Dear Samantha,
sorry it took so long to get back to you.
I think that the solution to your problem is to use a different
recipe. The recipe you are using is REDUCE_SCIENCE_NARROWLINE. In
performing QA checks, that recipe only masks out the central 10% of
the frequency range, and uses all the rest to determine the rms of the
spectrum. Obviously, the lines you see in your images extend well
beyond the central 10% and should be masked out before measuring the
rms.
REDUCE_SCIENCE_BROADLINE does exactly that: it masks out the central
50% of the bandpass before measuring rms.
I tried a reduction of your data forcing the use of
REDUCE_SCIENCE_BROADLINE and the central pixels of your source were
not clipped at all.
If you want to have a very fine control on what is going on, you can
have a look at the primitive _QA_WITH_MASKING_. It lists a parameter
called method, which can be set to NARROW, STANDARD, WIDE, and AUTO,
with an explanation of what this does. You can set this parameter if
you edit the main recipe. Look at REDUCE_SCIENCE_BROADLINE and you
will see that the call is:
_QA_WITH_MASKING_ METHOD=WIDE
Let me know if this solves your problem.
Cheers,
Luca
On Mar 17, 2011, at 11:48 PM, Samantha Walker-Smith wrote:
> The recipe parameters that i changed are:
> [REDUCE_SCIENCE_NARROWLINE]
> BASELINE_ORDER = 1
> RESTRICT_LOWER_VELOCITY = -60.0
> RESTRICT_UPPER_VELOCITY = 50.0
>
> The files being used are:
> a20110201_00018_01_0001.sdf
> a20110201_00019_01_0001.sdf
> a20110214_00023_01_0001.sdf
> a20110214_00024_01_0001.sdf
>
> I dont think there was a problem when i did a quick reduce at the
> JCMT.
> Samantha
>
> On Mar 17 2011, Tim Jenness wrote:
>
>> On Mar 17, 2011, at 5:12 AM, Samantha Walker-Smith wrote:
>>
>>> Hi,
>>> Ive been using the oracdr pipeline to reduce cubes taken of the
>>> orion filament including orion A. A recurring problem is that the
>>> centre of the brightest part of Orion A keeps getting removed
>>> somehow in the pipeline and replaced by blank pixels, and I dont
>>> really know how to stop it, as im not sure which parts of the
>>> recipe i should alter and how. Ive included an example of the
>>> integrated intensity cube to illustrate.
>>> <ga20110201_18_1_integ.sdf>
>>
>> For people looking at the FITS headers you may get a bit confused
>> to find that the PIPEVERS indicates an orac-dr from 2009. This is
>> not correct, what's happening is that our /star rsync tree has an
>> oracdr.version file created back in 2009 which is used as a
>> fallback if git is not installed on the host system.
>>
>> Tim
|