NB Translation problem from MaximDL captures into PI

[jkmorse]

Hey folks,

I have raised this problem before but just experienced it again over the weekend using the updated MaximDL 6.  I hope someone has an idea of how to fix it or at least can tell me what is going on.  I like the ease of imaging with MaximDL with capture and autoguiding all in one package and when I shoot RGB there is no issue when I transport the images into PI for calibration, integration and processing.  But with exactly the same settings in both MaximDL and PI, namely shoot and read as unsigned integer 16bit FITs files, things always go wonky with my narrowband images and come out all weird in PI and can't be worked on (image blown out, PI can't see any stars, etc, etc.). The issue seems to be that the Raw MaximDL image is being truncated somehow when it gets to PI but I don't know how or why. I have tried any number of fixes, including every possible setting tweak I can think of in MaximDL and PI, and have hit on one that actually works but just makes the mystery deeper.  I can't even remember how I hit on this (obviously trial and error), but if I take the MaximDL image, then save it as a 32 bit floating point image in CCDStack 2.0, THEN take that into Nebulosity 3.0 and save that as a FITs file, it will now work in PI.  Note, just taking the saved image from CCDStack to PI does not fix the problem, I need the Nebulosity step too.

The easy solution is to simply stop using MaximDL for capturing narrowband images and I do not have this problem when I use Sequence Generator Pro, for example, but why this issue keeps cropping up and the weird way I can fix it is simply driving me crazy.  I have raised this with MaximDL, but without any response that helps.  So I am turning to you good folks to see if any of you have any ideas.

Thanks,

Jim

[RickS]

Jim,

I use MaximDL to acquire narrowband data for processing in PI and I've never had any problems.  It's very weird that your RGB is OK and narrowband is not.  It's not obvious how changing an optical filter would mess up the data from the camera 

Can you put one of the "bad" subs up somewhere I can download it?

Cheers,
Rick.

[jkmorse]

Rick,

Here is the Google Drive link to the original RAW FITS image (you will need to download and open in PI):

https://drive.google.com/file/d/0ByrTommGQM20b2lJQzlXeWdMeHM/edit?usp=sharing

Here is the same FITS image, but washed through CCDStack and Nebulosity as described in my opening note:

https://drive.google.com/file/d/0ByrTommGQM20d1ZZWGFyNzgzdDA/edit?usp=sharing

Anything you suggest, I am all ears.

Best,

Jim

[RickS]

Jim,

The data in the original raw FITS file has a great big pedestal.  Clip it off with HistogramTransfer and it looks just fine.  I'll take a closer look at what's going on later tonight when I get home and have a bit more time.

Cheers,
Rick.

[topboxman]

It looks like you saved the file as "signed" instead of "unsigned" numbers. If you open with Nebulosity, you will see a minimum of -32463 instead of 0 or any number greater than 0.

You said you updated to Maxim DL6, maybe it defaulted to "signed" instead of "unsigned" numbers.

Peter

[RickS]

Yes, according to the BZERO and BSCALE values in the FITS header (0 and 1 respectively) the original files are signed 16-bit integer data.  The actual data values reflect this.  If you go to Format Explorer and tick the "Signed integer images store physical pixel" box in the FITS Format Preferences then the files will load correctly.  This will mess up reading old PI signed integer FITS files but you probably don't have any of those.

I still don't understand why Maxim appears to be saving RGB files as unsigned int (from what you said this appears to be the case) and narrowband files as signed int.  Maybe there are some per filter Maxim settings but I don't recall ever seeing them.  Ideally you'll figure out what is going on with Maxim and fix it but if not it looks like the FITS Format Preference change will do the trick.

Cheers,
Rick.

[jkmorse]

Thanks all,

I will test tonight and report out, but you all seem to have hit on the solution.  I swear I tried fixing as recommended in the past and that did not work but will retry.

Best,

Jim

[jkmorse]

Guys,

Thanks, that switch change in the FITs preferences did it.  Thanks for helping the hopeless 

Best,

Jim

[topboxman]

I would still find out what happened during capture of NB using Maxim DL. Double check the capture/save settings. Make sure Maxim saves raw images as "unsigned" numbers or integers.

You can easily test it anytime. Capture it indoors with camera covered or not or take flats and make sure it saves as "unsigned" numbers.

Peter