Need help - calibrated lights have no detail - Forums

Topic starter

Aug 1, 2021 · Edited: Aug 1, 2021

Hello to you all,

I do have a problem processing my latest images and really need some help or a hint what went wrong with my processing workflow.

On friday night, I took some images of the wizard nebula NGC7380 using my TS Photon 8" f/4 and my ZWO ASI183MM. I used PixInsight 1.8.8.8 to generate the master lights for the images taken with my Ha, green and blue filter. While green and blue are processed well (exposure time 300s with gain 178 ), the Ha data is unusable. Maybe there are some settings I have to change, that I don't know. But first things first. Here are some imaging and processing details:

exposure time 600s, gain 300, offset 5, temperature -25°C
try to integrate 10 frames with 600s each
during the night, I used EZ Live Stack script to view the image building up (but with no dark selected) - the typical amp glow was visible (no problem yet) but data looked well - nothing to worry about
dark frames came from a library I took a while ago - all my recent images were processed with that data and no problem so far
used PI WBPP script to integrate all data (bias, flats, darks and lights) using the standard setting but did turn off the optimization window for the darks (optimization window = 0), left everything else on default

I have to say, that I am quite new to PI and didn't have any real knowledge about it, yet. I learned a bit about some advances techniques but those are used after the integration process. For image integration itself, I did not find the need to get any further into that topic, because all went well until today.

Here are some samples of what I'm talking about. All images are stretched and downsampled to 2000 Pixel width.

Single Raw frame:

calibrated raw frame:

master dark:

So my guess is, that there may be something wrong with my calibration process. So far I realized, that the histogram of the original single frame is quite compressed to the dark side. Images of other targets do have more data in the lower third of the histogram. That means, that the SNR in the images here is worse than that in my provious images, isn't it? Maybe the target is fainter, which I couldn't believe at 7,2mag.

Long story short: are you able to help me with my problem? Is there a "magic slider" to save the data of the night?

By the way: I also tried DSS and the result is unusable, too.

Thank you for your help!

Clear skies!

Christian

Helpful

Reza Hakimi

Aug 1, 2021

Hey Christian
That problem once happened to me and it was when I accidentally misplaced my flats in my dark folder and my dark in my flat folder. after calibration the calibrated subs looked exactly like yours. so check that one probability out. make sure you apply the right masters in calibration.

Concise

Fritz

Aug 1, 2021

Hi Christian,

there is something wrong for sure. To trace this down, I would suggest to narrow it down by using only one single light frame to speed up things. Try to calibrate only and use flats (no bias, no darks) in one attempt and darks (no bias, no flats) in the second attempt. Make sure, to use no optimisation. If both attempts run fine, try one more attempt with flats and darks (no bias). This way you will find put what type of calibration causes thius headache. I use a different camera (1600mm pro) and had similar problems now and then. I found, that im my case the bias frames caused the trouble, but your case may be differnent.

Good luck & CS
Fritz

PS: Wenn du mir einen light frame und die master calibration frames (dark, flat, bias) irgendwo hochladen kannst, kann ich mir die Sache gerne einmal annsehen.

Helpful Supportive

Björn Arnold

Aug 1, 2021 · Edited: Aug 1, 2021

One can immediately see from the dark frame that you might accidentally have mixed different calibration frame types (flats into the set of darks). The dark shows vignetting which is impossible, unless your sensor‘s broken.

Edit: By the way, histograms will always have their peaks close to the dark side (if image is linear) because most of the data is background and most of the details are extremely faint and only become visible by stretching.

CS!
Björn

Helpful Insightful Concise

Christian Großmann

Topic starter

Aug 1, 2021

As @Reza Hakimirecommended, I tried to rerun the WBPP script and carefully selected all the frames again. This didn't solve the problem. I was quite sure, that now everything was right.

I spent some hours this afternoon to get into manual image integration with PI. Basically I tried to follow @Fritz 's advice to break down the process into basic operations. I manually created a master dark using the suggestions in the book "Inside PixInsight". I opened my lights in PI and moved the view of the MasterDark over them to see the alignment of the amp glow. It looked fine. Then I calibrated the lights using only the master dark (no flats or bias yet). The result is the same again. The calibrated light looks exactly as in my first post.

I also took another look at the master dark. It looks the same as in my first post, too. I am not really sure if @Björn Arnold is right. The dark corners may look like vignetting. But maybe they are not. As I understand the image content, I recorded the heavy amp glow, which is different on every part of the whole sensor. The corners are "colder" and therefore must be darker. There also is no "darkness" on the lower right corner. The amp glow becomes more visible as the exposure time increases. If I stretch the darks that were taken using shorter exposure times, the corners become less visible as I expected. With a longer exposure time, the overall histogram shows more lighter tones and the stretch of the corners is not as heavy as with shorter exposure times (I hope I'm right.). Another fact against this theory are the "light beams" that can be seen in the corners of the left side.

I broke down the problem to the dark frames and their substraction from the lights. But my problem isn't solved yet. Maybe someone has another idea. In between, I will record some new darks and try to rerun the process. Hopefully, I will get some new findings.

Thank you for your help, so far.

CS

Christian

Helpful Respectful

Björn Arnold

Aug 1, 2021 · Edited: Aug 1, 2021

Hi Christian,

I'm sorry, but I need to object. Here is a 300 second dark frame downloaded for the ASI183MM Pro from ZWO's website: (Heavily stretched in APP and downsampled to a 16th of overall size to upload it here).

Here is a 900second master dark, created with my Altair AA183M Pro (also stretched - but not that much - in APP and downsampled to a 16th of overall size to upload it here):

As you can see, the amp glow explodes with time but still, the whole sensor is evenly "illuminated". There is no reason to assume the corners are colder.

How do you take your darks?

Another note on your camera settings: I strongly recommend against using gain 300. The sensor may exhibit artifacts if the gain is driven towards maximum. It fully suffices to use unity gain for NB imaging. In addition, there is no need to cool it down to -25C. This sensor produces very little thermal noise. You may take a look at my HOO of Sh2-72 image, taken at sensor temperatures of 15 to 20 deg C (no, I didn't miss the negative sign). Minus 5 or -10 C are fully sufficient for low thermal noise.

Therefore, I strongly recommend to take another look at your darks as they look incorrect.

Cheers,
Björn

Well Written Helpful Insightful Engaging Supportive

Ruediger

Aug 1, 2021

Hi Christian,

two things come to my mind:

1. Make sure that the option “optimize darks” is not checked in PI resp WBPA.

2. you should take (for testing) a new dark. Especially you should take the dark with the same acquisition SW you used for the lights. Had ones such a problem, when I took darks in APT and lights in NINA.

3. calibrate a single sub step by step as suggested above, and see at which step the problem occurs.

CS
Rüdiger

Helpful Concise

Christian Großmann

Topic starter

Aug 1, 2021

@Björn Arnold
Ok. That's an interesting point. All my darks tend to have those dark corners. I am quite sure, that I took the images with just the lens cap on the camera. But I may be wrong.

Some minutes ago, I placed the camera in a dark cellar and put the protecton cap on the Coma corrector (I still have the filter wheel and the coma corrector connected). But I put a dark cloth over the equipment with just the fan and the heat sink out to enable a good cooling. Tomorrow I will check the darks for any changes.

@Ruediger
Oh... That's what I did. I know I took the darks using APT. And recently, I switched to NINA to record my lights. At the moment, I use NINA to record my new darks. So maybe that's a solution.

Thank you both for your help...

Christian

Björn Arnold

Aug 1, 2021 · Edited: Aug 1, 2021

@Christian Großmann
One problem that can arise is that the lens caps allow transmission of infrared raditation and I assume the camera's glas is not IR protective. Therefore, make sure that no IR reaches the sensor. What I recently did when I built the dark library for my new ASI is that I've placed the B filter in front of it. The lens cap will take care of the visible light and the B will kick out the IR photons. (One could also use a NB filter)

Another point: are you sure about the offset = 5? With that offset, an exposure time of 600s and gain 300, you might clip the some dark pixels at black.

You don't need to recreate the whole dark library for the diagnostics here a single dark frame will do it to see if something has changed.

Björn

Helpful Insightful Concise

Ruediger

Aug 1, 2021

Christian Großmann:
Oh... That's what I did. I know I took the darks using APT. And recently, I switched to NINA to record my lights. At the moment, I use NINA to record my new darks. So maybe that's a solution.

I had built the dark library with APT. It turned that the offsets were not matching. I had re-built my dark library with NINA and the problem was solved.

CS
Rüdiger

Christian Großmann

Topic starter

Aug 2, 2021

I'd like to thank you all for your help! During the night, I took a new set of darks and tried to rerun the WBPP script. Now all works fine. The result is very pleasing…

I did some further tests and found some more facts that may help someone with similar problems. If the subs have enough detail, the calibration process worked even with the bad darks. That's why I had no problems so far that seems to be related to the darks. But if there was NOT enough detail in the image (the histogram was really close to the dark side) the calibration failed.

There's a lot of stuff to learn in astrophotography. Solving those problems is a real progress for me…

So Thank you and Clear Skies to everyone.

Christian

Helpful Respectful Supportive

Related discussions

Having trouble setting up Windows 10 GPU acceleration for BXT or SXT?

Hello all, I'll acknowledge up front I'm no expert with the CUDA stack. I had quite a bit of trouble setting up my windows 10 laptop with NVIDIA GPU for BXT and SXT. I couldn't get the latest version of CUDA Toolkit (v12.0) to work under ...

Dec 26, 2022

Both posts are from astronomy forum users seeking or providing help with technical processing issues they encountered while working with specialized astrophotography software and equipment.