Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

AstroShader iOS AP app: Intro + beta-testers wanted

Smartphone AP Software Astrophotography
  • Please log in to reply
731 replies to this topic

#476 joezerneem

joezerneem

    Sputnik

  • -----
  • Posts: 37
  • Joined: 31 Jul 2023

Posted 13 April 2024 - 01:07 PM

You're welcome!

 

I'm pretty sure the focus was optimal. I also checked the collimation of the Dobsonian before I started, which was perfect.

 

The max. exposure time with 40x magnification (lowest I can get with my 10 inch Dobsonian telescope) before getting star trails (untracked setup), seems to be around 0.333 sec.
However as a test, I can try using longer exposure times next session and see if this will prevent these alignment issues.

 

I didn't use the editing tools here, because all multiple exposure images (2x and 5x) failed with alignment.

 

I haven't tried using a different eyepiece with higher magnification (60x or 120x) yet, but this will give even less alignment stars (and therefore I assume more alignment issues). However, they may show better contrast between stars and background, so it may be worth trying anyway.

 

Any additional suggestions for testing are welcome.


  • Seawild likes this

#477 joezerneem

joezerneem

    Sputnik

  • -----
  • Posts: 37
  • Joined: 31 Jul 2023

Posted 14 April 2024 - 09:14 AM

Yesterday I got an unexpected opportunity to do some additional tests, despite the many veil clouds.
First I wanted to test whether an increased exposure time (ignoring star trailing) would improve upon alignment / stacking issues.

 

To prevent losing time, I started to point my telescoop pretty randomly to (a clear looking part of) the sky, with a few stars visible in the eyepiece.
Then I ran 3 series (exposure time 0.333, 0.666, 0.999 sec.) of 3 images with a single exposure, 2x exposure and 5x exposure.
You can see the results below. All the images seem to fail with 2 or 5 exposures.
For this forum, I had to compress the images to 400x300 jpg this time, I hope it is still usable.

 

After that, I tried the binning option (5 exposures with 0.333s exposure time) which appeared to be a great improvement (see image below).

 

I continued testing using binning with more exposures and this still worked well. I noticed Ursa Major looked pretty clear, so I pointed my telescope towards M51. I increased the number of exposures a few times with mostly good results and in the end, I managed to take a decent picture with 500 exposures total, with the reposition interval set to 100 (see picture below, edited with the Astroshader tools). The two galaxies M 51 and NGC 5195 are clearly visible as fuzzies.
Just in time before bigger clouds forced me to end my session.

Attached Thumbnails

  • iso_16000_0.333s_x1_13-Apr-2024_22.51.19.634.jpg
  • iso_16000_0.333s_x2_13-Apr-2024_22.51.52.342.jpg
  • iso_16000_0.333s_x5_13-Apr-2024_22.52.28.310.jpg
  • iso_16000_0.666s_x1_13-Apr-2024_22.57.06.690.jpg
  • iso_16000_0.666s_x2_13-Apr-2024_22.57.41.709.jpg
  • iso_16000_0.666s_x5_13-Apr-2024_22.58.24.698.jpg
  • iso_16000_0.999s_x1_13-Apr-2024_22.59.03.098.jpg
  • iso_16000_0.999s_x2_13-Apr-2024_22.59.48.013.jpg
  • iso_16000_0.999s_x5_13-Apr-2024_23.00.45.145.jpg
  • iso_16000_0.333s_x5_13-Apr-2024_23.02.43.578.jpg
  • iso_16000_0.333s_x500_13-Apr-2024_23.58.27.616.jpg

  • GSBass and Seawild like this

#478 TX Dave

TX Dave

    Lift Off

  • *****
  • Posts: 3
  • Joined: 02 Apr 2024

Posted 14 April 2024 - 10:14 PM

I am a total newbie to astronomy and astrophotography. I stumbled upon an AstroShader video on youTube. So glad I found this app. Super easy to use and works very well. I took this shot last night with my iPhone 13 and my SVBony MK105 telescope with a SVBony 20mm eyepiece and moon filter. 200 exposures. Sorry to interrupt this thread. I'm just excited to share my first image on Cloudy Nights.

moon3

Edited by TX Dave, 14 April 2024 - 10:16 PM.

  • GeezerGazer, Seawild, Domdron and 1 other like this

#479 GSBass

GSBass

    Voyager 1

  • *****
  • Posts: 10,519
  • Joined: 21 May 2020
  • Loc: South Carolina

Posted 15 April 2024 - 04:06 PM

This was a photo i took last night with AstroShader and hestia, I think I had it set to stack 5,  no post processing,thought it looked nice, just resized to fit here so I sure it will rake a hit

Attached Thumbnails

  • IMG_4189.jpeg

  • GeezerGazer, Seawild, joezerneem and 1 other like this

#480 Josephus Miller

Josephus Miller

    Ranger 4

  • *****
  • Posts: 369
  • Joined: 16 Feb 2023
  • Loc: New Mexico, USA

Posted 27 April 2024 - 06:22 AM

I've had this on my phone for like a year now but haven't done much with it. However, now that I have a) gotten a tracking mount, and b) read the instructions... wow! 

 

M3 from last night:

iso_494_6.999s_x5_26-Apr-2024_21.44.36.124.JPG

 

So much more than I could see in the eyepiece. I'll definitely be coming back to this soon. 


  • GeezerGazer, GSBass, Seawild and 3 others like this

#481 joezerneem

joezerneem

    Sputnik

  • -----
  • Posts: 37
  • Joined: 31 Jul 2023

Posted 10 May 2024 - 08:39 AM

I captured the globular cluster M13 with higher magnitude (120x) this time. I managed to capture the images without using binning, but had to zoom in at least a bit, to get decent images.
Also M57 was captured this way.

 

The reposition interval was set to 25.

 

A small thing I noticed, is that the number of exposures in the file name stays the same when finalizing an image prematurely. For example, when I set the total number of exposures to 1000 and finalize after 200 exposures, the filename still shows x1000.

 

The images are edited with the Astroshader tools and GIMP.
Converted to .jpg and 25% compressed to fit on the forum.

 

Both images contain stars with magn. above 15! Visually, 13.5 to 14.0 is the absolute max. I can reach under these weather conditions.
I'm very pleased with the results, M57 even shows the red color in the edge of the ring, which I can't see visually!

 

- 10 inch Dobsonian (manual)
- Celestron Nexyz smartphone adapter
- iPhone 6S
- 10mm plössl eyepiece (magn. 120x).
- Bortle 6 sky

 

M57

- iso_16000_0.333s_x250_10-May-2024_01.54.26.032

 

M13

- iso_16000_0.333s_x500_10-May-2024_00.22.29.836

 

 

Attached Thumbnails

  • iso_16000_0.333s_x250_10-May-2024_01.54.26.032_M57_25perc.jpg
  • iso_16000_0.333s_x500_10-May-2024_00.22.29.836_M13_25perc.jpg

  • GeezerGazer, Pinac, Seawild and 2 others like this

#482 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 10 May 2024 - 03:10 PM

I captured the globular cluster M13 with higher magnitude (120x) this time. I managed to capture the images without using binning, but had to zoom in at least a bit, to get decent images.
Also M57 was captured this way.

Joe, 

Those are excellent phone images, especially from a non-tracked mount!  Very nice job.  waytogo.gif

 

I was out last night too, and managed a couple of images at 1s/300 @ ISO 16000 of both M13 and M57.  A tracking mount permits more latitude in settings, but because of mount malfunctions last night, I limited the exposure to 1s.  frown.gif   The image of M57 clearly shows the central star at magnitude 15, which is impossible to see visually with only 8" of aperture.  This was taken through an APM eyepiece at 15.4mm, and this image was cropped quite heavily for this scale.  Like you, next time, I'll use a bit more magnification.  bangbang.gif   Click on the image for a bigger rendition.

 

iso_16000_1.0s_x300_09-May-2024_23.59.54.200 2.jpg


Edited by GeezerGazer, 10 May 2024 - 03:20 PM.

  • Bill Jensen, rlmxracer, Seawild and 3 others like this

#483 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 11 May 2024 - 12:00 PM

Here is a pretty good example of why AS is my goto night sky iPhone camera app.  The other night I took a photo of M13 with 100/1sec images through a C8 and 15.4mm eyepiece, on a tracking mount.  I compared it to photos I took in 2023 using NightCap in Stars Mode, which takes an automatic 10sec image, also from a tracking mount.  Both previous photos, taken with Stars Mode through an 8" Newt and a 125 EDL refractor looked very nearly identical, with the center of the cluster shaded with white light.  This does not happen with AS, which shows individual stars to the core for a more realistic image of the globular.  All of the previous photos I took of globs using Stars Mode in NightCap look about the same with a white core.  Stars Mode is convenient and fast, but it's not nearly as good as AS!  

 

Both of these images were converted to B&W for an equal appearance and put into the same orientation for a side-by-side comparison.  Click on the image to see detail. 

Ray

 

Screenshot 2024-05-11 at 9.29.43 AM.jpeg


  • rlmxracer, lwbehney, eyeoftexas and 6 others like this

#484 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 12 May 2024 - 09:34 PM

Last night I was out taking photos using AS with a C8 and an APM Zoom at 15.4mm (no filter) until transparency fell off a cliff, but was able to get this decent image of the Dumbbell Nebula, M27, stacking 40, 8sec exposures at ISO 3026, post processing the PNG file of this image in AS to save as a JPEG.  It shows good detail with the central star visible.

 

Click on photo for details:

iso_3026_7.998s_x40_12-May-2024_01.07.03.245.jpg


  • eyeoftexas, Seawild, astrshdr and 3 others like this

#485 astrshdr

astrshdr

    Developer - AstroShader App

  • *****
  • Vendors
  • topic starter
  • Posts: 186
  • Joined: 30 Nov 2022

Posted 16 May 2024 - 07:13 AM

I captured the globular cluster M13 with higher magnitude (120x) this time. I managed to capture the images without using binning, but had to zoom in at least a bit, to get decent images.
Also M57 was captured this way.

 

The reposition interval was set to 25.

 

A small thing I noticed, is that the number of exposures in the file name stays the same when finalizing an image prematurely. For example, when I set the total number of exposures to 1000 and finalize after 200 exposures, the filename still shows x1000.

 

The images are edited with the Astroshader tools and GIMP.
Converted to .jpg and 25% compressed to fit on the forum.

 

Both images contain stars with magn. above 15! Visually, 13.5 to 14.0 is the absolute max. I can reach under these weather conditions.
I'm very pleased with the results, M57 even shows the red color in the edge of the ring, which I can't see visually!

 

- 10 inch Dobsonian (manual)
- Celestron Nexyz smartphone adapter
- iPhone 6S
- 10mm plössl eyepiece (magn. 120x).
- Bortle 6 sky

 

M57

- iso_16000_0.333s_x250_10-May-2024_01.54.26.032

 

M13

- iso_16000_0.333s_x500_10-May-2024_00.22.29.836

Nice images!

 

Thanks for highlighting the exposure number vs filename issue - I'll get that fixed...


  • joezerneem likes this

#486 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 02 June 2024 - 04:59 PM

Saturday morning I took photos from midnight until 04:30, when Luna popped up.  At 03:05 I was using my C8, with Night Vision for 76x, and my iPhone 15 ProMax with AstroShader set to stack 260 - 1sec images at ISO 2756.  I was tickled to see such detail in the unprocessed image  of the Pillars of Creation in the Eagle Nebula when the image was complete with 4.3 minutes of total exposure.  

 

The original of this photo was saved as a PNG, with a file size of 60.1 MB... this image is 75 KB as a JPEG.

Ray

 

Click on this photo and it will open in a larger rendition to see detail.

IMG_5647.jpeg


Edited by GeezerGazer, 02 June 2024 - 07:33 PM.

  • Seawild, vtaDan, astrshdr and 3 others like this

#487 Thrifty1

Thrifty1

    Viking 1

  • *****
  • Posts: 716
  • Joined: 24 Feb 2006

Posted 11 June 2024 - 06:13 AM

I had my C6 telescope out for a quick 30 minute session last night. I wanted to try the AstroShader app on a star cluster.

This is M13 Hercules Cluster taken with my iPhone 14 Pro on a Meade 26mm Plossl eyepiece.

ISO 2537
2 seconds exposure
20 frames stacked

Processed further with Lightroom on iPhone.

Attached Thumbnails

  • IMG_8377.jpeg

  • GeezerGazer, eyeoftexas, GSBass and 5 others like this

#488 Seawild

Seawild

    Explorer 1

  • -----
  • Posts: 75
  • Joined: 15 Oct 2020
  • Loc: Mariana Trench, Indiana

Posted 11 June 2024 - 10:04 AM

Howdy!  Has anyone had any success using longer exposures like 5 seconds x 1200 seconds?  Or longer?  I'm going after the Lagoon Nebula tonight.  Need some advice!

 

I have read that using AS @ +1sec x ???, might reduce saturated noise.  Any feedback/info appreciated!

 

My new favorite cluster. M4. 

 

(9mm EP + 102x600 Refractor = 66x Magnification) + Digital zoom and Crop  / 1sec x 345 = 5.75 mins / ISO around 4280. / Tracking. / iPhone 8+

Attached Thumbnails

  • iso_4280_0.999s_x345_10-Jun-2024_00.03.18.229 M4 CL.jpg

Edited by Seawild, 11 June 2024 - 11:41 AM.

  • GeezerGazer, eyeoftexas and joezerneem like this

#489 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 11 June 2024 - 02:28 PM

Howdy!  Has anyone had any success using longer exposures like 5 seconds x 1200 seconds?  Or longer?  I'm going after the Lagoon Nebula tonight.  Need some advice!

Seawild, yes Lagoon can be imaged with AS.  In AS, the bulk of photo averaging takes place with longer exposures of anything more than 1sec.  At one second or less, minimal averaging takes place as every exposure is aligned and stacked.  The longer the exposure setting (2 sec or more), the more averaging takes place.  Averaging is what removes noise from the image... small bits of sensor noise that appears in one image, will be cancelled out by the next image, because sensor noise changes randomly from one image to the next.  With an image of the dark sky, the sensor is working very hard, while being photon starved, so noise can be a huge distraction.  Photo averaging cleans up all or most of that noise.  But in AS, the bulk of averaging takes place with multi-second exposures.

 

As an additional benefit, AS brightens multi-second exposures.  There is a significant, although incremental image brightness increase as you move the exposure slider from 1 to 100 seconds.    This was done to simulate a single long exposure such as can be taken with an Android phone or DSLR.  I typically use exposures of 1 to 10 sec with ISO's as high as 16000 on my iPhone 15.  

 

Exposures of 1sec or less, are not averaged before being aligned and stacked.  So why would I ever use an exposure of 1 sec or less?   I actually use 1 sec exposures regularly, because of wind gusts at my observing location.  Exposures of 1 sec or less are ALL aligned and stacked.  With an exposure of 10 seconds, you rely on your mount's tracking to keep the images aligned while averaging is taking place.  A gust of wind on a 8" Newt can ruin the 10 sec exposure.  But if every short exposure is aligned and stacked, there is a much higher chance of completing a useable image.  

 

I'd suggest starting with exposure set to 10 sec with number of exposures set to 10, just to see how the image comes out.  If the noise level in that final image is OK, increase the number of exposures to 50, 60 or even 150... however long you are willing to give the camera to define resolution.  If the image is too noisy, increase the exposure time to 12, 15 or even 20 sec.  Do not hesitate to use high ISO settings... even your maximum ISO.  You will not need a filter for the Lagoon because iPhones do not have an IR blocking filter built in.  When you've taken an image of M8 that meets your standards, move up to M-20, Triffid.  It is quite bright too.  

 

There is a photo of M8 in my gallery "Phonetography" album, but it is severely vignetted because of the zoom eyepiece I was using to take the photo.  It was taken in 2022 using NightCap, before I switched to AstroShader.  But that image will show about what you can expect with just 15 seconds of averaging 1 sec exposures:  https://www.cloudyni...59-geezergazer/

 

I look forward to seeing your images here.  Good luck and Clear Skies.

Ray


  • Seawild likes this

#490 Domdron

Domdron

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 24 Jan 2023
  • Loc: Thogoto, Kenya

Posted 12 June 2024 - 01:06 AM

In AS, the bulk of photo averaging takes place with longer exposures of anything more than 1sec.  At one second or less, minimal averaging takes place as every exposure is aligned and stacked.  The longer the exposure setting (2 sec or more), the more averaging takes place.  

I don't think that's correct. From what was previously discussed, I believe it works like this:

  • For <= 1 second exposures, they are exposed "normally" with the specified exposure time.
  • For > 1 second exposures, AS takes multiple <= 1-second exposures for each "nominal exposure", and then and averages/processes them together, but without aligning, to produce each "nominal exposure".
  • In both cases, the resulting (nominal) exposures are aligned and then stacked by averaging.

I.e. averaging always takes place, whether using more or less than 1-second exposures.

 

edit: changed terminology of sub/exposure to better match how it's called in the app.


Edited by Domdron, 12 June 2024 - 01:10 AM.

  • GeezerGazer and Seawild like this

#491 Domdron

Domdron

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 24 Jan 2023
  • Loc: Thogoto, Kenya

Posted 12 June 2024 - 01:11 AM

NB: Because of this, I don't really see the point of using > 1-second exposures vs. just increasing the number of exposures, except if I want to take >1000 seconds, the max. number of exposures currently available.



#492 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 12 June 2024 - 12:03 PM

Dom, I appreciate what you are saying, but let me explain with photos why I believe what I'm saying is correct.  Perhaps I'm misinterpreting something.  

 

In the photo comparison below, are two photos taken in a closet with minimal lighting on one side from the door cracked open.  The left photo was taken at ISO 16000, with a stack of 10 - one second images.  On the right, is the same lighting, same ISO, but with one 10-second exposure.  There is an obvious exposure difference because George has extended brightness control into multi-second exposures, as I explained in my longer post.  

Screenshot 2024-06-12 at 9.35.33 AM.jpeg

 

In this second comparison, I took the darker image (10 one-second images that are stacked and aligned) on the left and brightened it in post processing until it was slightly brighter than the single, 10 sec exposure, which is only averaged... not stacked and aligned.  This comparison shows a distinct difference in the amount of noise that is present in the two images.   

 

Please click on this photo to open and then expand it to see the differences.

 

Screenshot 2024-06-12 at 9.36.02 AM.jpeg

 

I have one more comparison to do, but will be home in a few hours to complete it.  It will show both images being brightened.  


Edited by GeezerGazer, 12 June 2024 - 05:41 PM.

  • Seawild and Domdron like this

#493 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 12 June 2024 - 05:36 PM

Sorry for the delay, I had important business that couldn't wait... I had to take my grandson kayaking!

 

OK, in this last comparison, BOTH images were brightened 1.00 by maxing out the Exposure and Brightness settings in Photos editing tools on my Mac.  So the left photo is the same as the left photo in the second comparison, and now, the right photo has been brightened by an equal amount.  

 

Click on this photo to open and then expand it to see the difference in noise.  

 

Screenshot 2024-06-12 at 3.11.35 PM.jpeg

 

The photo on the right that average 10 one-second images has a great deal less noise visible, even with the severe brightening.  The aligned/stacked photo on the left is both darker and noisier.  

 

Dom, there might be another explanation, but it is clear to me that 10 - 1 second photos that are aligned and stacked are not equal to one 10 sec exposure that is simply averaged, both in terms of brightness and the elimination of noise.  I have seen this when using AS for my own photos, especially those taken with night vision.  Perhaps it is time for a clarification from George on this point.  

 

Your thought?  

Ray


Edited by GeezerGazer, 12 June 2024 - 05:38 PM.

  • Pinac, Seawild and Domdron like this

#494 Seawild

Seawild

    Explorer 1

  • -----
  • Posts: 75
  • Joined: 15 Oct 2020
  • Loc: Mariana Trench, Indiana

Posted 13 June 2024 - 01:22 AM

M8 Lagoon Nebula

 

Thank you all for the great info!  I do not have a star-lock and am trying to keep everything as low budget as possible.  So my 2 second+ exposers all have trails.  (I did not Polar align or fine tune, I'm still learning how to drive a GEM.)

I tried a few settings and bumped up my ISO quite a bit.  Here is the pick of the litter.  No processing, just extraction and brightness bump. Sorry about the CA. 

Over all, I'm pretty stoked.  Thank you for the help!

 

15mm EP + 102x600 refractor = 40x / iso_8211 / 0.999sec x 800 / 13-Jun-2024_01.04.13 / iPhone 8+ / AstroShader

Attached Thumbnails

  • iso_8211_0.999s_x800_13-Jun-2024_01.04.13.418 copy.JPG

Edited by Seawild, 13 June 2024 - 01:24 AM.

  • GeezerGazer, Domdron and Josephus Miller like this

#495 Domdron

Domdron

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 24 Jan 2023
  • Loc: Thogoto, Kenya

Posted 13 June 2024 - 01:52 AM

Dom, I appreciate what you are saying, but let me explain with photos why I believe what I'm saying is correct.  Perhaps I'm misinterpreting something.  

 

In the photo comparison below, are two photos taken in a closet with minimal lighting on one side from the door cracked open.  The left photo was taken at ISO 16000, with a stack of 10 - one second images.  On the right, is the same lighting, same ISO, but with one 10-second exposure.  There is an obvious exposure difference because George has extended brightness control into multi-second exposures, as I explained in my longer post.  

attachicon.gif Screenshot 2024-06-12 at 9.35.33 AM.jpeg

Thanks for testing this Ray! The brightness difference should be expected since the left is an averaged version of ten 1-second exposures, so should have the same brightness (but less noise) than a single 1-second exposure. The 10-second exposure OTOH, while doing some averaging/processing as well (as previously discussed), mostly adds the 10 1-second actual exposures. So it should have about the same brightness as a "real" 10-second exposure.

 

 

 

In this second comparison, I took the darker image (10 one-second images that are stacked and aligned) on the left and brightened it in post processing until it was slightly brighter than the single, 10 sec exposure, which is only averaged... not stacked and aligned.  This comparison shows a distinct difference in the amount of noise that is present in the two images.  

Please click on this photo to open and then expand it to see the differences.

Attachment Screenshot 2024-06-12 at 9.36.02 AM.jpeg

I have one more comparison to do, but will be home in a few hours to complete it.  It will show both images being brightened.

This I find surprising, i.e. the increased noise in the left image. If the 10-second image (on the right) is really composed of 10 added (+ averaged/processed) 1-second exposures, noise -- especially read noise -- should be the same for both versions. So maybe the extra processing in the 10-second image somehow removes some noise? 

 

Sorry for the delay, I had important business that couldn't wait... I had to take my grandson kayaking!

That's some great business you have there grin.gif

 

 

 

OK, in this last comparison, BOTH images were brightened 1.00 by maxing out the Exposure and Brightness settings in Photos editing tools on my Mac.  So the left photo is the same as the left photo in the second comparison, and now, the right photo has been brightened by an equal amount.
Click on this photo to open and then expand it to see the difference in noise.
Attachment Screenshot 2024-06-12 at 3.11.35 PM.jpeg
The photo on the right that average 10 one-second images has a great deal less noise visible, even with the severe brightening.  The aligned/stacked photo on the left is both darker and noisier.
Dom, there might be another explanation, but it is clear to me that 10 - 1 second photos that are aligned and stacked are not equal to one 10 sec exposure that is simply averaged, both in terms of brightness and the elimination of noise.  I have seen this when using AS for my own photos, especially those taken with night vision.  Perhaps it is time for a clarification from George on this point.
Your thought?
Ray

As mentioned, the 10 1-second exposures (left side) should be aligned and stacked using averaging. Stacking always requires some aggregating function, and that's usually the average, possibly with some rejection of outliers (e.g. sigma-rejection) which gets rid of things like satellite trails.

 

The 10-second exposure, if it's still composed of 10 1-second exposures, should mostly add those 1-second exposures but does some additional processing. 

 

However, if true, this shouldn't explain the difference in noise you see in the 2nd comparison, i.e. the brightened 10 x 1-second exposure stack vs. the unbrightened 10-second exposure. This difference should really only happen if the right side was a true 10-second exposure. I think we need @astroshdr to explain this.


Edited by Domdron, 13 June 2024 - 01:52 AM.

  • Seawild likes this

#496 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 13 June 2024 - 10:52 AM

Dom, 

It is obvious to me that some averaging does occur when the 10 one-second images are aligned and stacked, but the single 10 second exposure does benefit from a higher level of noise reduction... even though it is 10 separate exposures.  I have arguably assumed that the difference is from a higher level of averaging, but I am not technical enough to know how else that noise reduction could be occurring.  Yes, I think we need an explanation.  

Ray


  • Seawild and Domdron like this

#497 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 14 June 2024 - 12:40 PM

I got to wondering if a greater number of 1 sec exposures would make up for the difference in image noise compared to an averaged 10 sec exposure.  So back to my closet for more images at 16000 ISO, one 10 sec exposure to compare to 6 different 1sec exposures aligned/stacked for 1, 10, 20, 30, 40 and 50 seconds.  At one second, the amount of noise in the image was substantial, so I eliminated that one.

 

This I found strange and I have no explanation.  The screenshot photo below from left to right, is a stack of 10, 20, 30, 40, and 50 one second images, aligned & stacked.  It appears that with longer intervals of aligning and stacking, the image becomes slightly darker.  

IMG_5744.jpeg

 

I had to brighten the stacked images by 1.00 with the Exposure slider in Photos on the Mac.  Then I opened the images in Previewer, zoomed in and took a screenshot for a side-by-side comparison of the stacked exposures (each on the right) with the single/10 second averaged photo (always on the left).  At 10 and 20 seconds, it was clear that the single averaged photo on the left presented less noise.  With a stack of 30 one-second photos, the amount of noise was closer, but not equal.  See below, click on and expand the photo for detail.  

Screenshot 2024-06-14 at 10.01.04 AM.jpeg


  • Domdron likes this

#498 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 14 June 2024 - 12:54 PM

This comparison presents the stack of 40 one-second images on the right; the 10 second average photo remains on the left.  When I cropped these images, I left the tags at the top to show which image was which so I wouldn't mix them up.  

Screenshot 2024-06-14 at 9.53.16 AM.jpeg

 

 


  • Domdron likes this

#499 GeezerGazer

GeezerGazer

    Surveyor 1

  • *****
  • Posts: 1,890
  • Joined: 06 Jan 2005
  • Loc: Modesto, CA

Posted 14 June 2024 - 01:00 PM

This last comparison was the stack of 50 one-second images on the right.  This image in my original file of 1.45 MB, actually has some hot (red) pixels that were not present in the other stacked images.  But overall, the amount of noise is very similar to the 10 second averaged image.  

Screenshot 2024-06-14 at 9.55.19 AM.jpeg

 

There's lots of processing here to show the results, but it appears to me that to reach the level of noise reduction of a 10 second exposure that's been averaged, the stacked one-second exposure need 5x the number of subs.  That is a lot more than I anticipated.  

Ray


Edited by GeezerGazer, 14 June 2024 - 01:06 PM.

  • Domdron likes this

#500 astrshdr

astrshdr

    Developer - AstroShader App

  • *****
  • Vendors
  • topic starter
  • Posts: 186
  • Joined: 30 Nov 2022

Posted 14 June 2024 - 06:44 PM

Hi Ray and Domdron,

 

Firstly just to clear up the point about the extended exposures, e.g. <=1 second vs >= 1second (0.3s in older devices). Domdon your explanation here #490 is correct. Any exposure time beyond the 1 second limit is made up of smaller unaligned exposures. The only alignment and averaging occurs between "exposures", e.g. if the exposure slider is 1 then no alignment and stacking will occur—a single 50 second exposure will be a combination of x50 1 second exposures but they are not aligned and they are not averaged they are added * together. (* It's not really a pure addition but more analogous to this than averaging.)

 

Image noise varies in how it presents itself depending on the source of the noise and which part of the histogram it presents itself. The noise can also make different contributions to the image e.g. absolute vs relative effect on the signal, especially given some of the nonlinearities in image processing. The10s exposure has a lower apparent (and possibly actual) noise from the way the data is combined.

 

Ray you also noticed the image getting darker with the number of exposures and what you also can probably see is the contrast and colour preservation improving. This is where the largely positive contributions of noise in shadow regions are being slowly corrected with a processing technique during averaging that allows more shadow detail to be recovered in editing/post-processing. (What is really happening is that the image is being lightened less, as opposed to darkened. Testing is very tricky as ground truth is very hard to know, cameras always make assumptions about scenes).

Domdron you mentioned sigma-rejection, the previous technique and some others, relate to ways of achieving this without knowing the data in advance, e.g. being able to calculate or μ, as it has to be done on the fly and not retrospectively as with normal astrophotography software.

 

...In slightly funny timing this will change significantly next week. Part of the big update I have been working on will improve the shadow detail. The way it works is to combine Apple's regular image pipeline that uses their fancy AI with the existing AstroShader one. The results so far in testing have been very promising, all the detail from AstroShader but with the rich colour data Apple's magic AI is capable of producing.


Edited by astrshdr, 14 June 2024 - 06:44 PM.

  • GeezerGazer, Seawild, vtaDan and 1 other like this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: Smartphone AP, Software, Astrophotography



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics