BatchPreProcessing script not handling meridian flip

[jcinpv]

I read in other places here that the BPP script would properly handle flipping images that were involved in a meridian flip.

I took five images, flipped the mount at the meridian, and took five more images. The integrated result showed both orientations instead of flipping one set to match the other.

What do I not understand?

Computer: MacBook Pro w/Mac OS X 10.9.1
Camera: Prime focus Canon 60D w/SCT Corrector and Hutex Galaxy filter
Mount: Losmandy G-11 w/Gemini-II controller
OTA: Meade 12" LX200
Guided with SBIG SG-4

John C.

[pfile]

i think the only thing BPP does not handle is different flats for a different position angle - people with rotators often rotate 180 degrees after a meridian flip to acquire the same guide star when using an OAG. but then you need 2 sets of flats.

anyway, it's not so much that BPP can or can not handle meridian-flipped images, but that StarAlignment (what BPP is built on) should not care and is absolutely capable of discovering that an image is rotated 180 degrees vs. the reference image.

did you do the integration by hand? if so did you select the _c_r.fit images (and not the _c.fit images?)

otherwise what you are describing is that either StarAlignment failed to register an image and simply wrote out the input image (which to my knowledge does not happen) or BPP somehow integrated the _c.fit images which i have not heard of either...

rob

[Juan Conejero]

i think the only thing BPP does not handle is different flats for a different position angle

Actually, you can define two "filters" with Add Custom to group frames by rotation. This way you can calibrate and register everything in a single script run.

[pfile]

whoops, i stand corrected.

rob

[jcinpv]

i think the only thing BPP does not handle is different flats for a different position angle - people with rotators often rotate 180 degrees after a meridian flip to acquire the same guide star when using an OAG. but then you need 2 sets of flats.

anyway, it's not so much that BPP can or can not handle meridian-flipped images, but that StarAlignment (what BPP is built on) should not care and is absolutely capable of discovering that an image is rotated 180 degrees vs. the reference image.

did you do the integration by hand? if so did you select the _c_r.fit images (and not the _c.fit images?)

otherwise what you are describing is that either StarAlignment failed to register an image and simply wrote out the input image (which to my knowledge does not happen) or BPP somehow integrated the _c.fit images which i have not heard of either...

rob

Different set of Flats for Meridian flip? You just made the AP acquisition process more complicated than I anticipated. Maybe I should have kept my fork mount.

I let BPP do the integration, and I've since learned to not do it that way. And I probably did select the wrong set of images.

Thanks, everyone. I'm learning.

John C.

[dzso.bacsi]

Different set of Flats for Meridian flip? You just made the AP acquisition process more complicated than I anticipated. Maybe I should have kept my fork mount.

I am definitely a beginner here, so I am basically trying to understand the process for my DSLR workflow.

I think there is a misunderstanding either in my knowledge about flats or the conclusion you draw. If you don't change your camera position related to the telescope you use during a session (even if it includes a meridian flip) I think you do not need different flats for before/after meridian flip. BPP recognize that your images have an appr. 180 degree difference after the flip, and according to my experience will handle it without problem. If you have a rotator or you manually rotate your camera after the flip, then you can/should make another set of flats, use them separately for calibration, but still the BPP script (more precisely the StarAlignment module inside) will recognise during stacking that you have rotated images relative to your reference image.

So far only once I had a two day session when I took images of the same target two nights consequently. It obviously inluded two meridian flips, but camera/OT position did not change. I made darks for each nights, but only one set of flats at the end of the second night. Flats worked OK, and StarAlignment (with BPP) worked perfectly.   

So I think in a normal case there is no need for different set of flats for meridian flip. Or did I seriously misunderstand something?

[Phil Leigh]

You ONLY need two sets of flats IF you rotate the camera after a meridian flip (or indeed at any time).
There is however no need to rotate the camera after a flip UNLESS your framing demands it.

You certainly do not need to rotate the camera just to make stacking work... Pi Star Alignment handles images that are 180 degrees apart just fine.

[pfile]

i think people missed what i said and what phil just reiterated. i'm talking about setups where you have a motorized rotator. it's certainly not necessary to rotate your camera after a meridian flip, but if you are using an OAG and a guiding software package that expects to find a guidestar in the same place after a meridian flip, you're going to want to rotate the camera 180 degrees.

or, maybe since you have a rotator, you set up for some non-right-angle position angle and after the flip you want to preserve the framing. in that case too you rotate the camera 180 degrees after the flip.

if you don't use a rotator, your post-flip images come out upside-down relative to the pre-flip images. if you do use a rotator, they come out in the same orientation but now you need new flats since the camera to OTA orientation has changed.

of course what i said about BPP was wrong, you can define 2 sets of flats which correspond to the pre-and post meridian-flipped images.

rob