×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Wrong geometry of FITS files captured in EKOS

I was able to take 30+ frames with ZWO ASI178MC-Cool without any issues with frame dimensions. I opened them PixInsight also without problems. The only warning I received is "Warning: Deprecated Format: FITS". However, I ran into an issue when trying to debayer the image. It gave an error "Missing or invalid CFA description property". The header has BAYERPAT set to RGGB so I'm not sure what the issue is with PixInsight regarding this.
5 years 1 month ago #34424

Please Log in or Create an account to join the conversation.

  • Posts: 20
  • Thank you received: 9
Corazon

There is a way to salvage those frames with 4 pixels chopped off. I have experienced this problem recently with my asi071 camera. I've found a simple work around to this is to crop all images to match the smaller image size using DynamicCrop and image containers in pixinsight. Its a quick way to batch crop all your images to the correct size so they will stack. You have to do your flats, bias, and darks also this way for it to work.
The following user(s) said Thank You: Jose Corazon
5 years 1 month ago #34425

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160
Same here, Jasem. Tested locally USB2, locally USB3, remote USB2, remote USB3, lights, darks, flats. No faulty images.
Last edit: 5 years 1 month ago by Alfred.
5 years 1 month ago #34430

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182


Yes, I know, but that's painful. I am trying to find out what causes this. It only affects my light frames, not the flats or the darks (I don't use bias frames for my ZWO, too much city noise for that to make any difference).
5 years 1 month ago #34432

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182

It only affects my light frames, not the flats or darks. That's why I am suspecting the problem originates from within EKOS, not the ZWO.
5 years 1 month ago #34433

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160

I know, same here. However, I cannot rule out that the first light frame after a dark or flat is affected. Therefore, I have to take darks and flats first.
5 years 1 month ago #34434

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182

I can tell you that that is NOT the problem. Embarrassingly, I am forced to admit that I am reusing my darks and flats for weeks or even months. As long as I don't change anything in the light train the flats continue to remove all the artifacts.

Therefore, the problem is unrelated to darks and flats. But your observation is correct: The FIRST light frame is most often, but not always, affected. Then it happens again infrequently throughout the sequence. As I wrote, on average it affects ~10% of my frames. I wonder whether it could be related to the data transfer rate or the USB hub? I am using a ZOTAC miniPC running UbuntuMate through a USB3 hub. EKOS runs directly on that machine, which I control via VNC.
The following user(s) said Thank You: Radek Kaczorek, Alfred
Last edit: 5 years 1 month ago by Jose Corazon.
5 years 1 month ago #34435

Please Log in or Create an account to join the conversation.

  • Posts: 983
  • Thank you received: 375

I would kindly disagree. This approach assumes you know which lines are missing. Whereas these might be: 4 top lines, 4 bottom lines, 2 top + 2 bottom, 1 top + 3 bottom, ..., whatever you choose.
Cropping images changes physical mapping of pixels and may render unpredictable results during stacking phase. It would also influence calibration phase due to possible shift of light vs dark or light vs flat.
The following user(s) said Thank You: Greg
5 years 1 month ago #34437

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160

:lol: That's good to know as it reduces the amount of test work. In my case it was much less than 10%. I was testing all night and just 2 images were bad.

I think in these cases the image file would be corrupted. What we see here though is the same number of pixels missing. But then, what do I know! I don't use any external hub at all. I removed all potential bottlenecks, everything is connected directly to a fully fledged desktop computer running indiserver. No wireless connections either. I find the combination USB3, decent server, Gigabit Ethernet, decent client is a pleasure to work with. Now, if only we could find out what has been going wrong recently...
Last edit: 5 years 1 month ago by Alfred.
5 years 1 month ago #34438

Please Log in or Create an account to join the conversation.

  • Posts: 20
  • Thank you received: 9

Using Pixinsight's DynamicCrop process, the process will display where the center anchor is in the image, thus you can derive what side the pixels are being cut off. If that's not a good enough indicator, then a visual toggle using CTRL+PGDN will show you the edge that has been chopped off when comparing it with a normal sized image. It has always been the bottom 4 pixels in at least 60 of the last 200 lights i have imaged in January. This problem started happening just recently with my asi071.

My stacks turn out just fine, flats are mapped great and so are hot pixels. As long as you KNOW where the pixels are being cut off, then all is good.

One thing to make note, I am running an older stable version of stellarmate 1.2.9.
5 years 1 month ago #34460

Please Log in or Create an account to join the conversation.

  • Posts: 424
  • Thank you received: 66
I've had this problem no matter what software I use. Us Pentax users suffer from lack of device support on all astro software. The geometry problem comes up unless one uses native files from the camera. But even then you have to be careful. Raw files captured to the camera SD card are not the same size as raw files captured by tethering - the tethering seems to give truly raw files, including the "guard" bands. This problem has dogged me from the first time I tried to use calibration frames and the stacking software - in this case Nebulosity - coughed it up.

It makes sense that calibration frames are not affected. The software doesn't really care about the CFA or dimensions for these , only that they are the same as the lights.

Solution:

1. Use Native file format for DSLRs. In the case of Pentax, the fits format does not put the correct CFA type in the header! I raised this issue before. (And anyway, you have to manually enter the CFA, its not read from the camera.)

2. For CCD cameras, I think this may work as its similar to the DSLR solution and I was trying it before the finding the CFA problem.
- Capture light frames using Native format with Kstars - if this is fits then I think your problem is solved - if not and your stacker doesn't understand the format, then you can try batch converting to fits with something like Nebulosity, or continue below...
- Find out the true geometry with something (again) like Nebulosity. (I've used Raw Therapee but even that does not report the correct dimensions.)
- put those dimensions in Kstars camera settings as max dimensions and do not try to do subframe the fits output. Keep it set to the max dimensions. You may be able to use fits format capture at this point with consistent geometries and CFA info.
5 years 1 month ago #34462

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
FWIW, I also see the same issue--4 rows missing from some images on my ZWO ASI 1600mm pro using Ekos.
It is usually, but not always, the first image from a set of images I start taking from the camera tab.
I don't use the scheduler (yet) so it's not that. Since I typically take 120s exposures, I just throw these away, not a big deal for me.
I only discover the issue in post when running Blink on PixInsight.
I'm sure I could recover several of these images (and companion images that were not "cropped") if that would be useful to someone who might be looking into this.

Hy
5 years 1 month ago #34472

Please Log in or Create an account to join the conversation.

Time to create page: 0.630 seconds