Canon 80D - Milky-way and stars Photography at Cherry Springs Park - PA

[dudeinwashingto]

Hello Canoners

Cherry Springs State park has been rated as "gold" by the dark sky association and i plan to visit it this memorial day weekend

I have a canon 80D and EF-S 18-55mm IS STM & EF-S 55-250mm IS STM Lens

Ravelli APGL4 70" Tripod with Adjustable Pistol Grip Head 

Celestron SkyMaster Giant 15x70 Binoculars

and Lots of coffee beans and an excellent percolator 🙂 🙂 

Would love to hear from you all of the optimal settings for photos and videos.

thank you
Paul

[TCampbell]

---

@Sandeepvirk wrote:
T campbell any suggestions for 80d 50mm 1.8

---

A few things come to mind... but mostly what comes to mind is that 50mm isn't a very wide focal length for taking long exposures and if you take short exposures you'll need to get aggressive with boosting the exposure and that means you're going to be dealing with noise issues.

The overall technical challenge for astrophotography is all about the "signal to noise ratio" (SNR).  

If you put a camera on a tripod and point it at a non-moving subject ... but in very little light ... use ISO and just take a very long exposure (if the camera isn't moving and the subject isn't moving you can do that) you'll end up with a very clean image (virtually no noise).

The problem with astrophtography is that it isn't a stationary subject... it's moving slowly (or more accurately... the Earth is moving slowly... about 15 arc seconds per second).

The 80D sensor performance seems to have extremely good dynamic range (much better than any other mid-range model or Rebel model) and it's graph of ISO vs. dynamic range suggests it's probably using an ISO-invariant sensor.  

What's that mean?

You've probably heard about the "exposure triangle" but in reality that's a bit of a myth.  The idea suggests that ISO is part of exposure and that increasing the ISO increases the exposure.   I can certainly see how people would get the idea since cranking up the ISO but leaving the other settings alone results in a brighter (but noisier) image.

But if you consider how a digital camera actually works... you'll realize ISO isn't really part of the exposure.

The sensor has it's base ISO (some level of sensitivity that it naturally has without applying any "gain" or amplifying the signal in any way).  When you take a photo... you (or the camera) select some apeture value and some time value and this allows some amount of light through the lens for some amount of time ... you've collected your photons of light on the sensor.  

At this point, the shutter closes and the exposure is complete.  Notice I didn't say anything about ISO.

THEN (after the exposure is complete) the camera looks at your ISO setting and it applies amplification to the data.  Most Canon models apply a certain amount of analog gain (not digital) before the information is converted from analog to digital (ADC = analog to digital conversion).  How much they do depends on the camera model.  At some point the camera has reached the practical limits of analog gain and switches to digital gain (post ADC it starts multiplying the numbers).  But this isn't really much different than just cranking up the exposure with computer software.  

An ISO-invariant sensor does only digital gain... there is no analog gain at all.

The catch is that this is a "digital" camera.  If you capture your images as RAW files (.CR2 ... and for astrophotography that's the only reasonable choice) then your camera is limited to 14 bit values (2^14 = 16,384) roughly 16 million values per pixel.  If you want things to be twice as bright, you just multiple the math values by 2 (easy enough).  But this creates a problem... anything that was already greater than half the max brightness (brighter than 8192) will get clipped.  Suppose the value was 8193... and you multiply that by 2... you get 16386... and you can't hold a value greater than 16384... (I'm doing a bit of fudging here becuase the values are really 0-16383... not 1-16384).    That means you lose some of your dynamic range.  

As you boost ISO, you also boost "noise".  Noise is always present in every image (even at ISO 100).  But the noise at that ISO is so very subtle that you can't see it ... your eyes wont detect it.  But as you amplify the "signal" you also amplify the "noise" right along with it.  The only way to keep the noise low is to avoid needing to amplify the image values in the first place.

But you can improve an image by doing a non-linear "stretch" of the data.  In other words bring up the darks and mid-tones more than you bright up the highlights... that way things don't "clip".  "levels" and "curves" adjustments in photo editing software let you do this... (non-uniform or non-linear "stretching" of the brightness values.)

With a 50mm f/1.8 lens, the longest exposure you can take on a stationary tripod without a "tracking" head is found by

1.  Multiply the focal length of the lens by the crop-factor of the camera to compensate for the crop.

e.g. 50 x 1.6 = 80

2.  Divide 500 by the result (80).  

e.g. 500 ÷ 80 = 6.25 

That's the number of seconds you can expose without needing to use a tracking head.  Notice that's not a lot of time.

You are going to get a lot of under-exposed images ... that's going to be tough because it means you're going to either need to crank up the ISO in the camera... or crank up the exposure in your processing software (e.g. Photoshop).  This is going to result in a fair bit of noise.

There are three choices that come to mind:

1)  Select a lens that will allow you to track longer and collect more photons.

Ideally that would be the Canon EF 24mm f/1.4L USM II... but that's an expensive lens.  Rokinon makes a manual 24mm f/1.4 lens that is inexpensive.  But as their lens is completely manual (no auto-focus and the camera cannot control the aperture) you probably wouldn't want to use that lens for regular photography ... so it's a purchase just for astrophotography.  

However... at 24mm the exposure time increases to 500 ÷ (24 x 1.6) = 13 seconds (just over double of what you're getting now... in fact you could probably just take it 15 seconds).  At the same time it's an f/1.4. lens... collecting nearly (but not quite) double the light as your f/1.8.  This means that the net of it is that you'll get about 4x more light gathered because it collects light nearly twice as fast... and collect it for about twice as long.

Canon's lens is just over $1500... Rokinon's lens is about $550.  

What about other options?

2)  The next option is to get a star-tracker.  Assuming you already have a sturdy tripod, you could buy a star-tracker such as a Sky Watcher "Star Adventurer" or "Star Adventurer Mini"... or the competing brand is iOptron and they make the "Sky Tracker Pro" and "Sky Guider".  

The Star Adventurer Mini and the Sky Tracker Pro are about $300.

The Star Adventurer and Sky Guider are their higher-end trackers (beefier and handle more weight) but run more like $400-450).  Technically you can get these for a bit less but they have accessories which you really do want and buy the time you have a decent package... it's in the $400-450 range.)

Remember... this ASSUMES you already have a sturdy tripod (if you don't... add that to the price.)

With this type of gear you can take much longer images and the whole problem of maximum exposure duration goes away.  There is the notion that the land will blur (because the tracking head rotates to compensate for the spin of the Earth) but you can take a single long exposure of the "landscape" and a long exposure of the sky and composite the two.

But this still represents an investment of at least another $300.

So what else can you do?

3)  Learn to process the noise.

When you take a short image (or at least what astrophotographers think of as "short"... most other photographers think a 6 second image is "long") you wont collect much light... and then you'll crank up the exposure in your post processing software (e.g. Photoshop).  This is going to result in "noisy" images.

You can process down the noise... but it comes with trade-offs.  The way the noise processing software works is by analyzing the value of each pixel in comparison to its neighboring pixels... and then "average" them out.  But doing this almost means that where pixels SHOULD deviate from their neighbors (real information) that ALSO gets averaged out.  The result is an overall "softening" of the detail you want... in addition to the reduction in noise.

What if you could have both?

There are some clever de-noising programs that are a bit smarter about trying to selectively reduce noise rather than just do a global de-noise to the whole image.

Another technique is to use Photoshop (or Photoshop Elements... I don't have Elements but I think you can do this) and build a "star" mask.

In Photoshop, you can "mask" out parts of an image that you don't want touched.  Imagine if you could take a pice of card-stock... and cut holes in it... and then lay it on top of a photo knowing that anything you adjusted could only happen to the bits that were exposed via those holes.  In a Photoshop mask, anything "black" in the mask means it blocks those pixels from being affected.  White is like the "hole" ... anything below a white area in a mask will be affected.  If it's gray... it'll be moderately affected.  You get the idea.

In Photoshop, you build the star mask using the following steps:

1. Dupliate your base layer
2. Add a mask to the duplicated layer (when you do this, Photoshop selects the mask as the "active" area by default)
3. In your new masked layer, click the "image" (not the mask) to make the "image" section active.
4. Select the whole image (Ctrl+A on Windows... Cmd+A on Mac)
5. Copy the image (Ctrl+A on Windows... Cmd+A on Mac)
6. Click back on the "mask" section in the layer list to make it active.
7. Alt+Click the "mask" section to display the mask in the main image area of Photoshop
8. Paste the image you just copied to the mask by using Ctrl+V on Windows or Cmd+V on Mac (you will notice it pastes a black & white version of your image... masks can ONLY be gray-scale)
9. It's a good idea to make a "back up" copy of this mask.  To do this, Ctrl+C / Cmd+C to "copy" the mask, then Ctrl+V / Cmd+V to paste the mask to a new layer.  You can de-activate that layer since we don't need it right now.. it's just a backup.
10. The stars in the mask image might be a bit too sharp (the mask works better if they are slightly softened... it improves the blending between masked and unmasked areas).  Blur it by picking "Filter" -> "Blur" -> "Gaussian Blur.." (from the main menu) and set a radius of perhaps .5 pixels then apply the blur.
11. Unfortunately this mask has "white" stars and a "black" background... we want the opposite.  Click Ctrl+I on Windows or Cmd+I on a Mac to "invert" the image. (you can also do this via pull-down menus)

At this point you now have a copy of your image with a layer mask that is protecting the stars... but not the background.  You may want to adjust this mask a bit ... for example you can select the mask with Alt+Click on the mask and then do a "levels" or "curves" adjustment to it ... if for example you want to eliminate the milky way and just have the stars.)

NOW you can attack the image more aggresively becuase your stars are protected.  If you want to apply a more significant boost to the Milky Way glow without blowing out the stars... you can do that.  You can also apply much more aggressive de-noising without worrying that you are going to soften the stars.

Anyway... that's the general idea.  Masks are your friend.

[Sandeepvirk]

Thnx so much for ur knowledge
Really helpfull