APOTY Results

[tomato]

Some great images here:
Astronomy Photographer of the Year: Huge plasma arc wins https://www.bbc.co.uk/news/in-pictures-66807814

The standouts for me are the Jupiter and Mars set images, how were they taken from the bottom of our atmosphere?

[ollypenrice]

Excellent images and the standard of judging has been transformed. To be honest, the Drechsler team just had to win this with their consistently ground-breaking work.

Olly

[Paul M]

Brilliant images. My favorites are:

1) The Aurora

2) Sprites, (a particular interest of mine.)

3) Moon v Mars, but it's surely contrived. At that image scale, motions will have prohibited any kind of stacking for a single frame? Or it's one "lucky" image!

[vlaiv]

2 hours ago, tomato said:

The standouts for me are the Jupiter and Mars set images, how were they taken from the bottom of our atmosphere?

Jupiter was taken with 21" telescope. and Mars is in line with about 10" of aperture. Fact that it is behind the Moon is just good timing.

[CraigT82]

15 hours ago, Paul M said:

Brilliant images. My favorites are:

1) The Aurora

2) Sprites, (a particular interest of mine.)

3) Moon v Mars, but it's surely contrived. At that image scale, motions will have prohibited any kind of stacking for a single frame? Or it's one "lucky" image!

For the Mars occultation image it will be two separate stacked images, one for lunar surface and one for Mars, then combined in post. 

[BGazing]

14 hours ago, vlaiv said:

Jupiter was taken with 21" telescope. and Mars is in line with about 10" of aperture. Fact that it is behind the Moon is just good timing.

But the apparent motion of Mars vs Moon would prohibit it.

Puzzled by the h-alpha winner, image is overprocessed

[BGazing]

Now browsing through planets, and the winner is a false color combination of UV and IR on venus. The sort of combination that takes informaton out of both sets of images...sigh.

Jupiter shot is simply amazing.

Edited September 15, 2023 by BGazing

[vlaiv]

10 minutes ago, BGazing said:

But the apparent motion of Mars vs Moon would prohibit it.

Not sure what you mean by this?

this is Stellarium simulation on the date the image was taken.

 

[BGazing]

1 hour ago, vlaiv said:

Not sure what you mean by this?

this is Stellarium simulation on the date the image was taken.

 

How much would be their apparent movement against each other in 6 minutes? How do you process it in AS - surface or planet?

[vlaiv]

Just now, BGazing said:

How much would be their apparent movement against each other in 6 minutes? How do you process it in AS - surface or planet?

Good point.

I think one should be really clever in how it approaches the problem.

First thing is ensuring short imaging time. This puts constraints on aperture size, and I was probably wrong to estimate it to around 10" telescope. If you choose not to do critical sampling, you can use larger telescope for shorter period of time.

For example - if you use 20" telescope for 30 seconds run - you will get the same SNR (or even greater because of read noise contribution) as using 10" telescope for two minute run, as long as you match the sampling rate.

Second trick that can be employed is to use very short section of the whole recording - say only one second or so for reference frame, at decent FPS like 200 FPS - that would mean 200 subs - which is plenty to get reference frame. Then you create two images. One image is stacking against the reference section of recording prior to that time you've chosen. This stack will provide data for Mars.

Other image is created by stacking second half of video against this reference - this will create Moon image.

You then combine the two images using Mars from first stack and Moon from second stack.

In any case - it can be done with some trickery, but I did not really think about it at all until you pointed out the transient nature of event.

This just makes recording much more "valuable" in my eyes if its done how I described above.

Of course - one could image Mars and the Moon before occultation event and then just combined them in Photoshop to match what it would look like at the beginning of occultation. That is much easier to do - but not as "true" as above method.

 

[BGazing]

IIRC, images like these are composite, combining separate shots of the Moon and of a planet involved. I recall watching a 'how to' tutorial.

A good Mars image requires for a color camera a 6 minute video in at least a fair seeing, and then relatively strict selection of frames (normally under 30 percent). In 6 minutes your Moon is gone. These objects are of different brightness, and, frankly, I am not sure how it would be possible for AS to handle them together.

[vlaiv]

I just measured in Stellarium.

It takes 30 seconds from first to "second" contact - or to fully occluded mars.

At the time Mars had 17" of apparent diameter. This makes it roughly 0.5"/s. If image was sampled at say 0.2"/px - this makes 2px per second drift. You can easily create reference image in one second.

Good Mars image requires good SNR - it does not have to be 6 minute video for that. Only if you sample at critical sampling frequency, but you can use larger telescope that gathers more light and you will shorten your imaging time that way.

Seeing and local conditions play a part as well, however - for above technique, it does not matter how long the video is.

AS!3 has interesting feature - it can use local quality estimator instead of whole frame.

It also allows you to choose how to build reference frame - and even use external reference frame (previous stack as reference).

Local / AP quality estimator means that only those APs that are of sufficiently high quality will get stacked.

If you take part of video prior to reference point - whole mars will be seen, but only APs on whole mars that can be matched against reference (where half of mars is occluded) will get stacked. Similarly - same goes for the moon.

Moon is not that much brighter than the Mars and they can be easily captured in single exposure. Exposure difference is not even x10 or so and with 12bit depth - you can easily capture both.

[tomato]

17 hours ago, vlaiv said:

Jupiter was taken with 21" telescope. and Mars is in line with about 10" of aperture. Fact that it is behind the Moon is just good timing.

Ok 21” aperture helps a lot, but it looks sharper than images taken with the HST, I remain impressed.

[vlaiv]

18 minutes ago, tomato said:

Ok 21” aperture helps a lot, but it looks sharper than images taken with the HST, I remain impressed.

Yeah, not sure about that

vs

[BGazing]

1 hour ago, vlaiv said:

I just measured in Stellarium.

It takes 30 seconds from first to "second" contact - or to fully occluded mars.

At the time Mars had 17" of apparent diameter. This makes it roughly 0.5"/s. If image was sampled at say 0.2"/px - this makes 2px per second drift. You can easily create reference image in one second.

Good Mars image requires good SNR - it does not have to be 6 minute video for that. Only if you sample at critical sampling frequency, but you can use larger telescope that gathers more light and you will shorten your imaging time that way.

Seeing and local conditions play a part as well, however - for above technique, it does not matter how long the video is.

AS!3 has interesting feature - it can use local quality estimator instead of whole frame.

It also allows you to choose how to build reference frame - and even use external reference frame (previous stack as reference).

Local / AP quality estimator means that only those APs that are of sufficiently high quality will get stacked.

If you take part of video prior to reference point - whole mars will be seen, but only APs on whole mars that can be matched against reference (where half of mars is occluded) will get stacked. Similarly - same goes for the moon.

Moon is not that much brighter than the Mars and they can be easily captured in single exposure. Exposure difference is not even x10 or so and with 12bit depth - you can easily capture both.

Well, something does not add up.

https://www.astrobin.com/1pnfjk/D/

Apparently 100 frames of Mars could be used originally, so the image was dim and noisy .At 18ms exposures. So that is why he made a composite image.

Here's the original https://www.astrobin.com/r5196p/E/

Whilst the description on the website says

Equipment used: Celestron EdgeHD 14 telescope, iOptron CEM70 mount, Astro-Physics BARADV lens, ZWO ASI462MC camera, 7,120 mm f/20, multiple 15-millisecond exposures 

So the 'real' window was 100 frames (without any selection), e.g. 1800 ms. Too short for anything as crisp on Mars as you see on the final photo.

[vlaiv]

1 minute ago, BGazing said:

Too short for anything as crisp on Mars as you see on the final photo.

I'm not suggesting one should use such a short period of time to create whole image.

I'm just saying that you can create reference frame that will be used to stack other subs onto (to calculate AP position / transformation).

In other parts of video there will be usable sections of both Moon and Mars that can be stacked on to this reference image - provided that you don't attempt to stack whole sub, but you do it AP by AP - or take small section of sub where Mars and Moon are apart - one that covers piece of Mars and then stack that to appropriate place in reference image - something like this:

left is some frame in video before occultation began and right is reference frame created out of those 100 subs. You can take alignment point on left frame and stack it to appropriate place on right point - but you can do that only for APs that are visible in right image - but that is enough - that is what you want.

[vlaiv]

I've been reading explanations on AstroBin - and this is essentially what he did - except he did it manually

He did create "reference" image and then stacked mars from previous video and blended the two in Photoshop.

My proposal above was to let stacking software do the composition - but general idea is the same.

 

[BGazing]

20 minutes ago, vlaiv said:

I've been reading explanations on AstroBin - and this is essentially what he did - except he did it manually

He did create "reference" image and then stacked mars from previous video and blended the two in Photoshop.

My proposal above was to let stacking software do the composition - but general idea is the same.

 

And there is probably a reason why he chose the pedestrian way. I suspect that letting AS do it would inevitably create artefacts and/or overlaps...

[Elp]

I actually recorded the occultation as it happened but I did it in widefield as I wanted to capture the whole moon. No double exposures or anything, captured it as it was seen, the point of contact and reveal was in real-time:

https://flic.kr/p/2o5iHHf

 

[tomato]

2 hours ago, vlaiv said:

Yeah, not sure about that

vs

Nah, I think the top one has been over processed.😉