Go Back   IceInSpace > Equipment > Astrophotography and Imaging Equipment and Discussions

Reply
 
Thread Tools Rate Thread
  #21  
Old 26-07-2016, 05:54 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Yep, you can use shorter subs at higher gains, but only because the read noise decreases at high gain. if you calculate everything in electrons, the signal does not change with gain, the shot noise in the signal does not change with gain, but the read noise reduces as the gain increases. As a result, the subs can be shorter at high gain.

You will be amazed (at the low noise) and bemused (what the heck should I set it to ??) in equal parts when you try it out. It is not a whole lot like ordinary chips, but it works beautifully.

Last edited by Shiraz; 26-07-2016 at 06:27 PM.
Reply With Quote
  #22  
Old 26-07-2016, 06:22 PM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Sounds good to me
Reply With Quote
  #23  
Old 26-07-2016, 08:24 PM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
I have just finished some test subs of M8 taken at various Gain settings and durations. As Ray has indicated you can sort of take your pick as far as settings are concerned. I was playing around with Dunk's '3X' rule and it worked out pretty well. First I did a 60" sub at Gain 139 as a baseline, then a 120" sub at Gain 100, then a 180" sub at Gain 50. I was watching the histogram in SGP as they downloaded and was able to see that the best of the three seemed to be the Gain 100, 120" sub which exhibited the best fine detail for the exposure length.
I brought them into Photoshop CC and did a couple of things to them to bring out the fine detail: an auto tone, shadow/ highlight adjustment, nothing extreme, and I did exactly the same thing to each one.

I have attached small versions below for your amusement. From left to right, they are: the Gain 50, 180" version, the Gain 100, 120" version, and the Gain 139 60 sec version.
Attached Thumbnails
Click for full-size image (Camera-Test-of-Gain-50_180sec_1x1_R_frame1-small.jpg)
174.0 KB92 views
Click for full-size image (Camera-Test-of-Gain-100_120sec_1x1_R_frame1-small.jpg)
183.7 KB105 views
Click for full-size image (Camera-Test-of-Gain-139_60sec_1x1_R_frame1-2-smalll.jpg)
176.5 KB89 views
Reply With Quote
  #24  
Old 26-07-2016, 08:30 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
Still seriously considering buying one of these cameras First though I need money and a digital focuser
Reply With Quote
  #25  
Old 26-07-2016, 08:52 PM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Glen, the gain on the camera follows a logarithmic scale, so going from gain 139 to 100 is 3.9dB gain ~1.57x ... so a comparable exposure is 1.57x 60s. Exposing for 120s is collecting more signal, which should explain why it looks brighter. Nice shots for short exposures

Obviously, like Ray says, you only really need to exposure your desired feature over the read noise. Pretty exciting stuff with this camera having such low read noise.
Reply With Quote
  #26  
Old 26-07-2016, 09:23 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by glend View Post
I have just finished some test subs of M8 taken at various Gain settings and durations. As Ray has indicated you can sort of take your pick as far as settings are concerned. I was playing around with Dunk's '3X' rule and it worked out pretty well. First I did a 60" sub at Gain 139 as a baseline, then a 120" sub at Gain 100, then a 180" sub at Gain 50. I was watching the histogram in SGP as they downloaded and was able to see that the best of the three seemed to be the Gain 100, 120" sub which exhibited the best fine detail for the exposure length.
I brought them into Photoshop CC and did a couple of things to them to bring out the fine detail: an auto tone, shadow/ highlight adjustment, nothing extreme, and I did exactly the same thing to each one.

I have attached small versions below for your amusement. From left to right, they are: the Gain 50, 180" version, the Gain 100, 120" version, and the Gain 139 60 sec version.
Thanks Glen. the problem with trying to evaluate the performance by looking at individual subs is that you miss out on the effects of stacking. For example, for a total exposure of 3 minutes, you will have just the single 180"@50 sub, but you will get three 60@139 subs in the same period. If you want to really compare apples with apples, you would need to stack the three 60@139 subs and compare the result with the single 180@50 sub. And then you would need to somehow stretch the two images so that the background noise looked the same in each - could be done, but not easy.

edit: just looked again - the subs are all pretty heavily black clipped - did you deliberately do that in Photoshop, or perhaps you need a larger offset?
Reply With Quote
  #27  
Old 27-07-2016, 07:06 AM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
Sorry, those shots were simply put there to demonstrate the relative gain performance as I see it in single sub captures; my processing inadequacies will always be there. As far as the optimal sub lengths and Gain is concerned, for me, it hinges more on the "economies of time" and the complexities and computer memory and storage needs of stacking say 1000 subs. I would rather shoot fewer subs at 100 Gain for a longer period than run 1000 x Gain 50 subs at 35" and have to deal with the file mgt , editing/selection, and processing. As Ray's curve shows there really is not alot of difference around the top of the curve, so why not do it as efficently as possible?
As I have said before, I don't enjoy the processing side of imaging, I do it to get an image that I enjoy and don't need to get the last available photon coming from my target. That is why "Optimal" will be different for different people, and as Ray has pointed out, the flexibility of this camera provides for a variety of approaches. My approach may always be seen as odd to some but it delivers what I want from imaging.

Last edited by glend; 27-07-2016 at 07:44 AM.
Reply With Quote
  #28  
Old 27-07-2016, 07:49 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Yes, the table of "optimum" settings was only ever intended as a starting point for setting up a camera that has no obvious "best" way to use it and that is so different from other cameras that there can be no guidance from the old ways of doing things. Looks to me like your results show that gain 50 180" works fine with your system and sky. You might be able to scratch out a bit more dynamic range with shorter subs, but you will still have plenty for producing beautiful images. The cost of optimising things is the extra data and processing load - after processing a few thousand subs, I agree that it is a significant consideration.

Last edited by Shiraz; 27-07-2016 at 09:03 AM.
Reply With Quote
  #29  
Old 27-07-2016, 09:57 AM
ZeroID's Avatar
ZeroID (Brent)
Lost in Space ....

ZeroID is offline
 
Join Date: May 2010
Location: Auckland, NZ
Posts: 4,949
Bugga ,... c'mon refund. I gotta get me one of them's cameras ...

( I am awaiting an old super fund reimbursement from the financial crash back in the 90's, about $7,000 recovered from lost investments. Cost me big time in those days, over $250,000. This will bring recovery up to $23,000, better than nothing I suppose )
Reply With Quote
  #30  
Old 27-07-2016, 10:53 PM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Ray, would you mind refreshing my memory wrt TargetADU that you did before for CCDs using some example figures from your 1600?
Reply With Quote
  #31  
Old 28-07-2016, 09:23 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by Camelopardalis View Post
Ray, would you mind refreshing my memory wrt TargetADU that you did before for CCDs using some example figures from your 1600?
Had been reticent to post anything on this approach for the 1600 because
- it doesn't have a single "gain"
- it also has a hidden gain that is not obvious from the ZWO data
However, it turns out to be not so bad after all, so thanks for the prompting.

The basic formula (slightly rewritten) is:
TargetADU = Bias + 10*RN*RN/gain
where TargetADU is the sky background value you should try to get in your subs if you want the optimum balance between dynamic range and signal to noise ratio.

From the ZWO data, the ASI 1600 has variable gain (actually inverse gain) from 5 electrons/ADU to about 0.5 electrons/ADU, depending on the nominal gain setting (somewhere from 0-300). Then the software (or hardware) changes from 12 bit into 16 bit data by filling the bottom 4 bits with zeros. This is in effect an extra gain term of 16x and the modified inverse gain actually varies from about 0.3 to 0.03 ( ie it looks much more like other cameras).

Plugging in these values and using the published read values, yields the results in the attached table, where the "sky ADU above bias" column shows the calculated values of the second term in the above equation. To use it, read off the "sky ADU above bias" for the gain setting that you are using, add on the measured bias ADU for the offset/gain you are using (measure it in a bias frame - it will typically be a few hundred ADU) and the result is the TargetADU value for the background sky in a well exposed sub. Measure the sky background in a dark part of a sub and if you measure a higher value than the TargetADU, you could reduce your sub length - or if the measured value is lower, consider increasing the sub length (although you might like to deliberately choose underexposed subs for RGB to keep some colour in the brighter stars).

in summary
for your gain setting > read the "sky ADU above bias" from the table > add the measured Bias ADU to get the TargetADU > compare TargetADU with the background sky ADU in a sub > adjust the sub length if needed

For example, I use gain 100, so my Sky ADU above bias from the table is 380. Adding on my measured bias of 310 ADU gives a TargetADU of 690. I measured the average background sky brightness in one of my subs to be 758 ADU in a dark sky region with no hot pixels (image below), so I could have used slightly shorter subs than 1 minute - but 60 seconds is close enough and keeps the number of subs down a bit.

And if you don't want to bothered with all that, just adding 400 to your measured bias ADU will get you fairly close to the TargetADU for normal gain settings.

Regards Ray
Attached Thumbnails
Click for full-size image (1600sample.png)
73.1 KB81 views
Click for full-size image (1600table.jpg)
32.4 KB77 views

Last edited by Shiraz; 29-07-2016 at 09:22 AM.
Reply With Quote
  #32  
Old 28-07-2016, 10:54 AM
SimmoW's Avatar
SimmoW (SIMON)
Farting Nebulae

SimmoW is offline
 
Join Date: Jan 2014
Location: Tamleugh, Victoria, Australia
Posts: 1,383
Thanks for posting the ADU info Ray, much appreciated! I gotta look up how to see such stats via Pixinsight.

I assume this technique can be applied using NB filters too? Just longer times..
Reply With Quote
  #33  
Old 28-07-2016, 11:19 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by SimmoW View Post
Thanks for posting the ADU info Ray, much appreciated! I gotta look up how to see such stats via Pixinsight.

I assume this technique can be applied using NB filters too? Just longer times..
select a small preview without hot pixels or stars (use STF to view the image - don't stretch it). Use Statistics with the selected preview and the data type set to 16 bit. If you have chosen a good region, the mean and median should be fairly similar - use mean. (by the way, take no notice of the actual values in the attached example - this was 4 second sub)

the technique can be applied to NB, but as you say, exposures will be a lot longer.
Attached Thumbnails
Click for full-size image (1600pi.jpg)
165.2 KB63 views
Reply With Quote
  #34  
Old 28-07-2016, 11:29 AM
SimmoW's Avatar
SimmoW (SIMON)
Farting Nebulae

SimmoW is offline
 
Join Date: Jan 2014
Location: Tamleugh, Victoria, Australia
Posts: 1,383
Brill Ray, thank you. P.S. I forwarded your latest image to the ZWO FB page, they loved it
Reply With Quote
  #35  
Old 28-07-2016, 12:35 PM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
Quote:
Originally Posted by SimmoW View Post
Brill Ray, thank you. P.S. I forwarded your latest image to the ZWO FB page, they loved it
Re Ray's image of the NGC 6727 area, the camera and its performance is just one aspect of the final image. You could give it to 99 other people and they would probably all come back with something different, most of them far worse than Ray's wonderful image. There are years of experience, theory background, and the tools and systems under Ray's control that come together to make that image. The camera is great but Ray makes the image.
Reply With Quote
  #36  
Old 28-07-2016, 12:45 PM
multiweb's Avatar
multiweb (Marc)
ze frogginator

multiweb is offline
 
Join Date: Oct 2007
Location: Sydney
Posts: 22,060
These cameras are very interesting. Might put one on my Santa's list soon enough when I retire the old QHY8 going on 10 years now. Got to manoeuver around SWMBO. How's that called again? Compensation?
Reply With Quote
  #37  
Old 28-07-2016, 03:51 PM
SimmoW's Avatar
SimmoW (SIMON)
Farting Nebulae

SimmoW is offline
 
Join Date: Jan 2014
Location: Tamleugh, Victoria, Australia
Posts: 1,383
Sure Ray's work is excellent, it's good to see some uber good results coming out with it, to show its full potential and spread the word
Reply With Quote
  #38  
Old 07-08-2016, 11:00 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Great stuff, thanks Ray

Gave the 1600 a try a over AstroFest but I suffered some driver issues with my laptop...hopefully I'll work those out and get some decent subs soon
Reply With Quote
  #39  
Old 07-08-2016, 05:57 PM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
So...here goes

I managed to collect 1 set (32x 60s, unity gain) of luminance subs of our favourite watery gas ball when hooked up to my Esprit 100 and here's the result.

This is 1/4 of the original in x and y, even then I had to crop the centre to 640x640 to get under 200kb. If anyone's interested in any sections at 1:1 just let me know.

Edit: I should add that this is completely uncalibrated, I haven't taken any bias/darks/flats yet and against my better judgement I wasn't even dithering
Attached Thumbnails
Click for full-size image (Lagoon160032Lx60s.jpg)
183.8 KB102 views
Reply With Quote
  #40  
Old 07-08-2016, 06:31 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
hmmm, 32 minutes of uncalibrated 60 second subs taken with a 100mm refractor - that actually seems to be pretty darn good - well done Dunk.

what do you think - is the camera a keeper?
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 10:24 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
Testar
Advertisement