Flat frames and capture area

[lowenthalm]

When capturing dark frames for my dark library, I have to capture a set for each combination of gain, exposure, binning and capture area that I commonly use. For the longest time, I thought (and it seemed when using Sharpcap) that I needed to capture flat frames for each binning and capture area combination I used.

In recent releases of Sharpcap 4, the program no longer seems to complain about the flat when I switch from full frame to a smaller capture area, and accepts selecting the the full frame flat even if my capture area is smaller than full frame.

I tested dark selection to see if the same behavior is holds, and it does. However, if I select a new dark when I am have a capture area smaller than full frame, it will send me off into the dark library for the smaller capture area.

So, is it in fact true that I only have to capture a full frame flat at each binning and filter combination I use, and Sharpcap will crop as appropriate for each capture area I select? This will save me a some of time!

When it comes to darks, the behavior described above is inconsistent. It sounds good and safe for bin1x1, but might not be a good idea for smaller than full frame capture areas when binning at more than 1x1 in cameras that handle binning internally.

[admin]

Hi,

the new behaviour is meant to apply equally to both dark and flat frames, and was introduced a few months ago. Previously, the capture area (width*height) had to match between the dark/flat and the light frames being captured - there was no check within SharpCap that the offset within the full frame matched if you were using ROI however.

In the new version, you can continue to work by the same rules as before, OR you can use a full resolution dark/flat frame. By full resolution, I mean the maximum resolution available for your camera at the current binning settings. When a full resolution dark/flat is in use and you select an ROI then SharpCap will cut out the matching ROI from the dark and flat frames to match the active capture area of the camera. This should also work under binning (but was not something I tested extensively). Note that there is no protection against things like having 'flip vertical' on in the dark frame but not in the light frame, but then there wasn't in the original code anyway.

So I think that you should be good to go ahead using full resolution dark/flat frames for capturing on any ROI as long as you stick with the basics of 'other settings the same'.

I see what you mean about the dark library taking you to the folder that matches the current ROI rather than the full ROI. That's a tricky one, since if people have already created a dark library at various ROI settings then switching to browse to the full resolution will be wrong for them, even if it is right for you. I will have to think about that...

cheers,

Robin

[lowenthalm]

This an excellent change. One set of darks would be nice. Some sort of migration scheme does seem in order, but the more I think about it the fewer suggestions I have. Good luck figuring something out!

One additional thing in a similar vein. It would hugely helpful if ROI changes did not to clear the plate solve solution. It drives me crazy trying to get a plate solve on a reduced ROI image. Getting a plate solve at full frame and then reducing ROI to something appropriate for the target object would be a godsend for us folks with 2800mm focal lengths. I might even suggest that behind the scenes, a plate solve frame capture is always at full frame to make a good plate solve more likely!

[admin]

Hi,

the switch to the full camera resolution would have a whole range of pitfalls sadly (different for different brands of camera too). These are handled nicely for the current resolution/capture area change, but that in turn has side effects that would be undesireable for a simple plate solve.

However, the idea of being able to keep the plate solve data when the ROI size or position is changed makes sense - should be possible to implement, since by definition we have the data that maps pixels to sky co-ordinates and we know the initial ROI size/position.

cheers,

Robin

[lowenthalm]

In light of your comment above that selected darks and flats are preserved and applied when switching to a smaller ROI, I kept an eye on whether this was working. It works great for flats - you only have to switch flats when binning. However, darks do not work. I was at bin1, full frame ROI live stacking a 1 minute stack of 3 second images and switched to a smaller ROI and repeated the 1 minute/3 second live stack capture at the smaller ROI. It kept both the flat and dark that I had selected for the full frame ROI selected in the processing controls, but only applied the flat correctly. You could see in the lives tack histogram that the dark was not being applied properly, along with the preview showing some pretty wild coloring. I then changed the dark to "None" and acquired another 1 minute stack. It look the same, with the same histogram, so it appears that the dark is not applied at all under these conditions.

I checked the save files: the camera settings file showed the full frame dark was being applied in the reduced ROI image, even though it wasn't. You could also see in the 16 bit images I saved that the reduced ROI image with the full frame dark "applied" and the reduced ROI image with the dark set to none had identical histograms (in terms of where the dark sky peaks were), nothing like the full frame image that had the dark correctly applied.

So something is still amiss here. I guess I should have reported this in bugs, but it directly pertained to this thread we had conversed on.

[admin]

Ah, OK, thanks for that - just tracked it down and it was an incorrect check on the relative bitdepths of the dark and the actual frame when calculating whether the dark could be used at ROI.

I have a fix for next week's update. Currently I think it would work with the camera in 8 bit mode, which is probably how I convinced myself it worked in testing in the first place.

cheers,

Robin

[lowenthalm]

I see this has been fixed in a recent release. Is there any way that I can get it to work with just the full ROI flats automatically? When I rename the folder with the reduced ROI darks to hide it from SharpCap, Sharpcap just puts me in the darks folder and lets me find it myself. It would be nice if it just automatically put you in the appropriate gain/exposure folder of the full ROI darks.

[admin]

Hi,

yes, that would be good - let me have a dig into it and see what can be done.

cheers,

Robin