Could it work?

[AstroGS]

I wanted to share an idea that I have had in my mind for some time now. What if a few of us could come together, agree on a target, agree on the acquisition data that each person would be required to capture, share the data with those that are better on processing and jointly create an image that we will be happy with. And why not, if we are really happy with it, to submit it for an APOD or Astrobin POD as a joint effort.

I would assume that having more people acquiring data for the same target but, each person to focus on one filter (if it is a narrowband target) or one part of the night sky (if it is a mosaic), it would provide more data to work with. I have seen some great images from other people collaborating on Astrobin (for example), that they produce some genuinely stunning images.

Dear moderators - if you feel that this is out of context or placed at the wrong forum, please feel free to manage it accordingly. Thank you 🙂

Edited December 6, 2023 by AstroGS

[Ouroboros]

Citizen Imaging!  I like the sound of that. How would it be coordinated? 

[AstroGS]

Just now, Ouroboros said:

Citizen Imaging!  I like the sound of that. How would it be coordinated? 

Not sure yet - open to suggestions. Initially, I wanted to see if there will be some interest for something like this 🙂

 

[badhex]

There is precedent for this - I think Astrobiscuit on Youtube has a discord where they run something he calls the BAT - Big Amateur Telescope.

An SGL based one would be great though. The mods might be able to set a club up for you in the clubs section even, which would give you a dedicated place to coordinate your efforts.

[bosun21]

I pity the poor souls who have to process all the different data files. One is enough for me.

[ollypenrice]

This has already been done very successfully - more than 13 years ago, no less.

https://apod.nasa.gov/apod/ap101005.html

In the past it really needed Registar to be workable but, these days, other software can co-register and resize data from different setups.

I collaborate with other imagers fairly regularly and also collaborate with myself, in that I use very old data to enhance a new image. The Squid data for this new image are about 10 years old and from a setup I no longer have.

Olly

[sinbad40]

 

one of my work colleagues based in France have started to do this with members from a French forum.  They based it on focal length of the telescopes and not the aperture like BAT.  They decide on a target, setup multi panels to be shared around etc. 

[AstroGS]

This is one of the images that inspired me: https://www.astrobin.com/zzoau6/ + the fact that the weather has been atrocious this year and it feels as I never get the opportunity to finish the projects the way I have imagined them. I would genuinely be happy to invest 8-10 hours of data on a target that we would all agree on. The thought to collect 100 hrs of data seems very exciting although, I would agree that processing will be quite the task!

It would be great indeed if the mods could help/ advice how we could evaluate the potential of this project.

 

[tomato]

I’m pretty sure that the data is already out there for a host of collaborative projects. I was in at the start of the BAT project but I just couldn’t keep up with the social media savvy banter on their Discord forum, probably an age thing.

Having said that I’d be more than happy to contribute data, as a galaxy enthusiast, my suggestion for a project would be a multi panel image of the Virgo region that you could just keep zooming in on.

[ollypenrice]

10 hours ago, AstroGS said:

although, I would agree that processing will be quite the task!

Not necessarily. The better the data, the easier the processing. 

Let's assume that, in the first example, you have a number of contributions which fully overlap. That's to say a single image, not a mosaic. These should be linear at this stage. These might be RGB, one shot colour (the same thing), luminance, Ha and OIII. Using Registar, you would put them all into one folder with nothing else in it. One of these frames would be your definitive crop, meaning all the others can cover it fully with, probably, a bit to spare.

1 Open just your definitive crop (AKA Reference Image) in Registar.

2 Choose 'multiple source' and click once on 'Register Images.' The software will open all the images and register each one, resizing and re-aligning where necessary,

3 Go to 'Crop and Pad' and click once so that each image will be cropped to be a perfect fit on your reference image. Save these.

So... in just two clicks you have a full set of linear images which fit each other perfectly, just as if they all came from the same telescope. Absolutely nothing to it.

Next, you'd want to remove gradients on all these images individually. I'd use DBE or ABE in Pixinsight. I'd also run SCNR Green where needed (it usually is) and Blur Xterminator.

Any images with the same filter would then need to be blended together. Eg you have three Ha contributions. I think some software will read them and weight them according to S/N ratio but I don't know about that. I'd probably give the three Ha images a basic stretch, stack them as Photoshop Layers and adjust their opacities till I got the cleanest blend and flatten them. Someone will know a more mathematical way of doing that. It wouldn't be the end of the world if you just gave all three to Registar, weighted them equally and settled for that, assuming they were all worth having.

Now a multi-source mosaic is always going to be more difficult. They are hard enough when everything comes from the same rig. I think APP is the software of choice for mosaics, at the moment. The first and most vital step would be flattening each panel (ABE or DBE/SCNR Green) before asking the software to make the mosaic.

I'd start with a single panel project! If nobody in the group has Registar I'd be happy to do the Register-Crop bit and send them back out. (Possible now we have fibre internet... )

Olly

[AstroGS]

On 07/12/2023 at 08:58, ollypenrice said:

Not necessarily. The better the data, the easier the processing. 

Let's assume that, in the first example, you have a number of contributions which fully overlap. That's to say a single image, not a mosaic. These should be linear at this stage. These might be RGB, one shot colour (the same thing), luminance, Ha and OIII. Using Registar, you would put them all into one folder with nothing else in it. One of these frames would be your definitive crop, meaning all the others can cover it fully with, probably, a bit to spare.

1 Open just your definitive crop (AKA Reference Image) in Registar.

2 Choose 'multiple source' and click once on 'Register Images.' The software will open all the images and register each one, resizing and re-aligning where necessary,

3 Go to 'Crop and Pad' and click once so that each image will be cropped to be a perfect fit on your reference image. Save these.

So... in just two clicks you have a full set of linear images which fit each other perfectly, just as if they all came from the same telescope. Absolutely nothing to it.

Next, you'd want to remove gradients on all these images individually. I'd use DBE or ABE in Pixinsight. I'd also run SCNR Green where needed (it usually is) and Blur Xterminator.

Any images with the same filter would then need to be blended together. Eg you have three Ha contributions. I think some software will read them and weight them according to S/N ratio but I don't know about that. I'd probably give the three Ha images a basic stretch, stack them as Photoshop Layers and adjust their opacities till I got the cleanest blend and flatten them. Someone will know a more mathematical way of doing that. It wouldn't be the end of the world if you just gave all three to Registar, weighted them equally and settled for that, assuming they were all worth having.

Now a multi-source mosaic is always going to be more difficult. They are hard enough when everything comes from the same rig. I think APP is the software of choice for mosaics, at the moment. The first and most vital step would be flattening each panel (ABE or DBE/SCNR Green) before asking the software to make the mosaic.

I'd start with a single panel project! If nobody in the group has Registar I'd be happy to do the Register-Crop bit and send them back out. (Possible now we have fibre internet... )

Olly

I do agree that a project like this can be as small or big as we want. I will need to see what will be the best way to organise something like this and then see how many would wish to participate. I ma always open to ideas of course from the more experienced members than me 😉

[Elp]

In terms of organisation it's why chatrooms and platforms like Discord are so popular, real time discussion and all. No reason why it couldn't work here for agreeing projects, could used a shared system similar to SharePoint or a shared GDrive or OneDrive account, it'd need decent data storage but I think if everyone just uploaded their linear calibrated stacks with details in the filename to an agreed standard, it'd be a fairly simple thing to do, someone on here might have their own server which would be easy to FTP (upload files) to.

Edited December 8, 2023 by Elp

[Allinthehead]

On 06/12/2023 at 19:13, ollypenrice said:

This has already been done very successfully - more than 13 years ago, no less.

https://apod.nasa.gov/apod/ap101005.html

In the past it really needed Registar to be workable but, these days, other software can co-register and resize data from different setups.

I collaborate with other imagers fairly regularly and also collaborate with myself, in that I use very old data to enhance a new image. The Squid data for this new image are about 10 years old and from a setup I no longer have.

Olly

I believe I recognise that Squid data!

[ollypenrice]

9 hours ago, Allinthehead said:

I believe I recognise that Squid data!

Are you sure? It came from a very long time ago and originally looked like this:

By using star removal and massive doses of NoiseXt I was able to get it to what you see in the present image but this did involve completely erasing all background and applying only the nebula itself, which is a bit borderline, ethically, if I'm honest. Needs must. In our earlier version, prior to adding the Fireworks, we used decidedly better Squid data supplied by SGL member AstroGS and published jointly https://www.astrobin.com/yrz3x8/ but the final 4 panel version is all 'in house.'

Olly

 

 

 

Edited December 9, 2023 by ollypenrice
Clarification

[AKB]

On 07/12/2023 at 08:58, ollypenrice said:

Possible now we have fibre internet...

Ooh, that must have changed your online life!  Certainly did mine.  Just the thing for big images (like this) and system updates.

Tony

[Allinthehead]

25 minutes ago, ollypenrice said:

Are you sure? It came from a very long time ago and originally looked like this:

 

 

 

Is it not the Squid data you sent me for the image below?

Edited December 9, 2023 by Allinthehead

[ollypenrice]

7 minutes ago, Allinthehead said:

Is it not the Squid data you sent me for the image below?

It must be because it's the only Squid data I've ever shot. I'd forgotten about that. You did a very good job with it!

16 minutes ago, AKB said:

Ooh, that must have changed your online life!  Certainly did mine.  Just the thing for big images (like this) and system updates.

Tony

Certainly did, and even more so for the six people who have robotic scopes based here. Upload speed went up by one thousand, three hundred times.  You know the village, of course, and there are probably six people living here who actually use the internet. Lord knows why we were blessed with a fibre connection a mountainous 8km from the main line - but we were.

Olly

[AKB]

1 hour ago, ollypenrice said:

Lord knows why we were blessed with a fibre connection a mountainous 8km from the main line

... "because you're worth it", to quote an advertising campaign.

[AstroGS]

16 hours ago, Elp said:

In terms of organisation it's why chatrooms and platforms like Discord are so popular, real time discussion and all. No reason why it couldn't work here for agreeing projects, could used a shared system similar to SharePoint or a shared GDrive or OneDrive account, it'd need decent data storage but I think if everyone just uploaded their linear calibrated stacks with details in the filename to an agreed standard, it'd be a fairly simple thing to do, someone on here might have their own server which would be easy to FTP (upload files) to.

I would be happy to collate the master files and create one integrated image and then to forward it to whoever does the processing - I wouldn't consider that this is my strongest point. 

[AstroGS]

So, in order to see how many SGL members would wish to participate, may I suggest that we start sharing ideas on:

 

1. Possible Targets:

• Cepheus
• Cassiopeia
• Perseus

2. Focal length: 300 - 500 (?)

3. Mono or OSC or combination?

 

 

[ONIKKINEN]

I briefly took part in a collaborative effort last spring and found out that its really not my thing even though my contribution was not much more than sharing data that i had already captured. Had no real motivation to process the shared stacks in the end because it really wasn't my own data, even if that stack is the best data i will most likely ever get to touch again. Maybe someone can relate, maybe not.

In any case, the process for that project was that everyone shared calibrated data (16-bit for RGB, 32-bit for Ha, for space saving reasons) in a shared google drive folder from which the person who was responsible for stacking got all the raw data. If i recall correctly the stacking and processing part took several days worth of work because there was just so much to go through (i think a few hundred GB, of which 70gb was mine because of short subs). Sharing data this way is not too difficult from a convenience point of view, since everyone gets 15gb of free google drive space to use and you can just delete the data from your drive once the stacker lets you know they grabbed the data.

Obviously this method requires that everyone participating has good upload speeds and an uncapped data plan, but it gets the best image out of the input data unlike just stacking masters which is less efficient than stacking each sub to one master.