Help migrating from Astrotortilla to plate solving with SGP

Hi, I’m new to SGP have downloaded the trial today. I’m trying to understand how to configure plate solving. I have used AstroTortilla for a few years and find that pretty straight forward, so I thought using ansvr might be the simpliest way to go, but the help section on SGP seems to want me to use either Pinpoint, Elbrus or PlateSolve 2 with Astrometry.Net being configured as a blind solver as back up when one of those fails. Can I use Astrometry.Net local solver as my primary plate solver and if yes, can I use the already installed indexes used by Astrotortilla? I can’t find any simple set up steps in the SGP help files and when I click on the settings button on the Plate Solve Options screen it takes me to a page http://127.0.0.1:8080/api/config which my browser reports it cannot connect. What am I doing wrong please and where do I need to look to work this out? Regards, Geof

1 Like

best way is to set up PS2 as the primary and astrometry.net or ansvr as the backup. i think you can use ansvr as the primary solver but it’s been so long now since i switched to PS2/ansvr i can’t remember if ansvr can be configured as the primary.

ansvr is just another wrapper for astrometry.net’s solver programs (solve-field / backend) and so the indices that you already have will work fine. heck for a while i had the indicies installed on my mac, then shared the folder with my windows 7 VMWare VM and pointed both AT and ansvr to those indicies. there’s going to be a “backend.cfg” file that’s part of ansvr, and in that file you can set up the location of the index files. my observatory computer is shut down so i can’t look at the configuration this second or else i’d give more details.

is ansvr installed? http://localhost:8080/api/config is where the configuration page of ansvr will live when the ansvr web service is running.

rob

Thanks Rob,
A friend here told me to use PS2 so I downloaded that yesterday evening and think that I got it working - at least it will solve images previously taken with my QSI+C14 combo, but I haven’t hooked up to the gear in the obs yet to try it ‘live’. No I haven’t installed ansvr, so do I still need that as blind solver for back up? I don’t really understand why I need a blind solver (not that I understand what a blind solver is), as when using AstroTortilla the past 3-4 years it solved every time…
Lots still to work through to get anywhere close to comfortable with SGP… :expressionless:
Regards, Geof

So I’ve installed ansvr with a few indexes (duplicates to what I had for AT) but in the correct folder and I can blind solve images with it ok.
I’m now having problems getting PS2 to solve some of the images taken with my QSI583+TSAPO100Q. ANSVR returns the correct solve in about 10 seconds telling me the image scale is 1.91 (which pretty much corresponds with the 1.92 that I get from Ron Wodaski’s CCD calculator), but when I try to solve with PS2 using 1.91 it times out with the error message…
Failed to Plate Solve Image (PlateSolve: APM file found, but could not be opened.)!
What does this mean please and how do I fix it?
Regards,

Geof

hm, i am not sure what that error means beyond what it says. almost seems like the permissions on the file are wrong. maybe Jared or Ken can comment. not even sure what the APM file is - it might be an intermediate file that PS2 uses?

a blind solver is a solver that does not need any pointing or scale hints to solve an image. Astrotortilla, being just a front end for astrometry’s software, is a blind solver but of course giving a scale hint dramatically speeds it up. IIRC the scale hint is the most inportant hint for astrometry; the pointing hint is not as useful just given the way it works.

PS2 (and elbrus, and pinpoint) are not blind solvers. they require a pointing hint (and a plate scale) that’s reasonably accurate in order to solve an image. the trade off is that they are usually faster - if ansvr is taking 10 seconds then PS2 could probably solve the same image in 1 or 2 seconds.

rob

I have found PS2 is very picky and can time out. Especially on my 80mm. On my 8in RC PS2 works just fine MOST of the time. If it fails I switch to Astrometry. As jared will tell you , neither works all the time in all parts of the sky. Hence, the back up.

Hope this helps.

Dennis

Thanks Rob and Dennis,
I’ve not yet tried solving in a live situation, i.e. hooked up to the camera and scope in my observatory at night. I’ve just been trying to solve images that I’ve taken over the past year with two different scopes, to confirm that I’ve got the set up right (or at least close). One configuration is my C14+Optec (F7) focal reducer/field flattener which with the QSI583 gives me an image scale of 0.43. The other is my 4" TSAPO100Q also paired with the QSI583, which gives an image scale of 1.91. What I find strange is that PS2 will solve several images taken with either configuration, but then fail on some others. I even worked through a set of 600sec exposures for the same target and it would solve some of them, then time out with that error message on other subs from the same session. I could not see any significant differences between the images. AstroTortilla and ANSVR solve everytime (well almost anyway).
Probably it’s not a huge deal if I can configre PS2 with ANSVR as back up (when I’ve worked out how to do that), but it is weird and somewhat disconcerting that PS2 will sometimes fail to solve almost identical images to ones that it does solve - that makes absolutely no sense to me…!!
As for speed, yes, PS2 solves in a couple of seconds where AT and ANSVR maybe takes 10-20 seconds, however, I’d rather have a successful solve every time in 20 seconds, than save 10+ seconds when it works, but for it to completely fail other times… It’s not going to be a good start to the day if I wake up and find that a session has paused at 1am, becuase PS2 failed to solve an image after slewing to it or after a meridian flip, etc…
Thanks for the help, I’m confident that I can work through this, plus testing in a live imaging session has to be what really counts :slight_smile:
Cheers, Geof

well i do agree, i have seen PS2 fail somewhat “randomly”, usually on busy starfields in cygnus for instance. randomly in the sense that one particular pointing will not solve but another, almost the same pointing will solve successfully. if i had to guess i’d say that PS2 has a hard time when presented with too many stars. it may have a config setting to limit the # of stars it considers, but i’ve never looked into that (it seems that SGP is managing the configuration of PS2 and so i’m not sure if it has that knob, and if it does, how to change it such that it sticks…)

i don’t think i’ve ever hit a case where both PS2 and ansrvr both failed, so you’ll likely be good with PS2 as primary and ansvr as backup.

rob

1 Like

Thanks again Rob,
Yep, I’m sure that youre right about double teaming PS2 with ansvr, but I haven’t yet found a way to try that out without going ‘live’ in the obs. So far the only way I can solve images is to open them in SGP then right click and either solve with PS2, or blind solve with ansvr. There doesn’t seem to be a way to configure the PS2 with ansvr back up from the right click option, so is there a way to test the ansvr back up to PS2 from the comfort of my armchair without hooking up in the obs? I have images that PS2 wont solve, so I’d love to see the fail over to ansvr working with them.
Regards, Geof

Geof,

I am new at this too, relatively. People seem to swear by PS2 yet I fail to see its usefulness since you have to pretty much tell it where it is looking at so it can platesolve. And I tend to think platesolving IS a tool to help you find out where the mount is pointing. Alas, I am still trying to get onboard with it.

Ansvar has always worked for me. I slew to a destination in the sky and have it blind solve and sync. And it always does, and then I slew to an actual target.

Welcome to SGP - among the best programs I have ever used. And I strongly recommend you also consider the Frame and Mosaic module.

Farzad

the reason it makes sense is that once you’ve synced your mount to the sky once, then the mount’s reported position is going to be pretty close to it’s true position. PS2 and other non-blind solvers need a RA/DEC hint that’s close to reality, but not so close such that there’s no point in even running PS2. PS2 does a spiral search outward from the given mount coordinates so even if the hint is wrong if it’s close the image will solve.

so depending on how far off your mount is from the sky the first solve with PS2 may fail and then ansvr would step in. but after that PS2 will probably get the job done.

It is mostly about speed vs convenience. It should normally be OK if the desired center is within the FOV.

PS2 is free and it needs a helping hand and some tweaking. Same is true of PinPoint - it needs a strong hint to be effective (scale and center) and can fail too in dense starfields if the settings are not right. Its also true of doing solves in PixInsight. All in all, plate solving is remarkable and makes pointing a dream, like say, after an automatic meridian flip. Just need to have realistic expectations and use it as a fine tool, rather than as a hammer.
Over the years I have realized a philosophy that you cannot use technology to overcome poor technique. For instance, I found working on tracking using an 80/20 rule, where 80 is mount alignment and 20 is autoguiding is more consistently reliable than relying solely on one to achieve good tracking.

1 Like

Hi Farzad,
I agree that ‘blind’ plate solving is a tool to help establish where the mount is actually pointing, which is how I used to use Astrotortilla (AT) when I set up / broke down my mount every session. Now that its permanently mounted in an observatory, I’m much closer to target from the get go, but have continued to use AT, albeit with with me at the helm, not automated. Ansvr is essentially identical to AT in that it uses the locally stored Atrometry.net solve engine, but I do see that PS2 is much faster when it solves, so as Rob says starting with PS2 with ansvr as back up seems to offer the best of both. I have read up a bit on Frame and Mosaic and can certainly see me using that once I get everything else working, so thanks for the suggestion.

Thanks Rob,
I think that I’ve got the process right in my head, now I just need to get the set up right andtry it out for real.
Cheers,

Geof

Thanks Buzz,
So true, its always worth spending time getting things close to perfect. Perfection is not required, but the payback of a good set up is worth some effort, so the good old 80/20 rule is definitely worth applying in astrophotography…!!
Cheers, Geof

Hello, Geof.

Sorry to have mistaken you with a beginner. I guess what threw me off was
your statement about being new to SGP; and you have an observatory already
so you are lightyears ahead of me.

I am not going to argue “faster” because speed maybe more important in some
cases than in others. I am nearly 9 months into the “how” stage of my
journey and I don’t have fancy targets that might be restricted by time
enough to necessitate a faster plate solving. I can see that as being more
important with fields of view narrower than 26 arc-minutes which is what I
am setup for. Additionally, being established on a pier and a good home
position for the mount it is definitely a lot easier to be directing the
mount to go very near where you want to be observing or imaging.

I know that PS2 is more effective, and one of these days I also will be
using that.

I haven’t actually used the mosaic module to do mosaicing yet because as I
said, my targets are me at this point in time, but I do play around with it
occasionally and I have learned that even with single frames, it is a great
tool to help with composition and generation of a sequence.

Farzad

Hi Farzad,
No worries and trust me I still feel like a beginner…:confused:. Yes, I’m totally new to SGP, I just logged in and it tells me that I still have 42 days of the free trial period left :slight_smile:.
I agree that faster is not always better, but if I can use PS2 with ansvr as back as Rob suggests then that will work for me - just wish I could test that without having to hook up live in my observatory. What I’ve been trying to do with SGP (and is my usual modus operandi for any software) is to test it off line like in a ‘sand box’ implementation, but I’m not sure that is possible with all of SGPs features. Still I have another 42 days to continue testing… :sweat_smile:.
Regards, Geof

…SGP’s abilities with indoors plate solving experimentation is
unfortunately limited, especially the case with me since I use a DSLR and
they don’t have a simulator for that.

I don’t know if there is any other software that can do all that SGP can
do, and I am just scratching the tip of the iceberg on this “hobby”. I am
looking forward to expanding inside SGP as I get more gear included such as
focusers, rotators, filter wheels, and a dome.

Best of luck in your evaluation.

Farzad

Thanks Farzad,
Yep, I sometimes think that this hobby is more about problem solving than astronomy, but the rewards are great on those occaisions when it all comes together :smiley:. The more we practice the better we get - in theory anyway :wink:.
Good luck with your own astro journey, I too used a couple of DSLR cameras for the first 4-5 years that I’ve been taking astro photos and only got my first mono CCD camera (2nd hand) about 8 months ago so I’m still working out how best to use that.
Regards, Geof

1 Like