Another Offset Question

Discussion of using SharpCap for Deep Sky Imaging
cuivenion
Posts: 227
Joined: Wed Feb 15, 2017 7:37 pm

Re: Another Offset Question

#11

Post by cuivenion »

oopfan wrote: Tue Oct 16, 2018 6:29 pm Hi,

You said:
Thanks guys. At gain 0 my minimum value is 16. The histogram isn't very revealing though. It doesn't appear to be clipped.
Did you mean to say "At OFFSET 0 my minimum value is 16" ?

Brian
Sorry yes the offset was 0 as well as the gain being 0. For that bias frame.
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: Another Offset Question

#12

Post by oopfan »

I've never seen a Gain of 0. What camera is it? My minimum Gain is 100 which corresponds to the highest full well depth and highest read noise. Ordinarily I use Gain 100 only for star clusters where I want to capture very faint stars and bright stars without saturating the bright ones. For all other DSO's I use Gain 200 with LRGB filters, and Gain 400 with SHO filters. The gain I choose also depends on how bright the brightest star is in the field, a star that I can't crop out. It's a judgement call but for DSO-imaging I stick between Gain 100 and 400.

Brian
cuivenion
Posts: 227
Joined: Wed Feb 15, 2017 7:37 pm

Re: Another Offset Question

#13

Post by cuivenion »

It's the ZWO ASI224. I suppose it would be more accurate to say minimum gain, but the table on this page does give the range 0-400 for gain.

https://astronomy-imaging-camera.com/product/asi224mc
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: Another Offset Question

#14

Post by oopfan »

Ah, I see, so my Gain 400 is equivalent to your Gain 135. Looking at the graphs I would say that my Gain 200 is your Gain 60.

Do you generally keep within the range of Gain 0 and 135 for DSO imaging?

Getting back to Offsets, you need to increase Offset > 0 for Gain 0 according to the screenshot you posted. Increment Offset by 10 until you get a histogram that is not clipped.

Brian
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Re: Another Offset Question

#15

Post by turfpit »

I never use The Brain, I use my brain. Reason - I could be using another capture package such as AltairCapture or imaging with a Linux laptop. This means I have had to understand this stuff. I don't use SC's Polar Align for the same reason - I use ASPA in the Celestron mount.
M31-light-frame.jpg
M31-light-frame.jpg (68.54 KiB) Viewed 3130 times
This is a light frame for the M31 set. I have Min 2960, therefore not being black level clipped (as Brian says I err on the cautious side) and I am confident I am not losing faint data. Max 65504 leaves me not saturated. Ok on a 20 million pixel frame I get a few saturated pixels but the number is small. The actual numbers for offset/gain/exposure matter less that what is being shown in the FITS Liberator histogram & Image Statistics. My M31 was black level 150 (out of 400), gain 200 (out of 5000) and exposure 180s.

The nice fat histogram is brimming with data due to the 3m exposure.

I had too many wasted nights where I found out the next day I had bad settings. Over the last year or so, Brian and I have worked hard at being able to develop a systematic way to avoid those wasted nights. Now I spend a few minutes taking test frames and studying the FITS output.

Dave
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: Another Offset Question

#16

Post by oopfan »

Sorry I just saw your question just now...
No probs I'll try increasing the offset next time I image. One other question though, whenever I use the brain function it never increases the offset, I'm wondering why this is if increasing the offset (even at gain 0) is seen as good practice.
I've brought this topic up with Robin in the past. I think he wants us to get into the good habit of using the Smart Histogram to ensure that the shape of the histogram is ideal for what we are capturing (referring to Dave's M31 image and histogram.) I agree but I would be happier if he would set a safe value of 100 when we apply the settings. That way we can't accidentally make a mistake.

Brian
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Re: Another Offset Question

#17

Post by turfpit »

_straight-out-of-siril.jpg
_straight-out-of-siril.jpg (84.66 KiB) Viewed 3130 times
The picture shows the M31 stack from Siril loaded into GIMP. At this point the stack had been cropped, colour calibrated, green noise removed and a histogram stretch applied. I had very little to do within GIMP. The whole calibration, registering, stacking process + GIMP took around 30 minutes. The benefit of having good data (from the correct histogram ) means having less wrangling to do at the end.

Don't forget the image above was 20x180s unguided - demonstrating the importance of good polar alignment, mount alignment and focusing.

All the histogram understanding came from writing the manual - I had to learn it to churn it. When I wrote that, I had no background in astronomy or photography - Robin was very patient in answering the endless (dumb) questions.

Dave
cuivenion
Posts: 227
Joined: Wed Feb 15, 2017 7:37 pm

Re: Another Offset Question

#18

Post by cuivenion »

oopfan wrote: Tue Oct 16, 2018 7:05 pm Ah, I see, so my Gain 400 is equivalent to your Gain 135. Looking at the graphs I would say that my Gain 200 is your Gain 60.

Do you generally keep within the range of Gain 0 and 135 for DSO imaging?

Getting back to Offsets, you need to increase Offset > 0 for Gain 0 according to the screenshot you posted. Increment Offset by 10 until you get a histogram that is not clipped.

Brian
I generally do keep to unity and below when imaging. I will be increasing the offsets,I presume I'm OK as long is the minimum is in the 100 - 1000 range? When looking at the example histograms there isn't enough resolution to it to see if the left hand side is touching in Fits Liberator.
cuivenion
Posts: 227
Joined: Wed Feb 15, 2017 7:37 pm

Re: Another Offset Question

#19

Post by cuivenion »

turfpit wrote: Tue Oct 16, 2018 7:15 pm I never use The Brain, I use my brain. Reason - I could be using another capture package such as AltairCapture or imaging with a Linux laptop. This means I have had to understand this stuff. I don't use SC's Polar Align for the same reason - I use ASPA in the Celestron mount....
Unless I'm misunderstanding somewhere a bias frame will have a lot less signal than a light frame though. That being said I tested a light frame of gain and offset 0 and the minimum was only 256.

Sorry for all the questions. I have one more, does loading undebayered fits into Fits Liberator skew the results? Would they need to be debayered before loading them in.
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: Another Offset Question

#20

Post by oopfan »

Maybe Dave can stretch his FITS Liberator image of M31 to show the shape of the histogram.

I image with the minimum Offset for two reasons:
1. I want to mitigate clipping of bright stars on the right-hand side.
2. I hate capturing dark frames during precious sky-time.

I get maybe two or three good nights per month, so in the off-time I have the telescope outdoors capturing darks all night. I have a temperature-logger Arduino along side the camera that I later use to segregate the dark frames by temperature in buckets of 0.5 degrees Fahrenheit. Now, when I am fortunate to get clear skies I just concentrate on acquiring light frames. I keep the offset constant for the gain I use. Doing so allows me to use my dark frame library. If I got creative like Dave :D and adjusted Offset to make the histogram look right for every target then my dark library would be useless. That is why I stick with a fixed Offset for a given Gain -- so I don't have to capture darks during precious sky-time. Of course you don't have to use the minimum offset like me. You can choose to add 50 or 100 to whatever you discover using FITS Liberator. What I recommend doing however is sticking with it so that you can leverage a dark frame library if you choose to create one.

Brian
Post Reply