2 problems with V4.1.13789

[MikeB838]

Hi Robin

I found this 2 problems in the new version (V4.1.13789 64bit)

1: Tools: "Plate Solve (solve only)" => "Deep Sky Image Annotation"
the position of the galaxies are correct but Asteroids are shown far off their real position

2: "Auto Colour Balance" in live stacking does not work as before (the image does not become neutral, but is bluer than neutral)
and after "Colour Calibration" it seems to be more blue

I noticed peaks in the histgram at the left side in Live stacking after a clear before a new frame arrives.

Histogram live stacking after clear

live_stacking_histogram.jpg (90.07 KiB) Viewed 934 times

Hopefully easy to fix - or to explain

Thanks for your work and cheers,
Michael

[admin]

Hi,

I can't test the plate solve function properly for asteroids until I get some clear sky, but I can (and have) tested the calculation of the co-ordinates in the latest version, and that seems to be fine - the way to do this is to bring up the full 'Select a Target' window to choose an asteroid target - in that you can filter down to just showing asteroids and then see the calculated co-ordinates in the right hand column for each asteroid. It's then easy to compare those to the calculations made independently (for example by heavens-above.com - ie https://heavens-above.com/MinorPlanet.aspx?desig=16). Those co-ordinates all come out the same. The only thing that could throw this off (I think) is if the time of your computer is wrong.

So, the thing to check is to see if everything else in the annotation view is also off - if everything is out then that would imply that the plate solve information is out of date or wrong. At that point it is worth trying another plate solve to see if it fixes the issue.

On item 2, nothing relevant has really changed recently in the code relating to that, so it is most likely some change in the situation/setup/other settings on your system that has caused the change. I notice two things of interest in your screenshot...

1) The saturation adjustment is turned up quite high - if this is set higher than usual then any colour to the background could be much more noticeable than usual

2) The blue peak in the histogram is a long way left - almost at the zero point. This could be caused by a mismatched dark frame or by the blue slider in the live stacking 'background' tab being too far to the left, subtracting too much from the blue pixel values. This is perhaps where I would start looking, as if there is any significant offset between the background brightnesses of the 3 channels, it is *very* hard to get an adequate colour balance using just the sliders to the right of the histogram. The background subtraction tab was added to help allow any such offset between the channels to be adjusted out to make colour balance easier.

cheers,

Robin

[MikeB838]

Hi Robin,

Item 1): I saw this live at the telescope and now in "Folder Monitor" the Asteroids are even farer away. May be the time for the calculation is not the time the stack is taken, but the current time (at least in the "Folder Monitor")?
I checked the time of the laptop. It is correct. In previous SC versions the Asteroid positions were marked nearly dead center (only some '' off).
Another plate solve does not fix it (I need the plate solve anyway in "FolderMonitor" to show the annotations).

Item 2): I now see that after using "Colour Calibration", even after a "clear" the "Background subtraction" is selected (I did not see that). Unmarking it
removes the histogram peaks at left after a "clear". In the "Folder Monitor" loading frames from that night I can not repeat the "blue problem" and "Auto Colour Balance" works as before. I had played with "Background Subtraction", but normally do not use it for my Asteroid measurement stacks.
I use the same dark and flat in the "Folder Monitor" as in that night.
I will have to check it again live, as soon as weather and time permits. So sorry for that noise!

So thanks for your help and hopefully you find something for item1. If you need a stack with an Asteroid, I can provide it.

cheers,
Mike

[admin]

Hi Mike,

unfortunately correct asteroid positions for the folder monitor camera (or the test cameras - anything where you are loading old images) are not something that is dealt with by SharpCap - all the positions of asteroids are calculated for the current time as shown by the system clock, not the timestamp of the image that has been loaded. I did wonder about adding that feature when I was doing the asteroid annotation (it would have helped with my testing), but it turned out to be complex enough to do that it didn't seem worthwhile.

cheers,

Robin

[MikeB838]

Hi Robin,

I think I have seen the Asteroids correctly annotated in the "Folder Monitor". Checked it today with Version 13685 and you are right. No Asteroid to see.
I'm wondering what I have mixed up. Sorry. And it is no problem for me, that the Asteroids are not correctly annotated in the "Folder Monitor", even that
I make havy use of the "Folder Monitor".

Cheers,
Mike

[admin]

Hi Mike,

it would simply depend on how recently the images you are using in 'Folder Monitor' were captured... If you were looking at images from the same day or the previous day, the asteroids might not have moved much (especially for the more distant ones). For older images, the annotations will be much further out.

cheers,

Robin

[MikeB838]

Hi Robin,

sure and I thought, I had seen it at least 12 hours after the shot (after that the Asteroid should be outside of the image or at least not centered).
But who knows what I have imagined

cheers,

Mike

[MikeB838]

Hi Robin,

after a while I managed to check with the last SC version (4.1.13835.0-64bit).

Here are my results:

1. Annotating an Asteroid live at the telscope is working! I do not know, what I have seen, sorry. And as you said, the next day in folder monitor the annotation is off. But this is okay for me!

2. "Auto Colour Balance" in live stacking does now work as before (no more blue hue). I had read in the "Whats new thread", that you had tweaked it

And thank you very much for two other changes in former SC versions:
- I very much appreciate that SC now remembers the last directories for darks and flats !

- showing a noticeable colour on titlebars with important settings is great!

I think it would be a great help too on the titles in "Live stacking" - eg. if on the "Background" Tab "Background Subtraction" is selected and others. Everything not set to default.

May be you can add this on your feature list?

Thanks for reading this thread again and for releasing the corrected SC versions, of course!

cheers,
Michael

[admin]

Hi Michael,

glad that the new version is working nicely for you and that you are finding the highlighting useful.

I will think about further places for the highlighting, but I think the two we have already are the most important as they can either cause changes to your saved data that you weren't expecting (preprocessing) or cause you to mis-judge exposure/gain (display stretch) and also end up saving bad data. I have seen both of those happen to people. In the case of the live stacking adjustments, if it takes you a while to spot that they are wrong, there is no real loss - once you do realise they need adjusting you go and fix them and the image on screen is now corrected - the wrong settings haven't affected the actual data in the stack, so the impact is much reduced.

cheers,

Robin

[MikeB838]

Hi Robin,

on one side you are right that only if data can be corrupted the highlighting is most important, but a highlighted changed sub item
(I think about eg. the flip setting in the Preprocessing tab) - may be highlighted in a more pale colour - can be of help.
As I have flats and darks selected, "Preprocessing" is already highlighted, and the other changes can be missed - or one has to check all settings by eye all the time.
But - this is only my thought and I think that there are enough other things to do

Thanks, clear skies and cheers,
Michael