Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Grid pattern with QHY5III 178?

  • Please log in to reply
21 replies to this topic

#1 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 14 September 2017 - 09:57 PM

Hi all,

 

Noobie here, have had a long interest in astronomy but only recently getting into astrophotography.

Picked up a QHY5III 178 mono camera recently and took some test solar sequences at full resolution, was surprised to see a faint grid pattern superimposed (histo stretched to highlight this). This occurs whether I use FireCapture (gain 5, 2 ms exposure) or SharpCap (gain 1, 2 ms exposure, offset 10), which to my understanding use different drivers; grid is present whether I use 8 or 16 bit processing.

 

Both programs seem to recognize it as a mono camera. I am using the 170223 QHY5III drivers and the 1.0.4.0 NativeWDM drivers; there are no other filters in the optical path. If it makes any difference, the sequences were through a Lunt LS50THa, although I can see the pattern even just waving the camera around the room not attached to a telescope or other lens. Sorry about the smudge on the lens.

 

I didn't see this with the QHY5L-II. I borrowed a QHY5III 174 mono and didn't immediately see this with some quick tests, I'll try to grab some sequences tomorrow to verify.

 

Has anyone else experienced this? It's completely possible that I'm doing something outstandingly dumb somewhere.

 

Thanks for any help you can provide!

 

Patrick

Attached Thumbnails

  • Sun_104051_lapl4_ap3.jpg


#2 HockeyGuy

HockeyGuy

    Ranger 4

  • -----
  • Posts: 398
  • Joined: 11 Nov 2015
  • Loc: Vaughan, ON, Canada

Posted 14 September 2017 - 11:01 PM

Interesting. Have you tried bias calibration?



#3 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 15 September 2017 - 12:49 PM

I did not, I figured that for the sun even in H-alpha, the signal would dwarf any bias noise.

Sequences from today with the QHY5III 174 mono did not show the same pattern. Pull out the 174, plug in the 178, grid pattern again.

 

The pattern stays completely vertical / horizontal when I rotate the camera, so it's not related to the imagery itself.

 

I have a ticket open with QHY, waiting to hear back.


Edited by FlankerOneTwo, 15 September 2017 - 01:12 PM.


#4 HockeyGuy

HockeyGuy

    Ranger 4

  • -----
  • Posts: 398
  • Joined: 11 Nov 2015
  • Loc: Vaughan, ON, Canada

Posted 15 September 2017 - 03:26 PM

Could you provide a bias frame, or better yet a master bias? It probably isn't the cause but it is always good to rule things out. If it is the bias offset causing this pattern, the strong signal would dwarf the read noise but not the bias offset.



#5 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 15 September 2017 - 05:42 PM

Interesting. Hope I'm doing this right - this is a master bias from DeepSkyStacker from 20 frames at shortest exposure (0.1 msec) with the same gain of 5. The histogram in Photoshop is basically all 0. I don't know if that's normal or if it suggests some sort of offset issue. I guess I can try to generate some flats.

Attached Files


Edited by FlankerOneTwo, 15 September 2017 - 06:00 PM.


#6 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 16 September 2017 - 01:29 AM

And here's the very histo-stretched cropped left half of one from SharpCap with the shortest possible exposure and maximum gain, offset of 300, 100 image stack. I do see a hint of a horizontal pattern, but not a vertical one. 

Attached Files


Edited by FlankerOneTwo, 16 September 2017 - 01:30 AM.


#7 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 16 September 2017 - 01:30 AM

Here's the histo of the unstretched offset bias above

Attached Thumbnails

  • histo.jpg


#8 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 6853
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 17 September 2017 - 05:12 AM

I have both an uncooled QHY5III-178C and a cooled ZWO ASI178MM and have done a fair amount of lunar and DSO photography with both and I've never seen anything like in your image. I've also used SharpCap, Firecapture, and Sequence Generator Pro with each of those cameras and again nothing strange to report.



#9 kingjamez

kingjamez

    Surveyor 1

  • *****
  • Posts: 1637
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 17 September 2017 - 10:02 AM

Those are Newton's Rings that can come up when doing Ha solar imaging. Tilt helps and can be the reason some cameras have it and others don't.

-Jim
  • OrionSword likes this

#10 Tapio

Tapio

    Vanguard

  • -----
  • Posts: 2028
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 17 September 2017 - 10:15 AM

Yes, there are Newton's rings in first image but that grid pattern is something else.

Maybe something to do with sensor properties.

You could contact ZWO to find out.

They also have forum which you could send your images.

https://zwoug.org/



#11 kingjamez

kingjamez

    Surveyor 1

  • *****
  • Posts: 1637
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 17 September 2017 - 03:27 PM

Yes, there are Newton's rings in first image but that grid pattern is something else.

Maybe something to do with sensor properties.

You could contact ZWO to find out.

They also have forum which you could send your images.

https://zwoug.org/

Ah yes, I was on a phone when I responded first and couldn't see the underlying pattern.



#12 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 20 September 2017 - 06:32 PM

As an update, the pattern appears pretty clearly in flats taken with the Lunt 50 pointed at the sky, but covered with a white plastic bag. This is out of the camera, no histogram stretching.

Similar image taken with the QHY5III 174 shows clear horizontal row noise, as is apparently known for that line of sensors, but not gridding.

The QHY5L-II did not clearly show any specific pattern that I could identify.

 

QHY is sending me another camera to test. Their support has been great so far.

 

The flat actually worked really well for cleaning up the image in AutoStakkert, at least as far as I can tell visually. Newton rings still an issue, I'm trying to find a reliable way to tilt at a consistent angle.

 

Attached Thumbnails

  • Master Flat.jpg
  • Sun_082855 processed.jpg


#13 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 06 October 2017 - 08:26 AM

Another update - the replacement QHY5III 178 has arrived and produces exactly the same pattern. It is visible in white light images at certain brightness levels, not just hydrogen alpha images. I'm puzzled that nobobdy else has seen this. I'm hard pressed to believe that somehow the laptop is responsible for this, but I'll try installing capture software on a different computer as well.



#14 Akwilliams

Akwilliams

    Mariner 2

  • -----
  • Posts: 282
  • Joined: 27 Aug 2016

Posted 06 October 2017 - 09:30 AM

That looks like a Bayer matrix... which I think you already figured out.   If it's a mono camera, then somewhere in the software there there must be a setting to turn it on/off - and set the pattern. RGGB/RGRB etc.

 

Although slightly different, When I bought my OSC, I got the exact same result until I worked out what the correct matrix was, it then went away..

 

in Deep Sky Stacker, I know you can change the matrix options.   It should look something like this for a mono camera

 

Capture.JPG

 

Rgds

Aidan


Edited by Akwilliams, 06 October 2017 - 09:36 AM.


#15 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 07 October 2017 - 12:23 AM

Hi Aidan,

 

I thought that at first as well, but on closer inspection I think the grid is too large to be a Bayer matrix - it appears to be 8 pixels across in each dimension. Isn't that too large to be from deBayering? FireCapture correctly recognizes it as a mono camera and will not let me turn on the de-Bayer function. SharpCap likewise recognizes it as mono.

 

Still puzzled,

 

Patrick



#16 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 6853
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 07 October 2017 - 02:20 AM

That's a strange looking bias file ( MasterOffset_ISO0 stretched2 ), but I guess it depends upon your gain setting and how you stretched the original. Here is a clip from one of my master bias files from a ZWO ASI178MM-COOL camera (lowest read noise gains setting, exposure "0", -15C temperature set point). An exposure of "0" in SGP just uses the lowest  supported shutter speed for the camera. You can see a grid pattern, but I've never seen a similar result on the final images after calibration and processing. This master bias is from 100 subs, integrated in PI and then auto stretched.

 

In any case, when imaging DSOs I no longer use bias files for either of my Sony CMOS cameras (IMX178 and IMX174) and I don't use either bias or darks when photographing something as bright as the moon (very low gains with short exposures).

 

I wonder, could this be from underexposing your images? What do your non-stretched originals look like (of the sun)? I guess if you were underexposing to the point where you had a very weak signal then this kind of pattern noise might become more obvious. In that case, maybe a bias would help, but if low exposure is causing your problems then I think the proper solution would be to increase your exposure times.

 

Attached Thumbnails

  • ASI178MM Master Bias.jpg


#17 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 6853
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 07 October 2017 - 02:29 AM

Also, here is an unstretched image of the moon that I took using the ASI178MM-COOL with an exposure of 8ms (0.008 seconds) and a gain of zero. I just opened the original file in PixInsight and then saved it as a 16-bit TIFF (with no processing except for the format conversion). After that I converted it to 8-bit in Photoshop and saved it as a JPEG. However, I  capture planetary images using the .SER video file format, normally I don't use .FIT for planetary work.

