Smart Histogram always way too dark, no matter what

[Menno555]

@donstim
For me the drop off is at Gain 60 (of 600)
I do have the Logarithmic box checked. And the peak was so close to the left that I could't tell if it was free from the left side or not. But no red bar was visible, only a tiny amber part (like 5%) and the rest (95%) green. And this is with no blacks or flats.
Will follow Robin's advise and will post info and screenshots then.

@Dave
That box is unchecked.

[turfpit]

Menno

That box is unchecked.

Good - that is one off the list.

Can you post a screen snip of an unprocessed light frame loaded into FITS Liberator. For example:

M81-M82-single-frame.JPG (89.43 KiB) Viewed 1779 times

For this example it told me I had an offset of 592 at the left hand side (so faint data not being lost); max was 65504 (just on the edge of overexposed).

With the same frame and a severe stretch in Siril:

stretched-in-siril.JPG (63.35 KiB) Viewed 1779 times

Based on those 2 simple checks (I wasn't going to spend 150m capturing unless I had some idea what was in the frame) I considered the SharpCap settings good enough to capture (it was 99% moon), resulting in:
viewtopic.php?f=16&t=2625



Also, can you post a link to to an 'unprocessed, as captured' light frame.

Dave

[Menno555]

Hi Dave

Thanks for this.
I do not have Fits Liberator but I do have another FITS Viewer with graphs and I do know that with my own settings, the graph looks very similar like your screenshot. The values you mention though,
Tried to use SIRIL too a few weeks back but it's lay-out and way of use is very confusing, non-intuitive for me
For now I don't have screenshots or whatever: I throw away captures and data that I don't use.
But it's going to be clear tonight and will do a Smart Histogram Brain session and share data then.

Menno

[turfpit]

Menno

If you google fits Liberator, it is a free download. I have spent a lot of time understanding the single frame histogram to final result relationship. 5 - 10 minutes spent with test frames at the beginning of a session can save a lot of disappointment and wasted nights. I don’t know why I know this ... LOL.

If you can get something back before tonight’s session ......

Dave

[admin]

Robin

Will do that.
Btw, that first step. Till now I did choose the darkest spot near the object I wanted to capture. Seemed logical for me. Is that okay or must it be the darkest part of sky I can see, even when it is not even close to the object?

Thanks in advance,
Menno

Hi Menno,

It should be the darkest area in the field-of-view that you intend to image.

Cheers, Robin

[Menno555]

Hi Menno,

It should be the darkest area in the field-of-view that you intend to image.

Cheers, Robin

Ah, really in that field of view. Will do. Gonna try the Heron Galaxy again, that was still on my to-do list for here
Also discovered today that my secondary mirror was way too tight from the factory. Could collimate but startest always failed and stars deformed.
So loodsend up everything, pre-collimation with artificial star (which already was a huge improvement) and tonight collimating and star test. Doubt it has anything to do with the Smart Histogram Brain but you never know: maybe it's due to a combination of things. Kinda Murphy's Law

[Menno555]

Okay, my Smart Histogram Brain report
Did everything suggested, so also choosing a dark spot in the field of view of the object.
Made 10 screenshots, will describe them.

1) Brain screen with the settings I provided
2) Brain screen with the suggested settings
3) Main SC with suggested settings
4) Fits Liberator suggested settings
5) Photoshop start Curves suggested settings (10 captures stacked)
6) Photoshop start suggested settings image unprocessed
7) Photoshop start suggested settings image unprocessed after 8 Curve actions
8) Main SC with my settings
9) Fits Liberator my settings
10) Photoshop with my settings, 2 Curve actions (7 captures stacked)

Used no darks or flats. This would make it even darker.
With the suggested settings I couldn't use DeepSkyStacker, with my own settings I could.
So, this is what happens when I use the Brain.
This all in Bortle 8 with a Meade LX65 8" f/10 ACF OTA, Optec Lepus Standard 0.62x reducer, Zwo ASI385MC camera, Baader Neodymium (but filter doesn't matter, had the same with other filters and with no filter)

1)

1

1.jpg (65.39 KiB) Viewed 1747 times

2)

2

2.jpg (87.36 KiB) Viewed 1747 times

3)

3

3.jpg (274 KiB) Viewed 1747 times

4)

4

4.jpg (144.63 KiB) Viewed 1747 times

5)

5

5.jpg (39.19 KiB) Viewed 1747 times

6)

6

6.jpg (16 KiB) Viewed 1747 times

7)

7

7.jpg (934.57 KiB) Viewed 1747 times

[Menno555]

8)

8

8.jpg (341.04 KiB) Viewed 1747 times

9)

9

9.jpg (160.9 KiB) Viewed 1747 times

10)

10

10.jpg (558.6 KiB) Viewed 1747 times

[BlackWikkett]

Hi Menno555

I think I see the disconnect between the brain suggestion and the unexpected results you are achieving. Looks like you're completing the brain sky brightness measurement without issue and the brain is suggesting the correct results based on the measurement and your input. In the brain image you shared it has total stacking time of 2 hours and the brain suggests 411 images. Individual images and as in your example 10 images will be be very dark. Try setting the total exposure time shorter to 60 minuets or less. Keep in mind the brain is suggesting settings for total stacking time not individual frames.

Menno555-brain.jpg (100.31 KiB) Viewed 1738 times

If you make changes to the stacking time and amount of read noise in the brain settings you'll soon see brighter individual frames. Looking at the frame you share with your settings (10) there is a lot of noise (much more than 10%) as well as exposing all the hot pixels and amp glow.

I think you can get a brighter image and more signal in the histogram but you'll have to sacrifice noise / detail based on the sensor you're using and the Bortle 8 skies of your observing location.

Good luck
-Chris

[Menno555]

Thanks for looking into it, Chris.
But this was as example. I already did try the full thing: it was then set for 1 hour with 30 second exposures and 120 frames. That was also with a Gain of 61. I did that and there was almost nothing to work with: stacking was only possible by first enhancing the gamma of all the FITS in PIPP and save that again. Only then there was enough data to stack. And also the same Photoshop Curve, and so on. There was less noise but but also almost no object.
Maybe it does work for the Live Stacking in SharpCap but I can't use that, too little stars in the captures. So maybe this doesn't work for stacking in other programs where there is no stretching during stacking, while that does happen in SharpCap? ... if I understood that right.

About the noise: if I set that to 20% for example, that would result in brighter captures?

Menno