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

Photons to Photos

DSLR
  • Please log in to reply
9 replies to this topic

#1 AlexL

AlexL

    Vostok 1

  • -----
  • topic starter
  • Posts: 192
  • Joined: 29 Nov 2008
  • Loc: Espinho, Portugal

Posted 03 August 2021 - 02:30 AM

Sometime ago, I came across the website https://www.photonstophotos.net/
Do you know what information can I get and how to read it?



#2 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,711
  • Joined: 20 Mar 2013

Posted 03 August 2021 - 03:12 AM

For astrophotography, the most useful information is input-referred read noise and quantum efficiency. The latter is under the page for DxO derived sensor characteristics.



#3 AlexL

AlexL

    Vostok 1

  • -----
  • topic starter
  • Posts: 192
  • Joined: 29 Nov 2008
  • Loc: Espinho, Portugal

Posted 05 August 2021 - 03:52 AM

Hello Wei-Hao,

 

Thank you for your quick reply.

On the Photons to Photos one can find Read Noise in DNs and Input-referred Read Noise Charts.

What is the difference between both?

 

Many thanks,

 

Alex.



#4 the Elf

the Elf

    Soyuz

  • *****
  • Posts: 3,796
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 05 August 2021 - 12:48 PM

You want the Input-referred RN charts. Pure read noise is just the DNs while Input-referred are relative to the input signal which is what you will see as signal to noise in the image. Dynamic range charts is the other important one.

You can use the charts to tell which of a bunch of cameras is the best and which ISO to use. It is obvious that a higher dynamic range is always better when you have to pick one out of two cameras. Now you need to find the best ISO. The basic question is if sacrificing one f-stop of DR by going to the next higher ISO value helps to cancel out more noise. The sweet spot is where the flat slope of the DR chart bends down. That is for the older generation of not ISO invariant sensors. But there are other factors that you cannot see in the charts. Always ask other users of the same camera model what they recommend.


  • AlexL likes this

#5 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,711
  • Joined: 20 Mar 2013

Posted 05 August 2021 - 07:31 PM

I do not agree with the above. The choice of ISO should be based on read noise (in electron, not in DN), not on dynamical range. Dynamical range is not very important in astrophotography.

#6 galacticinsomnia

galacticinsomnia

    Apollo

  • *****
  • Posts: 1,495
  • Joined: 14 Aug 2020
  • Loc: Pacific Northwest - Oregon

Posted 05 August 2021 - 09:01 PM

I do not agree with the above. The choice of ISO should be based on read noise (in electron, not in DN), not on dynamical range. Dynamical range is not very important in astrophotography.

How do you come to a conclusion that Dynamic Range is not very important in astrophotography.  That seems to be exact opposite of what I believe I have observed.  Perhaps if you are considering dynamic range as applied to a specific color space, that there are not 16billion colors in space, then I can see where having dynamic range is a don't care, but when you are looking for the subtle changes in tone, contrast in a specific color, then the range or dynamic range is indeed important. 

Admittedly, I could be completely incorrect, but there has been a huge increase in my ability to capture the subtle changes in tone in various nebula due to what I believe includes the benefit of a decent improvement in dynamic range in comparison to what I have observed with my older sensors.

Clear Skies !!



#7 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,711
  • Joined: 20 Mar 2013

Posted 05 August 2021 - 09:39 PM

Dynamical range is essentially determined by two things: saturation level and noise level.

 

Let's look at the noise level first.  For deep-sky astrophotography, we should do our best to make the noise in each sub dominated by photon noise.  To do so, we need long exposure for photon noise to overwhelm read noise.  So the main enemy that we are fighting against is read noise.  And this is what I was saying in the previous replies.

 

For astro-landscape, we often don't have the luxury to expose the subs long enough.  So photon noise cannot overwhelm read noise.  So again, the main concern is read noise.  

 

Now on to saturation.  What's the last time when the image quality of your deep astrophoto or astro landsacpes was limited by saturated pixels?  How many of you are really bothered by saturation that often?  Do you image extremely bright objects (M31, M42, M8, and what else?) all the time?  I know a small number of people who are obsessed by saturated cores of very bright stars, but I don't believe they are the majority judging by the discussions that often show up here.  Here, people discuss about how to fight against light pollution and noise to image faint nebulas.  What's the last time you see people discussing about cores of bright stars?

 

Personally I never care about saturated bright stars.  Their cores may be saturated, but they have many many outer pixels with rich colors.  I care more about faint nebulas and galaxies.  Occasionally when I have saturated nebula cores (on the three Messier objects I mentioned previously), I just took a few short exposures to recover the very small numbers of saturated pixels.  I spend hours and hours of time to collect photons in order to fight against noise.  Noise is the main concern for me, and I believe it should be the main concern for you too, if you care about getting the most back from the time you spend on integration.


Edited by whwang, 05 August 2021 - 09:55 PM.

  • AlexL, sharkmelley, Miguelo and 1 other like this

#8 the Elf

the Elf

    Soyuz

  • *****
  • Posts: 3,796
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 06 August 2021 - 02:13 AM

 

Personally I never care about saturated bright stars.

This is a personal choice. Others might want to go a different route. When imaging a star cluster it may make sense to minimize saturation. When a very bright star is in the FOV (like Alinitak between Horse Head and Flame) more DR is also a plus. I love my T7i because I can use it at ISO 400 and have twice the DR compared to ISO 400. That in turn reduces the number of subs for one image from almost 200 to almost 100. That is quite a difference for the computer I have. It also reduces read noise. Indeed, when the object is faint noise is the thing to care about. I don't say for bright targets DR is more important but it cannot be neglected.

 

I've seen PI's photometric color correction fail when the cores are saturated, no matter how much color there is left in the perimeter. If someone is attempting "true" star color the photometry approach is probably the closest you can get because it is based on professional data. In that case you don't want to saturate stars. Of course you can take very nice images with saturated stars, no doubt. It is a personal choice, imho.


  • galacticinsomnia likes this

#9 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,711
  • Joined: 20 Mar 2013

Posted 06 August 2021 - 02:46 AM

You said more DR is a plus.  I can agree with that.  It's a plus, but if you make it your priority, chance is good that you are wasting your precious imaging time because you lose your signal to read noise.

 

When I teach students photometry in my graduate-level course, I told them never conduct aperture photometry on saturated objects.  If PI fails because of saturation, then I am really sorry to say that it's a poor implementation.

 

Overall, I think you are using rare applications to justify a general decision.  Generally, one should prioritize S/N, not DR.  Getting high S/N can require many hours of integration.  Recovering saturated cores of stars or nebulas using short exposures is often just a matter of minutes, if not seconds.


  • Miguelo and galacticinsomnia like this

#10 galacticinsomnia

galacticinsomnia

    Apollo

  • *****
  • Posts: 1,495
  • Joined: 14 Aug 2020
  • Loc: Pacific Northwest - Oregon

Posted 07 August 2021 - 03:59 PM

You said more DR is a plus.  I can agree with that.  It's a plus, but if you make it your priority, chance is good that you are wasting your precious imaging time because you lose your signal to read noise.

 

When I teach students photometry in my graduate-level course, I told them never conduct aperture photometry on saturated objects.  If PI fails because of saturation, then I am really sorry to say that it's a poor implementation.

 

Overall, I think you are using rare applications to justify a general decision.  Generally, one should prioritize S/N, not DR.  Getting high S/N can require many hours of integration.  Recovering saturated cores of stars or nebulas using short exposures is often just a matter of minutes, if not seconds.

Having general rules or guide in which one performs their processes is generally, a good thing.  I can agree generally, that prioritizing S/N over DR is a wise and reasonable, especially when their are processes to cover over exposed stars.  There are also processes to improve S/N.  I guess prioritization of either is really going to depend more on someones style and how they like to work their data?  Generally, I think we all like less noise in our images as well as nice color in our stars and nebulosity.
Seems balance is key.

Clear Skies !!




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: DSLR



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics