WBPP diagnostic message

Can someone tell me what combination from what I have listed will make Pixinsight happy?
Load them all (though it may not use the bias). Just go to the "+files" and select all of them.
The dark flats (which are just another set of darks with a different exposure) will be used to calibrate the flats; the darks and flats will be used to calibrate the lights; because the darks contain the bias (unless you select otherwise), it doesn't really need the bias; when in doubt, look at the WBPP control pannel - it tells you exactly what WBPP is going to do.
 
My lights and darks are both 120 seconds. There's 2 issues going on here.

If I put a 120 second dark in the script, along with a flat generated in the flat wizard in Nina, Pixinsight complains about the issue I posted in the first post. That's problem one.

If I do as Fred suggested and use a Dark/Flat in place of a Dark, I get the warning in the Control panel. That's a different issue.

That's why I posted the list above, I'll post it again.

Lights are 120 seconds.
Darks are 120 seconds.
Flats are 0.60 seconds
Dark Flats are 0.60 seconds
Bias are listed as 0.00 seconds.

What combination from this list, of my current lights, and calibration frames from my library, will work in the script?

Yes I can post screenshots, but I've posted the errors it gives under the 2 different circumstances.

If I use in the script:

120 second lights
120 second darks
.060 second flats
0.00 second bias

I get this error in the diagnostic tab. There is no error in the Control Panel.

"==== PRE-PROCESS CONFIGURATION

==== Flat frames (filter = NoFilter, binning = 1, exposure = 0.60s, keywords = [], mode = pre-processing, frames = 50 (50 active))

Warning: No Master Dark will be used to calibrate the frames, the Master Bias will be used instead..

==== POST-PROCESS CONFIGURATION"

If I use in the script:

120 second lights
.060 dark flats
.060 flats
0.00 bias

I don't get a warning in the diagnostic tab, but I do get this warning in the control panel.

"WARNINGS

Light frame's exposure differs from the Master Dark's exposure.

You can add a compatible Master Bias and optimize the Master Dark in order to achieve a better dark current estimation."

Here's the Control Panel screenshot. Again, this is with Dark/Flats, so of course the exposure does not match the lights. I now see the box for Calibration settings, it was mentioned earlier, and I couldn't find it.

wbpp control panel.jpg


All I need to know is what combination of these calibration frames will work, with my current set of lights at 120 seconds? All I want to do is fill those tabs at the bottom of the script.

Darks 120 seconds.
Flats 0.60 seconds
Dark Flats 0.60 seconds
Bias 0.00 seconds.
 
Last edited:
Load them all (though it may not use the bias). Just go to the "+files" and select all of them.
The dark flats (which are just another set of darks with a different exposure) will be used to calibrate the flats; the darks and flats will be used to calibrate the lights; because the darks contain the bias (unless you select otherwise), it doesn't really need the bias; when in doubt, look at the WBPP control pannel - it tells you exactly what WBPP is going to do.

Thanks again Fred, you posted while I was still typing. Sure, I can load all of them, and I basically have in different combinations. But I end up with warnings in one place or another. I'd like to just load my lights, and calibration frames, and have no warnings or errors anywhere.
 
You need all the darks. The dark flats (0.6s) will be used for the flats, the 120s darks will be used for the lights. The only ones you don't really need are the bias. This is because the master darks will contain the bias, so flat - master_ dark(0.6s) subtracts dark current and bias, similarly light - master_dark(120s) subtracts dark current and bias.
You should not get any warnings.
 
You need all the darks. The dark flats (0.6s) will be used for the flats, the 120s darks will be used for the lights. The only ones you don't really need are the bias. This is because the master darks will contain the bias, so flat - master_ dark(0.6s) subtracts dark current and bias, similarly light - master_dark(120s) subtracts dark current and bias.
You should not get any warnings.

So put my darks, and dark flats, in the darks tab, that's what your suggesting?

Hey, it worked, no errors, or warnings anywhere. I didn't load all the lights, just one night to test. I'll take the bias out as well.

Thank you very much Fred.

wbpp control panel 2.jpg
 
Last edited:
If you select the "lights" (62 frames) row, you will see the 120s darks highlighted as selected for calibration. If you then select the "show calibration diagram" it will show you exactly what it is going to do. Similarly if you select the "flats" (50 frames) row. It is well worth becoming familiar with the way the control panel works. You can override the automatic behaviour - but I'd watch Adam's videos before you try anything too smart :) !
 
Thanks for that. I'll snoop around the control panel more next time, following your suggestions. Pixinsight is chugging away at the moment.:) I do have a couple of Adam's videos here, I really need to dedicate some time to watching/studying them, to hopefully gain a better grasp of what's going on.

It's a lot to cram into a 45 day trial. Not complaining, I'm very thankful there is a trial, but it's a lot to try and grasp in that amount of time, before laying down a sizeable chunk of money.
 
Last edited:
So this is confusing for me at this point in the discussion. What happened to the bias master being used to scale the darks to the lights and that would, I have always thought, included the flat frames? And the darks would never be a match for sky flats as those are different for each exposure as the sky lightens/darkens between exposures.
 
So this is confusing for me at this point in the discussion. What happened to the bias master being used to scale the darks to the lights and that would, I have always thought, included the flat frames? And the darks would never be a match for sky flats as those are different for each exposure as the sky lightens/darkens between exposures.

Steve, I hope your not expecting an answer from me.:p Just kidding of course.
 
What happened to the bias master being used to scale the darks
(Hopefully) very simply:
The camera deliberately adds the same bias to all images (give or take a tiny amount of bias noise). This is to ensure that the digitised readout is definitely positive. Call this signal B.
With the aperture covered, the camera will typically generate some unintentional dark current, which will typically increase with exposure time, t. Call this signal D(t).
With the aperture uncovered, the camera will capture photons, measuring the incident light; this signal will increase with the exposure time. Call this signal L(t).
Then:
  • a bias frame contains a (more or less fixed) signal B;
  • a dark frame contains B + D(t);
  • a light frame contains B + D(t) + L(t).
To obtain the desired image with just L, you could:
  • bias correct the dark: DC(t) = (B + D(t)) - B = D(t);
  • bias correct the light: LC(t) = (B + D(t) + L(t)) - B = D(t) + L(t)
  • dark correct the light LC(t) - DC(t) = ( D(t) + L(t)) - D(t) = L(t)
Or you can just dark correct the light:
  • light - dark = (B + D(t) + L(t)) - (B + D(t)) = L(t)
This is what the folks mean when they say that the darks include the bias.
Note that explicit bias frames are not needed by the second method.
 
So this is confusing for me at this point in the discussion. What happened to the bias master being used to scale the darks to the lights and that would, I have always thought, included the flat frames? And the darks would never be a match for sky flats as those are different for each exposure as the sky lightens/darkens between exposures.

You can still do this of course. However, this time around WBPP is setup NOT TO DO THIS AUTOMATICALLY or assume this is OK. You really have to want to do this. I have another comment... the scaling of dark frames is not something to be taken lightly. Those that do this need to understand what to look for when it is not working properly. Scaling darks to match light frames is fine- and better results are obtained when there is significant (measurable) dark current.

Assuming you have a well behaved sensor, you would want to use a master bias frame to calibrate flats of different exposure times. If you are using a CMOS detector with other electronic signatures, then matching darks of some exposure time is better. You can likely have flat exposures to be calibrated with a single dark of sufficiently small deltas.

Please hear me out. When the dark current is small or negligible any error in the calculation that scales this value will result in poor calibration. This is especially true for flat field images which are short exposures, have small amounts of dark current, and may be susceptible to other sources of electronic signatures (amp glows and other CMOS effects). I would go so far as to say, it is not advisable to scale darks to calibrate flat field images that are short exposures. Scaling a 180 second dark frame to calibrate flat field images is asking for trouble.

I discuss these issues in detail here:

My experience in the day-to-day world of assisting users supports the above assertions.

-adam
 
Last edited:
Here's a link to the calibrated master light on my Google drive if anyone would like to have a look. Again, I have very little experience at this, started early fall last year, no imaging over the winter. It's just shy of 11 hours integration. All comments are welcomed, I want to learn. This is 329 lights at 120 seconds, 50 flat, 50 dark/flat, 30 darks. There is a Svbony 1.25" UV/IR cut filter used here as well. Rest of the equipment is in my signature, I used the 533mc pro.

I hope it's ok to post a link like this here?

https://drive.google.com/drive/folders/1NOkHUN-oBybdefXlPtSqjvB48C9wK1VR?usp=sharing
 
Last edited:
of course and it's actually encouraged because a lot of times there's no way to debug problems without looking at an image.

rob
 
Thank you Fred for the feedback. I had no idea what HH 555 was until I looked it up. I see now from some very nice images online what your talking about.

Am I on the right track at least for a newbie? I'm sure a lot of you guys/gals are fairly confident in what your doing, but as a newbie I question every move. If I'm at least heading down the path in the right direction, I'll keep doing what I'm doing, and hopefully learn, and improve, along the way.

I can't see investing in something like Pixinsight, if the data your feeding into it is not going to produce a nice result. Maybe some of you out there can make a nice looking image from poor data, using Pixinsight, that's certainly not me at this stage. I need to know that my acquisition skills are somewhat adequate enough, to justify the cost of this program.

If things are looking ok, I don't mind spending the money. I've already invested in some other programs out there, but wasn't really excited with the outcome. Most definitely could be me, and or my data, but also could be a limitation of those programs. That's something I won't know until I share some data, and you experienced Pixinsight people can say, ya, or nay. If you can work with the data, then it gives me a reason to move forward.
 
I'm not very experienced myself. Over a couple of years using PI I've picked up lots of ideas, mainly by keeping my eye on this forum and listening to the real experts. As a back-yard astronomer with fairly modest equipment, my images are often OK but rarely brilliant. Your image here is better than a lot of mine. What kept me interested is when I realised what a huge improvement I could make to apparently mediocre images with some determined processing.
I think different approaches work for different people. Living in the UK without a permanent set-up, I grab every opportunity I can to put up a telescope and grab a few images, then see what I can "persuade" out of them with a bit of processing. I get motivation because I work with my 15-year old son, and because we are both members of an active local astronomy club.
If you are enjoying it - keep going. Unless you have extremely good equipment, you can soon reach the limit of very deep imaging, so it may be better (at least while you are building experience) to aim for fewer frames of more varied targets - but that's really just saying that's what worked for me.
The image you have posted here is good enough that it would keep me going for a while :)!
 
This is an example of applying PI to your image.
No complicated processing, just:
  • ImageSolver (you could combine this with the next step, but I usually run it first in case there are any problems);
  • PhotometricColorCalibration (not that I'm desperate to get the star colours exactly right, it's also a robust way to set an initial "sensible" colour balance);
  • linked STF, fixed using HistogramTransformation (convert from "linear" to stretched "non-linear");
  • CurvesTransformation, mainly to enhance contrast and boost the colour saturation a bit (you can play with this for hours...).
This is not intended as an example of "expert processing" but rather what you can do youself quite quickly.
 
Thanks again Fred for all the help. I'm about half way through my processing of the Pelican. Still have to watch different videos to get a feel for a flow that works for me. Seems everyone processes with subtle differences, just need to find something I feel comfortable with, and more importantly, works.
 
If you haven't already, I would get Warren Keller's "Inside Pixinsight". He walks throught a basic processing workflow for mono and OSC cameras. I find it really valuable having a printed reference that I can look things up in.
 
Back
Top