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

More subs or longer exposure or both

  • Please log in to reply
12 replies to this topic

#1 BobW55

BobW55

    Mariner 2

  • *****
  • topic starter
  • Posts: 257
  • Joined: 06 Jun 2018
  • Loc: Melvin, Michigan (Sanilac County)

Posted 25 August 2019 - 10:01 PM

The following were all shot with a ES ED102mm FDC100 Scope and a ASI294MC Pro Camera, with a SV 0.8x focal reducer, NO filters.  Camera gain was set at 150.  I have found this gives the best results so far.

 

Both images were stacked and processed with Nebulosity 4 and Photoshop CS (still learning PS)

 

M51  shot with ten 180 second subs with matching blacks and 25 bias frames, 25 flats

 

M51Crop
 
M74 shot with ten 120 second subs with matching blacks and 25 bias frames, 25 flats.
 
M74Crop
 
Granted these are two different targets, but it would appear there is a big difference in exposure time and quality of the final rendered image.
so my questions are:
 
1) Where do you cross the line of too long of an exposure.
2) Where do you cross the line of too many subs?
 
I was hoping to hit M51 again tonight to try the same settings with 30 subs, and 30 subs with even more exposure time (**** clouds)

 

Links to other threads and things to read are helpful, I don't mind homework.

 

 

 



#2 1DegreeN

1DegreeN

    Vostok 1

  • -----
  • Posts: 152
  • Joined: 26 Aug 2013
  • Loc: Prachinburi, Thailand

Posted 25 August 2019 - 10:09 PM

A quick search of these forums will indeed find many threads on this subject and I also recommend you watch this video https://youtu.be/3RH93UvP358 by Robin Glover, the author of SharpCap, about optimal imaging strategies for CMOS cameras. 


  • davidparks and Dynan like this

#3 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23583
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 25 August 2019 - 10:44 PM

Both, although longer subs may depend on the camera and the skies. Beginners rarely get enough data in general anyway, so more data, longer exposures, longer integrations in general should be the rule of thumb. When a beginner gets their first 10 hour integration, I'm usually pretty proud. ;) It's a big milestone. Then there are 20, 30, 50 hour integrations... Then there are the multi-year long APOD projects and multi-panel mosaics and stuff to explore. ;D

 

Anyway. More integration, pretty much always. Two hours? Drop in the bucket. Four hours? Good start for a beginner. Eight hours? Ok, beginner is starting to push it. Ten hours? Key milestone, beginner is probably entering intermediate territory now. Twenty, thirty, fifty hours and beyond? Excellent!

 

As for exposure length. I like to explain it this way. There are two competing forces that determine sub-exposure length: Stellar clipping and read noise.

 

1) Clipping is the force that makes you want to use shorter exposures. It determines how long you CAN expose for.

2) Read noise is the force that makes you want to use longer exposures. It determines how long you MUST expose for, at a minimum.

 

You need to expose long enough to sufficiently swamp the read noise, but not so long that you are clipping too much signal. A few stars clipped (10, 20, maybe 30 - lightly) is usually fine. Most stars clipped? Probably exposing too long. On the flip side, if you are seeing banding, or other structures, medium and large scale blotching, etc. in your background? You are probably not exposing long enough.

 

General rule of thumb is expose as long as you can, until you start to feel you are clipping too many stars. Then back off just slightly. Check the background noise. If the background generally looks clean and you aren't clipping tons of stars, then you are probably fine. If you are not clipping anything? Well, if you are ABLE to expose longer, you might as well expose longer. Longer sub-exposures tend to be better, and even if you reach the effective limits (i.e. your shot noise so completely overpowers read noise that read noise becomes meaningless), if you are not clipping anything then there is no harm done.

 

There is more to exposure than just this. But, as a beginner, this should do. Once you get into asking questions like how you optimize resolution and get the most detail, then you might want to start thinking about optimizing sub-exposure length more.


  • psandelle, PeteM, Ken Sturrock and 4 others like this

#4 BobW55

BobW55

    Mariner 2

  • *****
  • topic starter
  • Posts: 257
  • Joined: 06 Jun 2018
  • Loc: Melvin, Michigan (Sanilac County)

Posted 25 August 2019 - 10:51 PM

Thank you... All good and fair answers.  And I will check out that video.



#5 BobW55

BobW55

    Mariner 2

  • *****
  • topic starter
  • Posts: 257
  • Joined: 06 Jun 2018
  • Loc: Melvin, Michigan (Sanilac County)

Posted 25 August 2019 - 11:47 PM

Excellent video... they should ship it with all cameras.



#6 RJF-Astro

RJF-Astro

    Explorer 1

  • -----
  • Posts: 89
  • Joined: 13 Aug 2018
  • Loc: Zeist, Netherlands

Posted 26 August 2019 - 04:56 AM

Also evualuate your guiding performance when increasing sub-exposure length. A cable snag or wind gust can ruin subs. If you have less long exposure subs, the penalty is higher.
  • psandelle likes this

#7 terry59

terry59

    Cosmos

  • *****
  • Posts: 9101
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 26 August 2019 - 06:07 AM

Also evualuate your guiding performance when increasing sub-exposure length. A cable snag or wind gust can ruin subs. If you have less long exposure subs, the penalty is higher.

Ensuring the ability to successfully guide through a complete worm cycle should be the first order of business with any mount to be used for AP IMO. Potential cable snags (operator error) or wind gusts (the environment) are part of astrophotography. What about clouds, dew, pop up rain showers, etc.? 



#8 1DegreeN

1DegreeN

    Vostok 1

  • -----
  • Posts: 152
  • Joined: 26 Aug 2013
  • Loc: Prachinburi, Thailand

Posted 26 August 2019 - 10:40 PM

To follow up on Jon's comments on clipping vs read noise, I've been using the long rainy season here to analyse the data I captured earlier in the year particularly in regard to clipping. I have a couple of rigs, the first is a 110mm f5.6 refractor with a QHY163M camera. To swamp read noise by e.g. a factor of 10, at my normal gain setting for LRGB imaging and with the L filter it requires an exposure of just 32 seconds. In those 32 seconds, every star brighter than about mag 10.6 will be clipped. My second rig is a QHY168C OSC behind a 103mm f5.5 refractor: at my normal gain setting a 153 second exposure will swamp the read noise by 10 and the clipping threshold will be about mag 10.1

 

So for my upcoming imaging season I've concluded that, for most of the objects I plan to image, I'll have to include some short exposures for an HDR composite. 


  • Der_Pit likes this

#9 WadeH237

WadeH237

    Aurora

  • *****
  • Posts: 4683
  • Joined: 24 Feb 2007
  • Loc: Snohomish, WA

Posted 26 August 2019 - 11:21 PM

As for exposure length. I like to explain it this way. There are two competing forces that determine sub-exposure length: Stellar clipping and read noise.

 

1) Clipping is the force that makes you want to use shorter exposures. It determines how long you CAN expose for.

2) Read noise is the force that makes you want to use longer exposures. It determines how long you MUST expose for, at a minimum.

That's one of the best explanations I've read on this topic.  It very concisely names the competing factors.


  • Jon Rista likes this

#10 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 16784
  • Joined: 27 Oct 2014

Posted 27 August 2019 - 12:22 AM

 

The following were all shot with a ES ED102mm FDC100 Scope and a ASI294MC Pro Camera, with a SV 0.8x focal reducer, NO filters.  Camera gain was set at 150.  I have found this gives the best results so far.

 

Both images were stacked and processed with Nebulosity 4 and Photoshop CS (still learning PS)

 

M51  shot with ten 180 second subs with matching blacks and 25 bias frames, 25 flats

 

 
 
M74 shot with ten 120 second subs with matching blacks and 25 bias frames, 25 flats.
 
 
 
Granted these are two different targets, but it would appear there is a big difference in exposure time and quality of the final rendered image.
so my questions are:
 
1) Where do you cross the line of too long of an exposure.
2) Where do you cross the line of too many subs?
 
I was hoping to hit M51 again tonight to try the same settings with 30 subs, and 30 subs with even more exposure time (**** clouds)

 

Links to other threads and things to read are helpful, I don't mind homework.

 

1)  Look at the histogram in a program that does not stretch the data.  PixInsight or IRIS, rather than a terrestrial program.  Look at the obvious peak, or the mean value, in ADU.  Subtract the ADU from a bias frame from the ADU of the light.  Convert the result in ADU to electrons, using the gain of the camera.  Compare it to the read noise of the camera (RN).

 

People shoot for different targets.   Common ones are 20 X RN, or 5 X RN squared, or 10 X RN squared.  The idea is to "bury the read noise in the sky noise", to make RN insignificant.

 

People often balance that against oversaturating too much and "clipping high".  You'll almost always saturate some stars, the idea is not too many.

 

All the above is discussed in much more detail in this superb book.

 

https://www.amazon.c.../dp/1138055360/

 

2)  Never, in terms of image quality.  The main thing to remember is that total imaging time is more significantly important than getting subexposure "right", provided subexposure is within the ballpark.  <smile>

 

My personal rule of thumb is one hour minimum, two is better, four is good.

 

It's not a case of diminishing returns, but it is a square root thing.  Improving things by a factor of about 2 means increasing total imaging time by a factor of four.

 

Strong bottom line.  I never do more than one target in a night anymore.  And try to get in more than one night on a target.  Platesolving gets me back to where I was.

 

Recommendation.  Get subexposure time in the ballpark by burying the read noise, and shoot more subs.  Your examples were 20 and 30 minutes total imaging time, that's too little.


Edited by bobzeq25, 27 August 2019 - 12:27 AM.


#11 mic1970

mic1970

    Viking 1

  • -----
  • Posts: 718
  • Joined: 09 Aug 2018
  • Loc: St. Louis Metro Area

Posted 27 August 2019 - 07:48 AM

Could you start another thread on this with pics and such.  For example, how do you tell when you have hit 10x data to noise?  

To follow up on Jon's comments on clipping vs read noise, I've been using the long rainy season here to analyse the data I captured earlier in the year particularly in regard to clipping. I have a couple of rigs, the first is a 110mm f5.6 refractor with a QHY163M camera. To swamp read noise by e.g. a factor of 10, at my normal gain setting for LRGB imaging and with the L filter it requires an exposure of just 32 seconds. In those 32 seconds, every star brighter than about mag 10.6 will be clipped. My second rig is a QHY168C OSC behind a 103mm f5.5 refractor: at my normal gain setting a 153 second exposure will swamp the read noise by 10 and the clipping threshold will be about mag 10.1

 

So for my upcoming imaging season I've concluded that, for most of the objects I plan to image, I'll have to include some short exposures for an HDR composite. 



#12 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 16784
  • Joined: 27 Oct 2014

Posted 27 August 2019 - 09:49 AM

Could you start another thread on this with pics and such.  For example, how do you tell when you have hit 10x data to noise?  

Post #10, part (1) talks about how to quantitatively figure out when you have swamped the read noise.  It's not 10X data to noise, it's the relative importance of read noise (which can be managed) compared to sky noise (which cannot).

 

The whole topic is complicated.  There are a number of websites.  There are threads here.  But I recommend the book referenced in #10.  It's readable, thorough and covers all the important pieces.

 

Here's an older thread with 4 somewhat theoretical pages on the topic.  "Stacking efficiency" is something of a buzzword that applies.

 

https://www.cloudyni...yfog-histogram/

 

In terms of improving signal to noise ratio overall, that's where total imaging time (the most important thing) is the key parameter.  Post #10, part (2).  Referenced below are some relevant pictures of how that works.  The 7 images of the red nebula, showing the improvement with successive doubling of total imaging time, by doubling the number of equal time subs.  Jon Rista has done a similar sequence, maybe he'll read this and provide a reference.  Or you could PM him.

 

http://www.samirkhar...-exposures.html

 

You might also find the video referenced in post #2 helpful.  This is the missing piece.

 

https://www.youtube....h?v=ub1HjvlCJ5Y


Edited by bobzeq25, 27 August 2019 - 10:41 AM.


#13 BobW55

BobW55

    Mariner 2

  • *****
  • topic starter
  • Posts: 257
  • Joined: 06 Jun 2018
  • Loc: Melvin, Michigan (Sanilac County)

Posted 27 August 2019 - 10:46 AM

After having a look at the suggested video.  I think I have about found the sweet spot for the equipment I have.  

Everyone was telling to use only a gain of 120, I have found 150 gives better results. 

So I am down to exposure time vs # of subs.   For my setup, I think 180 seconds may be the correct spot, but until the clouds clear I won't have a chance to test my settings.  I wanted to shoot 20 subs starting at 180 and increase the time by 30 seconds.  I can usually image M51 for hours at a time before my scope flips.

Was suggested in one read to use the same target so you are comparing data of the same object.

 

From the video and what I have read, for my setup, shorter exposure, but more of them 20++ should work the best.   I only have one cable dangling (my USB/Serial cable) but that too will soon be gone.  Going to take out my CEM60 polar scope and run a few extra wires through it for my dew heaters (soon as I finish building them) and the USB/Serial cable.

All in all it all boils down to lowest numbers of noise.

 

I have also seen that people who are really good at photoshop can turn a poor image into a great one.  That is not me.... 


Edited by BobW55, 27 August 2019 - 10:48 AM.



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