Our Beautiful Universe - Photos and Videos

[Keith_McClary]

I have heard about green filters for the Moon

Oh, I thought @chemisttree was making a cheesy joke.

 

[chemisttree]

If the atmosphere is unsteady, sometimes a red filter is best. The old moon filters usually had a deep green tint.

 

[DennisN]

Does the lens exhibit any chromatic abberation in daylight photos?

I will test it and check.

 

[Andy Resnick]

A break in the clouds last night- first time in what seems like ∞ weeks. Cigar (M82) and Bode's (M81) galaxies @ 800/5.6, 9300 s integration time, 20s subs:

This was, by far, the worst imaging conditions I have ever dealt with- the temp was about 15°F and with the strong, gusting wind chill, about 5°F. Very poor transparency and seeing conditions. And the wind knocking my setup about. Normally, I wouldn't bother on a night like that, but as I said it has been about ∞ weeks since I could see stars at night. So... yeah.

 

[Andy Resnick]

Last night was very clear, excellent transparency, but also very poor seeing. Rather than fighting a losing battle re: PSF FWHM, I used my 105mm lens; smaller aperture = less susceptibility to seeing conditions. Here's Orion @ 105/2, 8.5 hrs Total exposure using 30s subs:

At 1:1, the Flame and Horsehead nebulae:

Stopping down the lens gives those nice starburst diffraction patterns... off-axis, though, the lens performance starts to degrade (faster than I expect), it's most noticeable as a greenish tint around Rigel. Here's the Orion Nebula, the lens performance is not as good as I would like...

Barnard's loop and other molecular cloud features are mostly too faint (unfortunately- I was motivated by this image), but M78 and NGC 2112 look ok:

There's a lot I like about the lens (Nikkor 105/1.4), I think the problem is a slight angular misalignment between the image and sensor planes and I'm not sure if this is 'within spec' or not; not sure how to compensate either way- it's not a 'micro focus' adjustment.

 

[DennisN]

In memoriam/in memory of

I am going to tell a story I haven't told here before. This thread has been going for 7 years now.
There is a reason I started this thread, "Our Beautiful Universe", way back on March 1, 2015.

Believe it or not, I did not start this thread because I like astronomy. I do like astronomy.
But the reason I started the thread was because of something completely different.

On February 22 2015, I had to put one of my dear cats to sleep.
Long story short, it was very painful because it was the first time I had to do such a thing,
and I had to do it the very same day I took him to the vet (he had lethal kidney values).

I felt terrible. Even though I don't know how it feels to amputate anything, it felt like I had been amputated.

One week later I was still grieving, and felt I had to get my mind on something else.
I watched a couple of inspiring videos about the universe and the conditions for life, and shortly after that I started this thread as a celebration of our Universe, life and existence.

Now, 7 years later, this thread has 1,555 posts from countless of PF members, and the thread has been viewed 134'000+ times. We have shared countless of fascinating videos, photos and advice of various things related to astrophotography. And I'm not sure I would have started doing some astrophotography myself if it was not for this thread.

And all this because of this little cat who once existed, and his name was Nisse (2006-2015):

So when life gives you lemons, make lemonade.

And since this still is an astro thread and not a cat thread, here's the Cat's Eye Nebula:

 

[Devin-M]

Core of Andromeda Galaxy, 2.5Mly
2130mm f/14.2 29x 90sec (43.5min) 6400iso, Bortle 2

It’s been a while since I’ve taken any astro-photos (its been cold, cloudy and rainy for quite a long time.) I managed to find a break in the clouds last night while the moon was down and captured this photo of the core of Andromeda Galaxy, which is around 2.5 million light years distant. I shot using a 150mm diameter Meade Maksutov Cassegrain on a Star Adventurer 2i mount with some modifications to go probably 3x over the weight limit with a Nikon D800 camera body. I aimed the telescope by taking test shots and plate solving through a website on my phone and then making adjustments. I used the 29 best images although I think I got around 60 total, only about half had round enough stars with 90 second exposures at this focal length on this mount. My bortle 2 dark sky location is near Shingletown, California.

