What camera should I buy for night photography

[jsecordphoto]

Sony A7s

The A7s is no doubt amazing for night photography, and landscapes in general. What keeps me from wanting one is the relatively low megapixels. Cant print very large with 12mp.

 

[chuasam]

Sony A7s

The A7s is no doubt amazing for night photography, and landscapes in general. What keeps me from wanting one is the relatively low megapixels. Cant print very large with 12mp.

That's rubbish. I've shot posters with a D70s and billboards with a D700.


Sent from my iPad using Tapatalk

 

[jsecordphoto]

Billboards I understand, considering you view them from hundreds of feet away, they can be like 10dpi. Maybe I'm just more demanding of prints than you are, because I wouldn't print much bigger than 12x18 or maybe 16x24 (maybe) with a 12mp file

 

[Didereaux]

Best $ value right now is the Canon 6D a VERY versatile low noise body. Essentially a 5D mkIII with fewer focus points. Exceptionally low noise levels at high ISO's. This allows better IQ because you cna shoot stars at faster speeds, so no trails. And a superb portrait camera.

 

[runnah]

Astrophotography is very different and has many special considerations.

90% of the normal matter in the universe is composed of Hydrogen atoms. This means many deep-space structures are glowing specifically in the Hydrogen emission lines... of which the Hydrogen alpha line is the strongest. Unfortunately the sensors in a typical camera are designed to mimic the sensitivity of the human eye, which is mostly sensitive to the greens in the middle of the spectrum and less sensitive to the blues and reds. As such, just in front of the imaging sensor on the inside of every digital camera there's a filter. That filter trims light emission in a way intended to mimic the eye... which means roughly 80% of the hydrogen alpha light is rejected and that makes for extra-long exposure times when doing astrophotography.

You can use a typical camera (any camera with a "bulb" setting and manual focus would work) EXCEPT you'd need to run extra-long exposures. Exposure times are already long enough and tracking for several minutes on end can be a bit of a challenge (if tracking is off then the whole image is blurred.) The long exposure also generates sensor heat and sensor heat generates "noise".

Astro-imagers will either use special cameras that don't have filters at all (they use a filter-wheel to dial-in the filtering they want... typically R, G, B, and Luminance ... or they might go for narrowband filtering (Ha, Hb, OIII, etc.)

A lot of imagers will buy a normal (usually "used") DSLR and then modify the camera (The Gary Honis website has instructions on how to do this with a lot of Canon models: ASTROPHOTOGRAPHY & DIGITAL IMAGING by Gary Honis )

You can also buy a DSLR specifically pre-modified for astro-imaging. The Canon 60Da is currently the only DSLR on the market that does this (Canon replaces the normal filter in the camera with a special filter optimized for astro-imaging... but this means that if you use the camera for normal terrestrial photography then you'll get ESPECIALLY WARM images (lots of reds) -- though you an "white balance" that out.)

The other consideration is that imagers use the AC power supply which is optional for most cameras. When you buy a Canon 60Da, it actually includes the AC power adapter (normally that's a separate accessory for any other camera). They do this because astro-imagers typically take LOTS of sub-frames... e.g. I might take 16 "lights" at 4 minutes each (64 minutes of exposure time), plus a minimum of 8 "darks" at the same ISO and duration (32 minutes of exposure time), plus a bunch of bias frames, and possibly even a bunch of flat frames (those can be faster.) But this means I'm probably working the camera almost non-stop for 2 hours of image just to capture ONE final image. You can imagine how that'd kill your batteries in a hurry.

Since the image times are long, you'll need a way to "track" the sky as it moves. To do this, you'll need a tracking mount.

There are trackers designed to be used with ordinary photography tripods:

1. Losmandy StarLapse system (this is my personal favorite... this company makes VERY high quality products.)
2. AstroTrac
3. Vixen Polarie
4. iOptron SkyTracker

These are motorized heads that mount to a regular tripod, but you align them so that their axis of rotation is pointed at the north celestial pole (so it's parallel with Earth's axis of rotation). That means that as Earth rotates from West to East (yes... West to East... I did not get that backwards), the tracking head rotates at the same speed... but from East to West. This means any object being imaged will remain stationary in your field of view (as long as you did a good job aligning the tracking head when you set it up.)

You can also mount the camera to any motorized equatorial telescope mount (such as an iOptron SmartEQ or a Celstron Advanced VX mount.) but these will be heavier to transport and more expensive.

It is MUCH easier to image in wide field (wide angle of view) then long telephoto (narrow angle of view) because the narrower the angle of view, the more critical it is to have accurate tracking (you get to be slightly sloppy when the angle of view is wide). As you get better.. you can up the ante by using longer focal length lenses.

While I normally don't care what brand someone uses for their DSLR (I don't think it's very important), the support for Canon cameras in astro-imaging is HUGE. Just about everything supports Canon. Support for any other brand is extremely weak. You'd think support for Nikon would be about as common... but it's not.

There is one other nuance that I encountered while helping a camera club image the moon at our observatory... the auto-focus won't work on stars... they're not bright enough. So you'll need to manually focus the image. To focus, we'd normally pop the camera into "live view" mode, crank the magnification to 10x, point to a bright star, and then work on focusing that star as best as possible using manual focus (I prefer to use something called a Bahtinov Focusing Mask... but that's another thread.) On any Canon I've used, this is fairly easy because you can just max out the ISO and shutter duration and the live-view amplifies the image to "simulate" your exposure. When I did this with the camera club's Nikon cameras, I could not get this to work. The consensus was that most Nikon cameras don't support this feature. This made it EXTREMELY difficult to focus the cameras. Hopefully Nikon will add this feature. (Canon and Sony both support it, but I don't recommend Sony because if you think it's hard to find astro-imaging software that supports Nikon... wait until you try to find software that supports Sony. Support is pretty much non-existent.)

So the top DSLR for astro-imaging would be the Canon 60Da.
Below that a "used" Rebel (something you're not afraid to void the warranty when you modify it) are popular.
Ignore any features of the body & focus system, metering, auto-focus points, etc. etc. etc. You won't use ANY of those features when doing astro-photography (no metering... no auto-focus, etc.) so it just doesn't matter. You basically are buying the camera for the sensor and the ability to take pictures in "bulb" mode (and they all do that). Although an articulated LCD screen is VERY nice because the camera is going to be pointing up ... no articulated LCD screen means you'll be on your knees down on the ground trying to use the camera. That makes the swing-out articulated LCD screen a very welcome feature. Apart from that, older/used low end bodies are just fine.

Thanks for the advice! But sending me to that site has given me severe eye damage!

 

[liam jacobs]

5d mark iii is the best option for shooting at nights.

 

[W.Fovall]

5d mark iii is the best option for shooting at nights.

no...
old noisy sensor.
6D has less noise.
7dmii has the least noise on a crop c.
5dmiii is for fast focus and fast fps.

 

[scorpion_tyr]

I use a 60D and a wide angle lens (10-15mm's) for most of my night time landscapes. I shoot a lot of star trails when I go out at night, mostly because I'm too close to cities to shoot the milky way. Here's what I can tell you, if you're wanting to get good foreground in the shot you can light it up with a flash, paint it with a flashlight, or let the moon light it up for you. You'll want to learn as much as you can about how focal length and aperture affect depth of field and how to manually focus for the desired effect. You'll definitely need a GOOD tripod, a $60 Manfrotto at the bare minimum. You'll also want a wired remote so you don't cause camera shake by pressing the shutter release button. The attached picture was taken with a 60D and Sigma 10-20mm f/4-5.6. The same concepts apply to non-trail pictures except you'll be using a faster shutter speed and higher ISO than I normally do.

 

[jsecordphoto]

I had a 60d before I picked up my 6d, and man did that camera suck at handling noise. The norm for milky way shots, for me, is iso3200. At that ISO there was a lot of noise just in the sky, never mind my foreground. Is it possible to do with a crop sensor? Yes. But I quickly outgrew my crop body and wanted something more suited for high iso/long exposures. If this is the type of photography you want to get into, it's better to buy a full frame to start off if you can afford it, rather than buying crop and wanting to upgrade fairly quickly.

