Post your low light long exposures

[RockyNH_RIP]

I am not that clear of light pollution but I want to try anyway, just for the experience... I may work on a setup this weekend...

Thanks for the "motivation"!

Pat in GA

 

[Scott Murray]

You don't need to believe me and if you prefer shooting at 6400 more power to you but in case you want to know what actually happens, this site is rather interesting.

www.Clarkvision.com: ISO and Digital Cameras, ISO Myths

I tested it and it seems to be correct, at least for those shots during those conditions. Before I always shot high ISO for stars but now I tried both, low does a better job for me.

So you are telling me that all camera sensors record all the light that is available at ISO 100? And all we have to do is press 'Auto fix' in LR or the equivalent and we have an amazing photo ready to share to the FB world. But if we go to blow that photo up to even a wall sized print it would look crap?

 

[Scott Murray]

I have pics with light pollution that works

For example this single lightning bolt works great with the city lights. ( Using lightning trigger)

 

[Scott Murray]

Same lightning but using 30 sec exposure and no trigger.

Its amazing what a micro second does to an image

 

[hark]

[MENTION=9753]Scott Murray[/MENTION] what setting suggestions and focal length suggestions do you have for milky way shots/low light? I haven't tried any and don't know where to start. My main lenses are listed in my signature. Thanks for any advice.

 

[Scott Murray]

@Scott Murray what setting suggestions and focal length suggestions do you have for milky way shots/low light? I haven't tried any and don't know where to start. My main lenses are listed in my signature. Thanks for any advice.

I would suggest.

D610, ISO 3200 (30% moon, no light polution), F2.8, 20 sec (any longer and you get noticeable star movement with FX)

Un-edited file. ISO 4000




ISO 3200 un-edited



Edited 3200 ISO



I hope that helps

 

[J-see]

So you are telling me that all camera sensors record all the light that is available at ISO 100? And all we have to do is press 'Auto fix' in LR or the equivalent and we have an amazing photo ready to share to the FB world. But if we go to blow that photo up to even a wall sized print it would look crap?

Nope, I'm saying that at ISO 100 and ISO 6400 the exact same amount of photons hit the sensor. The only thing that changes is that ISO affects the readout of the meter that is converted to a digital signal.

You can easily show me that the sensor is truly adjustable in sensitivity. Take a shot of the night sky at 1 second f/2.8 and then each following shot you up the ISO one stop and decrease the shutter one stop. If the sensor can be adjusted, the ISO 6400 at 1/60s should not be different from the ISO 100 at 1 second.

If the sensitivity of the sensor increases, which should increase accuracy since it now can measure a weaker signal, why then does noise increase too?

Here; two shots ISO 100 and 6400, the 100 overexposed 3 stops and the 6400 underexposed 3 in post to bring them both to the same exposure. The 6400 should make the sensor much more sensitive if that would be true which logically should result into a better shot showing a lot more stars than the ISO 100. And it is ETTR as prescribed to get the better quality.

If ISO only affect the read-out, 100 should be the better shot and contain as many stars as the 6400.





Here it's explained very simple in a video.

https://www.youtube.com/watch?v=WEApLA-YNko

 

[hark]

I would suggest.

D610, ISO 3200 (30% moon, no light polution)....



I hope that helps

Yes, the info helps. Thanks, Scott! One quick question...when there is light pollution, what changes in ISO/exposure do you make to compensate? Any?

 

[Scott Murray]

Yes, the info helps. Thanks, Scott! One quick question...when there is light pollution, what changes in ISO/exposure do you make to compensate? Any?

I would normally decrease the ISO as I get too much light and it impacts on the final photo. Also you cannot capture the Milkyway during a full moon as there is too much light.

Edit: you can normally tell when there is a full or near full moon as the night sky ends up a blue colour

 

[hark]

Got it! Thanks again, Scott!

 

[Scott Murray]

Got it! Thanks again, Scott!

I know a little bit

Speaking of which, the best time to do star trails is during a full moon as normally you cannot photograph the milkyway. As you can see the moon was coming into frame at the last shot.

 

[SteveL54]

Amazing shots, Scott!

 

[Scott Murray]

Amazing shots, Scott!

Thanks Steve.

 

[thegaffney]

Could I ask why you suggest the moon being at 30%?

I tried taking some last night with my new D600 and my old 17-55 (in crop mode) and it didn't come out very good, the moon wasn't out yet, but it was supposed to rise in an hour or so, and it rises around the area on the left

You can kind of see it going across the middle, this was at ISO 3200 2.8 and 10 sec, if I decreased the shutter speed it just made everything brighter, not just the stars.

