Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Gain Settings for ASI1600MM-Cool

  • Please log in to reply
67 replies to this topic

#1 BillD17

BillD17

    Ranger 4

  • -----
  • topic starter
  • Posts: 365
  • Joined: 25 Jan 2017
  • Loc: Incline Village, NV

Posted 15 March 2017 - 01:05 AM

What is the optimal gain setting for the ASI1600MM-Cool.  I have been setting the gain to 10 in SGP and taking 2min subs for L and 5min for RGB and NB.    The histogram has most of its weight in the left 1/4, but bright stars are saturating.  I get the impression I might get a better image if I bumped the gain to 20 or so.

 

When I focus, I push it to 500 so I can loop 5sec exposures, but that's clearly too noisy.

 

I also don't see a place to specify the offset for the camera in the SGP control panel.



#2 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 22046
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 15 March 2017 - 01:15 AM

What is the optimal gain setting for the ASI1600MM-Cool.  I have been setting the gain to 10 in SGP and taking 2min subs for L and 5min for RGB and NB.    The histogram has most of its weight in the left 1/4, but bright stars are saturating.  I get the impression I might get a better image if I bumped the gain to 20 or so.

 

When I focus, I push it to 500 so I can loop 5sec exposures, but that's clearly too noisy.

 

I also don't see a place to specify the offset for the camera in the SGP control panel.

You are likely using exposures that are too long. This camera is a very low read noise camera. You don't need exposures that are very long unless you are imaging at a dark site. IF you are imaging at a dark site, and by that I mean a green bortle zone or darker, then you might be able to use 2-minute L subs and 4-5 minute RGB subs. However, if you are near or in the suburbs or the city, even at Gain 0 you will likely find that you can't use L exposures much longer than 30-60 seconds and RGB exposures much longer than 60-120 seconds.

 

I live at a red/white zone border, and a 60-second Gain 0 L exposure is actually WAY overexposed (I get a background sky level of 300e- to 500e-, which is swamping the read noise by 85-150x (and you only really need to swamp it by 20x!))

 

As for the "optimal" setting. If you truly want to optimize, then you'll need to do some experimentation. You can take increasing exposures for each filter, over many nights, at several different gain settings, to determine, for a given telescope and set of filters, how long you can/should expose to get optimal results. What is truly optimal is going to depend on your seeing, your sky brightness, and your optical system (primarily, your f-ratio and aperture). These will differ for every imager. As a general rule of thumb, you can either aim to swamp the read noise by 20x, or swamp the read noise squared by 3x. So, if your read noise is 1.55e- at unity gain, you would either want to expose until your background sky in a calibrated sub was 31 ADU (20xRN rule) or 7-8 ADU (3xRN^2 rule). Some ASI1600 imagers these days are using something midway between those two, so 15-20 ADU @ Unity Gain, to find a balance of ideal read noise swamping vs. minimal clipping. 

 

That said, most ASI1600 imagers are using Gain 0 and/or Gain 76 (Offset 15) for LRGB, Gain 139 (unity) or Gain 200 (Offset 30-50) for NB, and are using exposures ranging from 30s to a few minutes for LRGB, and 90-300s or so for NB. 


Edited by Jon Rista, 15 March 2017 - 01:16 AM.

  • newman likes this

#3 Ketut

Ketut

    Ranger 4

  • -----
  • Posts: 306
  • Joined: 19 Jun 2014
  • Loc: Bali

Posted 15 March 2017 - 03:52 AM

 

What is the optimal gain setting for the ASI1600MM-Cool.  I have been setting the gain to 10 in SGP and taking 2min subs for L and 5min for RGB and NB.    The histogram has most of its weight in the left 1/4, but bright stars are saturating.  I get the impression I might get a better image if I bumped the gain to 20 or so.

 

When I focus, I push it to 500 so I can loop 5sec exposures, but that's clearly too noisy.

 

I also don't see a place to specify the offset for the camera in the SGP control panel.

You are likely using exposures that are too long. This camera is a very low read noise camera. You don't need exposures that are very long unless you are imaging at a dark site. IF you are imaging at a dark site, and by that I mean a green bortle zone or darker, then you might be able to use 2-minute L subs and 4-5 minute RGB subs. However, if you are near or in the suburbs or the city, even at Gain 0 you will likely find that you can't use L exposures much longer than 30-60 seconds and RGB exposures much longer than 60-120 seconds.

 

I live at a red/white zone border, and a 60-second Gain 0 L exposure is actually WAY overexposed (I get a background sky level of 300e- to 500e-, which is swamping the read noise by 85-150x (and you only really need to swamp it by 20x!))

 

As for the "optimal" setting. If you truly want to optimize, then you'll need to do some experimentation. You can take increasing exposures for each filter, over many nights, at several different gain settings, to determine, for a given telescope and set of filters, how long you can/should expose to get optimal results. What is truly optimal is going to depend on your seeing, your sky brightness, and your optical system (primarily, your f-ratio and aperture). These will differ for every imager. As a general rule of thumb, you can either aim to swamp the read noise by 20x, or swamp the read noise squared by 3x. So, if your read noise is 1.55e- at unity gain, you would either want to expose until your background sky in a calibrated sub was 31 ADU (20xRN rule) or 7-8 ADU (3xRN^2 rule). Some ASI1600 imagers these days are using something midway between those two, so 15-20 ADU @ Unity Gain, to find a balance of ideal read noise swamping vs. minimal clipping. 

 

That said, most ASI1600 imagers are using Gain 0 and/or Gain 76 (Offset 15) for LRGB, Gain 139 (unity) or Gain 200 (Offset 30-50) for NB, and are using exposures ranging from 30s to a few minutes for LRGB, and 90-300s or so for NB. 

 

Jon,

How to measure the read noise at unity gain? And, how to know the read noise swamped by 20x? Screenshots please...

One of member here is an idiot who lives in the middle of nowhere...please remember that laugh.gif

 

Ketut


Edited by Ketut, 15 March 2017 - 03:56 AM.


#4 FiremanDan

FiremanDan

    Aurora

  • *****
  • Posts: 4675
  • Joined: 11 Apr 2014
  • Loc: Virginia

Posted 15 March 2017 - 05:41 AM

Gain of 10? Mine goes to 11.
Why not just make 10 louder?
...... this one goes to 11.


Sorry, spinal tap joke.
  • JoeR, Jeff Smith, newman and 6 others like this

#5 GeneralT001

GeneralT001

    Vanguard

  • *****
  • Posts: 2390
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 15 March 2017 - 06:07 PM

These are some suggested numbers provided by other users. Not sure if you use SGP. If you are then just have a look at the Image Statistics window on the left (after you've taken a pic) to see what your subs Mean ADU number is.

 

Optimal Exposure:
 

Median ADU shown in SGP:

 

Gain 0 Offset 10:         400 ADU

Gain 75 Offset 12:       550 ADU

Gain 139 Offset 21:     850 ADU

Gain 200 Offset 50:    1690 ADU

Gain 300  Offset 50 :  2650 ADU

 

Better to be a little high than low.

 

For other Gains or offsets use the formula Jon indicated:

 

MinDN16 = (((ReadNoise * 20) / Gain) + BiasOffset) * 2^16/2^Bits

 

WHERE:

 

ReadNoise is the read noise in e-

Gain is the gain in e-/ADU

BiasOffset is the ADC offset in ADU

Bits is the bit depth of the camera

 

So, as an example, for the ASI1600 @ Gain 0 with 3.5e- read noise and 4.88e-/ADU at 12-bit:

 

MinDN16 = (((3.5 * 20) / 4.88) + 10) * 16 = ((70 / 4.88) + 10) * 16 = (14.4 + 10) * 16 = 25 * 16 = 400

You would need a minimum 400 DN 16-bit background sky @ Gain 0 with Offset 10.

 

 

You can get a reasonable approximation of the Read Noise and Gain from the ASI Graph:

 

1600-Gain-RN-DR-FW-vs-gain-716x1024.jpg


Edited by GeneralT001, 15 March 2017 - 06:10 PM.

  • newman, HxPI, Ketut and 1 other like this

#6 BillD17

BillD17

    Ranger 4

  • -----
  • topic starter
  • Posts: 365
  • Joined: 25 Jan 2017
  • Loc: Incline Village, NV

Posted 17 March 2017 - 07:33 PM

With a gain of 10 I’m getting a mean ADU count of about 900 with a 5min sub.  Does that mean I should increase the gain and reduce the exposure time for each sub?

 

Also, how do I set the offset in SGP?  The camera tab of the control panel has four entries for gain (at different bindings), but no entries for offset.

 

Finally, as a simple photographer, how do I relate gain to ISO.  I understand ISO, but from the curves above, it appears that increasing gain from 0 to 139 only really increases gain by a factor of 5x - from 5e-/ADU to 1e-/ADU.  I'm confused.



#7 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 5345
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 17 March 2017 - 09:22 PM

 

What is the optimal gain setting for the ASI1600MM-Cool.  I have been setting the gain to 10 in SGP and taking 2min subs for L and 5min for RGB and NB.    The histogram has most of its weight in the left 1/4, but bright stars are saturating.  I get the impression I might get a better image if I bumped the gain to 20 or so.

 

When I focus, I push it to 500 so I can loop 5sec exposures, but that's clearly too noisy.

 

I also don't see a place to specify the offset for the camera in the SGP control panel.

You are likely using exposures that are too long. This camera is a very low read noise camera. You don't need exposures that are very long unless you are imaging at a dark site. IF you are imaging at a dark site, and by that I mean a green bortle zone or darker, then you might be able to use 2-minute L subs and 4-5 minute RGB subs. However, if you are near or in the suburbs or the city, even at Gain 0 you will likely find that you can't use L exposures much longer than 30-60 seconds and RGB exposures much longer than 60-120 seconds.

 

I live at a red/white zone border, and a 60-second Gain 0 L exposure is actually WAY overexposed (I get a background sky level of 300e- to 500e-, which is swamping the read noise by 85-150x (and you only really need to swamp it by 20x!))

 

As for the "optimal" setting. If you truly want to optimize, then you'll need to do some experimentation. You can take increasing exposures for each filter, over many nights, at several different gain settings, to determine, for a given telescope and set of filters, how long you can/should expose to get optimal results. What is truly optimal is going to depend on your seeing, your sky brightness, and your optical system (primarily, your f-ratio and aperture). These will differ for every imager. As a general rule of thumb, you can either aim to swamp the read noise by 20x, or swamp the read noise squared by 3x. So, if your read noise is 1.55e- at unity gain, you would either want to expose until your background sky in a calibrated sub was 31 ADU (20xRN rule) or 7-8 ADU (3xRN^2 rule). Some ASI1600 imagers these days are using something midway between those two, so 15-20 ADU @ Unity Gain, to find a balance of ideal read noise swamping vs. minimal clipping. 

 

That said, most ASI1600 imagers are using Gain 0 and/or Gain 76 (Offset 15) for LRGB, Gain 139 (unity) or Gain 200 (Offset 30-50) for NB, and are using exposures ranging from 30s to a few minutes for LRGB, and 90-300s or so for NB. 

 

How are you evaluating that background sky level on a given sub? Is this something that the statistics process gives, or is there some formula you use for this? I am working through verifying some data now, from a new (to me) scope. Is it just the minimum ADU? confused1.gif


  • astronate likes this

#8 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 22046
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 18 March 2017 - 01:39 AM

Depends a bit on the data. For most NB subs I just measure the mean and subtract the bias offset. Object signal areas often don't differ much from the mean, so it's a good enough measure. For objects in regions of sky that have much brighter and darker regions, I will usually spot measure the mean from a half dozen or so areas from the darker regions of a sub, and gauge my background sky level that way.
  • rockstarbill likes this

#9 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 5345
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 18 March 2017 - 03:22 AM

Depends a bit on the data. For most NB subs I just measure the mean and subtract the bias offset. Object signal areas often don't differ much from the mean, so it's a good enough measure. For objects in regions of sky that have much brighter and darker regions, I will usually spot measure the mean from a half dozen or so areas from the darker regions of a sub, and gauge my background sky level that way.

Sorry if this is robotic, but HOW?



#10 Dunkstar

Dunkstar

    Messenger

  • -----
  • Posts: 489
  • Joined: 26 Mar 2012
  • Loc: Australia

Posted 18 March 2017 - 03:33 AM

Open the FITS file in something like FITS Liberator and hover the pointer around an area of background.



#11 dkeller_nc

dkeller_nc

    Surveyor 1

  • *****
  • Posts: 1524
  • Joined: 10 Jul 2016
  • Loc: Central NC

Posted 18 March 2017 - 09:19 AM

 

Depends a bit on the data. For most NB subs I just measure the mean and subtract the bias offset. Object signal areas often don't differ much from the mean, so it's a good enough measure. For objects in regions of sky that have much brighter and darker regions, I will usually spot measure the mean from a half dozen or so areas from the darker regions of a sub, and gauge my background sky level that way.

Sorry if this is robotic, but HOW?

 

In SGP, simply open the image, click on the "image statistics" button (the capital sigma button on the menu bar), and use your mouse to hover over the image.  The approximate ADU of the pixel under the mouse pointer will be displayed below the window.  It'll be displayed in 16-bit space, so you will have to convert to 12-bit if that's your preference.


  • rockstarbill likes this

#12 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 5345
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 18 March 2017 - 11:33 AM

Sweet, thanks!



#13 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 22046
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 18 March 2017 - 01:59 PM

 

Depends a bit on the data. For most NB subs I just measure the mean and subtract the bias offset. Object signal areas often don't differ much from the mean, so it's a good enough measure. For objects in regions of sky that have much brighter and darker regions, I will usually spot measure the mean from a half dozen or so areas from the darker regions of a sub, and gauge my background sky level that way.

Sorry if this is robotic, but HOW?

 

I usually use PixInsight. In a pinch, I'll just use SGP's mean/median values for each sub. 



#14 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 5345
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 18 March 2017 - 02:12 PM

 

 

Depends a bit on the data. For most NB subs I just measure the mean and subtract the bias offset. Object signal areas often don't differ much from the mean, so it's a good enough measure. For objects in regions of sky that have much brighter and darker regions, I will usually spot measure the mean from a half dozen or so areas from the darker regions of a sub, and gauge my background sky level that way.

Sorry if this is robotic, but HOW?

 

I usually use PixInsight. In a pinch, I'll just use SGP's mean/median values for each sub. 

 

 

In PI, when I look at an images readout data, I get the X/Y coordinate and a value for K. Are you looking for a specific K value range?



#15 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 5345
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 March 2017 - 01:46 AM

Okay so I tested this on a 5 sec sub, and a 30 second sub, and I am more confused now than I was to begin with. In the 5 second sub, I am seeing the value for background sky as 300-ish when I mouse around in PI, and on the 30 second image it about 1000-ish. Based on what I am reading here, I have no clue what to conclude. :lol:



#16 ChrisWhite

ChrisWhite

    Skylab

  • *****
  • Posts: 4000
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted 19 March 2017 - 06:42 AM

Hi Bill,

 

To keep it simply you can simply look at the mean/median value for the entire image to get a good idea of what your BG sky level.  Mean and Median will probably be pretty close to one another for DSO.  The guide that General posted to above is a good place to start. 

 

I've even been experimenting with shorter than the 20x read noise exposures that is the general consensus with respect to AP.  For this camera I have been imaging at about 10x read noise and getting (what I would consider) great results with LESS clipping on stars for NB.  YMMV as I am imaging with an 8" aperture, so it takes VERY little time to clip stars.  For LRGB I'm closer to the 20x read noise.  Right now I'm using this camera at Gain 0 and doing 90 sec RGB and 30 sec Lum.  I am still clipping stars, but think this is a good balance. 

 

You don't want to use the minimum and maximum ADU values of an image as these could be representing malfunctioning pixels and not give you an accurate representation of exposure. 

 

To address your comment: "Okay so I tested this on a 5 sec sub, and a 30 second sub, and I am more confused now than I was to begin with. In the 5 second sub, I am seeing the value for background sky as 300-ish when I mouse around in PI, and on the 30 second image it about 1000-ish. Based on what I am reading here, I have no clue what to conclude."

 

What is your gain setting?  If you use the outline that General posted above, and you are imaging at Gain 139 (unity), you could conclude that your 5 second exposures are underexposed (Target ADU would be 850) and your 30 second exposures are overexposed a little bit.  You could either adjust your exposure time to get it at the "sweet spot" where you just swamp read noise, or you could reduce your gain and lengthen your exposure to be able to take longer exposures at a setting with improved dynamic range.

 

EDIT:  Make sure you remember that this is a 12 bit camera, and that the ADU's we are discussing are equivalent to 16bit.  So, 300 ADU mean as displayed by SGP is really only 18.75 (in 12 bit). 


Edited by ChrisWhite, 19 March 2017 - 06:45 AM.


#17 rkayakr

rkayakr

    Apollo

  • *****
  • Posts: 1431
  • Joined: 27 Oct 2010
  • Loc: Northeast Ohio

Posted 19 March 2017 - 08:06 AM

ZWO posted a piece on gain settings for the ASI1600 and ASI071 on their web site:

https://astronomy-im...n-ascom-driver/



#18 dkeller_nc

dkeller_nc

    Surveyor 1

  • *****
  • Posts: 1524
  • Joined: 10 Jul 2016
  • Loc: Central NC

Posted 19 March 2017 - 08:47 AM

Bill - Here's another way to measure (and think) about this that Chris alludes to.  If you're imaging faint nebula, the target is always going to be pretty low on the ADU scale (doesn't apply to something bright, like M42, of course).  So you can make your decision on the other end of the scale - your stars.  I suspect you'll find that the centers of your bright stars are almost always going to be saturated (i.e., 65,000 in SGP, or K=0.999 in PI) at 20x of the read noise.  So it's an aesthetic decision on exactly how much you want to blow out the stars in the image but still get enough SNR on your primary target to avoid stacking hundreds and hundreds of frames.

 

That's not the only way to do it, of course - you can stack a bunch of short exposures for star color, and long exposures for the nebula, and make an HDR composition.  But most don't do that unless their primary target requires it (i.e., M42).


  • newman likes this

#19 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 5345
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 March 2017 - 09:09 AM

Hi Bill,

 

To keep it simply you can simply look at the mean/median value for the entire image to get a good idea of what your BG sky level.  Mean and Median will probably be pretty close to one another for DSO.  The guide that General posted to above is a good place to start. 

 

I've even been experimenting with shorter than the 20x read noise exposures that is the general consensus with respect to AP.  For this camera I have been imaging at about 10x read noise and getting (what I would consider) great results with LESS clipping on stars for NB.  YMMV as I am imaging with an 8" aperture, so it takes VERY little time to clip stars.  For LRGB I'm closer to the 20x read noise.  Right now I'm using this camera at Gain 0 and doing 90 sec RGB and 30 sec Lum.  I am still clipping stars, but think this is a good balance. 

 

You don't want to use the minimum and maximum ADU values of an image as these could be representing malfunctioning pixels and not give you an accurate representation of exposure. 

 

To address your comment: "Okay so I tested this on a 5 sec sub, and a 30 second sub, and I am more confused now than I was to begin with. In the 5 second sub, I am seeing the value for background sky as 300-ish when I mouse around in PI, and on the 30 second image it about 1000-ish. Based on what I am reading here, I have no clue what to conclude."

 

What is your gain setting?  If you use the outline that General posted above, and you are imaging at Gain 139 (unity), you could conclude that your 5 second exposures are underexposed (Target ADU would be 850) and your 30 second exposures are overexposed a little bit.  You could either adjust your exposure time to get it at the "sweet spot" where you just swamp read noise, or you could reduce your gain and lengthen your exposure to be able to take longer exposures at a setting with improved dynamic range.

 

EDIT:  Make sure you remember that this is a 12 bit camera, and that the ADU's we are discussing are equivalent to 16bit.  So, 300 ADU mean as displayed by SGP is really only 18.75 (in 12 bit). 

I am using gain 76, so I think the first adjustment I am going to make is to switch to Gain 0. Otherwise, if I wanted to image at ideal exposure lengths I would end up with a bazillion files, and this would only become worse with the addition of the focal reducer for the FSQ. 


  • sparkyht likes this

#20 HxPI

HxPI

    Apollo

  • -----
  • Posts: 1300
  • Joined: 05 Sep 2013
  • Loc: Virginia Beach, VA

Posted 01 June 2017 - 05:17 AM

These are some suggested numbers provided by other users. Not sure if you use SGP. If you are then just have a look at the Image Statistics window on the left (after you've taken a pic) to see what your subs Mean ADU number is.

 

Optimal Exposure:
 

Median ADU shown in SGP:

 

Gain 0 Offset 10:         400 ADU

Gain 75 Offset 12:       550 ADU

Gain 139 Offset 21:     850 ADU

Gain 200 Offset 50:    1690 ADU

Gain 300  Offset 50 :  2650 ADU

 

Better to be a little high than low.

 

For other Gains or offsets use the formula Jon indicated:

 

MinDN16 = (((ReadNoise * 20) / Gain) + BiasOffset) * 2^16/2^Bits

 

WHERE:

 

ReadNoise is the read noise in e-

Gain is the gain in e-/ADU

BiasOffset is the ADC offset in ADU

Bits is the bit depth of the camera

 

So, as an example, for the ASI1600 @ Gain 0 with 3.5e- read noise and 4.88e-/ADU at 12-bit:

 

MinDN16 = (((3.5 * 20) / 4.88) + 10) * 16 = ((70 / 4.88) + 10) * 16 = (14.4 + 10) * 16 = 25 * 16 = 400

You would need a minimum 400 DN 16-bit background sky @ Gain 0 with Offset 10.

 

 

You can get a reasonable approximation of the Read Noise and Gain from the ASI Graph:

 

attachicon.gif1600-Gain-RN-DR-FW-vs-gain-716x1024.jpg

 

This thread helps me a lot in understanding what the ideal gain and exposure settings should be for the ASI1600.  I'm getting median values of 850 adu for 0 gain at 5 min exposure and 2200 adu for 139 gain at 2 min exposure. It seems I've been way overexposing for my setup and location conditions! It may also explain why my image backgrounds look so "noisy" and can't seem to get rid of it. Looking forward to trying this new info out at my next opportunity. Thanks for sharing.

 

Ciao,

Mel


Edited by HxPI, 01 June 2017 - 10:01 AM.

  • FiremanDan likes this

#21 miwitte

miwitte

    Viking 1

  • *****
  • Posts: 936
  • Joined: 28 Dec 2016
  • Loc: Charlotte NC

Posted 01 June 2017 - 08:18 PM

So I posted something a day or two back about "optimal" median ADU in that chart. I assume that number is based on a very dark site. For us in red zones that number is not realistic and we need much higher median ADU due to light pollution. I mentioned what would be awesome if we could poll folks to see what ADU they are using for the 1600 and their SQM numbers. Then we could potentially have a chart that also include recommended ADU based on you light pollution numbers.
  • 42itous1 and Bob_Huerbsch like this

#22 dangarnett

dangarnett

    Explorer 1

  • -----
  • Posts: 58
  • Joined: 28 Jul 2016

Posted 02 June 2017 - 04:22 AM

What is the problem we try to solve when we raise the gain at all? Is it that we raise Gain to reduce exposure time so we have less of a chance of tracking/seeing errors?  

If we assume we have perfect tracking and perfect seeing, would Gain 0 (10 offset) be the most optimal?

 

Looking at the ADU values above for optimal exposures, would all of those exposures be equal to each other or would the 0 gain one be superior and have more dynamic range than the others. 



#23 HxPI

HxPI

    Apollo

  • -----
  • Posts: 1300
  • Joined: 05 Sep 2013
  • Loc: Virginia Beach, VA

Posted 02 June 2017 - 06:03 AM

So I posted something a day or two back about "optimal" median ADU in that chart. I assume that number is based on a very dark site. For us in red zones that number is not realistic and we need much higher median ADU due to light pollution. I mentioned what would be awesome if we could poll folks to see what ADU they are using for the 1600 and their SQM numbers. Then we could potentially have a chart that also include recommended ADU based on you light pollution numbers.

