Dithering preperation

[PhotoGav]

With a DSLR it's difficult to have darks of exactly the same temp as the lights anyway as you have no temp control over the chip. I think that so long as they are within even as much as 5 degrees  + or -, you'll get away with it. I seem to remember I even used the same fridge darks with summer lights and I never noticed any problems.

I currently use 49 darks as standard. Seems to work well. The same is true across the board, the more subs you have the better, though the law of diminishing returns does apply!

[ianpwilliams]

And you think that would include having Darks of different temperatures in the same stack? I'm sure if I really went for it I could try to take maybe 40 Darks in the shed in one go, although even then the temperature would change, so I would be limited to 20 unless I was prepared to mix and match (or take another batch of Darks on another night at the same temperature I suppose).

[ianpwilliams]

19 hours ago, wxsatuser said:

If you have more than a dozen or so subs use kappa sigma or median kappa, this will remove planes, sats and colour mottle.

You don't have to tell DSS about dithering as it will register the stars and then stack them on top of each other.

The only other thing I do is use Mosaic Mode under the result tab.

 

Would you use kappa sigma/median kappa for the Lights only, or for all frame types? The basic guide I have says Average for Darks, Median for Flats, and Median for Bias.

[PhotoGav]

2 hours ago, ianpwilliams said:

And you think that would include having Darks of different temperatures in the same stack? I'm sure if I really went for it I could try to take maybe 40 Darks in the shed in one go, although even then the temperature would change, so I would be limited to 20 unless I was prepared to mix and match (or take another batch of Darks on another night at the same temperature I suppose).

Unless the darks are wildly different in temperature I wouldn't worry at all. So, if you do the darks on any number of averagely cold nights, they will be fine for any lights shot on a different averagely cold night. I suppose ideally you would shoot darks on the same night as the lights, but that is just so inconvenient, especially given the precious nature of clear dark skies. I wouldn't worry too much about it though, much better to have nearly right darks than no darks!

[wxsatuser]

I don't use darks so cannot say.
Have tried darks and bias in the past and could'nt see a difference with or without.

Being quite lazy, my workflow is now the Hallas method and I don't use any calibration frames.
My setup dithers by 12pixels and this seems to produce good results if I gather at least 2hours of exposure.

 

[ianpwilliams]

On 2/11/2016 at 20:55, PhotoGav said:

Thankfully the coordinate system of the sky doesn't change as we move around, so once you have RA & Dec for your particular target it is fixed in stone (well relatively, a few thousand years and it will have moved a bit, I digress...) and will work all over the globe, though it might be below the horizon!

Blind Solve - make sure that you have downloaded Plate Solve and set it in the Point Craft bit of APT. Open an image in APT (using the IMG tab), then click 'Blind' in the Point Craft window. It should eventually come back with the RA & Dec values. A really nice feature in APT is the 'Custom' tab of the Objects window, where you can store all your own targets with coordinates and other bits of info. Makes going back to a target really easy.

The PHD2 TOT value will change from session to session, as it does within a session. However, it shouldn't change massively, so your settings should remain good. Yes, the PHD2 value always seems to be a bit more than the APT value, not by masses though.

Hope that all helps.

I'm trying to blind solve today to get my coordinates, but when I click on "Blind" it says "Unknown FOV size! Select the camera model and Focal Length". Do I have to connect my camera for "Blind" to work? Update: I've selected my camera in Tools which got me a step further, but it seems I need ASPS installed! Downloading that now...another update: I got a timeout when blind platesolving in APT (I now have ASPS installed). I tried running ASPS to change the settings as per this thread:

 

http://aptforum.com/phpbb/viewtopic.php?f=24&t=618

 

But ASPS won't actually open when I double-click the shortcut.

 

Assuming I manage to get the coordinates, and store them in my Custom folder, would it then be a case of selecting those coordinates tonight and clicking GoTo? I'm also thinking that surely the accuracy of reaching the coordinates will depend on my polar alignment, and so a plate solve (GoTo++) would also be required, is that right?

 

And how do the coordinates work? The object travels across the Sky every evening, so how would coordinates allow for that? Does APT take those coordinates and calculate where they should be based on that current time? Because if I do a session tonight it will definitely be at a different time than my previous session, and I assume that you wouldn't have to be out there at the same time for each session?

[PhotoGav]

Good morning!

First off, have a read of this: https://en.m.wikipedia.org/wiki/Equatorial_coordinate_system Essentially the system is similar to longitude and latitude on Earth, but projected out onto the celestial sphere as Right Ascension and Declination. The points are fixed in that celestial sphere, so one particular point on the sphere has fixed RA & Dec coordinates, eg. Betelgeuse is always found at RA = 05h 55m 10s, Dec = +07° 24' 25". So, the points are fixed and the Earth rotates within that sphere.

Blind solve - I'm not entirely sure why you are not managing to solve your image. What might be useful is for you to send me the image you are trying to solve and I will give it a go. You could just post the image in this thread. It should work though and I am sure that it is just a 'simple' set up issue, once sorted it will work perfectly for ever more!

Next session - yes, once you have your image coordinates, you will need to select those in the Point Craft window and do a GoTo++. It will slew to the coordinates and then plate solve to check where you really are, move the scope accordingly and plate solve again until you are accurately on target. It's very powerful stuff when it works!

All in all, this is a prime example of the persistence and patience required to get this Astrophotography mullarkey to actually work!!

[Pompey Monkey]

5 hours ago, ianpwilliams said:

I'm trying to blind solve today to get my coordinates, but when I click on "Blind" it says "Unknown FOV size! Select the camera model and Focal Length". Do I have to connect my camera for "Blind" to work? Update: I've selected my camera in Tools which got me a step further, but it seems I need ASPS installed! Downloading that now...another update: I got a timeout when blind platesolving in APT (I now have ASPS installed). I tried running ASPS to change the settings as per this thread:

 

http://aptforum.com/phpbb/viewtopic.php?f=24&t=618

 

But ASPS won't actually open when I double-click the shortcut.

 

Assuming I manage to get the coordinates, and store them in my Custom folder, would it then be a case of selecting those coordinates tonight and clicking GoTo? I'm also thinking that surely the accuracy of reaching the coordinates will depend on my polar alignment, and so a plate solve (GoTo++) would also be required, is that right?

 

And how do the coordinates work? The object travels across the Sky every evening, so how would coordinates allow for that? Does APT take those coordinates and calculate where they should be based on that current time? Because if I do a session tonight it will definitely be at a different time than my previous session, and I assume that you wouldn't have to be out there at the same time for each session?

APT just sends the stored coordinates to EQMOD/ASCOM, which should slew the scope to what it "thinks" that position is, based on its current sky model. The next platesolve of the Goto++ then syncs EQMOD/ASCOM with the solved coordinates and resends the desired coordinates to EQMOD/ASCOM. That way it iterates towards pointing the telescope to the correct position. This is why you have the option of changing the maximum number of iterations in Goto++ in APT.

[ianpwilliams]

3 hours ago, PhotoGav said:

Good morning!

First off, have a read of this: https://en.m.wikipedia.org/wiki/Equatorial_coordinate_system Essentially the system is similar to longitude and latitude on Earth, but projected out onto the celestial sphere as Right Ascension and Declination. The points are fixed in that celestial sphere, so one particular point on the sphere has fixed RA & Dec coordinates, eg. Betelgeuse is always found at RA = 05h 55m 10s, Dec = +07° 24' 25". So, the points are fixed and the Earth rotates within that sphere.

Blind solve - I'm not entirely sure why you are not managing to solve your image. What might be useful is for you to send me the image you are trying to solve and I will give it a go. You could just post the image in this thread. It should work though and I am sure that it is just a 'simple' set up issue, once sorted it will work perfectly for ever more!

Next session - yes, once you have your image coordinates, you will need to select those in the Point Craft window and do a GoTo++. It will slew to the coordinates and then plate solve to check where you really are, move the scope accordingly and plate solve again until you are accurately on target. It's very powerful stuff when it works!

All in all, this is a prime example of the persistence and patience required to get this Astrophotography mullarkey to actually work!!

Thanks, here is one of my Lights if you can have a look.

IMG_2177.CR2

[ianpwilliams]

Yes astrophotography really does test you to the limits! But I enjoy that to be honest. The challenge of it is what makes the reward even more special when it works. Well, that and the fact that you've just managed to take an actual photo of a galaxy, nebula, star cluster etc that is millions of light years away!

What I would really like to do is establish exactly what does what when it comes to GoTo etc in APT if I can.

For the GoTo it seems pretty straightforward. You click Objects, select your object, click Ok, then click GoTo, correct?

 

 

But then there's PointCraft which I'm a bit lost with (despite the explanations that appear when you hover the mouse over buttons):

 

 

So you've got the "Current Image" section, the Plate-solving Results section, and the "Center FOV at position" section.

As far as I can tell, the "Center FOV at position" section is used to get the scope pointing at the target, and having the target centred. You would do the same as with GoTo above (objects, choose target), and then hit GoTo++, and instead of going roughly where the object is it would go exactly where it is by taking photos and solving them, correct? In which case, why would you ever use the GoTo above?

 

Then there's the Current Image section which I don't really understand. The fact that it says "Current Image" suggests that you are working with an existing image, presumably from a previous session. So my guess is you click Objects, choose whatever the object is that is in the image, and then maybe click Auto? Hovering over Auto says it'll do shoot>get scope pos>solve>sync. So that sounds like it's trying to get the scope pointed exactly where it was when that image was taken? And if that's the case, when I'm outside next could I not just select the image attached to my last post and click Auto, and then I should end up pointing at the same place again? Or is it better to use blind plate solving indoors in advance? I guess it might be quicker if I already have coordinates stored (although it would surely have to sync, solve etc)? Is the Current Image section basically only for working indoors with existing images (for whatever reason), and the Centre FOV section is for when working outdoors?

 

Then there's the Plate-Solving results section. "Sync" says that it synchronizes the telescope position with the coordinates of the image centre. But I'm not sure what that means. Shouldn't the scope already be synced if it's been shooting, solving and syncing? And Store seems to create a custom set of coordinates for future use. So if I were able to Blind solve then I could click Store and I would have my custom coordinates, so then next session I would open PointCraft, click Objects, click Custom tab, select my coordinates, then click GoTo++?

 

I'm also not entirely clear on the jargon to be honest.

 

Solve - what exactly is this? "GoTo++" is GoTo>Image>Solve, so given that GoTo is to go to a set of coordinates, and image is to take an image, to Solve must be to compare the image just taken with the data stored in the program, to work out how much the scope is out by and by how much, and then to send a command back to GoTo to start again. Is that right? In which case, solving an image from a previous session would be the same process, but having an image right from the start?

 

Sync - "synchronizes the telescope position with the centre of the image". Again I'm not sure on this one. GoTo++ doesn't mention Sync, but the GoTo++ process seems to do what Sync is described as. "Auto" seems to include Sync (Shoot>Get Scope Pos>Solve>Sync), so that makes me wonder what the difference is between Auto and GoTo++. Maybe Auto relies on an existing image and GoTo++ doesn't?

 

I think it's fair to say that I find PointCraft very confusing! And it also seems to me that in any given session I could probably actually just click PointCraft, click Objects in Centre FOV, choose my object, click Ok, then click GoTo++, and that would be it, done, object found and centred. Except that in tonight's session if I go for a second Andromeda session (weather permitting) I will need these precise coordinates stored in advance, and then my Object will be those coordinates. Although even if it is just a case of doing what I just said, I would still like to know how to use the other buttons too because they may be needed in the future.

 

Incidentally I looked at the user guide on the APT website but it seems to just skim over Point Craft. And I couldn't find any YouTube videos which go into detail either.

 

 

[PhotoGav]

1 hour ago, ianpwilliams said:

Thanks, here is one of my Lights if you can have a look.

IMG_2177.CR2

Well, you've got me scratching my head now... I tried to blind solve that image and it timed out. So, I tried a different method and input the field of view as 117 x 78 arcminutes via the 'Shift-Click'  the Solve+ button. I then entered the coordinates of the Andromeda Galaxy taken from Wikipedia into the Current Image approx RA & Dec fields. It solved the image in 13s and gave this result: RA = 00:42:50, Dec = 41:15:13. That's good, we have an answer, but why didn't the blind solve work? Mmm...

As regards your other post, I will have a look at that and try and answer some of the questions a little later on.

[ianpwilliams]

Thanks, at least I can get those coordinates into my APT.

 

I had a quick go at processing Andromeda, and it doesn't look good to be honest, but then I've always had trouble with it for some reason. It was a bit of a rush job, but still:

The centre is very saturated (probably because of the 5-min subs), and there's very little of the outer bits. In fact it's not really any better than my previous unguided efforts:

 

 

So although I can see the benefit of doing the same target on multiple sessions, I wonder if this latest session of mine might not be worth adding to? Or on the other hand adding another session to it might make it usable? I'm not sure which it is. What did you think of the sub? Did it look ok to you? The histogram peak was 60% which I know is a little high, but quite a few people have said that they will often push it to 60%. But then that's probably what gave me the saturated centre.

 

[PhotoGav]

OK, lots of questions about Point Craft. I will try and explain as best I can... The good thing is that you have understood quite a bit of it already!

Firstly, the basic principle of plate solving ('solve' from your post above) is exactly as you understand it - scope is pointing at coordinates X, take a picture, work out what the real coordinates of that image are by comparing with the downloaded catalog of stars and then establish how far out the scope is from where you want it to be. The scope is then moved to the calculated correct place and the process starts again until you are within your accepted error margin of the target coordinates.

'Sync' - this is when you have solved an image and you have the actual coordinates of where the scope is pointing. You sync the system so that it now knows that what it thought was coordinates X is really coordinates Y. This way, when you do your next GoTo it will point more accurately as it has a better map of the sky in the system. The more times you Sync, the better your GoTos become. To be honest, for imaging it is fairly irrelevant as, just as you say, all you ever do really is start everything up, open Point Craft, select your target from your your list of custom targets, click GoTo++ and the system whirs away until you are pointing within a few pixels of your target again. Hit go on the plan and you are hoovering photons!

Current Image - this is where you solve the image that is currently open in the image viewer of APT, so that might be the image you have just taken or it might be an image you have loaded into the viewer from your computer. Hit 'solve' and it will try and solve the image using the rough coordinates entered in that area, which might well be the current scope position if you are working with an image you have just taken. Hit 'blind' and it will solve the image with no coordinate info - allegedly... Not sure why it hasn't worked for either of us with your Andromeda Galaxy image. I have a feeling that it has something to do with an incorrect field of view calculation when Canon 1100D is selected and the focal length is 650mm.

Auto - this is just a quick method to take a picture of where the scope is currently pointing, solve it and sync it, rather than having to do each step individually.

I find that Current Image and solving the image of where the scope is currently pointing is most useful when framing a new target. You fiddle about getting the framing just right, then when happy, take a picture, solve it and store those coordinates in the custom target entry. You are now ready to go straight back to the target anytime.

Hopefully that demistifies Point Craft and APT a bit. If not, just ask again!

As for your image of M31, yes the core is blown out! The best way to deal with that is to shoot a set of subs with a much shorter exposure time and blend the two images together in Photoshop. The sub did look very saturated, but that should process out. Do you use a light pollution filter? Where are you imaging from and how dark are your skies? How many subs are in that stack? The more subs you can manage the easier it will be to process. I'm sure it will be worth adding more subs and seeing what you get.

Good luck!

[ianpwilliams]

Thanks for that, I've definitely got a better understanding of Point Craft now!

 

As for my stack, I use an IDAS D1 LP filter, which makes a huge difference when I see the orange glow of my Orion subs by comparison. I'm about a 10-min drive outside Edinburgh so obviously there is some light pollution. I'm not sure how dark my skies are. I took 14 5-min subs, 20 flats, 40 darks (20 each from two Darks sessions, both the same temperature), and 200 bias (from my library!).

 

I'm just not sure what it is when it comes to Andromeda, I'm sure it doesn't like me! I did take some 10-min subs at 100ISO not so long ago, maybe I could combine that stack. Or if I'm going to do a second Andromeda session, maybe I should do maybe 3-min subs, and lots of them? Although I guess that would just be to have a better middle bit, so I'd still have to do more 5-min sub sessions to try to bring out the rest of the galaxy more, because it's pretty poor at the moment strangely. Also I'm not sure where I would begin when it comes to layering processed stacks.

[PhotoGav]

What ISO are you using at the moment? I remember I used to favour 800 ISO with my DSLR.

Andromeda is a complete pig of a target, spectacular, but a pig! It took me four attempts before I got a version that I was happy with. I would suggest that you keep going with your 5min subs and make a big stack to tease out the outer areas. Dedicate an hour or so to some 1 min or even 30 sec subs specifically for the core. As for layering the two - do you use Photoshop?

[Marky1973]

Ian, my "best" Andromeda so far included over 3 hours of subs....and it is still very far from anything that I would be happy to compare to some of the "pros" on this site! ;-)

[ianpwilliams]

10 hours ago, PhotoGav said:

What ISO are you using at the moment? I remember I used to favour 800 ISO with my DSLR.

Andromeda is a complete pig of a target, spectacular, but a pig! It took me four attempts before I got a version that I was happy with. I would suggest that you keep going with your 5min subs and make a big stack to tease out the outer areas. Dedicate an hour or so to some 1 min or even 30 sec subs specifically for the core. As for layering the two - do you use Photoshop?

Yes I'm sticking with 800ISO from now on (which my latest session was).

 

I don't have Photoshop, but I do have PixInsight, and GIMP, which I know does layering, so hopefully that should do the trick?

 

If I send you a link to my unprocessed stack could you possibly have a quick look at it? I'd like to get confirmation that this one is on the right lines, and that it's going to be worth adding more sessions to it.

 

If I do carry on with it, then maybe I'll dedicate the next session to 1-min subs (hopefully an hour or more), and then the couple of sessions after that to 5-min subs again. If I did end up changing sub length mid-session, I take it I wouldn't need to change any settings would I? The dithering would just take care of itself?

 

I just hope that it'll all be worth the effort, because this is going to take a while considering how rare clear nights are right now!

 

[ianpwilliams]

13 hours ago, Marky1973 said:

Ian, my "best" Andromeda so far included over 3 hours of subs....and it is still very far from anything that I would be happy to compare to some of the "pros" on this site! ;-)

Crazy isn't it! There was me thinking that upgrading to a guided setup, getting a light pollution filter, and learning how to dither might be enough to get some great results, which undoubtedly it will be with other targets, but Andromeda seems to be very elusive. Having said that, the challenge of astrophotography is part of the attraction for me, so I'm certainly up for trying multiple sessions to see what I can manage to get.

 

[ianpwilliams]

When it comes to the multiple sessions and whether I would need to take flats each session or not - I did remove the camera after what I will now call session #1, in order to take some Darks in the shed (I took 20 on two nights, both same temperature as each other and the session itself, so I have 40). So I'm not planning on taking any more Darks, so in theory I shouldn't need to remove the camera at all. But because I removed the camera after session #1, I guess I will have to take flats for the next session.

 

But then I thought I also read that if the focusser on the scope is adjusted then flats would also have to be taken again. Given that I may well need to readjust the focusser for the remaining sessions (I do lock it when it's focussed but it rarely seems to keep focus session to session), I guess I would need to take flats each session anyway just to be safe?

 

I also have "session #0.5", where I gradually kept on increasing the sub length, but because it looked very saturated I wrongly kept reducing the ISO to compensate. So I ended up with:

 

7 x 100 ISO Lights at 10-minutes each (I got carried away!)

4 x Darks

8 x Flats

50 x Bias

 

I've attached a jpeg of the TIF after applying the histrogram (nothing else) in Pix. Even in this one you can see that the centre is saturated. I did wonder if, should I try processing this stack, whether I might be able to combine it with session #1 and any other sessions that I might do. But then I guess that would depend on whether the coordinates were the same. So would it be a case of doing a blind solve on one of the subs from session #0.5? Or is it extremely unlikely that the coordinates will be the same, and therefore maybe I should disregard it and put it down to (lack of) experience?

 

[PhotoGav]

I think that you are definitely heading in the right direction with all of this. As for the subs of the different sessions, they don't appear to be massively different, so definitely worth stacking together. The worst that can happen is that you lose a bit from around the edges where the subs are significantly different. If it's a huge chunk, check through the subs and discard the ones that are too far off the majority.

With regards to Flats, you only need to shoot new ones if the focus is hugely different and you haven't removed the camera between sessions. If it is just a focus tweak, then I'm sure it won't make a huge difference. If you are in doubt, then just shoot a new set, it can never hurt!

Keep at it.

[ianpwilliams]

Good to know about the Flats. They are a pain to take after all. I do them by holding my laptop to my telescope as a light source, with a blank maximised notepad document open and a piece of A4 paper. Seems to do what's required, but a small laptop can quickly become very heavy after a short time!

 

So I suppose the next stage will be to apply the flats, bias and darks from the 10-min-100ISO-sub session to all of my subs in the 10-min-100ISO-sub session, and then do the same again for my 5-min-800ISO-sub session, and then stack all subs from both sessions together in one stack for processing? And if I did another couple of 5-min-800ISO-sub sessions I would do the same thing, apart from the 30-sec-800ISO-sub session which I would process separately and layer the processed stack instead?

 

Looks like I'll need to learn a bit more about DSS so I can apply calibration frames to subs from each session without stacking, and then stacking all calibrated subs together. Hopefully it won't be too hard to learn.

 

And would you suggest 30 or 60 seconds for my shorter-sub session? I'm actually considering 30 seconds because I could get a hell of a lot of those over the course of an hour or so.

[PhotoGav]

Yes, that's right with regards to the processing / stacking and keeping the short subs separate to all the long subs.

As for 30 or 60 second subs - give it a go and see what you get, but my guess is that you will plump for 30 second subs.

[ianpwilliams]

OK I'll see what I can work out in DSS. 

And if the forecast stays clear for Thur and Fri night then maybe they can be the short and long sub sessions. Fingers crossed!

[ianpwilliams]

I never know when to end these threads, because they inevitably lead onto other subjects because of the nature of astrophotography! It seems only right to post what will hopefully eventually be a decent Andromeda image, and I will no doubt need help stacking separately and combining stacks etc, so I think it's best if I see how my next session goes and then come back to it.

[PhotoGav]

Indeed, it is a never ending story. Good luck and I look forward to seeing the image.