 

[The_Traveler]

I am puzzled by what tcampbell said about using dark images to mask out noise.
I had always assumed that noise was random across a sensor.
Am I wrong and are certain pixels more liable to noise thus generating a specific noise pattern for each sensor?

 

[qleak]

I am puzzled by what tcampbell said about using dark images to mask out noise.
I had always assumed that noise was random across a sensor.
Am I wrong and are certain pixels more liable to noise thus generating a specific noise pattern for each sensor?

Long exposures tend to have biases that appear that are not modeled by white noise.

Think about water flowing over rock for long periods of time, eventually groves are formed in particular patterns. Similarly using a sensor for long exposures introduces electron flow pathways.

You can also think of the toaster effect. More heat => More resistance => More heat. So by running current for a longer period of time, certain pixels may get hotter.

These are highly sensitive to initial conditions, which is why people produce sometimes multiple dark frames per session.

 

[AsmaMohd]

Check out the 17 Best Cameras For Digital Low-Light Photography: Tested and Compared! at http://www.adorama.com/alc/0012810/article/15-Low-Light-High-ISO-All-Stars

 

[TCampbell]

I am puzzled by what tcampbell said about using dark images to mask out noise.
I had always assumed that noise was random across a sensor.
Am I wrong and are certain pixels more liable to noise thus generating a specific noise pattern for each sensor?

There is both "random" noise and "pattern" noise. Some pixels are just particularly sensitive and noise will show up on those pixels over and over. The dark frames server two purposes... one is to detect pattern noise, and the second is to assess just how much "noise" there is and this is where the "bias" frames come in.

A true "astrophotography" imaging camera has the ability to take a "bias" frame. Here's the idea. The sensor needs electricity in order to function (it's not solar powered). So the camera has to charge up the sensor to take an exposure and then do a read-out. Imagine then that if all you did was charge up the sensor, NOT open the shutter, and immediately read out the sensor (with a 0 second exposure time) that you'd actually "data" back. But none of that data actually represents image signal (because you never took an image.)

So you take a bunch of "bias" frames (basically set your camera to a fast shutter speed and take a bunch of frames without removing the lens cap) and the computer can assess how much "charge" reads out of a sensor that never actually took an image. THEN you take a bunch of "dark" frames. Any amount of charge over and above the "bias" charge represents "noise" in the image. So using "darks" with "bias" allows the computer to assess how much is actually "noise" vs. how much is just "bias charge".

And then of course you take a pile of "light" frames too. There are algorithm to average or use sigma clipping, etc. to statistically find the outliers and reject those to remove noise. But the "pattern" noise won't show up as an outlier because it'll be consistent. Fortunately the "dark" frames should help you detect the pattern noise. I may have mentioned we also sometimes use a process we call "dithering" where the telescope is moved enough to relocate stars by just a FEW pixels between each frame. The stacking software will use the stars to re-align but the cool thing is that since we nudged the scope a very tiny amount, all "real" data will move relative to the amount of our nudge... but all "pattern" noise will stay in the same spot. This allows the computer software to easily detect the difference between "noise" and just really really dim stars (that might be mistaken for noise.)

 

[qleak]

I may have mentioned we also sometimes use a process we call "dithering" where the telescope is moved enough to relocate stars by just a FEW pixels between each frame. The stacking software will use the stars to re-align but the cool thing is that since we nudged the scope a very tiny amount, all "real" data will move relative to the amount of our nudge... but all "pattern" noise will stay in the same spot. This allows the computer software to easily detect the difference between "noise" and just really really dim stars (that might be mistaken for noise.)

I heard once that dithering was discovered during world war II when they noticed the mechanical bomb trajectory devices were more accurate when on an operating airplane. The small movements and vibrations of the plane helped average out the errors in the mechanical computers.

This sounds more similar to your explanation of dithering in astronomy than the usual use of dithering I'd heard in changing between bit depths in audio / visual media files

 

[weddinglight]

What is the process for the shots????