 

[Devin-M]

Quick question... Sometimes I shoot dark calibration frames to remove noise from the final image on the way home (for example I might take 10x shots of 5min each at 6400iso with the lens cap on). Because I'm too impatient to wait around at the cold, dark sky location, the camera is heating up from for example 35F (outdoor temp) to 70F in the car while I drive. This results in the amount of noise in each dark frame increasing as the temperature of the camera increases, as seen in the animation below. My conceptual understanding is the photodiodes in the CMOS sensor are operating much like solar panels, and are somewhat sensitive to infrared light while taking exposures even when the lens cap is on. My question is can solar cells or CMOS sensors convert infrared light into electricity, and if so how does that reconcile with the 2nd Law of Thermodynamics? For example, I thought I shouldn't be able to extract useful work from a single temperature reservoir... but if a solar panel can convert infrared light into electricity, isn't it extracting work from a single temperature reservoir (suppose it's immersed in an insulated tank of water as a heat source)?

https://www.speakev.com/attachments/13c208ea-b300-4a24-a496-9e9f9fae4c25-gif.156383/

 

[Drakkith]

Sometimes I shoot dark calibration frames to remove noise from the final image on the way home (for example I might take 10x shots of 5min each at 6400iso with the lens cap on). Because I'm too impatient to wait around at the cold, dark sky location, the camera is heating up from for example 35F (outdoor temp) to 70F in the car while I drive.

You will almost certainly get better results by shooting your darks in concert with your lights. Shoot a couple of light frames, then shoot a dark, then a few more lights, then a dark. Rinse and repeat. The better the temperature match, the better the dark frame subtraction.

My conceptual understanding is the photodiodes in the CMOS sensor are operating much like solar panels, and are somewhat sensitive to infrared light while taking exposures even when the lens cap is on.

My understanding is that it's the thermal motion of the electrons in the sensor that leads to the generation of dark current, not IR radiation. A single IR photon doesn't have enough energy to cause an electron to jump the energy gap, but a lucky collision/interaction between several electrons/ions can give an electron enough energy to jump the gap and into the area of the pixel well that stores the photoelectrons prior to readout.

My question is can solar cells or CMOS sensors convert infrared light into electricity, and if so how does that reconcile with the 2nd Law of Thermodynamics? For example, I thought I shouldn't be able to extract useful work from a single temperature reservoir... but if a solar panel can convert infrared light into electricity, isn't it extracting work from a single temperature reservoir (suppose it's immersed in an insulated tank of water as a heat source)?

No. Prior to exposure the pixel wells undergo a charge separation process that puts them in a high-energy state. Photons, or random interactions from thermal motion, cause electrons to jump an energy gap and get caught in this charged well. Given enough time or photons the well becomes completely saturated and can no longer collect charge. You won't extract more energy out of this than it took to separate the charges in the first place.

A solar panel operates somewhat differently and I don't really know enough to explain it well. However, I do believe that the solar panel needs to be at a lower temperature than the emitting object it is capturing radiation from. Besides, the solar panel itself and the surrounding environment is a temperature reservoir, so there's more than one.

 

[collinsmark]

Quick question... Sometimes I shoot dark calibration frames to remove noise from the final image on the way home (for example I might take 10x shots of 5min each at 6400iso with the lens cap on). Because I'm too impatient to wait around at the cold, dark sky location, the camera is heating up from for example 35F (outdoor temp) to 70F in the car while I drive. This results in the amount of noise in each dark frame increasing as the temperature of the camera increases, as seen in the animation below.

An alternative that you might want to try in your off time (i.e., a cloudy night), is to realize that you can reuse DARK frames (for the given camera). Take your camera and an accurate thermometer outside and let the camera the acclimate to the outside temperature. Once the camera is in thermal equilibrium, start taking DARKs. Lots of DARKs.

Vary the exposure times in a controlled, roughly exponential way -- exposure times that you might likely use for your LIGHT frames. For example, 40 sec, 60 sec, 90 sec, 120 sec, 180 sec, 240 sec, 300 sec, etc.

The whole while, keep meticulous records of the ambient temperature. Also keep a record of the camera's ISO setting you are using for each dark (only use ISO settings that you would use for LIGHT frames). Sit down, have a beer in the cold. This will take a while. Maybe have two beers. Repeat as much as you can. As the night continues, the temperature is likely to drop, so you should expect to have several different temperature points for your DARK frames.

