Nikon Z 6/7 success?

Started by Aelfwine, April 23, 2019, 04:08:38 PM

Previous topic - Next topic

Aelfwine

Does anyone else here use a Nikon Z 6 (or 7) for astrophotography? I'm not having much luck with mine in Observatory. My darks look fine in camera and in Lightroom, but in Observatory they are bright magenta or purple and very noisy. My lights meantime, which likewise look fine outside of Observatory, have a pronounced magenta cast. (See the attachments.) Nor does this cast go away after calibrating and stacking.

Has anyone had better luck than me?

I should specify that I shoot RAW. Converting to DNG made no difference.

Sander Berents

Does it improve if you apply the Background Neutralization adjustment?
Sander Berents
Code Obsession, LLC
https://codeobsession.com

Aelfwine

I _think_ I tried that, and no.

Anyone who wants to try themselves, my darks and lights are here:

http://www.thehostetters.net/public/Observatory_help/


Sander Berents

It makes a big difference if I try that with one of your darks and lights.

Also please be aware that Observatory adjusts the display black and white points for every image you import, unless it is an 8 bits image. Its default stretch settings work well for well exposed grayscale images, but clearly not so much for your camera. I think that if you reset it to "All" with the Histogram cogwheel at the bottom right of your library window, that it will look much more like in Lightroom. You can also drag the black and white points in the Histogram (clicking left or right of them is an even easier way to nudge them).
Sander Berents
Code Obsession, LLC
https://codeobsession.com

Aelfwine

#4
I can't say I know what that means right now (e.g. I'm definitely NOT shooting "greyscale images", but rather default in-camera RAW images), nor am I sure what the "it" is that I'm supposed to "reset to All", but I'll take this under advisement and see what I can do. So, thanks!

EDIT: Ok, now I think I know what you're suggesting, and will give it a try. I'll let the forum know how it goes.

Aelfwine

So wait, do I need to do the suggested histogram adjustments and background neutralization to 1) both my darks and lights, and 2) to each image individually, i.e before creating a master dark, and before any stacking of lights?

Sander Berents

The black/white point adjustments are for displaying the images only. You can do that any time. Select all your images, and then choose "All" in the histogram cogwheel for example. The default is "Low".

You can apply adjustments to multiple images at once, by selecting them and then adding the adjustment. Don't apply Background Neutralization to your darks though if you want to use them for calibration, as that will lead to bad calibration.

You could apply Background Neutralization to all your lights, before stacking, but I typically just calibrate & stack, create a snapshot of the stack (Image ▸ New ▸ Master), and apply Background Neutralization to that.
Sander Berents
Code Obsession, LLC
https://codeobsession.com

Aelfwine

#7
"Select all your images, and then choose "All" in the histogram cogwheel for example."

Doing that for the darks causes Observatory to hang (spinning beachball that never ends). And doing it on individual darks doesn't seen to make any noticeable difference. So I'll skip that. Working on the rest...

EDIT: Adjusting the histogram to "All" on just the master dark does work.

Aelfwine

#8
OK, so with 1a) background neutralization and 1b) heavy background flattening of the lights, and 2) a lot of histogram adjustment on the master, I was able to get a good-looking stacked image.

However, now I can't get Observatory to plate-solve at all. I've specified the brightest star nearest the center (Alioth, in the handle of the Big Dipper), and the pixel ratio (30 arcsecs, calculated based on the angular distance from the tip of the bowl of the Big Dipper to the end of the handle, all told about 25 deg., times two because the Big Dipper spans about half the frame, so 50 deg. = 180000 arcsecs, divided by 6000 pixels, the horizontal resolution of the frame), and reduced the frame size. I've also downloaded and pointed Observatory to UCAC4. But it still fails to solve.

Any ideas on that?

I've put both the master and my histogram-massaged images in the same place as my darks and lights:

http://www.thehostetters.net/public/Observatory_help/



Aelfwine

#9
Thinking about this more, I might be able to make sense of it.

I assumed that the plate-solving system would use my histogram-massaged image, which had a suitable dark background, and clearly displayed the Big Dipper and environs. But that's not true, is it? Instead, it uses my NON-histogram-massaged MASTER, which has a pronounced bright background (with chromatic ring), which of course would baffle any matching system.

If so, then I have to create a corresponding set of flats to further calibrate lights, and hopefully thus remove the bright and circular bias in my lights (even though those biases are NOT visible in Lightroom, and even though I did a heavy linear background flattening of this purely starfield imaging sequence).

(Then again, if you can show me another way with the darks and lights I linked to, I'd be very happy to know of it!)

EDIT: Astrometry.net had no problem solving my histogram-massaged TIFF:

http://nova.astrometry.net/user_images/2801584#original


Aelfwine

On that note: I think it would be really helpful if the CodeObsessors would provide a guide to creating the various bias frames that Observatory can use. I had to rely on other sites for that, before I understood their nature and use. Many newbies would benefit from an Observatory-specific discussion/tutorial on creating these.

Sander Berents

Observatory will indeed ignore the black/white points for plate solving. These are for displaying only. If plate solving with astrometry.net works, then plate solving using Observatory's blind plate solver may work as well, since it uses astrometry.net. However, I suspect you will still need to calibrate your images with flats first.
Sander Berents
Code Obsession, LLC
https://codeobsession.com