But about a year ago, last March I took this photo below with my D7000, same lens,

But this was at ISO 400 2.8 and 60 seconds, BUT the moon had set about 3 hours earlier in the opposite direction

Was the rising moon in last nights photo maybe washing everything out? Bother were in the same spot, with the same amount of light from the city.

Maybe it was just hazy out?

 

[thegaffney]

I wish the north star was that low here, it seems to make it much more dramatic, looks great

 

[Scott Murray]

Could I ask why you suggest the moon being at 30%?

I tried taking some last night with my new D600 and my old 17-55 (in crop mode) and it didn't come out very good, the moon wasn't out yet, but it was supposed to rise in an hour or so, and it rises around the area on the left

You can kind of see it going across the middle, this was at ISO 3200 2.8 and 10 sec, if I decreased the shutter speed it just made everything brighter, not just the stars.

View attachment 134107


But about a year ago, last March I took this photo below with my D7000, same lens,

But this was at ISO 400 2.8 and 60 seconds, BUT the moon had set about 3 hours earlier in the opposite direction

Was the rising moon in last nights photo maybe washing everything out? Bother were in the same spot, with the same amount of light from the city.

Maybe it was just hazy out?

View attachment 134108

Sorry the moon at 30% is great for my lost city shot but it needs to come over my shoulder that is how I made the lost city shot. But yes if going for silhouette etc not advised.

 

[J-see]

Here's another I shot some days ago using different ISO to blend but it can serve as an example.

I shot one at ISO 100 for 30 seconds fully open to get as much light in as possible ignoring the star blur that caused. With those settings it is underexposed.
The other I shot at a faster shutter to avoid star trails, I closed down 2/3th of a stop but increased ISO to 3200. It is as much exposed to the right as was possible without clipping more than needed.






Here are both after processing them to a rather similar version. The one increasing exposure while the decreasing the other.
In terms of quality that's the worst and best approach in regards to noise and other imperfections.






Now let's look at a 100% detail of both shots. They both still have noise; when light is low photon shot noise is very visible. ISO 100 does not magically make that disappear.





While the 3200 shot has a lot of stops more exposure than the 100 in the RAW it captured, when it comes to true sensor exposure, it is the 100 ISO that had more exposure to the light. That's why even when both are fairly similar when processed, it is the better of both shots in regards to noise. 3200 only made its sensor exposure disadvantage worse.

And then there's a severe loss in DR too.

 

[Eyelight]

So, [MENTION=31330]J-see[/MENTION] you figured out that keeping ISO low produces a better image? Or is there more to it?

I've been studying low light long exposure a bit of late so perusing the threads on such.

Can't say that I've managed to follow all the thoughts in the last few pages though. Some of it sounds like shutter and f/stop have no effect, but I'm guessing I'm just missing something.

 

[J-see]

So, @J-see you figured out that keeping ISO low produces a better image? Or is there more to it?

I've been studying low light long exposure a bit of late so perusing the threads on such.

Can't say that I've managed to follow all the thoughts in the last few pages though. Some of it sounds like shutter and f/stop have no effect, but I'm guessing I'm just missing something.

It's actually the opposite; shutter and f/stop are the important parts for exposure.

I've always been shooting high ISO for stars, as I read somewhere, tried and since it worked, kept doing. (<-- the downside of tutorials) But because it works it is not necessarily the better or best method. My other night shots I used ISO 100 since they were mainly landscapes and for those shutter is not limited like stars. But during those I often noticed my sky and stars to be pretty ok even at this low ISO. I didn't understand why that was the case since high ISO increased my sensitivity and this should make a massive difference there.

To know what went on I had to inform myself about the very fundamentals first: how does this cam work? Not the settings but the hardware; what goes on when I expose? While reading more and more about it I stumbled upon that site linked in a previous reply and read that ISO does not increase the sensitivity of the sensor since our sensors don't have multiple sensitivity settings. That's a myth but about everywhere being sold as truth. Our sensors pixels have a certain efficiency and that's all. You can read everything about it in that link.

This implies that you could say there are two kinds of exposures; the exposure of the shot we take and the exposure of the cam, which is the actual exposure of the sensor to light. The only thing that affects the sensor's exposure is the aperture and shutter duration. ISO and everything else affects the exposure of your shot which is post sensor.

During the daytime this isn't very important but for night shots it is very relevant. The more light the better the signal, the less noise but as many of us assume, ISO does not add more light; it multiplies the light already present since it only changes the read-out of the sensor pixels. But while doing, it multiplies the noise too. The only light that makes the difference is the light that hits the sensor and here only shutter and aperture can improve.

Upping the ISO also lowers the brightness levels and in that affects the DR. You gain in the shadows but each ISO stop that gain is less. If we up the ISO we tell the cam that now it should read sensor pixels that are less saturated as if they are more saturated and in that we lose more and more levels that are pushed into 255, fully saturated which shows as white clipping in our shots.

Buckets of water are often used to explain this. If you imagine each sensor pixel to be a bucket filled with water, when we change ISO, we use a smaller bucket instead. We don't measure the water but the amount this water fills the bucket. What was "half a bucket" first, now is "a full bucket". What was "a quarter of a bucket" becomes "half a bucket". What used to be "a full bucket" before now actually becomes "twice the bucket" but since "a full bucket" is our maximum for these measurements, "twice the bucket" is considered as equaling "a full bucket". In all of this there is one constant; the volume of water. Regardless how I measure the water present, that measurement does not change its volume.

We gain at the other end since the multiplication now creates a difference where first was too little. But for these kind of shots the shadows matter less. It's all about the light.

So in the end when you use a fixed shutter and aperture, the only difference that ISO makes is an increase in noise while lowering the DR the cam can grab. If it is possible to shoot at ISO 100, do it, even at night. ISO 100 will not magically make night shots picture perfect nor will it be noise free. Compared to high ISO it is only noise less.

The trick is to maximize the sensor's exposure by increasing the shutter duration as long as possible while opening the lens as wide as possible. The more light hits the sensor, the better the SNR. Once you hit the limits there, do the rest in post.

This evidently is not the case when there is massive clipping to begin with. Why the cam does not store the read-out of those clipped levels I don't know yet but it doesn't and there's next to nothing I can do about that. Or maybe it does but I need to get that read-out somehow. Evidently when pixels have no signal read out or are fully saturated at the native ISO, there's nothing more to get. It'll require some more digging to find the information I lack and understand that part.


Here are two identical shots when it comes to sensor exposure. They only differ in shot exposure because of ISO.

In post I increased the exposure by 3 stops for the 100 while lowering by 3 for the 6400. After that, they're exposed equally in terms of shot exposure and sensor exposure. Then I killed the shadows (+100) and highlights (-100) and set the black and white at the lowest opposite. (+100 black, -100 white).





Look at the 100%.





Of course I can process them in such a manner the differences will be less noticeable but never until they're identical.

Here is another area which shows the exact gain and loss between low and high ISO. This is almost black in the normal shot but I killed black/white shadows and darkness to show what is not there compared to the others.





As you can see, low ISO I pay a price in the darker regions while with high ISO I pay a price in the light one. For these shots the loss in the brighter parts is much worse than the shadows.

 

[Eyelight]

TMIOP << Too Much In One Post. Hard to converse when there''s a novel to read. I'm going to divvy it up and toss a comment in here and there in blue.

It's actually the opposite; shutter and f/stop are the important parts for exposure.

I've always been shooting high ISO for stars, as I read somewhere, tried and since it worked, kept doing. (<-- the downside of tutorials) But because it works it is not necessarily the better or best method. Eureaka! My other night shots I used ISO 100 since they were mainly landscapes and for those shutter is not limited like stars. But during those I often noticed my sky and stars to be pretty ok even at this low ISO. I didn't understand why that was the case since high ISO increased my sensitivity and this should make a massive difference there. Stars are suns and they are very bright, so a low ISO makes perfect sense. I've found stars in nightscapes that I didn't know were there until I was processing.

To know what went on I had to inform myself about the very fundamentals first: how does this cam work? Not the settings but the hardware; what goes on when I expose? While reading more and more about it I stumbled upon that site linked in a previous reply and read that ISO does not increase the sensitivity of the sensor since our sensors don't have multiple sensitivity settings. That's a myth but about everywhere being sold as truth. Our sensors pixels have a certain efficiency and that's all. You can read everything about it in that link. I glanced at the page. Can't say how accurate all the info is, but my understanding is a sensor is just that. It senses what is allowed to get to it. Changes in ISO are changes in how the recorded signal is amplified.

This implies that you could say there are two kinds of exposures; the exposure of the shot we take and the exposure of the cam, which is the actual exposure of the sensor to light. (This is where we need to be careful. Better to come up with a new word or phrase to describe what you're thinking. I almost (maybe did) got lost here.) The only thing that affects the sensor's exposure is the aperture and shutter duration. (OK, makes perfect sense) ISO and everything else affects the [SUB]exposure[/SUB] processing** of your shot which is post sensor. Using good old ISO 400 film, the shutter and f/stop were the only tools of exposure. The film was the film, just like the sensor is the sensor. We could boost the film in processing to get more speed and to do this we shot at the speed we were planning to boost to. Sensors seem to work the same except the boosting or amplifying is done in camera and not in a lab. **less confusing word

During the daytime this isn't very important but for night shots it is very relevant. The more light the better the signal, the less noise but as many of us assume, ISO does not add more light; it multiplies the light already present since it only changes the read-out of the sensor pixels. But while doing, it multiplies the noise too. The only light that makes the difference is the light that hits the sensor and here only shutter and aperture can improve. Need to read what I highlighted in red as it does not mean what you intended it to mean. I think I'm still following the thought though.

Upping the ISO also lowers the brightness levels and in that affects the DR. You gain in the shadows but each ISO stop that gain is less. If we up the ISO we tell the cam that now it should read sensor pixels that are less saturated as if they are more saturated and in that we lose more and more levels that are pushed into 255, fully saturated which shows as white clipping in our shots.

Buckets of water are often used to explain this. If you imagine each sensor pixel to be a bucket filled with water, when we change ISO, we use a smaller bucket instead. We don't measure the water but the amount this water fills the bucket. What was "half a bucket" first, now is "a full bucket". What was "a quarter of a bucket" becomes "half a bucket". What used to be "a full bucket" before now actually becomes "twice the bucket" but since "a full bucket" is our maximum for these measurements, "twice the bucket" is considered as equaling "a full bucket". In all of this there is one constant; the volume of water. Regardless how I measure the water present, that measurement does not change its volume. This is when folks start thinking your bucket is only half full. <<<<Whoa, I just had to toss that in for a little mid post humor, since you walked right into it with all those buckets.

Seriously though, too many buckets.

We gain at the other end since the multiplication now creates a difference where first was too little. But for these kind of shots the shadows matter less. It's all about the light.

So in the end when you use a fixed shutter and aperture, the only difference that ISO makes is an increase in noise while lowering the DR the cam can grab. If it is possible to shoot at ISO 100, do it, even at night. ISO 100 will not magically make night shots picture perfect nor will it be noise free. Compared to high ISO it is only noise less. All digital signals have noise. This is true, but like so many things photographic, it's the acceptable noise we are after and low ISO will always have less noise than higher ISO. When higher ISO setting amplifies the signal, it amplifies everything.

The trick is to maximize the sensor's exposure by increasing the shutter duration as long as possible while opening the lens as wide as possible. The more light hits the sensor, the better the SNR. Eureka! Once you hit the limits there, do the rest in post. I think there is a balance in what we let the camera process and what we do in post and trade-offs.

This evidently is not the case when there is massive clipping to begin with. Why the cam does not store the read-out of those clipped levels I don't know yet but it doesn't and there's next to nothing I can do about that. Or maybe it does but I need to get that read-out somehow. Evidently when pixels have no signal read out or are fully saturated at the native ISO, there's nothing more to get. It'll require some more digging to find the information I lack and understand that part.


Here are two identical shots when it comes to sensor exposure. They only differ in shot exposure because of ISO. 100 ISO should always produce a better shot for an equivalent exposure. I'm not sure it really proves anything when we purposely amplify a signal that did not need amplifying. But, experimentation can be illuminating

In post I increased the exposure by 3 stops for the 100 while lowering by 3 for the 6400. After that, they're exposed equally in terms of shot exposure and sensor exposure. Then I killed the shadows (+100) and highlights (-100) and set the black and white at the lowest opposite. (+100 black, -100 white).

Of course I can process them in such a manner the differences will be less noticeable but never until they're identical.

A sensor needs a certain amount of exposure to record a good image. When we cannot achieve that exposure due to light level, the need for more DOF, stop action shutter-speed, or a combination of these, we can raise the ISO or boost the exposure in post. Raising the ISO is using the cameras processing to amplify the signal and improve the image. I suspect the camera does better when the need is beyond a couple stops, but if within a couple stops it makes sense post would do the trick.