Hi David,
I have also tried running this on the raw data files (the same files as
previously listed, but as downloaded from CADC, without any processing),
and get the same problem. So, it is not to do with me chopping bits off
the spectral axis. Is it because data is no longer being written for the
receptors that are switched off? Is the currently released makecube
routine able to cope with raw data where the dimension of the receptor
axis (axis 2 in the raw file) varies? Or is it assuming that data from
all 16 receptors are written (with bad/blank data from the receptors
that are turned off), and so gridding incorrectly if there are only 12
in the raw data file?
Cheers,
Jane
David Berry wrote:
> Doh! Forgot the attachments...
>
> David
>
>
>
> 2009/5/12 David Berry <[log in to unmask]>:
>
>> Hi Jane,
>> The problem you report seems not to occur with the current
>> development version of the starlink software. Running the attached
>> script on the data you sent me, I get the attached pictures
>> corresponding to your pictureA and pictureB. What differences remain
>> seem to be mainly due to the use of mfittrend - if you miss out the
>> mfittrend stages the results are even closer. I suppose it makes sense
>> since the two approaches use mfittrend at different stages in the
>> processing.
>>
>> We are planning a new release of the starlink software collection in
>> July. If you need the fix before then, it may be possible for you to
>> get the current JAC development version via rsync - it depends on
>> what operating system you are using.
>>
>> David
>>
>>
>> 2009/5/11 David Berry <[log in to unmask]>:
>>
>>> 2009/5/11 Jane Buckle <[log in to unmask]>:
>>>
>>>> Hi David,
>>>>
>>>> I've put them here, so let me know when you've picked them up, as its taking
>>>> up my home space :
>>>> http://www.mrao.cam.ac.uk/~buckle/GBS/Orion/
>>>>
>>> Thanks Jane - I've got them.
>>>
>>> David
>>>
>>>
> >
--
Dr Jane V. Buckle
Astrophysics Group, Cavendish Lab, J J Thomson Avenue, Cambridge CB3 0HE
Tel: +44 (0)1223 337298
|