Later, organize these DARK frames. Rename each DARK frame such that it has

• Specific camera taking the photos
• ISO setting,
• Exposure time,
• Temperature for that particular dark frame, and
• Some unique identifier so you don't accidentally overwrite existing DARK frames

in the file name.

Repeat the next cloudy night. Put the camera in the refrigerator and repeat there too. If you live in a cold climate, and foresee yourself taking astrophotos in really cold weather, try the freezer as well.

Eventually, you'll have a DARK library with many frames for each particular ISO setting + Exposure time + temperature. Ideally, you'll have many DARK frames for each combination. Organize these on your comptuer. Once you have that, all you need to do when taking LIGHT frames is to record the temperature at the time (stash your thermometer in your camera bag), and your corresponding DARK frames will be waiting for you at home.

(Of course, the data is specific to the specific camera. You cannot share DARK frames between cameras.)

---------------

For personal reference, I have a few cooled cameras (dedicated, astrophotography cameras). Having a temperature controlled camera makes the process a lot easier. But even then it still takes a while. Whenever I get a new cooled camera, the first thing I do set the camera next to the computer and let it take DARKs for days. Literally days. But once it's done I don't have to take DARKs again for maybe a year or two. And I only use 2 temperature settings (0 deg C and -5 deg C).

 

[Devin-M]

My understanding is that it's the thermal motion of the electrons in the sensor that leads to the generation of dark current, not IR radiation. A single IR photon doesn't have enough energy to cause an electron to jump the energy gap, but a lucky collision/interaction between several electrons/ions can give an electron enough energy to jump the gap and into the area of the pixel well that stores the photoelectrons prior to readout.

No. Prior to exposure the pixel wells undergo a charge separation process that puts them in a high-energy state. Photons, or random interactions from thermal motion, cause electrons to jump an energy gap and get caught in this charged well. Given enough time or photons the well becomes completely saturated and can no longer collect charge. You won't extract more energy out of this than it took to separate the charges in the first place.

A solar panel operates somewhat differently and I don't really know enough to explain it well. However, I do believe that the solar panel needs to be at a lower temperature than the emitting object it is capturing radiation from. Besides, the solar panel itself and the surrounding environment is a temperature reservoir, so there's more than one.

Could you comment on the graph on this webpage… I might be misinterpreting it but I believe graph (a) shows a particular detector which is at 300k (~80F) operating temperature generating current from mid-infrared light up to 4000 nanometers with zero bias, which I take to mean the detector is operating in photovoltaic mode with no outside voltage applied…

https://www.researchgate.net/figure/a-g-The-spectral-responsivity-measured-at-zero-bias-ie-photovoltaic-mode-for-the_fig3_346511011

“(a)-(g) The spectral responsivity measured at zero bias (i.e. photovoltaic mode) for the Te-hyperdoped Si photodetector at different temperatures. The room-temperature spectral responsivity of a commercial Si-PIN photodiode (model: BPW34) is included as a reference (brown short dot). (h) Illustration of the below-bandgap photoresponse in the Te-hyperdoped Si photodetector. Te dopants introduce deep-level states (intermediate band) inside the Si band gap, which facilitate the absorption of photons with sub-bandgap energies. Process I: VB to CB (Eph ≥ Eg); Process II: VB to IB (Eph ≥ Eg-ETe); Process III: IB to CB (Eph ≥ ETe, only measurable at low temperatures where the thermal contribution is neglected).”

 

[Drakkith]

Could you comment on the graph on this webpage… I might be misinterpreting it but I believe graph (a) shows a particular detector which is at 300k (~80F) operating temperature generating current from mid-infrared light up to 4000 nanometers with zero bias, which I take to mean the detector is operating in photovoltaic mode with no outside voltage applied…

Not really. I'm not an expert in the area of photodetectors and solid state physics and such. I'll try to remember to give it a read tomorrow or the next day if I can, but I might not have time.

 

[Drakkith]

Could you comment on the graph on this webpage… I might be misinterpreting it but I believe graph (a) shows a particular detector which is at 300k (~80F) operating temperature generating current from mid-infrared light up to 4000 nanometers with zero bias, which I take to mean the detector is operating in photovoltaic mode with no outside voltage applied…

Quickly skimmed through the article just now. I come to the same conclusion as you.
Note that at 300K an object barely emits any radiation in the 1-5 micrometer range. You have to get warmer for that. You can use the calculator here to see the spectrum emitted by an object at a given temperature. Use 1 micrometer as the upper limit and 5, 10, or 20 as the lower limit to get a good looking graph of the region of interest.

 

[Devin-M]

Quickly skimmed through the article just now. I come to the same conclusion as you.
Note that at 300K an object barely emits any radiation in the 1-5 micrometer range. You have to get warmer for that. You can use the calculator here to see the spectrum emitted by an object at a given temperature. Use 1 micrometer as the upper limit and 5, 10, or 20 as the lower limit to get a good looking graph of the region of interest.

Thanks for the calculator! According to its output, with inputs for the emissivity of water (0.96) at 300k (~80F), ordinary room temperature water is emitting some blackbody infrared radiation from 3-4 micrometers— at wavelengths the “Te-hyperdoped Si photodetector” also @ 300k can generate current from in photovoltaic mode… I must be missing something because why couldn’t I just generate a small amount of electricity by submerging these room temperature photodetectors in room temperature water to harvest the 3-4 micrometer infrared black body radiation photons by photovoltaic means? Wouldn’t that conflict with the 2nd Law of Thermodynamics? I shouldn’t be able to generate any useful work from a single temperature reservoir, was my understanding.

 

[Drakkith]

I must be missing something because why couldn’t I just generate a small amount of electricity by submerging these room temperature photodetectors in room temperature water to harvest the 3-4 micrometer infrared black body radiation photons by photovoltaic means? Wouldn’t that conflict with the 2nd Law of Thermodynamics? I shouldn’t be able to generate any useful work from a single temperature reservoir, was my understanding.

That I can't answer. I'm certain the 2nd law isn't being violated, but I couldn't tell you how or why it isn't.

 

[collinsmark]

I must be missing something because why couldn’t I just generate a small amount of electricity by submerging these room temperature photodetectors in room temperature water to harvest the 3-4 micrometer infrared black body radiation photons by photovoltaic means?

Maybe I'm missing something myself. But It's my understanding that you couldn't generate any electricity by simply submerging the photodetector in water because there wouldn't be a light source in that situation.

It's my understanding of the test setup that the photodetector is placed and held at a given temperature, then it is exposed to a light source of a specific wavelength and specific intensity (with a proportional power reaching the detector, measured in Watts) and the current of the photodetector is measured (measured in milliamps). That is used to generate a single point on a single graph. For any given situation (wavelength of the light source and temperature of the photodetector), the current of the photodetector is proportional to the power of the light source. Which is why the measurements are in units of mA/W.

At least that's my understanding. The power is ultimately coming from the light source. The 2nd Law is not violated. The current vanishes as soon as you turn off the light.

 

[Devin-M]

Maybe I'm missing something myself. But It's my understanding that you couldn't generate any electricity by simply submerging the photodetector in water because there wouldn't be a light source in that situation.

View attachment 296237

It's my understanding of the test setup that the photodetector is placed and held at a given temperature, then it is exposed to a light source of a specific wavelength and specific intensity (with a proportional power reaching the detector, measured in Watts) and the current of the photodetector is measured (measured in milliamps). That is used to generate a single point on a single graph. For any given situation (wavelength of the light source and temperature of the photodetector), the current of the photodetector is proportional to the power of the light source. Which is why the measurements are in units of mA/W.

At least that's my understanding. The power is ultimately coming from the light source. The 2nd Law is not violated. The current vanishes as soon as you turn off the light.

A 300k (80F) black body emits some infrared between 3 & 4 micrometers, which is in the detection range of the photodetector.

 

[berkeman]

Wouldn’t that conflict with the 2nd Law of Thermodynamics? I shouldn’t be able to generate any useful work from a single temperature reservoir, was my understanding.

Should some of this side-discussion be split off into the Thermodynamics forum?

 

[collinsmark]

A 300k (80F) black body emits some infrared between 3 & 4 micrometers, which is in the detection range of the photodetector.

View attachment 296248

Yes, but the photodetector is also emitting infrared too -- the same amount that it receives when its own temperature is at 300 K, along with everything else in the surroundings being at 300 K, and when no external light source is present. Without the presence of the external light source the net current is zero. At least that's my understanding.

I don't know the test setup, but here's how I imagine it:

A broadband blackbody radiation source is involved; an incandescent bulb will do. The light from the source passes through a slit followed by a diffraction grating, thus splitting up the light (including infrared light) into a spectrum. The intensity along the spectrum is measured and calibrated (perhaps with a small, calorimeter device). With this information, the light intensity along specific wavelengths of the spectrum is known.

The photodetector can then be placed along the spectrum for measurements. Changing the wavelength is just a matter of moving the photodetector spacially to a different part of the spectrum produced by the light source + diffraction grating.

But again, if I'm imagining the setup correctly, the current in the photodetector will vanish when the light source is turned off.

**** Edit *****
Reading into the research paper a little more (https://www.researchgate.net/publication/346511011_Silicon-Based_Intermediate-Band_Infrared_Photodetector_Realized_by_Te_Hyperdoping), it states in the Device Measurement seciton: "A Globar (SiC) source coupled with a TMc300 Bentham monochromator equipped with gratings in Czerny-Turner reflection configuration was used as the infrared monochromatic source. Its intensity is spatially homogenized and was calibrated with a Bentham pyrometric detector."

The "TMc300 Bentham monochromator" utilizes a diffraction grating turret. So my imagined setup, albeit a bit simplistic, was conceptually accurate.
****************

Should some of this side-discussion be split off into the Thermodynamics forum?

That sounds like a good idea to me.

 

[berkeman]

That sounds like a good idea to me.

Can you folks suggest which posts I should split off into the Thermo forum? I don't want to mess up the astrophotography part of the discussion.

 

[Devin-M]

What if we just post further comments in a new discussion?

 

[berkeman]

That would be good too. I'm only able to move posts, not copy/paste posts. So maybe start a new discussion in the Thermo forum based on the discussion here. It's a pretty interesting discussion, IMO.

 

[collinsmark]

Can you folks suggest which posts I should split off into the Thermo forum? I don't want to mess up the astrophotography part of the discussion.

What if we just post further comments in a new discussion?

That would be good too. I'm only able to move posts, not copy/paste posts. So maybe start a new discussion in the Thermo forum based on the discussion here. It's a pretty interesting discussion, IMO.

I would suggest, if @Devin-M agrees, for @Devin-M to create a new thread in the appropriate forum (Thermodynamics?) with basically a copy-and-paste copy of Post #1558 as the original post.

Then copy over posts #1561 - Onward, to the new thread. The other posts are good for the astrophotography thread.

Post #1559 by @Drakkith is a toughy though; that post could go either way.

 

[Devin-M]

Moved here:
https://www.physicsforums.com/threads/infrared-detectors-the-2nd-law-of-thermodynamics.1011817/

 

[Devin-M]

My most recent astro-photo-session was a terrible failure. Long story short I spent hours in the cold & this was the only remotely salvageable shot from the whole session and it was done with my iphone. I call it “Tree with iPhone in Bortle 2 conditions.” I was attempting to shoot a dim target at 2130mm focal length at the limit of exposure time (90 sec per subframe) I can do without too much tracking issues and also at the limit of iso sensitivity I can reasonable achieve (6400iso). Unfortunately that wasn’t enough exposure time so while the stars came out ok there was no nebula visible in the final photos. I thought I’d be able to bring it out in post processing but it was just a bunch of noise. I’m going to have to choose a brighter target, switch to a much shorter focal length with bigger aperture, or shell out thousands of $… for now I’m ruling out the 3rd option.

Edit: I was too embarrassed to show the final image taken through the telescope but here it is, for posterity:

Here are a couple more photos of the same region through a much more capable scope of NGC 2170:

Source:
https://www.astrobin.com/2nzzzj/0/

Source:
https://en.wikipedia.org/wiki/NGC_2170

 

[Drakkith]

@Devin-M what were you shooting with?