From my own practical point of view, I am getting higher median ADUs and the image backgrounds look horrible. That's even with camera cooling and after image calibration and integration! My image processing skills are sorely lacking which is also part of the problem with getting the background smoothed out! I wonder what I'm doing wrong?

 

I believe it is true what they say, it is always much better to go to a dark site. For those who just don't have that option, there is going to be a compromise in your images. I'm willing try try less exposure and hopefully the backgrounds will process better. Better data would make life easier for me, for sure! 

 

Ciao,

Mel


Edited by HxPI, 02 June 2017 - 06:10 AM.


#24 miwitte

miwitte

    Viking 1

  • *****
  • Posts: 936
  • Joined: 28 Dec 2016
  • Loc: Charlotte NC

Posted 02 June 2017 - 08:50 AM

The problem with that is number of exposures. In a red zone you need 5+ hours of exposure to equal 15 minutes in a dark zone. big difference in total sub count if your exposing for 5 sec at a higher gain than 60 sec at gain 0. For 5 hours of integration at 5 sec exposures that's 3600 frames to process vs 300 at 60 sec.

Also when I tried to image using the median ADU for gain 0 offset 10 I obtained very little detail in my DSO I had to go to 60 sec(way way over the recommended ADU) basically the light pollution is swamping the photons from the DSO. So the question is for a particular SQM what should the median ADU be to overcome light pollution? Obviously we lose dynamic range and clip some stars doing this but I don't see any other way in a red zone unless I am completely misunderstanding things.

#25 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 22046
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 02 June 2017 - 10:16 AM

What is the problem we try to solve when we raise the gain at all? Is it that we raise Gain to reduce exposure time so we have less of a chance of tracking/seeing errors?  

If we assume we have perfect tracking and perfect seeing, would Gain 0 (10 offset) be the most optimal?

 

Looking at the ADU values above for optimal exposures, would all of those exposures be equal to each other or would the 0 gain one be superior and have more dynamic range than the others. 

Increasing gain can cost you dynamic range. However, because at a higher gain you reduce read noise, you simply don't need long exposures. 

 

This camera uses a 12-bit ADC, rather than a 16-bit ADC. This means the quantization error is higher. With the average CCD, quantization error is around 0.1-0.3 ADU. With the ASI1600 at Gain 0, it/s 1.4 ADU! To average out this quantization error, you need more subs. Now, at a higher gain, the quantization error actually drops. At Gain 76 it is 0.6 ADU. At Gain 200 it is 0.14 ADU. So higher gains greatly improve the sampling of each electron worth of signal, giving you a more accurate result. 

 

The camera is still 12 bit though, and in practice stacking only a few subs at any gain results in posterization, sudden large jumps in levels, or a "gappy histogram." To smooth out the noise profile, you need to stack more subs. You can recover around 3 bits by stacking 80-120 subs. You can recover around 4 bits by stacking 250-300 subs. 

 

I recently just stacked 413x30s subs, Gain 76/15, with aggressive dithering every 4 subs, and the results were a very nice clean noise profile with ~16.4 bits of real-world precision:

 

ly2qB44.jpg

 

Closeup of the noise profile:

 

t4DI4BI.jpg

 

This was done under pretty poor skies, 18.1-18.3mag/sq" during the nights I imaged. I plate solved the field, and plotted objects as faint as 22mag/sq". The faintest in the entire field was over 21mag. The faintest within the crop here is 19.49mag, which is still quite faint considering how bright my skies were:

 

dqogvKT.jpg

 

Lower bit depth CMOS cameras work really well when you dither aggressively and stack a lot of frames. To stack a lot of frames, you want the lowest read noise you could possibly get. At Gain 0, the total read noise from 413 subs would be 71.13e-! At Gain 76, the total read noise is 40.6e-. If I had dark enough skies, I'd probably use unity gain, which would give a total read noise of 31.5e-. For the record...stacking 20 long exposure KAF-8300 frames would give you total read noise of 40.3e-. ;)


  • PrestonE, h2ologg, HxPI and 1 other like this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics