Hi folks,
it's been a while since I have posted in this section - recently I have been focusing on bug fixes and stability, which don't really need much in the way of announcement. However, there have been a few small changes, tweaks and improvements slipped in, so I will summarize those - for about the last 6 months - here...
Planetary Live Stacking
* Expert Sharpening mode available with 8 sharpening levels rather than 4
* Auto Sharpen buttons for a first go at sharpening an image, adjustable auto-sharpen strength in expert mode.
* Add a 'Freeze' button to allow planetary stacking to be frozen (then closed, then re-opened, then unfrozen) without stopping the stack or timelapse - allows for mount adjustments, flips, etc
* Add 'Time Travel' capture which buffers frames from the camera in memory and then saves recent frames if the stack quality peaks at a new high level
* ROI tracking option now compatible with use of the camera 'Flip' options where available
(See the Planetary Live Stacking thread for more details)
Deep Sky Live Stacking
* Add an option to perform adjustable background subtraction (separate for R/G/B channels) - this removes a light pollution background before colour balance adjustments are applied, making colour balancing much easier.
Focus assistance
* Improvements to star FWHM measurements used in the Multi-Star FWHM tool - now more accurate and less sensitive to changes in image brightness
* Improved options for solar system focusing, including 4x4 binning for noise reduction and measuring the average of the best 25% of frames
Camera Related
* Make Offset adjustable for ASCOM cameras if they allow it to be changed via the ASCOM driver API
* Improved GPS data support for Moravian GPS cameras
* Improved frame rates for very short exposures on Player One cameras
* Updated SDKs for almost all brands of cameras - these bring bug fixes and support for newly released models
Hardware related changes
* Add tool to allow plate solve corrections to GOTO movements (and push-to) to place the target at a selected off-center position in the image - this is useful when using a guide/finder scope and camera to try to center a target in the main scope if they are not perfectly aligned. Select the 'Offset Plate solve and resync/Push To Target Position' tool
* Add option to prompt for confirmation before parking the mount
* Show progress during mount park
* Show progress window when connecting to ASCOM hardware automatically while opening camera
* Setup/Properties button for Alpaca devices will try to launch the configuration web page for the device
* Add hotkeys for diagonal mount movement using CTRL+ Numpad 1/3/7/9 (CTRL+Numpad 2/4/6/8 already do the 4 compass directions)
File formats and saving
* Change SER file writing so that the files can be played (for example by SERPlayer) while SharpCap is still writing to them
* Add option to save FITS files in 'bottom-up' row order (SharpCap has traditionally used 'top-down', but this is actually incorrect according to the FITS specification. Take care if you change this setting, as processing a mixture of top-down and bottom-up files may go badly wrong in some processing applications).
* Add full plate solving mapping (CD headers) to FITS files saved after SharpCap has performed plate solving
* Add options to set up custom FITS headers that will be added to all saved FITS files
Testing/Virtual Cameras
* The 'High Speed' planetary test camera can now play back AVI files as well as SER files (the AVI files must be in simple RGB format)
* The Planetary test camera supports a simulated ROI mode
* Fix timing issues that artificially limited the frame rate that could be achieved with the planetary test camera
Recent changes...
Re: Recent changes...
Thanks for all your ongoing work, Robin. Greatly appreciated. Can you comment on how this feature is different from the Background Subtraction under Preprocessing?
"Add an option to perform adjustable background subtraction (separate for R/G/B channels) - this removes a light pollution background before colour balance adjustments are applied, making colour balancing much easier."
"Add an option to perform adjustable background subtraction (separate for R/G/B channels) - this removes a light pollution background before colour balance adjustments are applied, making colour balancing much easier."
-
admin
- Site Admin
- Posts: 15625
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: Recent changes...
Hi Mike,
the background subtraction in preprocessing has to be fairly conservative about the target level it aims for when removing a background to avoid any chance of black clipping the image - that means that a substantial uniform background level is typically left behind by that sort of background subtraction. Of course it will be removing gradients if you have selected one of the non-uniform modes, and removing the colour cast in the background too, so it does have value.
The live stacking background subtraction is implemented as part of the processing calculations when converting the raw data in the stack into the displayed image. Combine that with the fact that the user can adjust the amount of subtraction and the code can be much more aggressive with making the remaining background left very small. If there is black clipping then the amount of subtraction can just be dialled down and the problem goes away.
Taking away nearly all of the background brightness from the stack using this new tool before applying the normal live stacking colour balance adjustments makes the colour balance *much* more controllable - no need to make really tiny adjustments to the colour balance sliders any more. Now you actually have the tools to properly fix the colour balance in the situation where you have maybe a yellow white background due to light pollution that appears green in the image due to the high green sensitivity of the camera. You can now *subtract* away the bulk of this light pollution background, leaving just the astronomical part of the image left to be adjusted by the more traditional colour balance controls as needed.
However... the new background subtraction in live stacking does *not* do gradients in background at all - if you are fighting those then using both types is probably called for - the image processing sort to level out the gradients, then the live stacking sort to get the sky glow background wiped down to practically zero.
The 'Auto Adjust' in the new background subtraction usually seems to work pretty well - it's designed to try to align the three channels histograms so that the left hand slopes rising towards the peak are pretty closely aligned, putting the peaks as far left as possible while keeping the chances of black clipping low. The alignment of the graphs for the left hand slope of the peak corresponds to subtracting so that the darker parts of the background have a neutral colour, which should work well unless you have bad vignetting.