 

[Devin-M]

@Devin-M what were you shooting with?

It was a 2130mm f/14.2 Maksutov Cassegrain w/ Nikon D800 dslr fitted on a Star Adventurer 2i mount (slightly modified to go at least 3x over the weight limit).

 

[Drakkith]

Dear god, f/14.2?!
It'll be the heat death of the universe before you get a deep sky photo finished!
Get a focal reducer for that scope!

 

[Devin-M]

Dear god, f/14.2?!
It'll be the heat death of the universe before you get a deep sky photo finished!
Get a focal reducer for that scope!

I got this one of the core of Andromeda a couple days before with identical exposure settings...

 

[Drakkith]

I got this one of the core of Andromeda a couple days before with identical exposure settings...

M31 (Andromeda Galaxy) is amongst the brightest of the deep sky objects. You can get a decent picture with two tin cans and a string.

Seriously, my F/8 scope is something close to 4x 'faster' than yours and I still think it's too slow!

 

[Ibix]

You can get a decent picture with two tin cans and a string.

The bottom of a bottle, my optics prof used to say.

 

[Devin-M]

M31 (Andromeda Galaxy) is amongst the brightest of the deep sky objects. You can get a decent picture with two tin cans and a string.

Seriously, my F/8 scope is something close to 4x 'faster' than yours and I still think it's too slow!

My counterweight is a 600mm f/9…

 

[Devin-M]

Does anyone else think I've captured a very strange transient signal?

view in WorldWideTelescope

 

[Devin-M]

Here's a cleaner re-process... (ps this is a stacked image... 84x 90 sec (2.1hrs), 6400 iso, 2130mm f/14.2)

view in WorldWideTelescope

 

[Devin-M]

Mystery solved... it's not in the individual frames so it must be nothing more than a stacking misalignment.

 

[Devin-M]

Restacked... now you can see the nebulas... I think something very bad happened during the 1st stacking attempt...

view in WorldWideTelescope

 

[Drakkith]

Restacked... now you can see the nebulas... I think something very bad happened during the 1st stacking attempt...

Was this taken from a bortle 2 site?

 

[Devin-M]

Was this taken from a bortle 2 site?

Yes it’s in bortle 2 conditions just outside of Shingletown, California, USA. It’s literally the same image I already posted but restacked… I think some very bad stacking misalignment happened the first time but it was quite hard to notice for some reason. I think with 90 second exposures a lot of the individual shots come out fine but over time the field drifts in the viewfinder so first I cropped all the source TIFs to the same field of view and then I restacked them without any dark or flat calibration frames and finally histogram stretched the stacked TIF in Adobe Lightroom.

 

[Drakkith]

Okay. It just looks really noisy for 2 hours of data in a bortle 2 spot. But I don't have your equipment, so this may be entirely normal for your setup.

 

[Devin-M]

It’s not a cooled camera & shot at 6400iso instead of a more ideal 100-400iso.

 

[Drakkith]

It’s not a cooled camera & shot at 6400iso instead of a more ideal 100-400iso.

I don't use DSLR's, so I'm not familiar with ISO settings. Why shoot at 6400 instead of 400ish?

 

[Devin-M]

I don't use DSLR's, so I'm not familiar with ISO settings. Why shoot at 6400 instead of 400ish?

As I understand it, it narrows the dynamic range of the sensor but puts the minimum detected values within the range of detection of the 14bit RAW files (on the Nikon D800 body). (otherwise certain various analog-to-digital detectable values will be normalized to 0 on the raw files). Above 6400iso (as I understand it, on the D800), the improvements are entirely digital as opposed to improvements in the analog to digital conversion).

 

[Andy Resnick]

As I understand it, it narrows the dynamic range of the sensor but puts the minimum detected values within the range of detection of the 14bit RAW files (on the Nikon D800 body). (otherwise certain various analog-to-digital detectable values will be normalized to 0 on the raw files). Above 6400iso (as I understand it, on the D800), the improvements are entirely digital as opposed to improvements in the analog to digital conversion).

That is not my experience (I shoot with a D810). I always shoot with as low an ISO as possible (ISO 64) to maximize the dynamic range, especially color information. Shooting with higher ISO values only increases the amount of noise in my stacked image.

 

[Devin-M]

That is not my experience (I shoot with a D810). I always shoot with as low an ISO as possible (ISO 64) to maximize the dynamic range, especially color information. Shooting with higher ISO values only increases the amount of noise in my stacked image.

Right, increasing the ISO lowers the dynamic range.

narrows the dynamic range of the sensor

But there are only so many brightness values you can store in the 14 bit raw file.

When shooting a very dim object would you rather accurately store in the RAW file the brightest analog to digital values or the dimmest analog to digital values? Turning up the iso means you are more accurately storing the dimmest A to D values in the raw files but that’s why it also makes the ordinarily dim sensor noise more visible.

Essentially by turning up the iso I am “throwing out” the brightest values recorded by the sensor (like the brightest stars) by recording them as simply “100% brightness” in favor of more accurately recording the dimmest A to D values (like the nebulas) rather than “throwing them out” by recording the dimmest values to the 14 bit raw file as “0% brightness”

 

[Devin-M]

Essentially, if you shoot a long exposure dark frame at 100iso with the lens cap on, all you are recording is the sensor noise, and you will barely see any noise in the final image it will just look almost fully black.

Now shoot one at 6400iso with the same exposure time (also with the lens cap on). Now the noise itself has more dynamic range because it takes on a greater range of values in the raw file.

So increasing the ISO reduces the overall dynamic range the sensor is able to record by “clipping” the brightest parts of the image if you have some bright areas without the lens cap on but it will increase the dynamic range in the raw file of the dimmest objects.

 

[collinsmark]

That is not my experience (I shoot with a D810). I always shoot with as low an ISO as possible (ISO 64) to maximize the dynamic range, especially color information. Shooting with higher ISO values only increases the amount of noise in my stacked image.

A good rule of thumb for deep sky (not planetary) astrophotography is once you decide on an exposure time (usually determined by how long your equipment can maintain proper tracking, or how long, on average, you can go without a stray cloud messing things up), you should generally increase the gain (ISO) as high as you can go without saturating stars. This generally minimizes the read noise, measured in electrons.

Check your camera's sensor's specs to be sure for that particular sensor, but its almost always the case that higher gain (i.e., higher ISO) reduces read noise.

As @Devin-M mentions, increasing gain (ISO) sacrifices dynamic range, but it reduces read noise and increases resolution. So long as you're not saturating stars, you don't need the dynamic range anyway.

Keep in mind that whichever gain (ISO) you choose, you should take DARK and FLAT (and BIAS and/or DARKFLATs) frames with that same gain (ISO) setting.

Here's a spec sheet of a typical camera:

Note that when the gain (ISO) increases, the read noise in units of electrons, decreases. Also note that for this camera (above figure), there is a discontinuity in some of the curves right around a gain of 60, above and below different circuitry is switched in or out. This is why you should check the specs of your particular camera's sensor, so you know what to expect regarding different gain (ISO) settings.

Because higher gain means finer resolution (electrons per ADU), sub-frames with higher gain might show higher noise in units of ADU, but the signal is also greater in units of ADU. And actually, the signal to noise ratio is generally better at higher gain (higher ISO). That's why read noise is typically specified in units of electrons rather than ADU units, because units of electrons more closely tracks signal to noise ratio.

Just make sure you're not saturating stars. If you're saturating stars, lower the gain (i.e., lower the ISO).

==============

If you were like me, you might be saying, "Woah, now. That's crazy. I've done terrestrial photography, portraits, landscapes, etc., for years. And using higher ISO always makes noisier photos. Lower the ISO if you want less noise!"

The flaw in that argument is when doing terrestrial photography, the tradeoff is between ISO and exposure. If you increase the ISO you must correspondingly decrease the aperture or increase the shutter speed (i.e., decrease the exposure time). (Here, "exposure" is aperature \times shutter speed). Increasing the ISO means you'll have to decrease exposure to avoid blown highlights.

Most of the noise in that case is ultimately the result of reducing the exposure. Fewer photons hit the sensor, increasing shot noise. Less signal is allowed to reach the sensor, so that decreases signal the noise ratio. That, and as previously mentioned, the read noise in units of ADU can increase due to the finer electrons per ADU resolution.

However, in deep sky astrophotography, there is no such tradeoff (planetary astrophotography is a different matter -- let's leave planetary astrophotography out of this discussion). For deep sky astrophotography exposure times are kept as long as you can, given your equipment and cloud cover. And there's almost no sane reason why would ever reduce your aperture. Thus there is no tradeoff regarding ISO and exposure.

So for deep sky astrophotography, the rule of thumb is once you pick your exposure time, pick your gain setting (ISO) such that stars are not saturated. There's usually no need to reduce the gain (ISO) much lower than that.

(And don't forget that your DARKS, FLATs, DARKFLATs, and/or BIAS frames must use the same gain (ISO) setting as your LIGHTs.)

 

[Devin-M]

Except the Nikon D800 series uses 14 bit raw files which is fairly good to great by DSLR standards but not necessarily Astro-specific camera standards, which might be 16 or 32 bit per color channel raw files… so if you keep the stars from saturating with a low iso setting on a dslr you might be sacrificing some of the dynamic range recorded of the nebulas which are much dimmer than the stars. Sometimes you’re shooting through a clip in narrow band filter which reduces the photon count drastically, and you’re going to edit out the stars anyway before compositing with an rgb image of those same stars so you might as well get even more dynamic range on the nebulas by sacrificing the dynamic range of the stars which you’re going to edit out with the dust and scratches filter for compositing with rgb.

 

[Andy Resnick]

A good rule of thumb for deep sky (not planetary) astrophotography is once you decide on an exposure time (usually determined by how long your equipment can maintain proper tracking, or how long, on average, you can go without a stray cloud messing things up), you should generally increase the gain (ISO) as high as you can go without saturating stars. [snip]

That's true- and with my setup, I saturate pretty quickly. Typically, the brightest stars within a field of view saturate between 6s and 20s exposure times (ISO 64).

Edit- I forgot to mention your comment "And there's almost no sane reason why would ever reduce your aperture.", because for me, there are at least 2 good reasons. First, (slightly) stopping down the aperture makes the images less susceptible to poor seeing conditions. Second, (slightly) stopping down the lens improves the images by decreasing aberrations.

 

[Devin-M]

That's true- and with my setup, I saturate pretty quickly. Typically, the brightest stars within a field of view saturate between 6s and 20s exposure times (ISO 64).

Edit- I forgot to mention your comment "And there's almost no sane reason why would ever reduce your aperture.", because for me, there are at least 2 good reasons. First, (slightly) stopping down the aperture makes the images less susceptible to poor seeing conditions. Second, (slightly) stopping down the lens improves the images by decreasing aberrations.

This was 9 shots * 5 minute exposures per shot @ 6400iso @ 600mm f/9 through a 6nm clip in narrowband filter on a D800, stacked and histogram stretched in Adobe Lightroom...

https://www.physicsforums.com/threads/our-beautiful-universe-photos-and-videos.800540/post-6464705

 

[collinsmark]

Edit- I forgot to mention your comment "And there's almost no sane reason why would ever reduce your aperture.", because for me, there are at least 2 good reasons. First, (slightly) stopping down the aperture makes the images less susceptible to poor seeing conditions. Second, (slightly) stopping down the lens improves the images by decreasing aberrations.

Yes, that's right, I'll clarify. If you're using a fast lens that's primarily designed terrestrial photography in mind, then yes, stopping down the lens a little can be a good idea. But if using a telescope designed specifically for astronomy (cheapy* telescopes possibly excluded), stopping down the aperture is pure sacrilege!

*(And to be clear, I'm not necessarily knocking cheapy telescopes. A cheapy telescope can still be better than no telescope, if you know what you're using. Case in point: chromatic aberrations don't matter when you're imaging in narrowband. [Edit: as in actual narrowband with separate SII, Ha, Oiii filters and a monochrome camera; not necessarily the filters designed for one shot color (OSC) like the Optolong L-Extreme.])