896 replies to this topic

[sharkmelley]

what actually happens when you tape over the lens contacts? does i completely eliminate the issue or allows us to shoot at any ISO level? 

Taping over the lens contacts has no effect.  This is because the internal processing that causes the rings is applied to the image data whether or not a recognised lens is attached. 

[JS04_Astro]

Taping over the lens contacts has no effect.  This is because the internal processing that causes the rings is applied to the image data whether or not a recognised lens is attached. 

so when i shoot flats as far over to the right of the histogram as possible, my flats look like rainbow splotches (my D5300 is Ha modified). i may have to play around with shutter speeds to find the right settings. I dont see colored centric rings in my lights at 60s ISO 200, only when i add flats into the stack i do.  

[Infomastr]

It's much harder to see in a single light compared to a single flat.  For the Z6ii, I tested a whole range of ISOs (from 100-6400) and EVs (from -3 to +3) to see where rings would be minimized.  Higher ISO (1600+) and +2EV helped in my case.  If you post some flats, a master stack, or do Mark's test protocol, you can get a better idea.

[sharkmelley]

so when i shoot flats as far over to the right of the histogram as possible, my flats look like rainbow splotches (my D5300 is Ha modified). i may have to play around with shutter speeds to find the right settings. I dont see colored centric rings in my lights at 60s ISO 200, only when i add flats into the stack i do.  

The rings on the D5300 are caused by Nikon's lossy compression (which can't be switched off).  The only way to avoid them on the D5300 is to keep the exposures very low.  Alternatively use my software that repairs the damage caused by the lossy compression. 

More information here:

https://www.markshel...ompression.html

[JS04_Astro]

The rings on the D5300 are caused by Nikon's lossy compression (which can't be switched off).  The only way to avoid them on the D5300 is to keep the exposures very low.  Alternatively use my software that repairs the damage caused by the lossy compression. 

More information here:

https://www.markshel...ompression.html

You dont have an option for Macs do you? i can dig out the windows PC if need be. 

[JS04_Astro]

It's much harder to see in a single light compared to a single flat.  For the Z6ii, I tested a whole range of ISOs (from 100-6400) and EVs (from -3 to +3) to see where rings would be minimized.  Higher ISO (1600+) and +2EV helped in my case.  If you post some flats, a master stack, or do Mark's test protocol, you can get a better idea.

I can’t share everything since my Mac is at home. this image below is a master stack with flats applied 

 

https://share.icloud...UiyxzDlHzfbmFaw

 

this is the only other one I have master stack before flats 

 

https://share.icloud...fDSyyTcLHkEZFTg

[sharkmelley]

You dont have an option for Macs do you? i can dig out the windows PC if need be. 

No, I'm not able to build a Mac version

 

 

I can’t share everything since my Mac is at home. this image below is a master stack with flats applied 

 

https://share.icloud...UiyxzDlHzfbmFaw

 

this is the only other one I have master stack before flats 

 

https://share.icloud...fDSyyTcLHkEZFTg

To me these don't look like the usual rings caused by Nikon's lossy compression.  More likely it's an optical effect such as stray/scattered light

[james7ca]

Okay, here are some results from a nearly new, unmodified NIkon Z50 II camera and a Voigtlander 50mm APO lens at f/2.8. I used an ISO of 1600, an exposure of 1.6 seconds, and a blank-screen iPad (dark gray TIFF image) with a sheet of flashed opal glass between the iPad and the camera lens (I also used several sheets of white paper stacked flat between the flashed opal glass and the iPad). The results are from a master flat that included 48 flat frames and 64 darks. I rotated the camera after every 8 flat frames so any non-radial variation in the iPad's display should have been reduced (mostly, if any existed). I used PixInsight for the calibration and flat integration and then PixelMath to produce the red divided by green and blue divided by green output images (from the debayered and channel separated master flat).

 

IMO, these results will vary with the exposure level, ISO, vignetting, and level of stretch applied to the resultant images but I think my copy of the Z50 II really isn't showing much in terms of concentric rings. There is certainly variation across the frame, but I suspect it is below what could ever be detected in a standard astro image. My actual images haven't shown any significant issues with colored rings or artifacts, but I am imaging from a Bortle 7 zone and haven't pushed the camera much beyond one hour of total integration time. Plus, in many cases I've used PixInsight's MultiscaleGradientCorrection tool and I think that process can actually remove complex color artifacts from the final results. So, the post processing on my deep space images may have removed any remaining trace of artifacts from the so-called Nikon colored concentric rings.

 

Of course my test results (shown below) did NOT get any such treatment, just a calibration, integration, debayer, and channel separation followed by a historgram transfer of a boosted screen transfer function in PixInsight. Without a stretch the flats looked completely uniform except for a smooth, radial vignetting from the lens. After the above processing I then rescaled each result down to 25% in PixInsight followed by another reduction of 50% using Photoshop. These 16-bit TIFF images where then output as 8-bit converted JPEGs using Photoshop

 

Interestingly, the blue divided by green result is showing thick horizontal bars at the top and bottom of the frame. I've done this same set of tests using a radically different lens (a NIKKOR Z 180-600mm zoom) with a completely different light source and the results look quite the same. So, these artifacts seem to be real although I don't know if they would be present in each and every sample of the Nikon Z50 II when subject to the same tests. In any case, I haven't noticed any evidence of these artifacts in my processed DSO images so again I think in actual practice these variations may be too slight to be seen.

 

There is certainly more work that could be done on the Z50 II to confirm these artifacts and I also have a Z30 that apparently uses the exact same sensor as the Z50. So, I could run tests using the same lens on both cameras and see what happens.

 

First the blue divided by green result (with the horizontal bars) and then the red divided by green.

Attached Thumbnails

• 
• 

Edited by james7ca, 03 June 2025 - 06:03 AM.

[james7ca]

And here is a normal and unlinked screen transfer stretch on the debayered master flat from my Z50 II. The stretch emphasized the darkening in the corners caused by the vignetting on the Voigtlander lens and there appears to be a slight color gradient going from side-to-side across the frame. But, it's still pretty uniform.

Attached Thumbnails

• 

Edited by james7ca, 03 June 2025 - 07:50 AM.

[james7ca]

I looked at some flats that I took in September 2024 with a Nikon Z30 and an Askar FMA180 telescope and while those SEEMED to show a slightly more pronounced set of concentric rings the blue divided by green channel showed a very similar set of horizontal bands at the top and bottom of the frame as did the Z50 II with both the Voigtlander 50mm and NIKKOR Z 180-600mm zoom. Thus, since I saw basically the same horizontal bands with two different cameras when using three different lenses and with two different light sources I'm fairly certain that those artifacts are in the sensor or the Nikon firmware (i.e. RAW preprocessing by Nikon).

 

Of course, it could have something to do with the way I'm taking the flats and I'm using the fully electronic shutter feature on both of these cameras. For diagnostic purposes I may switch to using the mechanical shutter and with both cameras using the same lens. I may be able to try some sky flats since if these horizontal bands show up in sky flats it must be something in the camera.

 

That said, I did NOT see these kinds of horizontal bands on the testing I did on the Nikon Z8 using this same process of taking the flats ( __shown here in this thread__ ).  However, on the Z8 I used ISO100 and a 105mm NIKKOR Z macro lens.

Edited by james7ca, 03 June 2025 - 11:24 AM.

[otoien]

It does not explain it, but the area inside those bars has an exact 16:9 aspect ratio, so in all likelihood somehow related to the sensors video mode, unless the PDAF rows have a similar distribution.

[sharkmelley]

The horizontal bars you are seeing with the Z50II are quite common with Nikon mirrorless cameras and are related to which rows contain the PDAF pixels. The rings certainly look like the usual Nikon rings caused by the hardcoded correction but I would need to look at a raw file to be more certain.

[Michael Covington]

It does seem that over the years, DSLRs have gotten less suitable for astronomy as the manufacturers find other fancy things to do with the sensor.  I still like my Canon 60Da for its lack of quirks despite having an older-style sensor.

[whwang]

It does seem that over the years, DSLRs have gotten less suitable for astronomy as the manufacturers find other fancy things to do with the sensor.  I still like my Canon 60Da for its lack of quirks despite having an older-style sensor.

It's sad and true.

 

I still use my hacked D800, which gives images as clean as possible.

[JS04_Astro]

No, I'm not able to build a Mac version


To me these don't look like the usual rings caused by Nikon's lossy compression. More likely it's an optical effect such as stray/scattered light

Stupid question, do we need to convert the fixed files back to a format Pixinsight can read? I’m getting errors now on my wbpp. Not sure what I did wrong now

[sharkmelley]

Stupid question, do we need to convert the fixed files back to a format Pixinsight can read? I’m getting errors now on my wbpp. Not sure what I did wrong now

The fixed files are in DNG format.  PixInsight should understand DNG.

[JS04_Astro]

The fixed files are in DNG format.  PixInsight should understand DNG.

Hmm maybe my flats were not taken properly. 

 

https://drive.google..._tm?usp=sharing

 

See if you can reach my link, I may need to provide access. I would appreciate if you had a moment to take a look. I included a light frame, Flat frame, and a fixed flat 

[sharkmelley]

See if you can reach my link, I may need to provide access.

No-one has access to the file(s) at that link.

[JS04_Astro]

No-one has access to the file(s) at that link.

please try again 

[sharkmelley]

Hmm maybe my flats were not taken properly. 

 

https://drive.google..._tm?usp=sharing

 

See if you can reach my link, I may need to provide access. I would appreciate if you had a moment to take a look. I included a light frame, Flat frame, and a fixed flat 

It appears to me that you are mixing NEF (and DNG) raw files with the FITS files recorded by ASIAir.  Don't do that because they are not compatible.  Don't mix ASIAir files with genuine raw files.

[JS04_Astro]

It appears to me that you are mixing NEF (and DNG) raw files with the FITS files recorded by ASIAir.  Don't do that because they are not compatible.  Don't mix ASIAir files with genuine raw files.

This could be why wbpp is failing. I haven’t had clear skies to retake new lights. Should tonight. I just wanted to be sure I had taken my flats properly.  

[primeshooter]

Okay everyone, we could have a chance to fix this. I don't know if you are aware, however Thom Hogan has reached out to his readership about firmware updates / upgrades. I have emailed him about this issue and linked to Mark's page. He hasn't responded to my email yet however he will be busy I expect. If others perhaps email him regarding it, perhaps it'll make the list and be put in front of Nikon. Remember Thom helped solve the star eater issue in the early 2000s...

 

Link is here to the suggestions: https://www.zsystemu...-wish-list.html

Ps: I popped in the (randomly) missing crop mode, 4:5.

Edited by primeshooter, 14 June 2025 - 11:45 AM.