Direct focus plate solve

[keving]

I have been successfully using the 'regular' autofocus and, quite independently, the plate solving. Last night I tried the 'Direct' autofocus but received an error where plate solving failed. This was odd as I had 'manually' solved the same view (under the analysis menu). This happened after a star had been found and slew made. In fact, the error was received very quickly after the first image was captured making me think the plate solving hadn't even been attempted?

Any ideas?
Thanks

[Hamza Touhami]

Hi Kevin

Typically, a plate solve ( ctrl+u) will fail if the stars are bigger than 5 pixel and/or the focal length is incorrect. try taking an image at binning 2x2 and do a direct focus, does it work better? binning effectively reduces the size of the stars in your image and this is what I advise doing in the automatic observation.

[keving]

This is a plate solve initiated automatically by the Direct autofocus (Incidentally, not sure why it does this?) so the whole direct autofocus fails. Quite independently, I have solved successfully using the option under the analysis menu (single image solve) so I have focal length and everything else OK there. Is there another place to enter parameters especially for the Direct focus?

Thanks

[keving]

Seen this again recently (so doesnt appear to be a 'glitch'). In Direct Autofocus, there is a slew to the star, an image is acquired, and then an attempted plate solve - which fails. I can, however, solve the acquired image manually using Analysis/Plate-solving/single-image...
I am using 10.3.42.385

????

[Hamza Touhami]

Hello Kevin

When try the direct focus and it fails, can you please save this same image and send it to me for inspection?

Thanks

[keving]

Hi Hamza

Yes will do...

But another clue here is that the failure message occurs quite quickly... When I run into problems with manual plate solving (poor initial position for example) the computation is happy to sit there for minutes and work in ever increasing circles around the notional target (On the few times this has happend I eventually pressed cancel so I could give it better starting point). This leads me to think that the 'failure to plate solve' is not really that as such, but rather, a failure to *connect* with the plate solver in the first place. I installed the UCAC4 catalouge (I thought it mught have been reliant on this, as I had previoulsy used the GSC) but its makes no difference.

Anyway - will send an image on next time I am out.

cheers

Kevin

[Hamza Touhami]

I would really like to get to the bottom of this, awaiting your images.

[keving]

HI Hamza

I have been trying out my spectroscopy rig under Prism, and the Direct focus is not really applicable there... But will be back imaging as soon weather permits...so images to follow soon I hope!

Thanks

[Hamza Touhami]

I have built an entire section of this forum to show off your images so please do

[stellar.skies]

Hamza,

I am having this issue as well. Can plate solve perfectly fine independent of this process. Here are a few screenshots.

Showing that I can plate solve (UCAC4 catalog):

My direct focus settings (can auto focus just fine by the way):

Get this message right after it captures the initial image for the Direct Focus process:

Not sure why it is failing?!?

Best Always,

Daniel

[Hamza Touhami]

Good evening Daniel,

Thanks for the screenshot, I have warned Cyril about this issue and we will look into it further, stay tuned.

[Cyril Cavadore]

My advice ->
There is a lot of noise in this image .. increase exposure time or bin 2x2

[stellar.skies]

Cyril,

It was set for a 6-second exposure with 2x2 binning, same as the photo that was plate solved fine in the first screenshot. This is a QHY163M camera which at the time was set to -15C so there is little noise. I will say it sure shows an image like there is noise but if I take the same shot in another software package I don't see that noise??

The moon was up during this testing, would that have a impact?

Best Always,

Daniel

[keving]

So I now have some evidence of my experience with this...
Here is the snapshot taken by Direct focus after slewing to its auto-detected star...

[I have stretched the fits file and saved as jpg as the former is over 20MB...]

I get the message about failure to solve almost immediatley afterwards.

Here is the solved version of that image done using Analysis tools (or right context click? )

Anyway, the point is that I think this has nothing to do with platesolving per se, unless the Direct focus uses a comeletly different routine than those available via the menus..?

Kevin

[stellar.skies]

Kevin,

It does not work on any of my setups. Maybe it is a temporary bug. I was going to get a log but we have clouds tonight. Tomorrow I will try to get a log.

Best Always,

Daniel

[Hamza Touhami]

Thanks gentlemen, I appreciate the inrush of information. In fact, the more logs you get us the better we can understand and help.

[Cyril Cavadore]

If you get with the same exposure time an image, and just plate solve afterwards, does it work ?

[stellar.skies]

Cyril,

Yes it sure does. That is what baffles us as to why when plate solving in direct focus it fails. Same exposure and binning used for both.

Best Always,

Daniel

[keving]

The image in my screenshot was the very same one taken by the direct focus. As you can see, it’s been solved using the prism analysis tools. Further the solve was very easy ( almost immediate) but the focus routine gave up immediately. My guess is that the focus routine is not being passed the image at all...(some kind of lock in place?)

Kevin

[Cyril Cavadore]

And if you try to bin 2x2 ?
Also may be pointing accuracy is not very good also ... and center coordinates are too far way from expected position.

[keving]

"And if you try to bin 2x2 ?
Also may be pointing accuracy is not very good also ... and center coordinates are too far way from expected position."

This would only make a difference (surely?) if we *couldnt* plate solve the *exact same* image using the menu options in Prism (see screenshots supplied earlier which shows this indeed possible - and easy).
The only conclusions I can come to are:
(i) The solving routines in Direct focus are not the same as those available directly from the menus (which work very well with the very images taken by Direct focus). They may have been optimised for a 'quick and dirty' approach in focussing - which I could understand.
(ii) There is some problem in linking the solve routine to the image in Direct focus.

I am inclined to believe (ii) because the message comes up almost immediatly - there is no real attempt to solve.

Please do correct me if there is something wrong with the argument above - but I can't see how it can be otherwise at this stage.

Given the multitude of possible Windows variants there are, with all kinds of sharing, security options etc, I could see how there may be some 'lock' on a file if it doesnt comply with some draconian Windows imperative. (This might also explain inability to focus while sequence window is up?)

many thanks

Kevin

[stellar.skies]

Cyril,

I will also say that my captures made in PRISM look much worse than they do in other programs. I can't understand why my ZWO ASI1600MM-C and QHY163M images look so noisy and grainy in Prism vs. Sequence Generator Pro for example. See the attached screenshot comparing a 9 sec flat dark and how it is displayed in Prism vs. Sequence Generator Pro. Is there something wrong with how Prism interprets data from these cameras?

Maybe this is why direct focus fails for me?

[keving]

I have noticed this too... Maybe subject for another thread?

But while I'm here...I see Prism renders in only 8-bit resolution. This is not the resolution of the underlying image, but I can't see why you can't opt to render at the native scale? Maybe you can and there is a setting somewhere?

Back to focus thread... I had a thought that maybe things only work if you run the programe as adminsitartor (comes under possibility (ii) of my last post).

cheers

[Hamza Touhami]

Good morning gents and I appreciate the inrush of information. Indeed the discussion about the image display is better moved to its own thread to remain focused. Cyril is aware and working on this issue.

[keving]

So, summarising a more detailed post from elsewhere...
The direct focus only works when used from the 'Automated observation' section - it does not work when used from the Obervatory/focus menu..

[Hamza Touhami]

Thanks Kevin, we are testing this as I have a new operational observatory today. I will keep everyone posted.

[Cyril Cavadore]

Hello,
Log file can help to tackle the issue.
Also this direct focus has two possible catalog : GSC and UCAC4, have you tried both ?

Yes plate solve has been optimized in this case ...

[Cyril Cavadore]

Hello,

8 bits resolution is the screen resolution, that is mandatory and unavoidable
The image data behind is either 16 bits or 32 bits float.

Log file can help to tackle the issue.

Also this direct focus has two possible catalog : GSC and UCAC4, have you tried both ?

Yes plate solve has been optimized in this case ...

[Cyril Cavadore]

After deep investigation it turned out the plate solving can fail with very sharp or undersampled images. A fix has been written, version 10.3.47.