I'm also working on a fully automatic '1 click' auto colour calibration for deep sky - a lot of it works already, but it needs fine tuning and lots of error handling and so on before I can release it.
cheers,
Robin
the background subtraction in preprocessing has to be fairly conservative about the target level it aims for when removing a background to avoid any chance of black clipping the image - that means that a substantial uniform background level is typically left behind by that sort of background subtraction. Of course it will be removing gradients if you have selected one of the non-uniform modes, and removing the colour cast in the background too, so it does have value.
The live stacking background subtraction is implemented as part of the processing calculations when converting the raw data in the stack into the displayed image. Combine that with the fact that the user can adjust the amount of subtraction and the code can be much more aggressive with making the remaining background left very small. If there is black clipping then the amount of subtraction can just be dialled down and the problem goes away.
Taking away nearly all of the background brightness from the stack using this new tool before applying the normal live stacking colour balance adjustments makes the colour balance *much* more controllable - no need to make really tiny adjustments to the colour balance sliders any more. Now you actually have the tools to properly fix the colour balance in the situation where you have maybe a yellow white background due to light pollution that appears green in the image due to the high green sensitivity of the camera. You can now *subtract* away the bulk of this light pollution background, leaving just the astronomical part of the image left to be adjusted by the more traditional colour balance controls as needed.
However... the new background subtraction in live stacking does *not* do gradients in background at all - if you are fighting those then using both types is probably called for - the image processing sort to level out the gradients, then the live stacking sort to get the sky glow background wiped down to practically zero.
The 'Auto Adjust' in the new background subtraction usually seems to work pretty well - it's designed to try to align the three channels histograms so that the left hand slopes rising towards the peak are pretty closely aligned, putting the peaks as far left as possible while keeping the chances of black clipping low. The alignment of the graphs for the left hand slope of the peak corresponds to subtracting so that the darker parts of the background have a neutral colour, which should work well unless you have bad vignetting.
I'm also working on a fully automatic '1 click' auto colour calibration for deep sky - a lot of it works already, but it needs fine tuning and lots of error handling and so on before I can release it.
cheers,
Robin
Re: Recent changes...
Hi Robin,
Thanks so much for a great program and your frequent updates. I absolutely love it for doing a solar time lapse and also for solar single image capture. I have a basic question about live solar/planetary stacking and how it integrates with the time lapse feature. Perhaps I’m overthinking it. Is the time lapse feature using the target stack image for whatever amount I’ve set at as far as stack length to make the individual time lapse frames. This would require me to be constantly resetting the stack as the solar features change, solar flares et. Or is the time lapse basically just recording a video and saving a clip every 30 seconds or whatever time I set to save a clip that is independent of the stacked image ?
To phrase my question another way when doing a four hour time lapse of the sun and I set the time lapse interval for every 30 seconds is it just recording my target stacked image every 30 seconds and if I don’t click on reset stack I end up getting the same recorded stacked image or is the video time lapse totally separate from the target stack and I don’t have to hit the reset stack button every 30 seconds to get a different image?
Thanks again for an amazing software program!
Bob
Thanks so much for a great program and your frequent updates. I absolutely love it for doing a solar time lapse and also for solar single image capture. I have a basic question about live solar/planetary stacking and how it integrates with the time lapse feature. Perhaps I’m overthinking it. Is the time lapse feature using the target stack image for whatever amount I’ve set at as far as stack length to make the individual time lapse frames. This would require me to be constantly resetting the stack as the solar features change, solar flares et. Or is the time lapse basically just recording a video and saving a clip every 30 seconds or whatever time I set to save a clip that is independent of the stacked image ?
To phrase my question another way when doing a four hour time lapse of the sun and I set the time lapse interval for every 30 seconds is it just recording my target stacked image every 30 seconds and if I don’t click on reset stack I end up getting the same recorded stacked image or is the video time lapse totally separate from the target stack and I don’t have to hit the reset stack button every 30 seconds to get a different image?
Thanks again for an amazing software program!
Bob
Re: Recent changes...
Playing with it some more I realized that the reset stack after every time lapse frame is exactly what I need to do. What was confusing me was it mentioned how it helped with image rotation issues on alt/az mounts in the description -so I just ignored it since I a 10 micron equatorial amount. 
Bob
Bob
Re: Recent changes...
Just wanted to thank you for the FWHM in the file naming structure! It's working great!
-
admin
- Site Admin
- Posts: 15625
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: Recent changes...
Hi Bob,
the stack image will evolve over time without resetting, since the way the stacking works means that as new frames are added from the camera, the effects of older frames are gradually removed from the stack. The 'Target Stack Length' setting, along with the stacking frame rate controls how long this process takes, with a 1000 frame target stack length and 10fps being stacked, the approximate lifetime of data in the stack is 100s (1000 frames / 10fps).
However, the resetting can help - when you reset the stack, the 'alignment reference image' is captured from the initial few frames and then the subsequent frames are aligned to this. So, if features are moving around/changing shape significantly, this could help ensure they are shown as moving correctly in the timelapse, particularly if you are using the 'multi point' stacking mode.
cheers,
Robin
the stack image will evolve over time without resetting, since the way the stacking works means that as new frames are added from the camera, the effects of older frames are gradually removed from the stack. The 'Target Stack Length' setting, along with the stacking frame rate controls how long this process takes, with a 1000 frame target stack length and 10fps being stacked, the approximate lifetime of data in the stack is 100s (1000 frames / 10fps).
However, the resetting can help - when you reset the stack, the 'alignment reference image' is captured from the initial few frames and then the subsequent frames are aligned to this. So, if features are moving around/changing shape significantly, this could help ensure they are shown as moving correctly in the timelapse, particularly if you are using the 'multi point' stacking mode.
cheers,
Robin