 

Also, an STF AutoStretch (boosted) crop from this same image showing the pattern noise in the dark sky surrounding the moon. Obviously, the sky would be considered underexposed in a shot like this.

 

Here is a link (on CN) where you can see a finished image of the moon taken with the above settings (single image):

 

  https://www.cloudyni...18#entry7995999

 

And here is a link where you can see a stacked (400 frames) and processed image of the moon that was taken with the ASI178MM-COOL:

 

  https://www.cloudyni...16#entry7797447

Attached Thumbnails

  • Moon_0.008sec_1x1_610nm_frame36.jpg
  • Moon_0.008sec_1x1_610nm_frame36 Stretched.jpg

Edited by james7ca, 07 October 2017 - 02:55 AM.


#18 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 07 October 2017 - 02:12 PM

That's a strange looking bias file ( MasterOffset_ISO0 stretched2 ), but I guess it depends upon your gain setting and how you stretched the original. Here is a clip from one of my master bias files from a ZWO ASI178MM-COOL camera (lowest read noise gains setting, exposure "0", -15C temperature set point). An exposure of "0" in SGP just uses the lowest  supported shutter speed for the camera. You can see a grid pattern, but I've never seen a similar result on the final images after calibration and processing. This master bias is from 100 subs, integrated in PI and then auto stretched.

 

In any case, when imaging DSOs I no longer use bias files for either of my Sony CMOS cameras (IMX178 and IMX174) and I don't use either bias or darks when photographing something as bright as the moon (very low gains with short exposures).

 

I wonder, could this be from underexposing your images? What do your non-stretched originals look like (of the sun)? I guess if you were underexposing to the point where you had a very weak signal then this kind of pattern noise might become more obvious. In that case, maybe a bias would help, but if low exposure is causing your problems then I think the proper solution would be to increase your exposure times.

Hi james7ca,

 

Thanks for your response. Interesting that you have some gridding also with a  deep stretch. I generally do not see it in unprocessed images, mainly back when I was doing a lot of histogram stretching trying to see if I could get any prom detail out of the 8-bit captures. More recently I've been capturing (with the 174, though) in 16-bit mono with relatively low gain and exposures set to yield about a 85% histogram - this seems to work pretty well to allow me to pull out prom detail from a single capture.

 

I can recreate it in defocused raw captures in white light, but have to tilt the camera at just the right angle in a certain range of brightness.

 

I'm just getting in to AP so was looking to do something that didn't require complicated image capture/calibration workflows. Using flats seems to remove the pattern pretty effectively and I'm getting the hang of the workflow, so I'll probably just stick to capturing flats and live with it as a type of fixed pattern noise.



#19 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 07 October 2017 - 09:39 PM

Hi james7ca,

 

Had a chance to play with your lunar image a bit after work - with a deep histogram stretch I see what I think is the same pattern in your image - I've attached a Photoshop screen shot below. The grid dimensions appear the same, 8x8 pixels. I get the same result in PixInsight, so it's not a Photoshop problem. The effect on your lunar image is a bit different from the discrete gridlines that I see in my images, I'll have to grab some lunar sequences and see what they look like. It has me still wondering if this is a readout / driver problem.

Attached Thumbnails

  • screenshot.jpg

Edited by FlankerOneTwo, 07 October 2017 - 09:40 PM.


#20 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 6853
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 07 October 2017 - 10:22 PM

Those are JPEG compression artifacts. JPEG compression works on blocks, so any stretch or reprocessing of that data is almost certain to show square artifacts. Also, to get under the posting limits on CN I very often have to apply a pretty high compression rate. In any case, you should never use JPEG as a storage medium for images that are going to be processed.



#21 FlankerOneTwo

FlankerOneTwo

    Viking 1

  • *****
  • topic starter
  • Posts: 629
  • Joined: 30 Aug 2017
  • Loc: Vegas, baby!

Posted 07 October 2017 - 10:56 PM

Ah, that explains why it looks different. My sequences were AVIs / SERs and I only stored in lossless TIF during processing, so doesn't explain the pattern. Lots to learn about CCD imaging and processing! Oh well, I guess I'll wait to hear back from QHY.


Edited by FlankerOneTwo, 07 October 2017 - 10:59 PM.


#22 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23089
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 08 October 2017 - 01:48 PM

Would be nice if planetary imaging software had some built in dithering capabilities. Periodically do quick, short shifts in RA and DEC so that the pixels the planet falls on changed over time. You might lose a few frames each time, but I think that would be far preferable to having FPN cause problems in the stack. 




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics