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

Shorten total integration time with new equipment

astrophotography equipment
  • Please log in to reply
133 replies to this topic

#76 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 09:45 AM

Just to add to how wrong that position is, here is some work done by this guy named Roland.

 

https://www.astrobin...6uf7/C/?nc=user

 

Diminishing returns.  Figure it out for your scope, focal length and f stop.  At F6 with higher read noise 8300 sensor and depending on your filters pass through you may need long exposures for target SNR.

 

On the 6200, you will hit diminishing returns much quicker.  Not sure what the point/debate is here.

 

Measure and do your own thing. COMPARATIVELY SPEAKING one would take more shorter exposures with the 6200 vs a KAF *.* but no one stops you from fighting diminishing returns if that's your thing.



#77 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 August 2020 - 12:03 PM

Frank I fully agree. The point was made here (incorrectly) that people should aim for the shortest exposures with their given gear and the counter to that claim was that is more optimal to expose as long as possible without clipping. This example shows the results of the same gear but with a stack vs single exposure with the same total integration time.

This wasn't intended to get people to shoot one hour exposures. There is a good discussion about this test specifically on Groups.io.

#78 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 12:16 PM

There are a few people doing really short exposures and stacking hundreds if not thousands and getting amazing results on a CMOS/6200 system. (half-lucky-imaging)

 

I do my exposure time based on diminishing returns. These new CMOS sensors are fast in that regard. It's been my experience, and experience from people I know that anything above diminishing returns with modern CMOS doesn't seem to have the same benefit as the cited example from a CCD.

 

i've said it before, it's probably time we let CCD and CMOS stand on there own rather than beat the drums of comparison for comparison sake.

 

I'll refigure my diminishing returns as i move to an f3.3 scope and optimize around that too. My new scpe has an image circle for 6200 so that's why i'm optimistically hanging around and discussing it. It's an eventuality. (similar tech to my 2600)



#79 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 August 2020 - 12:24 PM

Can you quantify what you mean by diminishing returns? On it's own that doesn't mean a whole lot nor really add to the discussion.

#80 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,124
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 19 August 2020 - 12:25 PM

I was not convinced by Bill's example because of several facts. First of all I don't know the optical system being used. Second of all, that's an old noisy CCD camera. Third, it appeared to me that he was shooting narrow band images. Fourth, he was apparently shooting in a location with good seeing but lots of skynoise. When I put them all together, I wonder if the example actually would come out the same if I used a modern CMOS low noise camera with the ability to adjust gain (and offset).

 

I guess I'll just have to try it! All of those differences made me wonder if the hour long exposure was actually long enough. And,  had it gone on longer would the skynoise etc made the situation worse.  Now if the skies will just clear!

 

Rgrds-Ross



#81 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 12:31 PM

Here is one guy that started a trend: 

 

https://astronomy-im...tostakkert.html

 

He did some "lucky imaging" of galaxies.  There is one of his whirlpool that is linked elsewhere (on an asi1600) where he stacked 2000 1 second subs. Even the one linked to his blog post there is amazing for what it is. (edit, posted the image i found on an astronomy discord server)

 

It looks so fluffy like you can almost squeeze it like a toy... there are a few others but my astrobin search fu isn't working and some of them were probably wiped in the data loss.

 

(edit not sure if this link works, but hope so)

 

20160505_M51_2000x1s_AutoStakkert_ASI160

 

i'm tempted to try some similar stuff on my epsilon if UPS ever finds it.


Edited by sn2006gy, 19 August 2020 - 12:35 PM.


#82 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 August 2020 - 12:35 PM

I was not convinced by Bill's example because of several facts. First of all I don't know the optical system being used. Second of all, that's an old noisy CCD camera. Third, it appeared to me that he was shooting narrow band images. Fourth, he was apparently shooting in a location with good seeing but lots of skynoise. When I put them all together, I wonder if the example actually would come out the same if I used a modern CMOS low noise camera with the ability to adjust gain (and offset).

I guess I'll just have to try it! All of those differences made me wonder if the hour long exposure was actually long enough. And, had it gone on longer would the skynoise etc made the situation worse. Now if the skies will just clear!

Rgrds-Ross


That's not my data. It's data Roland from Astro-Physics shared. The one hour is just what was used for a test. If you read the description data on Astrobin he's showing that fine details were captured in the longer single exposure vs a stack of shorter exposures, using the exact same gear. Ignore what the sensor is. That's just going to muck up the real meat of the result. That is - exposing longer without excessive clipping results in more fine detail captured.

Now where that exposure point is, depends on the system. Lower noise cameras would expose for less time. However, you should expose them without time being the only factor.

If I took a stack of images at 3*RN^2 vs a single image at 20*RN^2, if the total exposure time was the same which would provide finer details?

#83 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 02:10 PM

That's not my data. It's data Roland from Astro-Physics shared. The one hour is just what was used for a test. If you read the description data on Astrobin he's showing that fine details were captured in the longer single exposure vs a stack of shorter exposures, using the exact same gear. Ignore what the sensor is. That's just going to muck up the real meat of the result. That is - exposing longer without excessive clipping results in more fine detail captured.

Now where that exposure point is, depends on the system. Lower noise cameras would expose for less time. However, you should expose them without time being the only factor.

If I took a stack of images at 3*RN^2 vs a single image at 20*RN^2, if the total exposure time was the same which would provide finer details?

 

Mr Rockstar, i only ask these questions and join these discussions because I am interested and I am learning. Just want to clear that up since it feels things have gotten adversarial in some weird way.

 

To answer this question though, it feels contrived to not include the sensor differences. (my honest answer)

 

Isn't 20*RN^2 Well beyond diminishing returns?  i mean, you've blown out noise no matter the sensor surely... (well blown out on a 6200)

 

Couldn't you do shorter exposures, dither and integrate for same time and achieve same results? With less risk?  I mean, a 6um or larger pixel is fairly forgiving for tracking so 1 hour on these wouldn't be impossible, but it can be impractical for many.

 

Is this persons data available for others to try and integrate?  Maybe i'll shoot the same area with my color :)

 

I know read noise compounds when you stack more images, but the read noise on modern sensors is super low when cooled in great conditions - for almost any sensor.

 

Isn't QHY working on a 2x read for their QHY600 camera?  Further reducing read noise - for what intends to be a reason for more sampling vs longer samples as such an optimization would be useless for longer subs.



#84 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 August 2020 - 02:33 PM

I still don't get what you equate to "diminishing returns". In the case of an image exposed 20*RN^2 you have exposed 20 times the read noise of the camera squared. For broadband, this is fairly easy to do. For Narrowband that would be a challenge and not one someone is likely to attempt. In the case of my most recent usage of the ASI6200 camera, for narrowband I was at 3*RN^2 with 15 minute subs at gain 100 offset 50 on the Wizard Nebula using a F6.7 10" iDK. I considered that to be all I was willing to do in terms of exposure time in the environment I image in. Would I consider going longer to have diminishing returns? No not really. 

 

Now if I were imaging in broadband, I would not pay attention to the swamping of read noise much. I would determine the right exposure time based on star clipping and back it off to the point where clipping wasnt an issue. I am still not in a place where any returns are diminished per se, and given that I know that my background median ADU needs to be 704 for that 20*RN^2 swamping, it is more likely I would far exceed that then not. If I am not clipping stars and I am swamping read noise by a good margin, therein lies my optimal exposure length. My 16200 is not likely to differ much from the 6200 in terms of where those exposures will be. They wont be identical but the delta is not going to be some massive amount. 

 

Where read noise becomes a real factor, is with narrowband. Outside of narrowband, modern cameras (which include CCD's, like the 16200 chip which was released in 2016) have such low noise that for broadband imaging even the KAI-11002 is incredibly viable.

 

Some people could target ADU specifically for their imaging. Once they hit 3*RN^2 or in the case of the ASI6200 at Gain 100 - 528 median ADU (28 ADU over bias), they could just stop there and stack a bunch of exposures up. In doing so it is possible some very faint detail is lost. It is also possible there is not any faint detail lost. It really depends on the target.



#85 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,124
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 19 August 2020 - 02:59 PM

I'm not personally interested in stacking thousands of images and that technique my well work for broadband with a small FOV camera and fast downloads. I don't really see that technique as generally practical. (Of course I don't have the right equipment and  may just be jealous.)

 

So, at my dark sky site the sky noise is .1e/second using this handy calculator for my F8 optical system and a reasonably narrow band filter.  My read noise (off the chart supplied by the vendor) is going to be 2e (QHY600). That means that to get 20x above the read noise^2 I have to expose for about until the sky noise is 80e. That will happen in 800 seconds, right?  If the read noise was 10, then I'd have to expose 5x longer or 4000 seconds to get the same result. Seems to me that 800 is a lot more practical than 4000. Seems to me that Roland's example always works but the specifics are tailored to that camera/chip which is fine, the principle is the same.

 

I remember buying a set of 2" square narrow band astrodon filters and discussing narrow band imaging with a CCD camera with about 10e of read noise that we both owned. He recommended 30 minute exposures. He said, exactly what Roland said. For the most detail in NB you need much longer exposures than you might think. 

 

Rgrds-Ross


  • rockstarbill likes this

#86 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 03:07 PM

I still don't get what you equate to "diminishing returns". In the case of an image exposed 20*RN^2 you have exposed 20 times the read noise of the camera squared. For broadband, this is fairly easy to do. For Narrowband that would be a challenge and not one someone is likely to attempt. In the case of my most recent usage of the ASI6200 camera, for narrowband I was at 3*RN^2 with 15 minute subs at gain 100 offset 50 on the Wizard Nebula using a F6.7 10" iDK. I considered that to be all I was willing to do in terms of exposure time in the environment I image in. Would I consider going longer to have diminishing returns? No not really. 

 

Now if I were imaging in broadband, I would not pay attention to the swamping of read noise much. I would determine the right exposure time based on star clipping and back it off to the point where clipping wasnt an issue. I am still not in a place where any returns are diminished per se, and given that I know that my background median ADU needs to be 704 for that 20*RN^2 swamping, it is more likely I would far exceed that then not. If I am not clipping stars and I am swamping read noise by a good margin, therein lies my optimal exposure length. My 16200 is not likely to differ much from the 6200 in terms of where those exposures will be. They wont be identical but the delta is not going to be some massive amount. 

 

Where read noise becomes a real factor, is with narrowband. Outside of narrowband, modern cameras (which include CCD's, like the 16200 chip which was released in 2016) have such low noise that for broadband imaging even the KAI-11002 is incredibly viable.

 

Some people could target ADU specifically for their imaging. Once they hit 3*RN^2 or in the case of the ASI6200 at Gain 100 - 528 median ADU (28 ADU over bias), they could just stop there and stack a bunch of exposures up. In doing so it is possible some very faint detail is lost. It is also possible there is not any faint detail lost. It really depends on the target.

My brain deals in integration time.

 

Perhaps I totally misunderstand SNR - but In integration time, diminishing returns can penalize you when you image above those diminishing returns as the SNR of your stack for fewer subs paying a higher penalty is lower than a stack of optimized subs at lower penalty is it not?

 

Lets say you normalize your SNR

 

Sub Length | SNR

1 Second  | 1.00

3 Seconds | 1.66

10 Seconds | 2.69

30 Seconds | 3.66

100 Seconds | 4.38

300 Seconds | 4.68

1000 Seconds | 4.79

3000 Seconds | 4.83

 

Why would I want to shoot over 300 seconds?  Why wouldn't I base my subs on diminishing returns especially in the context of integration time?

 

To yield the most improvement, i'd want more stacks of 300 second subs than fewer stacks of 3000 second subs.

 

If you know the signal rate of your skyglow, the signal rate of your target, the QE of your sensor, your f-ratio and you know how much each filter impacts the signal rate (e/px/s) can't you figure out the diminishing returns of your S/H/O  L/R/G/B or naked OSC or LPS filter and calculate what your optimal exposure time is? It varies on emission/filter/scope/camera/speed

 

when I think through these numbers and do my dumb google sheet math, a 6200 shines in more shorter subs... (or any of these clean modern BSE cmos)

 

I got most of my e/px/s calculations from sharpcap and its sky calculator where you can insert filters and see how it impacts the e/px/s and then figuring out a few targets to test it on... it's still best case scenario and napkin math, but i honestly feel like in integration time, diminishing returns on CMOS plays a critical roll in squeezing out value.

 

I do trade processing time/compute... but heck, my computer + camera costs less than most comparable cameras...

 

On CMOS swamping 20x the read noise SNR penalized by diminishing returns

on a CCD, you're still trending in higher SNR overcoming read noise.


Edited by sn2006gy, 19 August 2020 - 03:47 PM.


#87 ChrisWhite

ChrisWhite

    Fly Me to the Moon

  • *****
  • Posts: 6,142
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted 19 August 2020 - 03:13 PM

I'm not sure what is meant by diminishing returns either.  The signal from your DSO is rather constant.  Diminishing returns is a concept that when applied in this context would come to mean that additional seconds would yield less benefit for every additional second added to the exposure.  That's just not true.  Additional seconds acquire additional photons at the same rate regardless of how long the exposures are.  So, a 10 minute exposure will get you twice as many photons as a 5 minute exposure and half as many photons as a 20 minute exposure. 

 

The only problem you will run into when imaging for additional time is that you run the risk of clipping stars.  That's why in my earlier post I stated to expose as long as you can without clipping stars. 

 

In fact, I would say the opposite of this argument to be true (that longer exposure times has diminishing returns to shorter exposure times).  If you take ten exposures at 60 seconds, you must waste some of the time in each of those exposures to acquire enough signal to get above the read noise floor, whereas a single 600 second exposure wastes 1/10th of the amount of signal.  Again, circling back to my earlier point... the real limitation here with respect to how long you might want to go is determined by star clipping.

 

Trendy or not, thousands of 1 second (lucky-ish) imaging exposures are not the equivalent of the same amount of time spent taking long exposures.  I for one, am loathe to dealing with too many subs (and why I loathe the panasonic sensor that requires very short exposures to prevent star clipping and recover bit depth). 


  • rockstarbill likes this

#88 Peregrinatum

Peregrinatum

    Apollo

  • *****
  • Posts: 1,497
  • Joined: 27 Dec 2018
  • Loc: South Central Valley, Ca

Posted 19 August 2020 - 03:35 PM

from what I have learned, the diminishing returns relates to the equations that describe SNR... incremental time is always equal, but incremental SNR is not since it is described generally by a power equation... when an SNR increment starts to become less than the prior SNR increment you are no longer using your imaging time as efficiently... but other factors come into play like exposing long enough for faint signals, software overhead time, etc, I think diluting read noise is more of a floor for exposure time and not a ceiling... but thats just how I look at it


  • sn2006gy likes this

#89 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 03:38 PM

I'm not sure what is meant by diminishing returns either.  The signal from your DSO is rather constant.  Diminishing returns is a concept that when applied in this context would come to mean that additional seconds would yield less benefit for every additional second added to the exposure.  That's just not true.  Additional seconds acquire additional photons at the same rate regardless of how long the exposures are.  So, a 10 minute exposure will get you twice as many photons as a 5 minute exposure and half as many photons as a 20 minute exposure. 

 

The only problem you will run into when imaging for additional time is that you run the risk of clipping stars.  That's why in my earlier post I stated to expose as long as you can without clipping stars. 

 

In fact, I would say the opposite of this argument to be true (that longer exposure times has diminishing returns to shorter exposure times).  If you take ten exposures at 60 seconds, you must waste some of the time in each of those exposures to acquire enough signal to get above the read noise floor, whereas a single 600 second exposure wastes 1/10th of the amount of signal.  Again, circling back to my earlier point... the real limitation here with respect to how long you might want to go is determined by star clipping.

 

Trendy or not, thousands of 1 second (lucky-ish) imaging exposures are not the equivalent of the same amount of time spent taking long exposures.  I for one, am loathe to dealing with too many subs (and why I loathe the panasonic sensor that requires very short exposures to prevent star clipping and recover bit depth). 

 

OK... so Integration isn't just how i think - but SNR in integration... so that's why i think diminishing returns laugh.gif (just edited to update this)

 

I only cited the photo of the 1x2000 subs as an example. My other post was based entirely on more normal sub times.

 

It's my understanding that if you have a sensor with a read error rate, quantum efficiency rate and total well depth that doubling the exposure time does not double the SNR.

 

Just as doubling your total integration time doesn't double your SNR. Both I believe have diminishing returns...  but again.. maybe i'm wrong, i'm a computer nerd by day and a hobby astronerd by night.

 

Just as there is some possible overhead to shorter subs, there is a benefit to more samples. If i have a bad sub, it's a smaller fraction of my integration at risk. I can dither every x frames i get benefit during integration. Tracking may not be as much of a hassle too.

 

Star clipping isn't a problem when I image based on SNR napkin math

 

(replaced signal with SNR just so i wouldn't confuse myself)

 

This works for me since i can do some kind of fuzzy math target SNR for individual SUB and Stack. I know when to give up for both.

 

Plus, I can say - if i can get to bortle two vs my bortle 6, my sky polution rate changes, my SNR changes and WOWZWERS i can double the signal on mys tack in half the time! (Which is the reason in another thread I fought back against you Rockstarbill because i did the math)

 

If i just wanted to 20x swamp my sensor i guess i could as a rule of thumb and i'd probably be happy too


Edited by sn2006gy, 19 August 2020 - 03:51 PM.


#90 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 19 August 2020 - 03:45 PM

from what I have learned, the diminishing returns relates to the equations that describe SNR... incremental time is always equal, but incremental SNR is not since it is described generally by a power equation... when an SNR increment starts to become less than the prior SNR increment you are no longer using your imaging time as efficiently... but other factors come into play like exposing long enough for faint signals, software overhead time, etc, I think diluting read noise is more of a floor for exposure time and not a ceiling... but thats just how I look at it

This sounds like an interesting consideration. What variables are being considered and what calculation is being used? 



#91 ChrisWhite

ChrisWhite

    Fly Me to the Moon

  • *****
  • Posts: 6,142
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted 19 August 2020 - 03:48 PM

I only cited the photo of the 1x2000 subs as an example. My other post was based entirely on more normal sub times.

 

It's my understanding that if you have a sensor with a read error rate, quantum efficiency rate and total well depth that doubling the exposure time does not double the SNR.

 

Just as doubling your total integration time doesn't double your signal. Both I believe have diminishing returns...  but again.. maybe i'm wrong, i'm a computer nerd by day and a hobby astronerd by night.

 

Just as there is some possible overhead to shorter subs, there is a benefit to more samples. If i have a bad sub, it's a smaller fraction of my integration at risk. I can dither every x frames i get benefit during integration. Tracking may not be as much of a hassle too.

 

Star clipping isn't a problem when I image based on SNR napkin math

You are correct that doubling your exposure time does not double your SNR.  You must quadruple your exposure time to double your SNR.  But this applies to short subs as well.  The diminishing return you are describing has a more or less equal impact on short and long subs in relation to TOTAL integration time. 

 

So 64 one minute subs is double the SNR of 16 one minute subs.  Similarly, 4 subs of 16 minutes would have double the SNR of a single sixteen minute sub. 

 

But longer exposures (without clipping stars) can have plenty of benefit for capturing faint signals, EVEN if you have already swamped read noise adequately. 

 

Again, with the caveat that exposing too long can clip stars. 

 

Anyway, this is how I understand this and apply it to my own exposure length decisions....

 

EDIT-  I'd like to add that using shorter subs does not recover any of the diminishing returns you are talking about.  (Now that I understand where you are coming from).  You still have the SAME diminishing return using shorter subs as longer subs when looking at SNR in your total integration time.


Edited by ChrisWhite, 19 August 2020 - 03:53 PM.

  • rockstarbill and Umasscrew39 like this

#92 Midnight Dan

Midnight Dan

    Voyager 1

  • *****
  • Posts: 13,677
  • Joined: 23 Jan 2008
  • Loc: Hilton, NY, Yellow Zone (Bortle 4.5)

Posted 19 August 2020 - 04:04 PM

You are correct that doubling your exposure time does not double your SNR.  You must quadruple your exposure time to double your SNR.  But this applies to short subs as well.  The diminishing return you are describing has a more or less equal impact on short and long subs in relation to TOTAL integration time. 

Chris is correct.  If your integration time is the same, it doesn't matter if you take short subs or long ones.  You will always need 4 times the overall integration time to get a doubling of SNR.

 

The issues with short subs vs long are strictly in other factors as have been pointed out.  You want your read noise to be an insignificant piece of the puzzle, so you want your subs to be long enough to swamp that.  But if they're too long, you can start saturating too many stars.  Anywhere between those two extremes will show no difference in terms of diminishing returns. 

 

The problem I often run into with my OSC camera, is that those two "extremes" overlap.  In other words, in some cases if I expose long enough to swamp the read noise, I'm already saturating too many stars.  So I have to decide whether to accept a little less read-noise-swamping, or a little too many saturated stars, and set my exposure length accordingly.

 

Also there should also be no difference in the amount of fine detail you can detect.  I've done targets where I've taken over 1000 30-second subs, and others where I've taken 50 5-minute subs.  I see no difference in the detail I can pull out.  The fine details come from long integration times, not sub length.  I shoot for 10-20 hours of integration if possible and that makes all the difference.

 

As for sub length, I lean towards shorter  if I have the option, just due to practical reasons.  Guiding problems are not as evident in short subs.  And subs that are lost to satellite streaks or other issues are a much smaller % of the overall integration time.

 

-Dan


  • rockstarbill likes this

#93 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 04:04 PM

 

EDIT-  I'd like to add that using shorter subs does not recover any of the diminishing returns you are talking about.  (Now that I understand where you are coming from).  You still have the SAME diminishing return using shorter subs as longer subs when looking at SNR in your total integration time.

Smaller subs doesn't recover anything, they just hit the diminishing return wall and stop.  Anything longer and i'm not impacting the SNR rate of my integration as much as I could by having another exposure. So in essence, I am gaining SNR in integration time by not reaching for diminishing returns in sub time. (in terms of SNR)

 

Sure, that 5% may capture more faint nebula in single subs. BUT, i find this napkin math worst amazingly well because it helps me plan for what I'm doing and I can optimize my integration for that as needed. Especially if someone has already done the e/px/r of that emission.

 

Just so i can wrap this all up because I honestly feel cheated I was told i was wrong or lying, the basis for SNR math is what lead me to another conclusion.

 

I got into another debate with Rockstar in the 2600 image thread that i'd have a better image if i got to darker skies. Mathematically that is true. My SNR would DOUBLE or more for said time/exposure based on SNR math.  Moving to a Mono 2600 vs a OSC 2600 may only offer me a small performance increase - and most of that to me is based on simplicity of SHO palette vs trying to pull that out of an RGGB color. It wouldn't double my performance. I tend to beat dead horses, so i'll beat dead horses.  With my OSC, I take more smaller/shorter exposures based on SNR math. I have the math for no filter (2600 say it has UV/IR window), I have the math for my Triad Ultra and I have the math for the bortle sky rating and when possible, i try and see the emission rate of the object i desire to get based on scope/sensor/sky/emission. I dither between every frame and bayer drizzle so the penalty of the bayer matrix also isn't anywhere near the SNR benefit i get going to dark skies.

 

so yeah..  i may be beating this dead horse a bit and we're both probably right. No one size fits all... but i tend to use SNR math to do my planning and no lie, I could be doing it all wrong, but it works.

 

If you know your read noise, dark current rate, lp signal rate, target signal rate, integration time. then you don't have to guess if more exposures causes a dark current penalty when dark current on modern sensors is essentially non existent and like i mentioned above - QHY is doing 2x reads to reduce it more which to me sounds like a reason to use SNR math and more exposures but that's just moi.. dark current/read noise is moot once you've swamped it and hit your signal rate too.


Edited by sn2006gy, 19 August 2020 - 04:08 PM.


#94 Peregrinatum

Peregrinatum

    Apollo

  • *****
  • Posts: 1,497
  • Joined: 27 Dec 2018
  • Loc: South Central Valley, Ca

Posted 19 August 2020 - 04:05 PM

This sounds like an interesting consideration. What variables are being considered and what calculation is being used? 

I found a researcher who developed an SNR model as a function of:  FL, Aperture, Image Scale, FWHM, Sky Mag, Faint Signal Mag (LM), Read Noise, Dark Current, and Efficiency.
 

I was able to model it in Excel as Total Integration time as a function of Limiting Magnitude, and make determinations with it for total integration time, based on diminishing returns, using my imaging train and sky conditions, I discussed the details in this post:

 

https://www.cloudyni...urns/?p=9964656



#95 ChrisWhite

ChrisWhite

    Fly Me to the Moon

  • *****
  • Posts: 6,142
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted 19 August 2020 - 04:14 PM

Smaller subs doesn't recover anything, they just hit the diminishing return wall and stop.  Anything longer and i'm not impacting the SNR rate of my integration as much as I could by having another exposure. So in essence, I am gaining SNR in integration time by not reaching for diminishing returns in sub time. (in terms of SNR)

 

I dont think you understand this.  Ultimately the diminishing return discussion ONLY applies to TOTAL integration time.  You are gaining nothing by using shorter subs with respect to diminishing returns.  Nothing.  You will have the same diminishing return whether you use long subs or short subs once you analyze your TOTAL integration time. 

 

Whether you take your diminishing return in a single long sub or many short subs... you still have the SAME diminishing return if your TOTAL integration is the same between the two.


Edited by ChrisWhite, 19 August 2020 - 04:16 PM.

  • rockstarbill likes this

#96 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 04:14 PM

This is what i did my math off of, but Peregrinatum's may be even better...
 
Snr


#97 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 04:16 PM

I dont think you understand this.  Ultimately the diminishing return discussion ONLY applies to TOTAL integration time.  You are gaining nothing by using shorter subs with respect to diminishing returns.  Nothing.  You will have the same diminishing return whether you use long subs or short subs once you analyze your TOTAL integration time. 

I may be wrong, but I believe SNR on subs has the same diminishing returns.

 

I'll see if i can find the whitepaper i used



#98 Peregrinatum

Peregrinatum

    Apollo

  • *****
  • Posts: 1,497
  • Joined: 27 Dec 2018
  • Loc: South Central Valley, Ca

Posted 19 August 2020 - 04:17 PM

With my setup shorter exposures are hugely inefficient due to SGP overhead... for me to "break even" on overhead time, in regards to total integration time, I need at least 75s subs when mono imaging.


Edited by Peregrinatum, 19 August 2020 - 04:18 PM.


#99 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 909
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 19 August 2020 - 04:21 PM

With my setup shorter exposures are hugely inefficient due to SGP overhead... for me to "break even" on overhead time, in regards to total integration time, I need at least 75s subs when mono imaging.

Just so people know, when I say shorter, I mean, less than the hour cited here smile.gif

 

My subs are typically 300 seconds on my F5.5 scope... (with a triad ultra)

 

I have an Epsilon coming in at F3.3, so i'll do the math there and see when diminishing returns kicks in and experiment with that.

 

I myself, haven't gone so crazy to try shorter subs and massively more files to integrate, but.. .i may give it a whirl sometime. Just want to make it clear that the crazy example I cited to me, was just in opposites of the other end of time with 1 hour subs

 

SGP download times are criminal.

 

With NINA it asynchronously downloads, dithers and settles before SGP finishes downloading. 


Edited by sn2006gy, 19 August 2020 - 04:23 PM.


#100 Midnight Dan

Midnight Dan

    Voyager 1

  • *****
  • Posts: 13,677
  • Joined: 23 Jan 2008
  • Loc: Hilton, NY, Yellow Zone (Bortle 4.5)

Posted 19 August 2020 - 04:22 PM

With my setup shorter exposures are hugely inefficient due to SGP overhead... for me to "break even" on overhead time, in regards to total integration time, I need at least 75s subs when mono imaging.

Good point.  There is definitely a per-sub overhead time that needs to be factored in.

 

-Dan




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: astrophotography, equipment



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics