Problem with master bias

[Diane Miller]

I'm trying, after some time off, to use PI again.  Very frustrated with lack of information.  I'm shooting with a DSLR (now the Canon 1DX2) and hoping to use the Batch Preprocessing routine until I get more advanced.  I'm just trying to make a Master Bias, setting it to show the rejection files.  Sometimes I get the right 2/5 or so of the frame showing nothing, sometimes the pattern fills the frame.  Sometimes I see vertical stripes and sometimes horizontal.   But the horizontal ones seem to be when the fiale fills the frame -- some sort of camera orientation issue??  I'm attaching a screenshot after doing STF.  The camera was in horizontal orientation, files shot in a dark closet with lens cap on, ISO 800, shutter speed 1/8000 sec, f/32 to try to block out any stray light.  Have seen a similar problem with the Canon 7D2 I used previously.

Parameters were:
Combination: Average
Normalization: No Normalization
Weights: Don’t Care (all weights =1)
Scale estimator: Median Absolute Deviation from the median (MAD)
Check the two boxes for Generate Integrated Image and Evaluate Noise
Pixel Rejection (1):
Winsorized Sigma Clipping
No Normalization
Check all boxes
Pixel Rejection (2):
Sigma Low 4.0
Sigma High 3.0
Range Low 0.0
Range High 0.98

Are my parameters wrong for a DSLR?  A camera orientation issue?  WHAT SHOULD THESE FRAMES LOOK LIKE AFTER STF?

Thanks for any help!
I tried to attach screenshots but then I couldn't post, even after I downsized them as JPEGs.

[pfile]

if you use dropbox or google drive or such to post the xisf files, someone can take a look.

when working with CR2 masters by hand, make sure that the DSLR_RAW module (format explorer tab on left, double-click DSLR_RAW) is set to "pure raw".

rob

[Diane Miller]

Thanks Rob -- I had done that before but apparently it got changed with the last update.  But when I look there I don't see Pure Raw -- jut DSLR Raw.  Is that correct now?

[Diane Miller]

Apologies -- I didn't read it your answer right to double-click DSLR RAW.

[Diane Miller]

I've done that now -- assume I accept the defaults there?

[Diane Miller]

I accepted the defaults and ran the ImageIntegraton (parameters above) to get a master bias.  It ran much faster than before and the rejection file are now black and white.  Low is very dark, high is dark gray, integration is light gray, all show fine vertical line mottling.

Is that right?

[pfile]

without any screen stretch it should probably be black... but with STF turned on it would look like what you are describing.

an all white rejection image seems wrong though, what rejection method are you using and what are the settings?

thanks

rob

[Diane Miller]

Rejection low is very dark with lots of very thin vertical broken white lines.  Rejection high is dark gray with same lines.  Integration is medium gray with same lines.  (All stretched.)

Pixel Rejection (1):
Winsorized Sigma Clipping (for calib masters; use Linear Fit for Lights)
No Normalization
Check all boxes
Pixel Rejection (2):
Sigma Low 4.0
Sigma High 3.0
Range Low 0.0
Range High 0.98

I have about 100 bias files.

[Diane Miller]

Clarification:  I meant to say the files are now black and white as opposed to colored, as they were before I did the correct settings for Pure Raw.

[Diane Miller]

Next problem:  I tried just loading all the biases, darks, flats and lights and running the batch script.  It chugs along quite a wile and then says it can't integrate fewer than three frames!  But I have over 80 biases, flats and darks.  (Only 36 lights.)  Same result if I load all the biases or just the master bias.

What on earth am I doing wrong?  Must be missing a parameter somewhere.

Is this stuff covered in the $40 PDF that got panned for being too basic?  I'm going back to the Keller videos but they are pretty outdated.  Too bad a purchase gets outdated so fast with no updates.

[pfile]

yeah, that's correct, the files should be black and white when loaded with pure raw. they are the completely raw sensor data in "CFA" (color filter array) format.

more than likely something has gone wrong in the calibration of the lights, yielding a situation where the data has been mostly destroyed and staralignment can't see any stars. therefore there are no registered images (or at least fewer than 3) and so ImageIntegration is complaining.

i'd start by removing the flats and the darks; just put in the bias frames and the lights. it will probably complain about no darks or flats, but should run anyway. if not, get rid of the bias frames and only put in the darks and lights and try that.

i don't know what's in the PI Coffee pdf... i did not purchase it since, well, i've been using PI for 6 or 7 years now...

warren keller's forthcoming book will very likely address the BPP script.

rob

[msmythers]

Here are the basic settings I use with the Batch Preprocessing script and my Sony DSLR. This should work to get you a stacked image from raw images from your camera. These settings are not optimum but the first thing is to get an image without errors during the stacking process. I don't run Dark Optimization for a quick BPP. BPP is made for evaluation, a quick look at your image not for final production work. When your just starting to learn PI though it is a great way of getting your image quickly, well as quickly as your computer can process. 

For this example I am using 200 bias, 200 flats, 48 darks and 23 lights. If I had used different quantities of files I would chose the right rejection algorithm for those files.

I've included the DSLR-Raw settings on the bottom right. If I'm having issues using BPP I will reset the BPP settings and clear all files. One last thing is to run the diagnostic to check for issues.

Mike

[Diane Miller]

Thanks guys -- wonderful to see the screenshots Mike!  I've made notes from so many sources that I had some conflicting or questionable information.  Now have a doc I trust on the steps and parameters.

The previous error seemed to have something to do with darks so I looked at them with Adobe Bridge.  Some were made right after the lights but the battery ran down after about 20 so I shot another 50 the next night, same temp.  I had been loading all of them when I got the error.  I went to the second set only and everything worked. 

I think maybe the rejection files look OK -- low is a very black background but with. many tiny colored spots.  High has more tiny colored spots and a few larger and a bunch of fine meteor trails.  Here is a link to the files: 

https://dianemiller.smugmug.com/Photography/Files-for-PI-forum/n-FPknxD/i-jnhP99z

I'll use this file to see how much farther I can get, and work on learning to fine-tune the parameters later.  The shot was with a 50mm lens on a full frame DSLR of the core region of the Milky Way.  Basically the same composition that is on the cover of Jerry Lodriguss' Guide to the Deep Sky. 


I'll undoubtedly be back, but many, many thinks for the help getting this far!!