Our Beautiful Universe - Photos and Videos

[Andy Resnick]

Darn, they were quite cheap on Amazon. I'll buy me a pack of those, they could be fun to play with . Thanks for the idea!
And it was a very cool photo with overlayed spectra!

@Andy Resnick , by the way, I remember you said you bought a new hard drive (or drives?) on Amazon recently. Would you mind saying which one? I'm thinking of buying two new ones, I'm starting to run a bit low on disk space.

I got this one:

https://www.amazon.com/dp/B08XKMBTXH/?tag=pfamazon01-20

14 TB should last me a while :)

 

[Andy Resnick]

@Drakkith , @Tom.G , @collinsmark :

Thanks for the education! This weekend is supposed to be clear, so I'll have some time to try all this out.

 

[Andy Resnick]

Last night I had 7 moon-free hours of viewing and now I have a better (but still incomplete) understanding of what my mount driver is doing, PEC correction, etc. It's a long-ish post but the content may be useful to the community. I tried to experimentally validate parts of @collinsmark's thorough discussion above.

Here's what I did: first, setup and balance the lens on the mount. Next, sighting through the polar alignment reticle, obtain (approximate) polar alignment using the azimuth-elevation mount adjustment. After positioning the lens in the 'start' position (counterweight down, pointing at Polaris), turn on the mount driver and after deleting all the prior PEC data (only for the purposes of this experiment), generate a "mount model" by pointing to 4 (bright) reference stars well-separated by hour angles- this is what I was previously calling 'the GoTo alignment process'.

Then I imaged Neptune and Uranus, each for about 20 minutes before turning to M45 (Pleiades). At this point, there is no active PEC and I have a partial mount model. The first 200 frames of M45 (6 s subs, 10 sec intervals), when 'stacked' and z-projected, shows how the stars apparently move in my field of view due to all of the various tracking errors:

The star motion in the field of view has an AC and DC component: the AC component is RA-only with an amplitude of 155 pixels (corresponding to 193 arcsec in my setup) and period of 8 minutes. The DC component is drift in *both* RA and DEC in nearly equal amounts. My 'acceptance rate' of sub exposures is around 20%, which sounds low but is double was it was before replacing the gearbox.

Then I imaged M45 for 2 hours, doing no additional alignments or focus checks. After 2 hours, the 200-image z-projection looks like this:

It's almost exactly the same, with the primary difference being defocus, most likely from temperature changes over the 2 hour period. If focus was preserved, I would still keep 20% of the subs.

Then I performed a single PEC training run and switched to 400mm to image the Horsehead nebula (IC434). Because the lens focal length decreased, I scaled the next z-stack images accordingly.

When I removed the 2X tele, I had to start over: turn off the mount, remove the lens, re-mount and rebalance, etc. etc. After building the initial mount model, the 200-image z-projection looks like this (10s subs, 13s intervals):

The effect of PEC correction is barely noticeable- the AC component is 175 arcsec, and while the RA drift is more, the DEC drift is less. I should point out that during PEC training, there was no indication the star position would vary so wildly- the star did not move anywhere close to 140 pixels, more like 30 pixels (peak to peak).

I started imaging IC434 and after every 30 minutes, I stopped imaging, refined the mount model, and checked focus before resuming imaging. After 4 hours (1:30am!), the 200-frame z-projection looks like this:

Which is a significant improvement- the AC component amplitude is now only 43 arcsec, the periodicity has approximately halved, and RA drift is nearly absent. To reiterate, no additional PEC training runs were performed- only periodic refining/improving the mount model. At this point, my sub exposure acceptance rate is closer to 50%.

My conclusion from this is that the stepper motor rates are modified by both PEC data and mount model parameters.

I also conclude that pointing errors cause not just DC drift in RA and DEC, but also an oscillatory component in RA. I cracked open my copy of W. M. Smart's "Spherical Astronomy", but there's no discussion about this point.

Finally, I conclude that I can get significant performance improvement just by repeatedly refining the mount model, aligning to more and more stars (spanning an increasingly large hour angle) over time. The driver software must use the mount model to compute how the mount's local coordinate system maps to the reference coordinate system, and drives both RA and DEC motors accordingly. I realize that some driver software packages may not do this, but mine (apparently) does.

Woot!

 

[Drakkith]

Which is a significant improvement- the AC component amplitude is now only 43 arcsec, the periodicity has approximately halved, and RA drift is nearly absent. To reiterate, no additional PEC training runs were performed- only periodic refining/improving the mount model. At this point, my sub exposure acceptance rate is closer to 50%.

After rerunning the GoTo alignment, where did you place the target in the FoV? How does that affect your z-projection? I also don't understand how you could be 'refining' the mount's internal model by repeated GoTo alignments. Shouldn't it simply be as good as the last alignment procedure allowed?

 

[Andy Resnick]

After rerunning the GoTo alignment, where did you place the target in the FoV? How does that affect your z-projection? I also don't understand how you could be 'refining' the mount's internal model by repeated GoTo alignments. Shouldn't it simply be as good as the last alignment procedure allowed?

Model targets ('fiducials') are always placed in the center of the FoV, and it does not impact the z-projection. 'refining the model', according to Losmandy, means re-aligning to prior fiducials and aligning to new fiducials as they rise and become visible as the night goes on. For example, I started by creating an alignment model using Capella, Aldebaran and Jupiter. Later in the evening, I refined the model using Castor, Betelgeuse, and Rigel. The GoTo software has a list of 40-odd fiducials (bright stars), but I can use any object in the available catalogs- for example, Jupiter.

I suspect the model parameters are generated by some kind of eigenvalue/least-square fit equation- there is a matrix of reference coordinates for the various fiducials and a matrix of "mount coordinates" created during the alignment process.

Remember- the pointing direction of the telescope drifts as time passes; periodically re-aligning to the same set of fiducials apparently helps compensate for that drift.

Does that make sense?

 

[Drakkith]

Model targets ('fiducials') are always placed in the center of the FoV, and it does not impact the z-projection.

I mean your imaging target, not the GoTo alignment target stars (fiducials? Never heard that word before). I assume you have to move your scope off of your imaging target and onto the new alignment targets for every GoTo alignment. When you move your scope back to your imaging target, you place it back in the center of the frame, correct? I'm not sure how you make a z-projection, so I'm curious as to how re-centering the object affects it, if at all. Do you generate the z-projection from differences between each image, or from something else?

Remember- the pointing direction of the telescope drifts as time passes; periodically re-aligning to the same set of fiducials apparently helps compensate for that drift.

Does that make sense?

Somewhat. It makes sense that a new GoTo alignment would correct for any drift problems that the mount model couldn't compensate for, but how that would affect your tracking is beyond me given my understanding of how a GEM mount works.

 

[Andy Resnick]

I mean your imaging target, not the GoTo alignment target stars (fiducials? Never heard that word before). I assume you have to move your scope off of your imaging target and onto the new alignment targets for every GoTo alignment. When you move your scope back to your imaging target, you place it back in the center of the frame, correct? I'm not sure how you make a z-projection, so I'm curious as to how re-centering the object affects it, if at all. Do you generate the z-projection from differences between each image, or from something else?

Ok, let me break it down a little more. First- fiducial markers are, in the context of imaging, fixed reference points used to calibrate an imaging system: you can calibrate aberrations, magnifications, steering mirrors, wavelengths, any aspect of an imaging system by using appropriate fiducials. Many of the Webb telescope's alignment stages used fiducial markers (a bright star) to align the different mirror segments. In the context of my imaging setup, fiducial markers (bright stars) are used to compare the initially unknown rotation axis of my mount (the mount model) to the rotation axis of the Earth (the reference).

My imaging target is not one of these alignment targets. Creating the mount model ensures that the mount can slew to an imaging target and center it in my field of view. Refining the mount model, over time, reduces the uncertainty associated with the mount model.

the z-projection images should have perhaps be called 't-projection images', because successive images are taken at different times- I used terminology from microscopy (especially confocal microscopy). Imagine assembling a series of images into a cube, with axes labeled (x,y,t). I can cut the cube along various axes- in confocal microscopy, the image cube typically has axes (x,y,z), and it's standard to generate x-z or y-z images to get a 3-D sense of what you are imaging. I can also 'look through a face' of the cube. In this case, it doesn't make sense to generate x-t or y-t images, so instead I generated an x-y image where the brightness at any pixel is the maximum brightness along the t-axis.

If there is no drift/wobble/movement of the star during imaging, the t-projection would simply show the star. If the star is drifting or otherwise moving in the field of view, that movement appears as a streak in t-projection. If I generate a t-projection using a timespan that includes a mount model refining (or polar alignment adjust), the streak will have a large gap that corresponds to the correction of pointing error (and the post-realignment streak may partially overlap/obscure the pre-realignment streak), so I try not to do that.

Somewhat. It makes sense that a new GoTo alignment would correct for any drift problems that the mount model couldn't compensate for, but how that would affect your tracking is beyond me given my understanding of how a GEM mount works.

Yeah, this part is what I don't understand either. The best I can figure is, using the language of control theory, the Losmandy driver software computes the error signal by comparing the mount model and reference, then uses the error signal to generate a control signal for the motor drivers. Each time I refine the mount model, the error signal is re-computed and over time the control signal improves, reducing the tracking error.

How's that?

 

[Drakkith]

If there is no drift/wobble/movement of the star during imaging, the t-projection would simply show the star. If the star is drifting or otherwise moving in the field of view, that movement appears as a streak in t-projection. If I generate a t-projection using a timespan that includes a mount model refining (or polar alignment adjust), the streak will have a large gap that corresponds to the correction of pointing error (and the post-realignment streak may partially overlap/obscure the pre-realignment streak), so I try not to do that.

So what do you do? Do you put the imaging target back where it was in the frame before you moved to do realignment, or do you re-center it?

Yeah, this part is what I don't understand either. The best I can figure is, using the language of control theory, the Losmandy driver software computes the error signal by comparing the mount model and reference, then uses the error signal to generate a control signal for the motor drivers. Each time I refine the mount model, the error signal is re-computed and over time the control signal improves, reducing the tracking error.

I get what you're saying, I have just never heard of a GEM mount doing anything other than turning the RA axis at a set rate once you're on target, with perhaps some PEC thrown in.

 

[andrew s 1905]

I get what you're saying, I have just never heard of a GEM mount doing anything other than turning the RA axis at a set rate once you're on target, with perhaps some PEC thrown in.

Software Bisque Paramount mounts have had Protrack or many years. It uses feed forward control based on a pointing model. This is implemented in The Sky X pro software. It was developed by Patrick Wallace who has worked on many major telescopes pointing and tracking.

See here for so e of his work http://www.tpointsw.uk/index.htm

Regards Andrew

 

[Andy Resnick]

So what do you do? Do you put the imaging target back where it was in the frame before you moved to do realignment, or do you re-center it?

No, I don't re-center the imaging target prior to mount model refinement, I simply perform the mount model refinement and then go back to the imaging target.

I get what you're saying, I have just never heard of a GEM mount doing anything other than turning the RA axis at a set rate once you're on target, with perhaps some PEC thrown in.

Yeah, I know... that's the main reason it's taken me so long to figure out what is going on :)

 

[Drakkith]

No, I don't re-center the imaging target prior to mount model refinement, I simply perform the mount model refinement and then go back to the imaging target.

So you place the target back where it was on the last frame before refinement?

 

[phinds]

Interesting article from BBC

https://www.bbc.com/news/science-environment-64321824

outtake:

The number of stars that people can see with the naked eye has reduced dramatically over the last decade.

The cause is "Skyglow" from artificial lighting - the brightness of that glow has increased every year since 2011.

 

[Andy Resnick]

So you place the target back where it was on the last frame before refinement?

More or less- that happens all by itself. Remember, I am correcting for a few arcminutes of pointing error within a 2-degree FoV. The target doesn't need to cover the exact same pixels every frame, that's what the frame alignment step in stacking (as opposed to t-projection) accomplishes.

 

[Drakkith]

More or less- that happens all by itself. Remember, I am correcting for a few arcminutes of pointing error within a 2-degree FoV. The target doesn't need to cover the exact same pixels every frame, that's what the frame alignment step in stacking (as opposed to t-projection) accomplishes.

Sure, I just wasn't sure how not placing your imaging target back at the same spot in the frame would affect your t-projection.

 

[chemisttree]

It sounds to me like the AC component is a function of the parameters used to guide on the star. The star reaches a point in a “guide box” where the correction is applied until the star touches the other side of the “guide box” when the correction ceases and the constant error resumes. The mount then refines the size of the “guide box” and resamples the magnitude of the AC variation? Effectively changes the duty cycle of the correction input?

 

[DennisN]

An amazing photo of tall structures in Saturn's main rings (some "as high as 2.5 kilometers"):

Photojournal: PIA11668
Source: NASA/JPL/Space Science Institute
Published: November 1, 2010

Vertical structures, among the tallest seen in Saturn's main rings, rise abruptly from the edge of Saturn's B ring to cast long shadows on the ring in this image taken by NASA's Cassini spacecraft two weeks before the planet's August 2009 equinox.

Source: https://solarsystem.nasa.gov/resources/141/the-tallest-peaks/

Thanks to @BillTre who posted a link to it in the Random photos thread .

 

[DennisN]

Such a marvellous photo!

Jupiter and its rings from JWST

Jupiter and its rings from JWST JWST's Near-Infrared Camera (NIRCam) captured this image of Jupiter showing its auroras, rings, and moons Amalthea and Adrastea. [...]
Image: NASA, ESA, CSA, Jupiter ERS Team; image processing by Ricardo Hueso (UPV/EHU) and Judy Schmidt.

Source (with more nice photos): The best space pictures from JWST (planetary.org)

 

[jtbell]

Today I came home from my late-afternoon walk just after sunset, saw a nice crescent moon and (probably) planet setting over my house, and pulled out my iPhone XR. This is cropped and resampled in Photoshop, but with no other image adjustment.

I thought that must be Venus, so I looked online and sure enough, that's what it is. And Saturn is also in the picture! I didn't notice it before. It's just barely visible as a small smudge below and right of Venus, right next to a couple of small tree branches.

Here's a 1:1 pixel crop from the original image, with the levels adjusted to bring it out a bit better.

 

[Devin-M]

and (probably) planet setting over my house

That's Venus. If you draw a line from Venus to the Moon and then keep going past the moon about twice as far you can see Jupiter, then if you continue to follow the line and turn around about 180 degrees and look up you can see Mars. The very tiny dot just below and to the right of Venus right next to the tree in your picture is Saturn.

PS Edit: Tonight, Jupiter will be right next to the moon:

 

[Devin-M]

The Moon and Jupiter (bottom right) last night from my front yard, handheld with a 600mm f/9 lens on a Nikon D800 camera:

 

[DennisN]

a 600mm f/9 lens

Hmm... is it a mirror lens, or?

 

[DennisN]

This photo deserves it's own post too :

Exoplanet HIP 65426 b

Exoplanet HIP 65426 b

This image shows four JWST views of exoplanet HIP 65426 b in different bands of infrared light. The bar appearance in the NIRCam images is an artifact of the telescope’s optics. The white star icon in each image shows the location of the exoplanet's actual star, which was blocked by a coronagraph.
NASA/ESA/CSA, A Carter (UCSC), the ERS 1386 team, and A. Pagan (STScI)

Source: Exoplanet HIP 65426 b (The Planetary Society)

 

[Devin-M]

Hmm... is it a mirror lens, or?

It’s a Nikon 300mm f/4.5 + a Nikon TC-301 2x Teleconverter

 

[DennisN]

The Moon earlier tonight.

(it seems there was something strange going on during the stacking though, the left lower part seems a bit weird)

Gear: Sony A6000, Tokina 400mm f/5.6, Vivitar 3x Teleconverter, Rollei wireless intervalometer
Settings: 1/1000s exposure, ISO 800
Software: PIPP, AutoStakkert, Photoshop

 

[Drakkith]

Tried finding comet C/2022 E3 (ZTF) tonight, but couldn't seem to find it. Not sure if it's too bright here in Bossier City, or if I've just had a little too much to drink... (I'm a severe lightweight).

 

[collinsmark]

Speaking of which...

Here's Comet C/2022 E3 (ZTF), taken Friday morning (2023-01-27) from my back patio.

I've mentioned before that astrophotography is an exercise in failure. A celebration of failure! An onion of failure that you have to peel one layer at a time, with each peel revealing yet another layer of failure. It's OK though, as long as I learn a little along the way. Needless to say, the above image wasn't my first attempt (fourth?, fifth?). I could seriously go on and on about this one. I'm beat. Sooo much "learning."

Processing comets is a real challenge, since they move quite quickly against the background stars. Ya have to get quite clever in stacking techniques to avoid streaking in the stars and/or the comet itself.

Equipment:
Sky-Watcher EQ6-R Pro mount
Explore Scientific ED80-FCD100 telescope
Astronomik RGB filters
Orion Field Flattener for Short Refractors
ZWO ASI1600MM-Pro monochrome camera
Guide scope and ASI290MM-mini guide camera

Software:
N.I.N.A.
PHD2
PixInsight

Integration:
Bortle class 7 (maybe 8 ) skies
R: 57×40 sec
G: 57×40 sec
B: 57×40 sec
Subframes were acquired in a loop, alternating R, G, B, R, G, B..., and so on (this is important for later). Telescope tracking/guiding was done on the background stars (like normal). Total integration/exposure time: 1.9 hours. I would have acquired for longer but there was a fixed window from the time the comet rose above my building till morning twilight. I even moved my telescope to a part of the patio that makes no sense except for the comet (at the expense of most everything else. There's a whole story there).

Two phases of stacking/alignment were performed. Phase 1 is aligning on the background stars. When integrating that stack, heavy emphasis was placed on Winsorized Sigma Clipping for high valued pixels: The idea was the comet streak would be treated as a statistical anomaly, leaving pretty much only the background stars.

For phase 2, frames were aligned on the comet (same raw data on the previous phase, just aligned differently). But before alignment on the comet, the background stars (from the previous step) were subtracted from each frame, leaving pretty much only the comet after integrating that stack.

Finally, the stacked image of the stars were combined with the stacked image of the comet. Lots of other miscellaneous processing steps were performed, not mentioned here.

----

I have a slightly different configuration set up now as I write this. Hopefully I'll be able to get a wider field of view with it. Weather's looking iffy though. Clouds are coming in. Sooo much learrrning.

[Edit: clouds came in. No additional "learning" attempt for me today. I'm going to go munch on an onion.]

 

[DennisN]

Here's Comet C/2022 E3 (ZTF)

What a lovely color!

 

[DennisN]

Orion Field Flattener

Haha, when I read your post again, I read "Onion Field Flattener".

 

[collinsmark]

Comet C/2022 E3 (ZTF) taken from my back patio last night/this morning (morning of 2023-Feb-01).

I feel this is an improved one over my last post. 'New data, slightly new processing techniques, some new equipment, wider field of view, and longer integration/total exposure time.
'Same old light pollution though.

Equipment:
EQ6-R Pro mount
Explore Scientfic 80ED-FCD100 telescope
Orion 0.8x Focal Reducer for Refractors (first light)
Baader CMOS-optimized RGB filter kit
ZWO ASI2600MM-Pro main camera (first light)
Guide scope and guide camera

Acquisition software:
N.I.N.A.
PHD2

Processing Software:
PixInsight
GIMP

Integration:
Bortle class 7 (maybe 8 ) skies
All subframes binned 1x1
Comet:
R: 132×40 sec
G: 133×40 sec
B: 132×40 sec
Total Integration time of comet: 4.4 hours

The total integration time of of the background stars is closer to 5 hours, since I intentionally let the comet drift outside the field of view for awhile (helps with Phase 1, below).

Notes:
Processing was more-or-less the same as last time, Except that I was more careful with the stretching operations and color calibration. Brief summary:

• Phase 1: Aligned on the background stars. Integrated to get an image of background stars (letting the pixel rejection algorithm [Winsorized Sigma Clipping] do the heavy lifting of removing the comet artifacts).
• Phase 2: Re-aligned to create a new stack aligned on the comet's head. Integrated to produce a starless image of the comet.
• Phase 3: Combined the two images (stars and comet). Performed miscellaneous post processing.

I guess I'll call it a success. Close enough. That was my 5th? maybe 6th attempt at Comet C/2022 E3 (ZTF). I think I might move on now.

Goodbye green comet!

Oh, by the way, if you didn't know, its green color comes from diatomic carbon, apparently.

 

[DennisN]

I have a slightly different configuration set up now as I write this. Hopefully I'll be able to get a wider field of view with it. Weather's looking iffy though. Clouds are coming in. Sooo much learrrning.

[Edit: clouds came in. No additional "learning" attempt for me today. I'm going to go munch on an onion.]

Total Integration time of comet: 4.4 hours

I guess I'll call it a success. Close enough. That was my 5th? maybe 6th attempt at Comet C/2022 E3 (ZTF). I think I might move on now.

I admire your patience. And I am inspired by it.
And I like your shots a lot, I think they look cool!

I've never tried shooting a comet, I hope I'll come to do that.
I kept a close eye on the weather, and there was one (possible) opportunity for me for this comet, but I would have had to go at least down to the beach to try it.
And it was very cold and windy. And late.*
So I waited, but all I got afterwards was grey skies.

* Edit:

There are two types of people who would go down and stay on a dark beach on a cold and windy night in the middle of the winter: crazy ones and astrophotographers (and the two types are not necessarily mutually exclusive ).

 

[DennisN]

There are two types of people who would go down and stay on a dark beach on a cold and windy night in the middle of the winter: crazy ones and astrophotographers (and the two types are not necessarily mutually exclusive ).

Speaking of which, here's a photo I found when I went through old photos.
(shot with my phone about two years ago, before I had a system camera).

It was REALLY cold that winter night, so I put on a REALLY warm jacket (it's actually a really good jacket).
I was probably just testing shooting the night sky with my smartphone.
And what a crappy spot I chose, it looks like there are spotlights directed at the camera .
Despite this, somehow the phone managed to pick up Orion and some other stars.

 

[collinsmark]

The Flaming Star Nebula (also called IC 405, SH 2-229, and Caldwell 31), imaged from my back patio from late Dec 2022 through Jan 2023.

The Flaming Star is a reflection and emission nebula in the constellation Auriga. Most of what you see here is from the emission spectra, since it was imaged exclusively with narrowband filters.

We at the Shady Crypt Observatory like the Flaming Star Nebula because it kinda looks like a zombie.

Equipment:
Meade 10" LX200-ACF fork mounted on an equatorial wedge.
Starlight Instruments FTF2008BCR focuser modified for electronic focusing.
Off-axis guider (OAG) with ZWO ASI174MM-mini guide camera.
Baader 3.5-4 nm Ultra-Narrowband filter set.
[Edit: I might have switched to the Optolong 3 nm narrowband filter set at some point.]
ZWO ASI6200MM-Pro main camera.

I used to regularly have dreams about zombies. In most of them I was fighting the zombies, trying to survive and such. But in the really good dreams, not only did I lose the battle, I became a zombie myself.

Software:
Nighttime Imaging 'N' Astronomy (N.I.N.A.)
PHD2 guiding
PixInsight with RC-Astro plugins

It really wasn't bad being a zombie. It was kind of peaceful, really. Stress was low. Life was fairly simple. Most of existence consisted of wandering around, looking for brains and stuff.

Integration:
Bortle class 7 (maybe 8 ) skies
All subframes binned 3×3
SHO mapping
SII: 68×10 min = 11.33 hrs
Hα: 81×10 min = 13.5 hrs
Oiii: 71×10 min = 11.83 hrs
Total integration time: 36.67 hours

I wonder if that's what it's like working in Human Resources.

 

[Andy Resnick]

An improved stack of Comet C/2022 E3- fixed the registration process and did a better job combining the two separate stacks:

 

[bruha]

Excelent! Can I ask what is this bright spot lower right ?

 

[Andy Resnick]

Excelent! Can I ask what is this bright spot lower right ?

Thanks! That bright star is "HD 42818", one of the bright stars in the constellation Camelopardalis.