4.2 Enhancements/Improvements for DSO Live Stacking

[MarMax]

Robin,

I've been using 4.2b exclusively for the past three months. I'm also vigilant about checking for updates and downloading them. As a one trick pony DSO user of SharpCap I'm rarely using the planetary side of the application. I've always wondered if you have statistics on your user base and how many are planetary versus DSO.

As a deep space observer I'd like to see further enhancements for 4.2 which I've shared before. As an end user with zero understanding of what's under the hood I'd like to know if improved gradient removal and star reduction are feasible enhancements or just too complex to tackle. If feasible it would be nice to know if this is something on your to-do list.

Moving on to other suggestions, for the arcsinh stretch and the softening option it would be nice if the arrow buttons made 0.1 increments versus 0.2. I've found that 0.1 is the sweet spot and have to manually type it in versus using the arrow button.

Adjusting the color in the background tab after you've color calibrated seems to be hit and miss for me. Many times the handles for each color overlap and I'm not able to easily clink on the adjust tab. And then when I do it seems like mouse movements (trackball for me) are too extreme. I'd suggested the fine adjust arrows for the histogram a while back which I use constantly and love. Is there any way to do something like this for the background adjustments?

In the more recent versions of 4.2 I've noticed that in addition to closing something (like live stacking or FWHM) does not actually close it. That is, clicking the X in the corner of the feature dialog box used to close the feature. Now it seems like it only closes the dialog box yet the feature continues to run and I need to go to the main menu to close the feature.

EDIT to add one more suggestion regarding rejection of frames with a satellite trail. Typically when a frame with a satellite trail is stacked it's quite obvious with a straight white line across the entire frame. So this suggestion is for those obvious satellite trails, not the short dim or long dim trails that you don't notice without pixel peeping. Even when I'm diligent at watching the stack and unstack the bad frames there's always one that gets past me (a bathroom break, a binocular break, etc.). Can you detect these more obvious trails at the time of acquisition and delete or not add to the stack?

Thank you,
Mike

[admin]

Hi Mike,

my data suggests that use of deep sky stacking is about double the user of planetary/solar system stacking - at least in the last month of data. Mind you, that ratio may change significantly when Jupiter comes back into position...

4.2 has definitely ended up having more improvements for deep sky than planetary, in part because it's hard to see at the moment what other improvements for planetary are useful/possibly. There are definitely some still for deep sky that could in theory be added.

For the star size reduction, while the new stretch options help, it would be good to be able to go further. However, the normal approach for that is to create a star mask, separate the image into 'stars' and 'starless' and then process each separately before recombining. There are a significant number of steps there that would need to work reliably without significant user interaction to make that process workable inside of SharpCap live stacking - it's something I would like to try but it is significantly too big to consider for including into the initial release of 4.2

Background subtraction is probably an easier challenge - while you can do it in the 'preprocessing' tab, there is no way to undo/adjust that once it has been applied - putting the same sort of background subtraction algorithms into the live stack processing would probably help, since you could tinker with the settings and see the results. Something to consider...

For satellite trails, you have 3 current options (of which you can use any one or any combination)...

1) 'Satellite trail removal' in the preprocessing section - this is basically what you are asking for (or close to it) - detection and removal of the trails in the individual frames before they are added to the stack. It's a bit awkward to set up (best done with previously saved FITS images in the folder monitor camera), but once adjusted can detect and mask out the worst trails

2) The 'Unstack' button to remove the most recent frame if it added a nasty trail

3) The sigma clipped stacking option (as opposed to the default stack). This is good at excluding the pixels in the satellite trail that are significantly different to the normal pixel values for that area without any user interaction. It's also significantly improved in 4.2, allowing it to remove trails in the first 10 frames and the sensitivity can be adjusted more easily (although still best done with a previous set of FITS files to allow you to tinker with the values and then repeat the same data).

Both 1) and 3) work unattended, so you have some options there.

Finally, for the minor tweaks and potential issues, I will have a look to see what can be done.

thanks for the feedback!

Robin

[MarMax]

Thanks for the dso/planetary user info. I've always been curious and figured it was 50/50, and perhaps it is during planetary season.

I'll need to play a bit more with satellite trail removal. I've been using it consistently since you introduced it but must not have it set optimally. I've never been one to save a bunch of frames to play with later and perhaps that's something I need to do. There are just so many different types of trails. I'm assuming the test frames should be gathered without sigma clipping and satellite trail removal applied.

Mike

[admin]

Hi Mike,

yes, you want to gather the test frames 'raw' from the camera with no processing or stacking applied - one way to do this is to set up Live Stacking to save the raw frames while stacking - that way you can do your normal live stacking and grab some frames for later testing. The 'Folder Monitor' camera then lets you replay the captured frames again and again as if they were being grabbed from a camera, so you can see the effects of different satellite trail detection/removal settings, and at least get things set up to catch the worst ones.

I should also point to the 'new in 4.2' Satellite Forecast map (at the bottom of the tools menu) - this plots expected satellite density across the sky and lets you see how it will change over the night. You might think it is going to be pretty uniform, but that's not the case at all - there are big variations both in the number of satellites and in whether they will be visible or not due to the shadow of the earth. Picking the right target to observe in the right area can help a lot.

I've looked at some of the other points you raised -

* Closing the tool area at the bottom (live stacking, planetary live stacking, histogram, etc) does seem to fully turn off the tool in my testing - for instance when using one of the live stacking tools, the image goes back to the normal unprocessed image from the camera either straight away or when the next frame is captured from the camera.

* The background subtraction can be adjusted more finely in a few ways already - holding down <SHIFT> while dragging the sliders makes them move more slowly than the mouse. You can also click on one of them then adjust it using the left/right arrow keys (or <SHIFT>+ left right for smaller adjustments). I have however stopped the 'raw' markers (which cannot be moved) from interacting when under the mouse - that makes it easier to be sure of adjusting the other markers that can be moved...

* I have changed the default adjustment on the soften parameter from 0.2 to 0.1

cheers,

Robin