Is this a balancing act or seeing conditions?

[wimvb]

During the half decent nights we've had this past week, I tried capturing the Iris nebula (NGC 7023). On three consecutive nights I managed to get to know my new mount and test its limits. I thought I'd share some of my findings and ask you guys/girls for some comments.

This is common for the two images

• mount: AZ-EQ6 GT
• scope: SW 150PDS
• camera: Pentax K20D at ISO 1600
• sub frame exposure: 6 minutes
• no guiding
• Bahtinov mask for focus
• PA using the Synscan handset, until the reported error was less than 2 arc minutes
• no dithering was used (my imaging setup is stand alone without computer control, and I can't connect my ditherbox to the AZ-EQ6)
• in both cases I used bias frames, darks and flats as well as Cosmetic Correction
• both images stacked and stretched (same settings) in PixInsight

This is different for the two images

• The image taken on 28 Oct consists of 20 subframes, while the image taken on 30 Oct consists of 13 subframes
• The two nights were not completely cloudless. I had to discard frames due to clouds and due to tracking issues (single frames selected with the Blink tool; images with elongated stars were discarded from the stack)
• On the first night, the scope wasn't as accurately balanced (RA and DEC) as on the second night.

My conclusions from this experiment are

• the softer image from 30 Oct is due to fewer subs resulting in more noise as well as poorer seeing conditions
• the streaks in the first image are very likely caused by poorer balance resulting in slight tracking inaccuracies

Your thoughts on this, please. The main question being: are the streaks caused by the poorer balancing, or can they be caused by poorer seeing?

If caused by residual lukewarm/hot pixels, then seeing shouldn't be an issue.

Or am I missing something?

(And btw, guiding setup is ordered, and should be arriving tomorrow. Clouds and rain have arrived already)

[sharkmelley]

What happens if you integrate the subs without image registration - is more drift seen in first image than in the second image? 

Are you using dark optimisation during calibration of subs in both cases?

Mark

[wimvb]

2 hours ago, sharkmelley said:

What happens if you integrate the subs without image registration - is more drift seen in first image than in the second image? 

Are you using dark optimisation during calibration of subs in both cases?

Mark

There is always a slight drift between subs, but I believe it was more for the first image. I didn't create a log file when doing the image integration. If I had, I would have access to the offset values (dx, dy). I remember that when using the Blink tool on the subs for the second image, there was hardly any drift noticable between exposures. I may make a small movie of the calibrated but unregistered subs to show drift between exposures. Will post it here when done.

I did use dark optimisation in both cases. Normally I don't use darks, but rely on cosmetic correction and pixel rejection during integration. But I also have amp glow at the left edge of my subs, and this is partially corrected by dark frames. (My usual remedy for the amp glow is to crop the image.) So, this time I used darks.

[ChrisLX200]

Yes it could be fixed pattern noise moving across the frame (if so it should follow the drift bottom-left to upper-right). Why it's completely absent on the image from the 30th is a mystery, but I suspect focus is a factor in addition to any seeing artifacts in causing softness. In both cases you are looking down into the noise floor anyway with short 6min integrations, longer would be better I think (those on the 28th seem to be tracking OK, there is some elongation on the 30th). You'll find it much easier to use longer subs with the guider operational.

ChrisH

[wimvb]

Here are two video clips I created using the Blink tool in PI, that show the individual, calibrated but unregistered (unaligned) subframes.

Framerate = 5 fps. You can clearly see the advantage of proper balancing.

28 oct, poorly balanced

Blink28.avi

30 oct, properly balanced

Blink30.avi

 

BTW, there are also some nice cosmic rays and a satelite trail in the frames.

[wimvb]

7 hours ago, ChrisLX200 said:

Yes it could be fixed pattern noise moving across the frame (if so it should follow the drift bottom-left to upper-right). Why it's completely absent on the image from the 30th is a mystery, but I suspect focus is a factor in addition to any seeing artifacts in causing softness. In both cases you are looking down into the noise floor anyway with short 6min integrations, longer would be better I think (those on the 28th seem to be tracking OK, there is some elongation on the 30th). You'll find it much easier to use longer subs with the guider operational.

ChrisH

Thanks for your replies, Chris and Mark.

At 6 minutes exposure time, there is no visible fixed pattern noise in my subs. Light pollution takes care of keeping the histogram central peak well seperated from the lower values, and thermal and other noise take care of the rest. Besides, any fixed pattern should be equal in both sets, since fixed pattern noise is created by electronics, not by optics or external conditions (other than temperature, which was pretty much the same at both nights). Focus may have been an issue, although I use a Bahtinov mask to the best of my ability. Among my camera's many quirks is the unexplicable behaviour of live view. I focus by taking 3 - 6 second exposures, and adjust focus. I finish by tightening the focus lock screw, after which I check focus again, just to make sure that there is no shift. I don't always check focus during a shoot, but at those times I have, there was no visible focus shift. Since I don't have my camera attached to a computer, I can't use FWHM measurements on stars.

The guider will definitely make long exposures possible, but I'm already quite pleased with the performance of my mount as it is. My guess is, if I can get 6 minutes unguided exposures without much trailing, guiding shouldn't be that hard to get working. (Famous last words??)

Latest update on the guiding setup: it's in the courier's van. Hopefully he/she won't get stuck in the snow.

Cheers,

[sharkmelley]

Looking at your AVIs it is clear that the coloured streaking in the first image is caused by the drift of hot/warm pixels (i.e. thermal fixed pattern noise).  The effect is often referred to as "walking noise".

Mark

[wimvb]

Forgot about THAT fixed pattern. Of course, since hot pixels are fixed, they introduce a pattern. I was more thinking of the pattern that is associated with read noise, ie horizontal or vertical lines that are visible on bias frames and underexposed light frames.

In the mean time, my new guiding setup arrived. DHL managed to knock on my door during the few minutes I was out on an errand. Luckily the driver came back later this afternoon.