I'm looking for a program to enhance the resolution of my lunar images.... any ideas?

I use Registax for lunar/planetary imaging.  It can be a little quirky, but when it works it works very well, and it shines at processing video.  As an example, here's a raw frame from a video I took through a 32" scope:

After processing the video in Registax:


You can easily get similar improvement for the Moon -- I just don't have a before/after offhand.

Also, I dont believe stacking reduces light pollution, as stacking adds all light together, but it does improve the SNR.

That's right.  Every frame will have the same light pollution, and stacking just averages the information across all the frames, so what's the average of something that is constant?

What stacking does improve is the shot noise (from the discreteness of light -- in low light levels the statistics of individual photon strikes on each pixel on the sensor becomes important, hence the graininess in images).  The SNR for this type of noise improves with the square root of the number of frames.

 Is it true Wat that taking a lot of very short exposures is better than fewer subframes that have longer exposures?

For lunar-planetary imaging (LPI), this is true.  For deep sky object (DSO) or Milky Way imaging, "it depends", and in your particular situation it probably doesn't matter much.  (Also with Milky Way photography, do not fall for the misconception that lower ISO is always better.  It often isn't.)

The SNR improves with the square root of number of frames, but in low light (such as when capturing faint details in galaxies and nebulae) it also improves with the square root of exposure time, for the exact same reason.  Longer exposure = more photons striking the sensor = better statistics!  So what matters most is the total integration time.  If light from objects you are imaging was the only factor, then it doesn't matter if you use many short exposures or fewer long ones.

However, there's a subtlety for DSO imaging in that there are additional sources of noise from the camera itself (read noise and thermal noise or "dark current").  This noise is more significant in longer exposures in dim light, for which you may want to use calibration frames.  The more calibration frames you take, the better they themselves are averaged together, and therefore the better they are subtracted out to make the final image.  So it can be better to take more shorter exposures, but in practice this is most important with very dark (not light polluted) skies and when exposure times are very long anyway (several minutes or more).  Otherwise it probably doesn't matter, and I would simply opt for taking as long exposures as possible to get the most signal in each frame.  I would also pay close attention to your camera's ISO performance (is it ISO invariant or not?) and do some tests to find the best settings for your equipment and your sky.

-------------------------------------------------------------------------------------------------------------------------------------------------------------

For lunar and planetary imaging, the answer is very strongly that more shorter exposures is better.  There is a lot of light available, so there is no trouble getting proper exposures in much less than a second.  But the objects you want to capture are very small, so atmospheric turbulence becomes important.  That blurs out the details, as if you were looking through water or heat haze.  The distortions fluctuate very rapidly, but if you take a large number of very fast frames, then by random chance a few will be distorted less than others.  Then with programs like Registax, the computer can analyze the frames and sort them by how much they are distorted, allowing you to stack the highest quality ones and reject the rest.

This makes video capture a great method for LPI, since individual frames can be very short and you capture a large number of them very quickly.  But the most important thing by far is to do your imaging on nights when the turbulence is weakest.  The above image of Saturn was from about 12 seconds worth of video at 29fps, but I also took it when the seeing was as good as I've ever had.  Sites like cleardarksky.com show predictions for the atmospheric seeing, and you can also tell when seeing is good or bad by how steady the stars are.  Twinkling stars means bad views of the Moon and planets!

Bit of an update. Turns out that line in Orion neb was due to collimation issues so my mate got a laser collimator and I corrected the optics to be spot on, it was a bit out. I was hoping it would also fix the sharpness of the image which it did a little bit but not as much as I hoped for.

Any ideas why the focus isn't so great, and yes it was focused as best as possible, used the frame and focus in backyarkEOS to get the lowest number on a selected star.

It's a skywatcher 200mm imaging reflector scope, so it should be sharp but its not really that great. The finder scope is actually sharper.  Could it be a defective mirror shape?

My mate is looking to process the stars to bring them down in size in pixinsight. But it would be great to take sharp pics in the fist place.

Once he uploads the new pic I will post it here.

This is mine via deepsky stacker.

The image is a little sharper after collimation but still would like a sharper image but i think its a limitation of the mirror and maybe the CCD itself to some degree.

I'm looking for a program to enhance the resolution of my lunar images.... any ideas?

I use Registax for lunar/planetary imaging.  It can be a little quirky, but when it works it works very well, and it shines at processing video.  As an example, here's a raw frame from a video I took through a 32" scope:

After processing the video in Registax:


You can easily get similar improvement for the Moon -- I just don't have a before/after offhand.

Also, I dont believe stacking reduces light pollution, as stacking adds all light together, but it does improve the SNR.

That's right.  Every frame will have the same light pollution, and stacking just averages the information across all the frames, so what's the average of something that is constant?

What stacking does improve is the shot noise (from the discreteness of light -- in low light levels the statistics of individual photon strikes on each pixel on the sensor becomes important, hence the graininess in images).  The SNR for this type of noise improves with the square root of the number of frames.

 Is it true Wat that taking a lot of very short exposures is better than fewer subframes that have longer exposures?

For lunar-planetary imaging (LPI), this is true.  For deep sky object (DSO) or Milky Way imaging, "it depends", and in your particular situation it probably doesn't matter much.  (Also with Milky Way photography, do not fall for the misconception that lower ISO is always better.  It often isn't.)

The SNR improves with the square root of number of frames, but in low light (such as when capturing faint details in galaxies and nebulae) it also improves with the square root of exposure time, for the exact same reason.  Longer exposure = more photons striking the sensor = better statistics!  So what matters most is the total integration time.  If light from objects you are imaging was the only factor, then it doesn't matter if you use many short exposures or fewer long ones.

However, there's a subtlety for DSO imaging in that there are additional sources of noise from the camera itself (read noise and thermal noise or "dark current").  This noise is more significant in longer exposures in dim light, for which you may want to use calibration frames.  The more calibration frames you take, the better they themselves are averaged together, and therefore the better they are subtracted out to make the final image.  So it can be better to take more shorter exposures, but in practice this is most important with very dark (not light polluted) skies and when exposure times are very long anyway (several minutes or more).  Otherwise it probably doesn't matter, and I would simply opt for taking as long exposures as possible to get the most signal in each frame.  I would also pay close attention to your camera's ISO performance (is it ISO invariant or not?) and do some tests to find the best settings for your equipment and your sky.

-------------------------------------------------------------------------------------------------------------------------------------------------------------

For lunar and planetary imaging, the answer is very strongly that more shorter exposures is better.  There is a lot of light available, so there is no trouble getting proper exposures in much less than a second.  But the objects you want to capture are very small, so atmospheric turbulence becomes important.  That blurs out the details, as if you were looking through water or heat haze.  The distortions fluctuate very rapidly, but if you take a large number of very fast frames, then by random chance a few will be distorted less than others.  Then with programs like Registax, the computer can analyze the frames and sort them by how much they are distorted, allowing you to stack the highest quality ones and reject the rest.

This makes video capture a great method for LPI, since individual frames can be very short and you capture a large number of them very quickly.  But the most important thing by far is to do your imaging on nights when the turbulence is weakest.  The above image of Saturn was from about 12 seconds worth of video at 29fps, but I also took it when the seeing was as good as I've ever had.  Sites like cleardarksky.com show predictions for the atmospheric seeing, and you can also tell when seeing is good or bad by how steady the stars are.  Twinkling stars means bad views of the Moon and planets!

That shot was amazing, Wat!  How do you wield a 32" scope?  I have enough issues with my 8" one lol.  Did you use an LPI to take that?  I have that camera but haven't used it much yet because of focusing issues with the scope (I think I have to make whats known as a Bahtinov mask for it?)
So as far as light pollution and stacking is concerned, the amount of background light pollution in say a stack of 100 images will be the average amount of light pollution in those frames but the shot noise will be 1/10th that of a single frame?
Thanks for the highly interesting stuff regarding exposure times and ISO invariance.  I did  a series of test images where the exposures were all of the same length and I did ISO bracketing.  I found that up to ISO 3200 I could see colors in the stars and the background noise was negligible.  Above ISO 3200 however there was a sharp drop off in dynamic range and background noise in my images.  So would ISO 3200 be the ideal ISO for deep space imaging?
The camera I am using is an Olympus EPL-6.  I found a site that measures various sensor statistics which I will post.  They dont have that specific camera on their site but they have both the EPL-5 and the EPL-7 and they are very similar and use the same sensor.

Also, question, the new Olympus cameras have a feature called Live Composite, is this useful for AP?  They mention star trails, lightning, fireworks and light painting on their site, but I am doubtful it does anything for AP, since it only records light changes after the initial frame (but maybe it would be useful for shooting meteors?

Live Composite mode:

https://learnandsupport.getolympus.com/learn-center/photography-tips/fireworks-light-trails/7-ways-you-can-use-live-composite

https://learnandsupport.getolympus.com/learn-center/photography-tips/fireworks-light-trails/7-ways-you-can-use-live-composite


http://www.photonstophotos.net/Charts/S ... istics.htm

http://www.photonstophotos.net/Charts/S ... istics.htm

The sensor info on the cameras is on that site.

How do you wield a 32" scope?  I have enough issues with my 8" one lol.

Oh, I just pick it up and carry it around.  It only weighs 2.25 tons.


Actually a 32" Dobsonian telescope can be portable, but otherwise scopes in this size range are almost always in dedicated observatories.  This one is the 32" Ritchey-Chrétien scope at GMU.  To access it you take an elevator to the observatory on top of Research I, and you control it and the dome by computers.  It's a lovely scope to work with, though sadly after the 2011 earthquake the telescope acquired a severe jittering motion when tracking, which wasn't solved while I was still attending.  So doing any DSO imaging was absolutely out of the question.  But as you can see, it was still perfectly capable of producing a nice planetary image by aligning and stacking video frames.

Did you use an LPI to take that?

An LPI need be nothing more complicated than a webcamera.  I used my old Canon Powershot.

Thanks for the highly interesting stuff regarding exposure times and ISO invariance.  I did  a series of test images where the exposures were all of the same length and I did ISO bracketing.  I found that up to ISO 3200 I could see colors in the stars and the background noise was negligible.  Above ISO 3200 however there was a sharp drop off in dynamic range and background noise in my images.  So would ISO 3200 be the ideal ISO for deep space imaging?

Sure thing!  Your result sounds reasonable. Is roughly consistent with my experience with a Canon Rebel T3.

How do you wield a 32" scope?  I have enough issues with my 8" one lol.

Oh, I just pick it up and carry it around.  It only weighs 2.25 tons.


Actually a 32" Dobsonian telescope can be portable, but otherwise scopes in this size range are almost always in dedicated observatories.  This one is the 32" Ritchey-Chrétien scope at GMU.  To access it you take an elevator to the observatory on top of Research I, and you control it and the dome by computers.  It's a lovely scope to work with, though sadly after the 2011 earthquake the telescope acquired a severe jittering motion when tracking, which wasn't solved while I was still attending.  So doing any DSO imaging was absolutely out of the question.  But as you can see, it was still perfectly capable of producing a nice planetary image by aligning and stacking video frames.

Did you use an LPI to take that?

An LPI need be nothing more complicated than a webcamera.  I used my old Canon Powershot.

Thanks for the highly interesting stuff regarding exposure times and ISO invariance.  I did  a series of test images where the exposures were all of the same length and I did ISO bracketing.  I found that up to ISO 3200 I could see colors in the stars and the background noise was negligible.  Above ISO 3200 however there was a sharp drop off in dynamic range and background noise in my images.  So would ISO 3200 be the ideal ISO for deep space imaging?

Sure thing!  Your result sounds reasonable. Is roughly consistent with my experience with a Canon Rebel T3.

I've always wanted a Dob but was worried about exposing the mirror to sand and dust because I live near the ocean!
Canon Powershot- do you use it afocally (lens to the eyepiece)?  I have done that with my old Olympus C-7070
About the Live Composite feature, I find it intriguing because the way it works is that it takes an initial frame and then keeps the exposure going....but it doesn't add any light that's stationary, it only adds light that wasn't in the original frame.  Does this mean that light pollution wouldn't increase (since it's static) even with long exposures?  Although I realize stars would show trailing....but negating light pollution would be worth it!

Also, what is more important in collecting the most light, longer exposures or a lens with a greater aperture diameter?

the way it works is that it takes an initial frame and then keeps the exposure going....but it doesn't add any light that's stationary, it only adds light that wasn't in the original frame.  Does this mean that light pollution wouldn't increase (since it's static) even with long exposures?  Although I realize stars would show trailing....but negating light pollution would be worth it!

Your signal is the same in every frame, too.  If you're hoping for a magic way of getting rid of light pollution, that isn't going to be it.  (And if negating LP was that easy, don't you think everyone would be using it?)  Instead consider filters (depending on the type of light pollution and what you are imaging -- you get the best results with narrowband filters on emission nebulae), or if possible, getting away from the light pollution.

Also, what is more important in collecting the most light, longer exposures or a lens with a greater aperture diameter?

Doubling the diameter of aperture is equivalent to quadrupling the exposure time, since you collect light over 4x as much area.

the way it works is that it takes an initial frame and then keeps the exposure going....but it doesn't add any light that's stationary, it only adds light that wasn't in the original frame.  Does this mean that light pollution wouldn't increase (since it's static) even with long exposures?  Although I realize stars would show trailing....but negating light pollution would be worth it!

Your signal is the same in every frame, too.  If you're hoping for a magic way of getting rid of light pollution, that isn't going to be it.  (And if negating LP was that easy, don't you think everyone would be using it?)  Instead consider filters (depending on the type of light pollution and what you are imaging -- you get the best results with narrowband filters on emission nebulae), or if possible, getting away from the light pollution.

Also, what is more important in collecting the most light, longer exposures or a lens with a greater aperture diameter?

Doubling the diameter of aperture is equivalent to quadrupling the exposure time, since you collect light over 4x as much area.

I have an Astronomik UHC filter I bought after reading a review that it was really good in high levels of light pollution but there are two issues
1) it affects color balance
2) it increases exposure time by 6 stops!

About the signal being the same, wouldn't star trails be recorded since the camera wouldn't be following the earth's rotation?  But would  light pollution still be recorded in added frames even though the light pollution sources are stationary?

there are two issues
1) it affects color balance
2) it increases exposure time by 6 stops!

Yes, you should expect it to do both.  It blocks out a large fraction of the visible spectrum, and it doesn't block across the visible spectrum uniformly.  The manufacturer's website lays out the pros and cons and details of its performance for various setups.

The effect on color balance is something that could be solved in post process, while the long exposure time implies the use of a tracking mount.

About the signal being the same, wouldn't star trails be recorded since the camera wouldn't be following the earth's rotation?  But would  light pollution still be recorded in added frames even though the light pollution sources are stationary?

You will get star trails on a light polluted sky. 

What you're describing sounds the same as if you took many short untracked exposures and blended them with "lighten only" (choose the brightest value the pixel had in any frame), as opposed to addition (add all the values together.) This will result in a darker sky than if you captured the same length of star trail with a single exposure (assuming same camera settings otherwise like aperture and ISO), but you will not capture fainter stars.

The effect on color balance is something that could be solved in post process

Yes, but not trivially, I think, since amplifying the suppressed colours wiil also bring back the light pollution.  I think some kind of masking is necessary.
One can argue that if post processing is necessary, then it's better to photograph without a filter and suppress the light pollution post process.  Post processing can do more than a filter, so it might be better to preserve as much information as possible until that point rather than throwing away much information even before the light enters the lens.

Filters might still be useful in some cases, such as when post processing is not an option, or because of sensitivity or dynamics limitations of the sensor or film.  If you're absolutely sure that you don't need the information thrown away by the filter, it's surely better do to it at the earliest stage.  However, if you end up undoing much of what the filter did post process, then perhaps you shouldn't have used it.

Yes, but not trivially, I think, since amplifying the suppressed colours wiil also bring back the light pollution.  I think some kind of masking is necessary.

I think the idea is not to bring back colors that were suppressed (that information is gone anyway), but re-balance the colors that remain.  For example if the filter cut out more blue light, resulting in a redder image, then mapping more of the green channel to blue may help correct the color balance.  But really this depends on the subject of the image, the type of filter, and the photographic equipment itself.  For broadband targets especially this may be a poor solution since much of the signal would be rejected with the light pollution.

The Astronomik UHC filter transmits in two broad bands which are roughly blue-green and red, so it captures the important spectral lines of emission nebulae and rejects most of the rest:



Visually this can help improve the view of nebulae under light polluted skies, while for photography it requires an extra touch.  Another option is to shoot with a camera modified for astrophotography, where the internal filter which blocks much of the H-alpha line is removed.  According to the manufacturer, color balance with this filter is much improved then.

Ultimately the best way to deal with light pollution is to get away from it, or if capturing narrowband targets, narrowband filters work wonders as Phunnie and others have shown previously.

there are two issues
1) it affects color balance
2) it increases exposure time by 6 stops!

Yes, you should expect it to do both.  It blocks out a large fraction of the visible spectrum, and it doesn't block across the visible spectrum uniformly.  The manufacturer's website lays out the pros and cons and details of its performance for various setups.

The effect on color balance is something that could be solved in post process, while the long exposure time implies the use of a tracking mount.

About the signal being the same, wouldn't star trails be recorded since the camera wouldn't be following the earth's rotation?  But would  light pollution still be recorded in added frames even though the light pollution sources are stationary?

You will get star trails on a light polluted sky. 

What you're describing sounds the same as if you took many short untracked exposures and blended them with "lighten only" (choose the brightest value the pixel had in any frame), as opposed to addition (add all the values together.) This will result in a darker sky than if you captured the same length of star trail with a single exposure (assuming same camera settings otherwise like aperture and ISO), but you will not capture fainter stars.

Thanks, Wat, I tried out a few different filters before picking this one.  The thing I liked about it was that it allows passage of small portions of each of the RGB channels, so there is at least some information in all 3 channels to work with.  There are even stronger filters out there that dont allow the passage of any info at all in the red, and I wanted to preserve H-alpha data for nebulae.  I've also considered sending the camera in for modification to remove the blocking filter, right now the cameras IR/UV cutoff filter allows passage of about 35% H-alpha, which is better than average but not great.
So based on what you're saying Live Composite will give me a darker sky per total exposure time vs a stack of images taken with the same total exposure time in manual exposure mode?  That's what I seem to be getting when I tried it out.  But like you said, it did not reveal fainter stars than what I have been able to detect in a single exposure.  That is a major trade off!

Wat more details about Live Composite mode here:

https://www.thewanderinglens.com/setting-scene-star-trails-live-composite/

It's stated that it's a "game changer" but it seems like this would only work with star trails, not stacking pinpoint star images?  It takes a 15-60 sec image and after that only records light movement...can this be duplicated in software rather than using this mode?

Dug a hole in the snow in my backyard and managed to get a few minutes on orion before it got cloudy again. Focuser obstruction is pretty strong with this one, but I bought a moonlite focuser that should hopefully solve the issue.