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

ASI 533MC Pro Color Dark Frames?

CMOS Imaging
  • Please log in to reply
22 replies to this topic

#1 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 05 June 2021 - 02:46 AM

Hallo 533MC Owners,

 

Do you see a need for any dark frames with this camera? Here is some background information: My master dark (30 frames) is very clean, there is no sign of amp glow. This camera has quite a few hot pixels, but they do not show up in my images (I am dithering). In order to see the impact of darks, I generated two stacks (one with and one without darks). After cropping and background extraction for both staks I generated a difference frame. The stretched difference frame does not show any structure resembling my photographed object, it's pretty much noise. Now to me it looks like there is no need to use any darks when dithering. Am I wrong and there is a benefit of using darks or do you draw the same conclusion? Thanks a lot in advance - looking forward to learn something new.

 

Best and CS - Oliver


Edited by Berlinsky, 05 June 2021 - 02:48 AM.


#2 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 05 June 2021 - 05:44 AM

Hallo 533MC Owners,

 

Do you see a need for any dark frames with this camera? Here is some background information: My master dark (30 frames) is very clean, there is no sign of amp glow. This camera has quite a few hot pixels, but they do not show up in my images (I am dithering). In order to see the impact of darks, I generated two stacks (one with and one without darks). After cropping and background extraction for both staks I generated a difference frame. The stretched difference frame does not show any structure resembling my photographed object, it's pretty much noise. Now to me it looks like there is no need to use any darks when dithering. Am I wrong and there is a benefit of using darks or do you draw the same conclusion? Thanks a lot in advance - looking forward to learn something new.

 

Best and CS - Oliver

There's really no amp glow with this camera. So if your telescope's light gathering capacity can overpower hot pixels, then you don't need darks.



#3 AhBok

AhBok

    Gemini

  • *****
  • Posts: 3,127
  • Joined: 02 Dec 2010
  • Loc: Lakeland, TN

Posted 05 June 2021 - 09:02 PM

I’ve tried it a few times with my 533 relying on just dithering and cooling. Didn’t work very well for me. The background was pretty noisy.
  • elmiko likes this

#4 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 06 June 2021 - 01:21 AM

I’ve tried it a few times with my 533 relying on just dithering and cooling. Didn’t work very well for me. The background was pretty noisy.

I saw your pictures with the 533MC on Astrobin - very nice. Did you use bias as well as darks or only darks?

 

CS - Oliver



#5 AhBok

AhBok

    Gemini

  • *****
  • Posts: 3,127
  • Joined: 02 Dec 2010
  • Loc: Lakeland, TN

Posted 06 June 2021 - 02:37 AM

Thanks Oliver. I calibrate with darks, flats and bias frames.
  • elmiko and sbharrat like this

#6 sbharrat

sbharrat

    Viking 1

  • -----
  • Posts: 627
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 06 June 2021 - 07:11 AM

Hallo 533MC Owners,

 

Do you see a need for any dark frames with this camera? Here is some background information: My master dark (30 frames) is very clean, there is no sign of amp glow. This camera has quite a few hot pixels, but they do not show up in my images (I am dithering). In order to see the impact of darks, I generated two stacks (one with and one without darks). After cropping and background extraction for both staks I generated a difference frame. The stretched difference frame does not show any structure resembling my photographed object, it's pretty much noise. Now to me it looks like there is no need to use any darks when dithering. Am I wrong and there is a benefit of using darks or do you draw the same conclusion? Thanks a lot in advance - looking forward to learn something new.

 

Best and CS - Oliver

As one who has one of these on the way, might I ask why one would *not* use darks? For this particular experiment, it doesn't appear to help but it is presumably possible that it does in other (un-run) experiments (different targets, etc). So I am wondering if there is something "hard" to using darks here. I assumed (hoped) that I since I would be pretty much using the same gain and set temperature, I would only have to take the darks and create a dark master ONCE and then reuse it (ok, maybe redo it every 6 months of so). Have you found this to not be the case? 



#7 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 06 June 2021 - 10:01 AM

As one who has one of these on the way, might I ask why one would *not* use darks? For this particular experiment, it doesn't appear to help but it is presumably possible that it does in other (un-run) experiments (different targets, etc). So I am wondering if there is something "hard" to using darks here. I assumed (hoped) that I since I would be pretty much using the same gain and set temperature, I would only have to take the darks and create a dark master ONCE and then reuse it (ok, maybe redo it every 6 months of so). Have you found this to not be the case? 

I suppose it won't hurt using darks, but it does not seem to help as long as you are dithering; at least I could not see a difference. Also, according to the manual the dark current is 0.001e/(pix*sec) at -5°C. Therefore, if you are taking a 5min minute dark exposure at -5°C, this yields to less than 1 ADU at unity again 100.

 

CS - Oliver



#8 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 06 June 2021 - 10:21 AM

I suppose it won't hurt using darks, but it does not seem to help as long as you are dithering; at least I could not see a difference. Also, according to the manual the dark current is 0.001e/(pix*sec) at -5°C. Therefore, if you are taking a 5min minute dark exposure at -5°C, this yields to less than 1 ADU at unity again 100.

 

CS - Oliver

Could you explain the formula please?



#9 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 06 June 2021 - 10:33 AM

Could you explain the formula please?

0.001 e/(pix*sec) = 0.3 e / pix for a 5 min frame (300s), for unity gain 1 ADU = 1 e, thus < 1 ADU per pixel dark current contribution

 

CS - Oliver



#10 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 06 June 2021 - 10:51 AM

0.001 e/(pix*sec) = 0.3 e / pix for a 5 min frame (300s), for unity gain 1 ADU = 1 e, thus < 1 ADU per pixel dark current contribution

 

CS - Oliver

I am sorry, but that doesn't compute.

0.001e /(pix*sec) = 0.001e*300s /pix = 0.3 e*s / pix. It doesn't make sense to me.

What is your notation? What is "pix", what is "sec"? What are their units? I assume that "e" stands for electron...

Do you have an original reference to where you got this formula by any chance?

Thank you.



#11 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 06 June 2021 - 11:05 AM

I am sorry, but that doesn't compute.

0.001e /(pix*sec) = 0.001e*300s /pix = 0.3 e*s / pix. It doesn't make sense to me.

What is your notation? What is "pix", what is "sec"? What are their units? I assume that "e" stands for electron...

Do you have an original reference to where you got this formula by any chance?

Thank you.

Just pick the factor based on the temperature you are using, e.g. 0.001 e/(pix*sec), see 533 MC manual. In this example the dark current contribution per pixel is 0.001 electrons per second. After 5 min (300sec) this results into 0.3 electrons per pixel, because the photo site is integrating over a duration of 5 min. At unity gain 1 ADU corresponds to 1 electron. Finally, the dark current contribution per pixel for a 5 min frame is less than 1 ADU.

 

CS - Oliver



#12 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 07 June 2021 - 02:19 AM

Just pick the factor based on the temperature you are using, e.g. 0.001 e/(pix*sec), see 533 MC manual. In this example the dark current contribution per pixel is 0.001 electrons per second. After 5 min (300sec) this results into 0.3 electrons per pixel, because the photo site is integrating over a duration of 5 min. At unity gain 1 ADU corresponds to 1 electron. Finally, the dark current contribution per pixel for a 5 min frame is less than 1 ADU.

 

CS - Oliver

I see what you mean now. Thanks.

BTW, if you are capturing very faint objects, which DSOs typically are, a lot of your hot pixels will be comparable or stronger in signal. So dithering will only help so much. I just recorded my first nebula, Iris nebula, and I took 180x1min subs at 200 gain. Even with the darks and dithering, there were hot pixels present in a live stack view of Sharpcap, little islands of hot pixels in the shape of dithering pattern. Of course, DSS will remove those when stacking raw frames. 



#13 philinbris

philinbris

    Viking 1

  • *****
  • Posts: 512
  • Joined: 10 Jan 2020
  • Loc: Brisbane, Australia

Posted 09 June 2021 - 03:03 AM

Hi all.

This spreadsheet may be of interest to some https://www.cloudyni...the-zwo-asi533/

I have read that instead of Bias frames, Flat Darks are better for the 533 due to the extremely low noise. That's what I am using.

Also, as an aside,  I think the well known Dr Robin Glover (SharpCap dev) talks about exposure in this U Tube https://www.youtube....h?v=3RH93UvP358. I think he indicates long exposures (ie 180 sec mentioned above) is not really required if you have the HD space at least. This seems to go with the spreadsheet that suggests 45 secs. I have some reasonable shots of M104 using 30 and 45 secs.

Cheers



#14 kraegar

kraegar

    Apollo

  • *****
  • Posts: 1,450
  • Joined: 17 Apr 2012

Posted 09 June 2021 - 09:40 AM

With my 533, and now my 2600, I used Darks + Flat darks.  The subtraction is very low, but generating a master dark is quick, and it doesn't add much to the calibration to do it.  I generated a full dark library on a cloudy night, and then stacked them all the next night. Now I have a library at -5c for a range of exposure times for a gain 100. I plan to do the same for a gain of 0.


  • CCD-Freak and jdupton like this

#15 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 09 June 2021 - 10:10 AM

I plan to do the same for a gain of 0.

Hi Tony,

In what cases do you use gain 0? Since the dynamic range is almost the same as for unity gain, I could only think of longer exposure length for bright targets? Or are there other cases where it would make sense?

 

CS - Oliver



#16 kraegar

kraegar

    Apollo

  • *****
  • Posts: 1,450
  • Joined: 17 Apr 2012

Posted 09 June 2021 - 10:13 AM

I just switched to a RASA 8 from my WO Star 71.  At 15 second subs, my first image (bortle 6 skies) all of the brighter stars completely saturated.  Going under 15 seconds seems unreasonable, so I'm going to try a gain of 0 to try and recover the star color.  I tried every trick I know of to recover the color on the stars, they're just cooked.

 

get.jpg?insecure



#17 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 09 June 2021 - 10:18 AM

I just switched to a RASA 8 from my WO Star 71.  

I see, f/2 is insanely fast smile.gif



#18 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 10 June 2021 - 12:00 AM

Hi Tony,

In what cases do you use gain 0? Since the dynamic range is almost the same as for unity gain, I could only think of longer exposure length for bright targets? Or are there other cases where it would make sense?

 

CS - Oliver

The dynamic range at zero gain is three times larger than that at 100 gain - the full well at gain 0 is 50,000 and and the full well at unity gain of 100 is 16,384 (14-bit sensor.)


Edited by acrh2, 10 June 2021 - 12:01 AM.


#19 Berlinsky

Berlinsky

    Lift Off

  • -----
  • topic starter
  • Posts: 18
  • Joined: 18 Oct 2020

Posted 10 June 2021 - 12:45 AM

The dynamic range at zero gain is three times larger than that at 100 gain - the full well at gain 0 is 50,000 and and the full well at unity gain of 100 is 16,384 (14-bit sensor.)

I am afraid the dynamic range is not even 0.5 stops higher at gain 0, see here: https://astronomy-im...533mc-pro-color

 

CS - Oliver



#20 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 10 June 2021 - 02:24 AM

I am afraid the dynamic range is not even 0.5 stops higher at gain 0, see here: https://astronomy-im...533mc-pro-color

 

CS - Oliver

Are you familiar with the logarithmic scale? 0.5 on it is a factor of 3.



#21 bulrichl

bulrichl

    Ranger 4

  • *****
  • Posts: 360
  • Joined: 27 May 2018
  • Loc: La Palma (Canary Islands)

Posted 10 June 2021 - 06:21 AM

Are you familiar with the logarithmic scale? 0.5 on it is a factor of 3.

Are you?

 

What you are saying would be true if the scale was logarithmic to the base 10. However, Dynamic Range (in stops) is logarithmic to the base 2.

 

The plot says:

at gain   0: DR = 13.8 stops, corresponding to a dynamic range of 1 : 2^13.8 = 1 : 14263
at gain 100: DR = 13.5 stops, corresponding to a dynamic range of 1 : 2^13.5 = 1 : 11585

Thus by going from gain 100 to gain 0 the improvement of DR is 14263/11585 = 1.23 (approximately), not 3.

 

Bernd



#22 acrh2

acrh2

    Vostok 1

  • -----
  • Posts: 124
  • Joined: 16 Mar 2021

Posted 10 June 2021 - 07:40 AM

Are you?

 

What you are saying would be true if the scale was logarithmic to the base 10. However, Dynamic Range (in stops) is logarithmic to the base 2.

 

The plot says:

at gain   0: DR = 13.8 stops, corresponding to a dynamic range of 1 : 2^13.8 = 1 : 14263
at gain 100: DR = 13.5 stops, corresponding to a dynamic range of 1 : 2^13.5 = 1 : 11585

Thus by going from gain 100 to gain 0 the improvement of DR is 14263/11585 = 1.23 (approximately), not 3.

 

Bernd

Yeah, you are right. This dynamic range thing is to the power of 2. I'm curious, where did you get the numbers 13.5 and 13.8 - did you read them from the graph or is there a table somewhere? 



#23 bulrichl

bulrichl

    Ranger 4

  • *****
  • Posts: 360
  • Joined: 27 May 2018
  • Loc: La Palma (Canary Islands)

Posted 11 June 2021 - 06:14 AM

I'm curious, where did you get the numbers 13.5 and 13.8 - did you read them from the graph or is there a table somewhere? 

I am not aware of a table with measured values. I coarsely estimated the numbers from the graph.

 

Bernd




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





Also tagged with one or more of these keywords: CMOS, Imaging



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics