The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Shooting Milky Way shots with the 50mp CMOS backs

tsjanik

Well-known member
Thanks for all of the advice. In the end I used my XF/IQ150 and a Phase One 28D @ f/5. The iOptron StarTracker held the weight ok:

XF IQ150 1600 ISO 3m 30s @ f/5
Fabulous shot Graham. The distant hills and impressionistic, yellow blur of the wheat(?) really enhance the image.


Tom
 

GrahamWelland

Subscriber & Workshop Member
Yellow blur is actually the foreground of a lake bed. The iOptron StarTracker moves the camera to counter the spin of the earth which will blur anything that isn't in the sky. In this case I got away with it due to the shift of the scene up to leave a hard horizon.
 
Sigma 20mm f1.4, ISO 400, 30s

No obvious coma wide open

This is the power of f1.4 - no need for tracking mount / foreground is not blurred

_DSC2117.JPG
 

Paul2660

Well-known member
? No coma

The image is pretty small bit there appear to be plenty on wings on the Stars towards the edge? That is coma.

I have said lens and mine has the same amount wide open or worse. Takes f2.2 before totally gone.

Sure can be cropped out. And I agree about the range of available light wide open sad it marred by coma.

Paul C
 
? No coma

The image is pretty small bit there appear to be plenty on wings on the Stars towards the edge? That is coma.

I have said lens and mine has the same amount wide open or worse. Takes f2.2 before totally gone.

Sure can be cropped out. And I agree about the range of available light wide open sad it marred by coma.

Paul C
No that is not coma. It's star trails. If it was coma then you would expect the same for both corners.

The Sigma 20mm f1.4 was designed to correct coma towards some good extent (when compared against previous f1.4 lenses).

You do not observe obvious coma until you enlarge the picture but before you can observe coma noise would destruct the image quality first if it was shot with smaller aperture.
 

Paul2660

Well-known member
I fully understand star trial photography.

I would need to see your shot at higher res. Star trails are equal in volume of light your shot shows what appears to be coma with a bright star and wings and it only shows in the corners where coma is always the worst. Wings from the center of a brighter subject are not trails trails are the same in brightness even at 30 sec.

I realize at 30 sec you will show trails so I can't tell for sure from your image.

However your 30 seconds is too long an exposure as you will have bluer in the Milky Way at exposure times that long. It will now show at the size of your posted image but should show at full size. The longest exposure I have seen that works is around 15 seconds maybe 17 depending on the night.

As for the 20 Sigma 1.4. Do a bit of reading from others and please don't quote the marketing hype from sigma to me. I have followed this lens from day of annoucement and it not coma free wide open. No one who has shot one has reported this either and there are plenty of reviews showing the effects of coma. I had high hopes Sigma could produce such a fast wide without coma but they did not albeit the coma is better than what the Sigma 24 shows.

If your lens is coma free wide open you have an extraordinary copy with optical quality that no one else has seen. Just google coma and sigma 20mm 1.4 you will find plenty of examples.

Paul C
 
As for the 20 Sigma 1.4. Do a bit of reading from others and please don't quote the marketing hype from sigma to me. I have followed this lens from day of annoucement and it not coma free wide open. No one who has shot one has reported this either and there are plenty of reviews showing the effects of coma. I had high hopes Sigma could produce such a fast wide without coma but they did not albeit the coma is better than what the Sigma 24 shows.

If your lens is coma free wide open you have an extraordinary copy with optical quality that no one else has seen. Just google coma and sigma 20mm 1.4 you will find plenty of examples.

Paul C
The problem is that when you look into higher resolution, you would notice the noise of f2.8 shots before you notice the coma of f1.4 shots.

If you use tracking mount (which ruins the foreground) then it's worth using the f2.8 lenses to minimize coma. You can grab coma test results from lenstip and the Sigma 20mm f1.4 is indeed not bad at all when compared against other f1.4 lenses:

coma.jpg
 

GrahamWelland

Subscriber & Workshop Member
Btw, even though I didn't do it in my shot, the normal approach to using the tracking mount is to shoot a static foreground image and merge with the tracked sliding image that will typically mess the foreground due to movement. It's just a matter of time & technique.

Would I prefer ISO 6400 @ f/1.4 on MFDB? Sure I would. In the meantime you can do very well using a tracking mount and typical medium format f stops.
 
As one who has done this type of photography for years I fully understand the need for light transmission for the milky way. However there is nothing WORSE than a series of shots ruined by excessive coma as the stars will all have wings. If you feel this is not a huge issue then I would totally disagree. I currently don't know of any 1.4 lens from 14mm to 24 that does not produce excessive coma. And with either the Rokinon 24mm 1.4 or the highly esteemed Nikon 24mm 1.4 the coma from F 1.4 to 2.0 is so bad that you will end up cropping down to a APC C sized frame and that still won't get rid of all of it. By the time you adjust the aperture to where the coma is minimum, you have lost any advantage for light transmission. The 14mm 2.8 Samyang, is still considered by photographers who work at night as one of the best lenses on the market due to it's total lack of coma, and these same photographers understand the compromise they are making over light transmission for low coma.

Coma is the single most destructive problem there is after noise.

You can't imagine the time it takes to attempt to manually correct for this and the Rokinon 24 is not a good player wide open not even close. As I stated it only becomes useable after F2.2.

I'm my geo the wider the better so the 24 really won't work anyway. If you are shooting monument valley the 24 focal would work fine. For me just shooting the sky is not what I am after but instead I attempt to combine some landscape elements and 24 mm does not give enough sky after cropping out the coma.

Adam Woodworth has pioneered a lot of great techniques and his work is often shot with an F 2.8 lens wide open.

Paul C
I agree that the 14mm f2.8 lenses are good choices to get started with as they produce less coma and don't require excessive amount of work. However for more advanced Milky Way shooters, when taken stacking and stitching into consideration, then the 24mm f1.4 lenses beat the 14mm f2.8 lenses by a fairly good amount of margin.

First of all, when you do stitching, the coma regions would in most cases be automatically eliminated as shown below (because the corners would almost always be cropped away in the process of stitching):

7549576_1408426232.jpg

Secondly, let's do some simple calculations:

For a given interval of 430 seconds, shooting with these two different lenses leads to two different strategies:

a) 14mm f2.8 ISO 6400 15 seconds each frame; shoot 29 frames in the same direction
b) 24mm f1.4 ISO 2500 9 seconds each frame; shoot 8 frames in each direction, and shoot in 6 different directions for the stitch (i.e. 6x8=48 frames in total)

After stacking, the pixel-level SNR of a) is ISO 6400/29^.5 = ISO 1200, and the pixel-level SNR of b) is ISO 2500/8^.5 = ISO 800.

After stitching, both options provide the same angle of view (i.e. 114° diagonal), however the stitching option provides almost 6 times the number of pixels.

To sum up, for the same 8 minutes of shooting, 14mm f2.8 would provide you 36MP ISO 1200 image quality, while 24mm f1.4 would provide you 180MP ISO 800 image quality. Clearly the f1.4 lens is a winner here. This is why medium format lenses are not favored in this application as they are too slow.
 
Last edited:
Btw, even though I didn't do it in my shot, the normal approach to using the tracking mount is to shoot a static foreground image and merge with the tracked sliding image that will typically mess the foreground due to movement. It's just a matter of time & technique.

Would I prefer ISO 6400 @ f/1.4 on MFDB? Sure I would. In the meantime you can do very well using a tracking mount and typical medium format f stops.
I totally agree that you could do a composite of two images. However there are two disadvantages:

a) For my example post there the foreground is a tree and it would be difficult to do the layer mask work (even with sophisticated luminosity masks); you would have to shoot the sky in a different location to avoid the blurred tree ruining the sky.

b) It is quite usual that the color temperature shifts quickly during the whole process and there is a significant difference of color temperature between your foreground shot and your tracked sky shot. Surely you could manipulate the color temperature and do the composite anyway but that may be denied by some landscape photographers as it's no different from downloading Milky Way materials and doing the composite work with any foreground you shoot (or download).
 
? No coma

The image is pretty small bit there appear to be plenty on wings on the Stars towards the edge? That is coma.
No. Wings towards the edge are surely star trails. Coma is always in the tangential direction and is never in the radial direction. Coma will also appear in all corners not just in one corner.
 

Paul2660

Well-known member
At F1.4 the 20mm produces coma, period. As your lens tip shots show. Yes it may be better than the 24 1.4 but it's still there and from my testing, and others who shoot the night sky on regular basis, the 20mm 1.4 coma is still pretty harsh, especially when the lens is in exact focus. The astro community had high hopes for this particular lens and it is a great lens, but for night skies, not wide open. Most times in y area it's not wide enough anyway to really get a good composition.

As for your shot, many of the stars towards the edges of the frame appear to have taken on a triangular appearance, which is coma. But without being able to view your shot at higher res, then I can't tell for sure. But the Milky Way in the US is in a southern sky right now, and star trails looking south always take on a very definite pattern. But always a single line, nothing above and below, which is what I see. Please provide a 100% crop of the upper right, and left areas if you can.

The exposure is for 30 seconds if I read the meta correctly, thus you will show trails and your milky way will be slightly blurred when viewed at 100% against the trails, which to me is distracting. Again, can't tell if that is the case in your shot without viewing at 100%, but for me it's always been the case, past 17 seconds.

As for higher ISO, and noise, it's a compromise for sure, but there are way to stack, even for the Milky way, where the noise is less, shooting at ISO 6400 or 3200, not wide open, Adam Woodworth's techniques show how to do this. His compositions are the champs of night sky landscapes.

Paul C
 
At F1.4 the 20mm produces coma, period. As your lens tip shots show. Yes it may be better than the 24 1.4 but it's still there and from my testing, and others who shoot the night sky on regular basis, the 20mm 1.4 coma is still pretty harsh, especially when the lens is in exact focus. The astro community had high hopes for this particular lens and it is a great lens, but for night skies, not wide open. Paul C
For single exposure yes. For my method of stitching I mentioned above no there is no coma because coma is cropped away in the stitching procedure.

Most times in y area it's not wide enough anyway to really get a good composition. Paul C
For single exposure then yes. For my method of stitching I mentioned above then no. I can get even wider than 11mm's angle of view if I want.

As for your shot, many of the stars towards the edges of the frame appear to have taken on a triangular appearance, which is coma. But without being able to view your shot at higher res, then I can't tell for sure. But the Milky Way in the US is in a southern sky right now, and star trails looking south always take on a very definite pattern. But always a single line, nothing above and below, which is what I see. Please provide a 100% crop of the upper right, and left areas if you can.

The exposure is for 30 seconds if I read the meta correctly, thus you will show trails and your milky way will be slightly blurred when viewed at 100% against the trails, which to me is distracting. Again, can't tell if that is the case in your shot without viewing at 100%, but for me it's always been the case, past 17 seconds.
crop.jpg

Here is the 100% crop of the image. What you see in the smaller version is star trails instead of coma. As I don't own the lens I can't be sure but I tend to tell that the lens is a bit off-centered because the top left corner is out of focus. Yes you can observe a small amount of coma but it's simply not visible in smaller prints (before noise is visible while you enlarge).

As for higher ISO, and noise, it's a compromise for sure, but there are way to stack, even for the Milky way, where the noise is less, shooting at ISO 6400 or 3200, not wide open, Adam Woodworth's techniques show how to do this. His compositions are the champs of night sky landscapes.
This is not a valid argument as I could also apply your logic and claim that the Voigtlander 10mm f/5.6 Sony FE Mount (fullframe) is better than your 14mm f2.8.

My argument is that given the same amount of time, the 24mm f1.4 (wide open) produces a better final result when compared against the 14mm f2.8, if you shoot in the way I mentioned above (stacking + stitching). The 24mm f1.4 uses the same amount of time for exposure to produce the same angle of view for your composition (which is coma-free thanks to stitching), but gives you a better SNR and a significantly larger number of pixels.
 

Paul2660

Well-known member
Thanks for the crops, I agree traditional star trails. And with that much trailing, the milky way should be slightly blurred when you view it at 100% on a single exposure, since the trails will be in tack focus and the movement of the nebula makes it hard to view, as your eyes will focus on the sharp trails.

And sure you can use the 20 1.4 wide open for traditional star trail photography and the coma will not show, but since I tend to work with partial moon for illumination, wide open is too much.

The Milky Way, is another subject totally, and much hard to work with for sure.

The method that Adam Woodworth uses, combines higher ISO work, with F 2.8 glass, or higher and he is using even ISO 12800 on the D810A. The best Nikon for the Milky way for me is the D750 as it's higher ISO range is 3x cleaner than the D810, per Adam's research the D810A does much better but it's just to expensive for me to justify and the camera won't work well in normal lighting. http://www.adamwoodworth.com/ I would love to travel to some of his spots.


I agree with all your points on faster glass over tracking hardware, even though as shown by Graham, the tracking hardware can produce wonderful results. And the difference between what you can capture at F 1.4 and F 2.8 is also impressive. I have just resorted to cropping out the edges with shots from the 20mm 1.4 and using the center, also tend to use the 4:3 Nikon D810 crop when I have that camera on, or on the D750 as it takes out the vast majority of the areas with coma. There is also a coma brush you can easily make in Photoshop, that will do a wonderful job of removing the wings, just takes a ton of time, but works good on the brighter stars which display the greater amount of coma.

Sad to see the Sigma off center, but that is common on their lenses, as I had to purchase 3 24mm 1.4 lenses before I found one not off center on one side or the other and the lens I ended up with is still not quite perfect, but gets the job done.

The other issue of course with the F 1.4 is the extremely shallow DOF exhibited, so working in a combination shoot, attempting to pull in parts of the surrounding landscape can be difficult with out a focus stack, but it can be done.

All this just points out how much work is needed to be done to really capture such a shot, and sadly the vast majority of viewers tend to take it for granted and or feel it's faked in some way (general perception of most digital photography now).

Paul C
 
The method that Adam Woodworth uses, combines higher ISO work, with F 2.8 glass, or higher and he is using even ISO 12800 on the D810A. The best Nikon for the Milky way for me is the D750 as it's higher ISO range is 3x cleaner than the D810, per Adam's research the D810A does much better but it's just to expensive for me to justify and the camera won't work well in normal lighting. Adam Woodworth Photography I would love to travel to some of his spots.
Actually the Nikon D810A is the only camera I plan to keep (and I might just sell off all my medium format gear as well).

The main advantage of the Nikon D810A is being able to push shadow like crazy for high ISO long exposures without getting flooded by the purple noise. Below shows a comparison between the D800E vs the D810A, for ISO 12800, 8 seconds each frame, 16 frames each camera stacked, +5EV push:

d810a.JPG

I sold my D800E straight away after I tested the D810A. Thanks to the RAW files of the D810A I could abuse it and shoot the milky way when there is no moon, without any artificial light (so that the whole scene is natural), without having to capture any darkframes to generate the master darkframe before stacking. This makes the shooting extremely flexible, i.e. there is no requirement for the heavy tracking mount equipment, no requirement for same humidity and temperature, no need for extra battery life / extra time / extra storage to capture darkframes etc. I could just shoot whatever I want and try out lots of compositions on the fly. (example 1, example 2)

Many people claim that the D810A is not recommended for general photography, but I see it in different ways. I get better skin tone with the D810A and I also get better colors of the sunset/morning glow. Nikon has published some comparisons at their website: Nikon D810A - General Subjects

Another good side of the D810A is its slow speed of depreciation - normal cameras would update so fast, e.g. D800E is succeeded by D810, 5DM3 is succeeded by 5DSR, and IQ3 80MP is succeeded by IQ3 100MP etc. But specialized cameras like the D810A do not update so frequently and hence can be used for a long time.

The only downside of the D810A is that the lowest native base ISO is ISO 200 instead of D810's ISO 64. That means more than 1 stop of less dynamic range for low ISO landscape photography.
 

Paul2660

Well-known member
Getting away from MFD quick, but yes the D810A is a wonder, but it's flip is the point you mention, loss of DR at low ISO, I guess a bit similar to the D5, even though the 810A has a sony sensor.

For me the solution was the D750, I don't need the extra res for night work, and it's full frame, and if you look at Adam's comparisons, very close in high ISO performance to the 810A. The 810A is 3.8K now seems to have gone up possibly. And I just don't shoot that much Milky Way to justify it.

I keep my eyes open for a used D810A however.

Also I believe the 800E was the worst with the red channel noise, I don't see that as much with the regular 810, however I agree it's not as clean as the 810A

Paul C
 

Don Libby

Well-known member
Graham, care to share your thoughts on the iOptron tracker? I weighed my 28, XF and IQ and came up with 7.2 pounds which looks to be about 1/2 a pounder the weight limit. How difficult was it to set up? Any other thoughts would be appreciated.

Don
 
Top