis once the cube was created. I think I did it this way to only set
up the WCS once.
I've just tested in Hilo and it works fine now - I assume the fix
is already in place.
Andy
|