Stitching night sky?

[Michael Fontana]

Klaus
a external batteries in an isolated box with some small hand-warmers would have made sens, as my 4 batteries wen't down pretty fast in these conditions of cold and 1/2 h-exposure. Still 2. 7 GB of RAWS were possible whit the 4 batteries.

As for the the lines: Actually I don't mind the 30 min of exposure, resulting in longer star-traces. 10 Minutes is rather short, meanwhile at the moment I don't mind having these traces, and using a astronomy tracking-panohead is not the goal.

I uploading a RAW with these lines for inspection and hopefully hints how to get rid of it....

Download the RAW here

 

[Klaus Esser]

As for the the lines: Actually I don't mind the 30 min of exposure, resulting in longer star-traces. 10 Minutes is rather short, meanwhile at the moment I don't mind having these traces, and using a astronomy tracking-panohead is not the goal.

Download the RAW here

I didn´t mean the star-stripes to be shortened by a tracking had - i ment that very long exposures can cause color- and luminance-shift.

The download needs a password!

best, Klaus

besides: did you try the camera´s noisereduction for longtime-exposures? This will mask noise as well as shifts. But takes very long extra time.

 

[Michael Fontana]

I edited the link for the RAW; does it works now?

 

[nicolas claris]

I edited the link for the RAW; does it works now?

Bonsoir Michael
Yes it works …

 

[nicolas claris]

Bonsoir again Michael

I'm now emailing you the full res JPEG view extracted with C1 4.6

A little less "harsh" than C1 3.9

To me, when looking the image at 100%, they are plenty different stars stripes, because stars have not the same brightness… this should maybe explain why some seems very bright?

It doesn't bother me at all, au contraire!
But if this is a tech issue, then I agree, we need to know what happened.

BTW there is quite large green cast on the middle of the image. I did noticed it also on the stitch you did post some days ago here… Another isue?

 

[Michael Fontana]

BTW there is quite large green cast on the middle of the image. I did noticed it also on the stitch you did post some days ago here… Another isue?

Yep, that ones is bothering me. I thought it might a special reflection from snow and lens maybe a combination of it, therefore I had to try it out in summer, again.


After redoing last month, I have the same green cast - with another lens! Look at the vertical green cast at the left parth of each single image...

Must be the sensor, producing that cast when exposing 15 min - it doen't happend on 5 minutes exposures.(The two yellow rays at the left are probably sattelits, or planes...)

Anybody has a idea how to get rid of it?
Is there a possibilty of creating a °mask° of it that might be inverted and applied on these images in Photoshop?

If there is a pattern, I should be able to reproduce and correct it.

Any help?

It's hard to correct it in post - big danger of destructing the color of the small star lines

 

[nicolas claris]

hmmm
could be some marks of dirt on the sensor made by wipes during sensor cleaning… ?∞¿

 

[Bart_van_der_Wolf]

Must be the sensor, producing that cast when exposing 15 min - it doen't happend on 5 minutes exposures.(The two yellow rays at the left are probably sattelits, or planes...)

Anybody has a idea how to get rid of it?
Is there a possibilty of creating a °mask° of it that might be inverted and applied on these images in Photoshop?

If there is a pattern, I should be able to reproduce and correct it.

Any help?

It's hard to correct it in post - big danger of destructing the color of the small star lines

Hi Michael,

In Astronomy imaging it is common to eliminate as much as possible all signals that are not caused by actual exposure from stars and nebulae.

As part of a series of exposures, one also needs to collect a number of "Dark images". It is not uncommon to collect 10 or more "Darks" These "images" are each taken with the exact same exposure time as the actual exposures and at the same ISO and temperature, but with the lens and camera eye piece covered. From these "Darks" a master dark image is produced, by a process of averaging to reduce noise. The master dark image is then subtracted from the actual exposures, which will remove any pattern noise and hot pixels that have accumulated during the long exposure times.

In fact several other image series are also collected, e.g. so-called "flats" to eliminate vignetting and other non-uniformities in the exposures, and "offset" exposures with the shortest exposure time to reduce the read noise from the sensor.

An important thing is that those maniplulations need to be done when the data is still in linear gamma. space. Doing it right requires large numbers of Raw images that need to be processed in linear gamma space. For that reason there is dedicated software that helps to automate that process (e.g. IRIS and ImagesPlus, both Windows applications but I assume they could run on an Intel Mac under Parallels).

I have ImagesPlus, and I could try and make a "master dark" for you to see if that explains the effect you see. I would need a number of lens&eyepiece covered dark exposures from you (Raw files), all taken with the same 15 min. exposure time and ISO as the original exposures, and at the same temperature as the original exposures if possible. They are normally taken with the dark frame subtaction of the camera disabled. Ideally they should have all been taken approx. at the same time as the actual exposures, because the sensor may have developed new hot pixels and changed its characteristics a bit, but we can at least try what we still can do after the fact. We could start with 4 to 8 Darks, and see if that's revealing anything useful and take it from there.

Bart

 

[nicolas claris]

As always interesting post, Bart…
What is/are the main difference(s) with your method than the one one can choose in the Canon's menu (to 'eliminate' noise and hot pixels).
It lengthen the shoot by the same amount of time than exposure, and does the calculation in camera.
I understand that Canon's 'method' is too late for Michael, but what about future shots?

Once again thank you to share your brilliant knowledge

 

[Bart_van_der_Wolf]

As always interesting post, Bart…
What is/are the main difference(s) with your method than the one one can choose in the Canon's menu (to 'eliminate' noise and hot pixels).

It lengthen the shoot by the same amount of time than exposure, and does the calculation in camera.

Hi Nicolas,

The standard Dark frame subtraction that Canon and other manufacturers offer as an option, uses a single image and a single dark frame. That is a step in the right direction for the elimination of 'hot' pixels and pattern noise that gradually develop during longer exposures. However, there is a technical drawback as well.

To be specific, it significantly increases the overall noise floor. It may still be better to tolerate a higher noise floor but also eliminate hot pixels and pattern noise, because there are good noise reduction programs but fewer good hot pixel repair uilities. Also, hot pixels in a Bayer CFA influence the neighboring pixels during Raw conversion, so repair should ideally be done before demosaicing (otherwise artifacts like "black holes" may be created). Hot pixels and pattern noise are generally also more distracting than noise, so it's a trade off.

By averaging multiple dark frames, their noise is reduced so the dark frame subtraction will have a lower noise as well. Therefore , in scientific photography, it is common to not use the camera's dark frame subtraction, but to create a better master dark frame from multiple dark frames. The added benefit for Astrophotography is that the actual exposures take less time, and that allows to take many more exposures which can be averaged as well to reduce noise even further and reveal very weak star signals hidden in that noise.

Specialized programs like ImagesPlus (Windows only, but maybe works on an Intel MAC under Parallels), do their calculations on Raw linear gamma CFA data before demosaicing, and then do the actual demosaicing afterwards on the corrected Raw Bayer CFA data.

I understand that Canon's 'method' is too late for Michael, but what about future shots?

Michael's shots are slightly different from 'regular' Astrophotography, in that he shoots single exposure star 'trails' and landscape silhouettes. Other than that, the Astrophotography principles still apply and one can benefit from also shooting a number of dark frames (same exposure time, ISO, and temperature, but covered lens and eyepiece) and a number of bias or offset frames (same procedure as the dark frames but shot at 1/8000th second), during the same session. With more than 16 dark/offset frames we get diminishing returns, so 10 to 16 images for supporting dark&bias files is usually plenty.

Once again thank you to share your brilliant knowledge

You're welcome, I try to help others who may in return offer their particular skills. Synergy is my goal.

Bart

 

[Michael Fontana]

Bart - is there a thing that you don't know in photography?
We other need to stickt our heads together and start to dig .. ;-)

You named and showed exactly the solution I've been looking for, but I had a rough idea, only!

While the hotpixels aren't a big problem - RAW Developer has the ability to find and delete them, but the green cast had been annyoing me that much, that I didn't explored it further in winter.

I have about 5 panos plus about 30 singles shots with these cast, therefore your methode by creating that mask would make really sense. The one I'd like very much to see well done is the following - 10 frames and good stitch: in big size it looks really nice, with plenty of details, due to the long exposure of 20 minutes and the snow - plus a clear night! - apart from the nasty cast

Thanks for your kind offer, as I can't go up these days, I'll try at the week-end - with cooler wheater - to take the dark images her in town. Anything special to consider?

Is it really important to have the exactly same lens as with the shots? I mean there's no light entering the lens... therefore, here shouldn't be a difference.
I had been using 2 different lens for these nightstitches.

In October, I will probably go up again; and I' d like to take some other nighties, if we can create a dark master.

 

[Michael Fontana]

Nicolas

I' ll take Bart's dark image methode any moment over the camera's dark frame by practical reasons only:

apart from the problems with stitching - the stars will move in the 20 min of the dark frame as well as the wheater might change (please note that by 10 shots x 20 min plus 200 min of dark frames there will be 7 hours of total exposure.... )

in which time the moon or day might rise, therefore there's a high interest in having the exposures as short as possible, while beeing long enough for leaving nice star tracks... which is depending on the north/south-orientation as well....

Further, it's not that easy to handle panohead and cam etc in these conditions, per example in a clear (important for the sky) winter night at that sea level, you just can't stand for 5 hours outside in the night.... so you have to search shelter and warm up.
After 25 minutes you' ve to come back - and quite often there's no other way than to walk through the image.... you can't have a headlamp and you' re tapping blind like a mole to the tripod... ;-)

Seriously, last december I had to ski through some images, just to get it done.
From that reason too, you want to limit the numer of exposures as much as possible, while having enough images for allowing a nice big image.

 

[Bart_van_der_Wolf]

Bart - is there a thing that you don't know in photography?
We other need to stickt our heads together and start to dig .. ;-)

LOL, there probably is, but it's difficult to know what you don't know ...

Thanks for your kind offer, as I can't go up these days, I'll try at the week-end - with cooler wheater - to take the dark images her in town. Anything special to consider?

Is it really important to have the exactly same lens as with the shots? I mean there's no light entering the lens... therefore, here shouldn't be a difference.

Assuming the lens doesn't produce some sort of interference with the camera electronics, you can use the body cap instead of a lens. It would be safer and even avoid light leaking throught the edges of the lenscap.

When trying to mimick freezing conditions, one can always consider using a fridge, or even a freezer if necessary, and wait a while for the internal camera temperature to drop. Just wrap the camera tightly in a plastic bag or cling foil when it's cold to avoid condensation when you take it out for pressing/releasing the shutter release button and acclimatization. Since you've successfully used the camera outside/below it's recommended operating temperature range before, there should be no surprises. Temporarily the LCD will get very slow, the rubber sealings will get stiffer, and the metal parts will shrink, but the main thing is to keep condensation down while it warms up again.

If you want to do it really well, then you should take both dark frames (at the same ISO and exposure time as the original images), and bias frames (same conditions only at 1/8000th sec.). More than 16 images each would be a waste of time, 8 darks and 8 bias frames would probably be enough for now, afterall the non-uniformity is more significant than the last couple of bits of noise.

When the camera characteristics don't change much over time, then the resulting master files can be used on other occasions as well, thus saving a lot of time during the actual shoot. We'll see how much difference there is between then and now.

Bart


P.S. Looking at the stitch, there is probably also some vignetting involved, but that can be corrected for at normal temperatures, or with the Raw converter, or the stitcher's blending algorithm. Let's first collect the darks and bias frames.

P.P.S. Warning (not for Michael because he knows this). When changing in and out of significantly different temperature environments, the transition should be gentle, so let it take a bit of time. With a camera in a bag, the bag will act as a sort of insulator due to the pocket of air that doesn't move around much in the bag. However the air will condense on the colder surfaces at is cools down because cold air can retain less water vapor / moisture. So, before entering a warmer (higher absolute humidity) environment, one should contain the low humidity cold air in the camera as it warms up. The air inside will get relatively dryer as the temperature rises, and condensation will be avoided. Just wanted to make sure people don't do silly things.

 

[Michael Fontana]

Good evening, Bart

It's cold outside - we even have a bit of snow here - I start the serie for the "master dark". So we say 6 x 15 minutes?

I will go up in February again - for some night shots as well.

 

[Bart_van_der_Wolf]

Good evening, Bart

It's cold outside - we even have a bit of snow here - I start the serie for the "master dark". So we say 6 x 15 minutes?

I will go up in February again - for some night shots as well.

Hi Michael,

Sure, body cap on and eye piece of viewfinder shut, 6 x 15minutes at the same ISO you are going to use for the actual shooting. Make sure you have switched the long exposure noise reduction in the camera to off.

If you can also take some "read noise" bias shots (same ISO, shortest shutterspeed (1/8000th s), body cap on, eye piece of the viewfinder shut), let's say 8 before and 8 after the 6 long exposures, that would be helpful.

With that we'll see what we can come up with to improve the actual long exposures.

Cheers,
Bart

 

[Michael Fontana]

Yep, long exposure noise reduction is off; I did upload 6 RAWs with 30 min of exposure already, this might be more significant than the 15 min, which I'm taking right now.

Talking the 1/8000 in between the 15 min-serie.