Possible affect of slant correction on hot pixel removal?

[David Boyd]

After changing the slit in my LISA, I noticed that the calibration lines had a slant (-0.87) so used the slant command in ISIS to correct this. I then found that several hot pixels, which were included in the hotpixel file, were appearing in the spectrum. When I reset the slant value to 0.0 and reprocessed the spectrum, the hot pixels disappeared. Is it possible that applying a slant correction affects the removal of hot pixels, possibly because their coordinates are not correctly calculated?

I have since slightly adjusted the position of the slit plate in the holder to remove the slant and therefore the need to make a slant correction, but I thought it was worth asking the question in case anyone else encounters this problem.

David

[Robin Leadbeater]

Hi David,

That's odd. My LHIRES needs significant corrections for slant and the ALPY for slant and smile but I have not noticed any issues. (Though I have not actively looked at what is happening). The workflow report that ISIS produces on the "Go" page seems to suggest that all image calibration including treating hot pixels is done before geometric correction. The only thing I can think of is do you use the cosmic ray filter which works after geometric correction and might become confused due to smearing of any remaining single hot pixels caused by the geometric correction ?

Robin

[David Boyd]

Hi Robin,

Thanks for responding. I see what you mean about the cosmetic correction being applied before the slant correction.
This plot shows the effect of applying the slant correction. Nothing else changed between the two runs.

V1329 Cyg hot pixel.png (12.78 KiB) Viewed 4411 times

There is a hot pixel visible on the image at that location which was in the hot pixel file. With the slant correction set to zero the hot pixel is not present in the spectrum. When I apply a slant correction of -0.87 the hot pixel appears. I was not applying the cosmic ray filter. I then did apply the cosmic ray filter and it removed the hot pixel from the spectrum.

Curious. I've now corrected the cause of the slant so the problem has gone away.

David

[Forrest Sims]

Hi David and Robin,

Trying to learn from your experience.

This is a bit of a tongue twister. So if I understand your description correctly, you have ultimately eliminated the problem in your spectrum caused by the hot pixel by tuning out the slant angle to 0.00 in your LISA. (My most recent LISA spectrum data showed a slant angle of -.80). And it sounds like the ISIS Slant Correction reintroduced the hot pixel back into the data after it had been removed by the ISIS Cosmetic correction process in the case before you tuned the slant angle from -0.87 to 0.00. And then it appears that by chance the Cosmic ray filter removed the hot pixel again and saved the day.

So I am wondering what I should do with my LISA. Should I try to adjust the slant to 0.00? I am not even sure in the LISA Users Guide and Reference Manual exactly which section describes the steps to do this tuning, but most sections say “Be careful: this operation is complex and very precise…..”

And or, should I always have checked the Cosmic rays filter box on the General tab in ISIS?

Thank you

Woody

[David Boyd]

Hi Woody,

I would be surprised if the slant correction itself reintroduced the hot pixel. I just know the hot pixel was not removed when I applied the slant correction. I think the cosmic ray filter did what it is was supposed to do - it removed an unexpected bright pixel (or pixels) in the image. I normally don't tick the cosmic ray box and only use it if I see an obvious problem in the spectrum. I don't know what recommended practice is on that.

I removed the slant in the images by removing the slit holder, loosening the two screws holding the slit, rotating the glass slit plate and its holder very slightly and retightening the screws. Trial and error eventually got me to a slant of -0.1 and I quit at that point.

David

[Robin Leadbeater]

I normally don't tick the cosmic ray box and only use it if I see an obvious problem in the spectrum.

Same here. I like to go back to the image to see exactly what the problem is before applying the cosmic ray filter. The difference between a defect and a genuine spectrum feature is much more obvious in the image than in the binned spectrum profile and can often be traced back to a particular exposure. Also if you are anywhere near undersampled in the spectrum, particularly one with high ADU counts, I find you have to be very careful with the cosmic ray filter as it can "correct" whole rows of legitimate pixel values in some areas giving a very odd stepped effect in the spectrum.

Robin