What's new in SharpCap this week...
- admin
- Site Admin
- Posts: 14358
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
What's new in SharpCap this week...
Hi folks,
I am going to try out the idea of writing a short post roughly once a week to outline what's new in SharpCap (with a bit more detail than the bullet points on the downloads page) as well as what is likely to be happening next.
This week is a bit of a bumper update with more bug fixes and small changes than usual - largely because there was no update last week.
Bugfixes
In the bugfixes side of things, there are still a number of fixes going in to address small (and perhaps not so small) issues with the new UI option (viewtopic.php?t=4875). New UI bugs fixed this time around include a possible crash when changing the Windows screen scaling with SharpCap running and a loss of control/tool area content after repeated unpin/repin of the control/tool panel. There may still be some font sizing issues when running on a high DPI display, and a possible issue with the image display area freezing. If you are affected by New UI related issues, please post in the New UI thread (link above) or as a separate bug report with as much info as possible.
Other bugfixes include fixing a missing taskbar icon for the 32 bit version, fixing a problem selecting live stacking from Python scritping and fixing a crash if you tried to 'Save exactly as seen' using the button in the mini histogram before the camera has captured a frame.
Gradient Removal
The main feature this time around is the addition of a gradient removal option to the background subtraction control. When enabled, this will try to detect any linear gradient in the background brightness of an image and remove it. This correction is applied to each frame coming from the camera before stacking or saving (just like flat correction, dark subtraction, etc). How accurate will the gradient removal be? Well, it depends on the type of image you are taking...
* Image is mostly background, with small areas of galaxy/nebulosity, or larger areas of faint stuff - the detection of background gradient should be excellent, removing 99% or possibly even more of the gradient
* Image has large areas of bright nebulosity or galaxy, but there are good sized background areas on all sides - should give good results, likely to reduce the gradient by at least 90%
* Image is all galaxy/nebulosity, or galaxy/nebulosity covers one side of the image but not the other - use with caution... In these circumstances the detected and removed 'gradient' might be actually part of the image that you want.
Small Features
I've had a couple of questions recently about whether it is reasonable to continue to use the polar alignment results to make further adjustments through an imaging session - basically people tried leaving SharpCap Polar Align running as they imaged to watch the results to see if they change. Unfortunately this doesn't work - as soon as you move the mount in RA, the results are ruined. I have now added code to detect rotation of the image once you get to the adjustment stage of the Polar Align process, and deliberately stop producing alignment updates if significant further rotation is detected - you will also get a message suggesting that you should restart alignment or undo the rotation.
Other small usability improvements include adding hotkeys to zoom the image (CTRL + to zoom in, CTRL - to zoom out), tweaks to the row noise reduction to reduce the chance of it over-correcting and some changes to ensure that the polar alignment guidance text is always drawn entirely within the camera image, rather than partly out of view in some cases.
Finally, another usability tweak to the Satellite Trail Removal feature is a new notification that appears when the satellite trail removal mode is set to 'Remove'. This notification will remind you to check the sensitivity/minimum trail length/width settings in test mode before using the 'Remove' mode. In some cases where people have just turned the option to 'Remove' without making adjustments, parts of the image are incorrectly detected as trails, basically messing up the image. This seems to be most common currently with globular cluster images.
Coming soon
The next things planned are:
* Improvements to satellite trail removal - trying to deal with non-continuous trails in them and those that vary greatly in brightness along their length, along with trying to reduce the chances of non-satellite parts of the image being incorrectly detected as trails
* Autofocus improvements - a number of small but significantly helpful tweaks to this to make it more usable, more reliable and easier to launch
cheers,
Robin
I am going to try out the idea of writing a short post roughly once a week to outline what's new in SharpCap (with a bit more detail than the bullet points on the downloads page) as well as what is likely to be happening next.
This week is a bit of a bumper update with more bug fixes and small changes than usual - largely because there was no update last week.
Bugfixes
In the bugfixes side of things, there are still a number of fixes going in to address small (and perhaps not so small) issues with the new UI option (viewtopic.php?t=4875). New UI bugs fixed this time around include a possible crash when changing the Windows screen scaling with SharpCap running and a loss of control/tool area content after repeated unpin/repin of the control/tool panel. There may still be some font sizing issues when running on a high DPI display, and a possible issue with the image display area freezing. If you are affected by New UI related issues, please post in the New UI thread (link above) or as a separate bug report with as much info as possible.
Other bugfixes include fixing a missing taskbar icon for the 32 bit version, fixing a problem selecting live stacking from Python scritping and fixing a crash if you tried to 'Save exactly as seen' using the button in the mini histogram before the camera has captured a frame.
Gradient Removal
The main feature this time around is the addition of a gradient removal option to the background subtraction control. When enabled, this will try to detect any linear gradient in the background brightness of an image and remove it. This correction is applied to each frame coming from the camera before stacking or saving (just like flat correction, dark subtraction, etc). How accurate will the gradient removal be? Well, it depends on the type of image you are taking...
* Image is mostly background, with small areas of galaxy/nebulosity, or larger areas of faint stuff - the detection of background gradient should be excellent, removing 99% or possibly even more of the gradient
* Image has large areas of bright nebulosity or galaxy, but there are good sized background areas on all sides - should give good results, likely to reduce the gradient by at least 90%
* Image is all galaxy/nebulosity, or galaxy/nebulosity covers one side of the image but not the other - use with caution... In these circumstances the detected and removed 'gradient' might be actually part of the image that you want.
Small Features
I've had a couple of questions recently about whether it is reasonable to continue to use the polar alignment results to make further adjustments through an imaging session - basically people tried leaving SharpCap Polar Align running as they imaged to watch the results to see if they change. Unfortunately this doesn't work - as soon as you move the mount in RA, the results are ruined. I have now added code to detect rotation of the image once you get to the adjustment stage of the Polar Align process, and deliberately stop producing alignment updates if significant further rotation is detected - you will also get a message suggesting that you should restart alignment or undo the rotation.
Other small usability improvements include adding hotkeys to zoom the image (CTRL + to zoom in, CTRL - to zoom out), tweaks to the row noise reduction to reduce the chance of it over-correcting and some changes to ensure that the polar alignment guidance text is always drawn entirely within the camera image, rather than partly out of view in some cases.
Finally, another usability tweak to the Satellite Trail Removal feature is a new notification that appears when the satellite trail removal mode is set to 'Remove'. This notification will remind you to check the sensitivity/minimum trail length/width settings in test mode before using the 'Remove' mode. In some cases where people have just turned the option to 'Remove' without making adjustments, parts of the image are incorrectly detected as trails, basically messing up the image. This seems to be most common currently with globular cluster images.
Coming soon
The next things planned are:
* Improvements to satellite trail removal - trying to deal with non-continuous trails in them and those that vary greatly in brightness along their length, along with trying to reduce the chances of non-satellite parts of the image being incorrectly detected as trails
* Autofocus improvements - a number of small but significantly helpful tweaks to this to make it more usable, more reliable and easier to launch
cheers,
Robin
Re: What's new in SharpCap this week...
Extremely happy to see the Gradient Removal tool going into action. Thank you very much for that!
Re: What's new in SharpCap this week...
Hi Robin
Did test the new gradient removal feature. Here the results.
I used a saved set of FITS of my UMi Dwarf capture. One stack with gradient removal Off and one set On.
Both sets stacked in DeepSkyStacker with the same settings and same calibration files.
The gradient is not removed as you can see but it's divided in a more equal way. I don't know if that is expected but that's the result here. I suspect because the captures do not have a lot of gradient to begin with?
I use SiriL for background extraction and the set with gradient removal set to On has clearly less to do, I can work with a lower tolerance in the SiriL background extraction which is just what one wants there.
Here the stack with gradient removal Off.
Here the stack with gradient removal On
Did test the new gradient removal feature. Here the results.
I used a saved set of FITS of my UMi Dwarf capture. One stack with gradient removal Off and one set On.
Both sets stacked in DeepSkyStacker with the same settings and same calibration files.
The gradient is not removed as you can see but it's divided in a more equal way. I don't know if that is expected but that's the result here. I suspect because the captures do not have a lot of gradient to begin with?
I use SiriL for background extraction and the set with gradient removal set to On has clearly less to do, I can work with a lower tolerance in the SiriL background extraction which is just what one wants there.
Here the stack with gradient removal Off.
Here the stack with gradient removal On
You do not have the required permissions to view the files attached to this post.
- admin
- Site Admin
- Posts: 14358
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: What's new in SharpCap this week...
Hi,
perhaps I have over hyped the gradient removal by missing out a key work - 'linear'. The code will look for a constant gradient across the frame (it can be in any direction, including diagonal) and will remove that. If the gradient is more complex then the linear component will get detected and removed, leaving behind the higher order components. I think this is what is happening to Menno's images...
By the way, does Siril give any numbers for the amount of background variation that needs to be removed in each case? It would be interesting to know how much of the original variation the simple linear approach managed to deal with.
cheers,
Robin
perhaps I have over hyped the gradient removal by missing out a key work - 'linear'. The code will look for a constant gradient across the frame (it can be in any direction, including diagonal) and will remove that. If the gradient is more complex then the linear component will get detected and removed, leaving behind the higher order components. I think this is what is happening to Menno's images...
By the way, does Siril give any numbers for the amount of background variation that needs to be removed in each case? It would be interesting to know how much of the original variation the simple linear approach managed to deal with.
cheers,
Robin
Re: What's new in SharpCap this week...
Hi Robin
I'm not sure what you mean with this?
Menno
- admin
- Site Admin
- Posts: 14358
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: What's new in SharpCap this week...
Hi Menno,
I was imagining that there might be some measurement value in Siril that shows you how much irregularity there is in the background - a number of some sort which is maybe the root mean squared deviation of the background brightness from a constant or similar.
If that number was - say - 152 for the frame without SharCap's gradient removal and 62 for the frame with Gradient removal then you could say that SharpCap has removed about 60% of the variation in the background. That's the sort of number I was interested in - has it removed 10%, 30%, 70%, 90%?
cheers,
Robin
I was imagining that there might be some measurement value in Siril that shows you how much irregularity there is in the background - a number of some sort which is maybe the root mean squared deviation of the background brightness from a constant or similar.
If that number was - say - 152 for the frame without SharCap's gradient removal and 62 for the frame with Gradient removal then you could say that SharpCap has removed about 60% of the variation in the background. That's the sort of number I was interested in - has it removed 10%, 30%, 70%, 90%?
cheers,
Robin
Re: What's new in SharpCap this week...
Okay, that I understand a bit
There is no real important numbers other than the Tolerance as indicated in the screenshot. It has something to so with sigma units, whatever those are
It ranges from 0.01 to 4.00. But this is only apparent when using the automatic feature. I can add or remove blocks by hand but no number for that is given
Don't know if it is of any use, but this is from the previous screenshots:
The value for the stacked capture with gradient removal On I can easily set to around 0.15. There's only a bit in the right lower corner. so it's rather "clean".
With the gradient removal Off, the slider/number gets to around 0.8 before the huge patch on the lower left is extracted.
Bot of them of course being a kind of "false positives" in the sense that normally the parts that are not covered are stars and/or objects.
Menno
There is no real important numbers other than the Tolerance as indicated in the screenshot. It has something to so with sigma units, whatever those are
It ranges from 0.01 to 4.00. But this is only apparent when using the automatic feature. I can add or remove blocks by hand but no number for that is given
Don't know if it is of any use, but this is from the previous screenshots:
The value for the stacked capture with gradient removal On I can easily set to around 0.15. There's only a bit in the right lower corner. so it's rather "clean".
With the gradient removal Off, the slider/number gets to around 0.8 before the huge patch on the lower left is extracted.
Bot of them of course being a kind of "false positives" in the sense that normally the parts that are not covered are stars and/or objects.
Menno
You do not have the required permissions to view the files attached to this post.
Re: What's new in SharpCap this week...
Hi Robin,
The linear background subtraction looks good and I look forward to seeing its effect on autostacks.
A trivial point maybe but you explain above that in the case of more complex backgrounds it will subtract the lower more nearly linear frequency leaving higher order variations behind. Would it help a bit to have it routinely run successively twice?
Tim
The linear background subtraction looks good and I look forward to seeing its effect on autostacks.
A trivial point maybe but you explain above that in the case of more complex backgrounds it will subtract the lower more nearly linear frequency leaving higher order variations behind. Would it help a bit to have it routinely run successively twice?
Tim
- admin
- Site Admin
- Posts: 14358
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: What's new in SharpCap this week...
Hi Tim,
you are right that it will leave higher order variations behind, since it only removes the linear component. If you ran it twice then the second pass would do very close to nothing (not quite nothing, since the 'least absolute deviation' fit used to find the background is a bit more complex in its response than the 'least squares' fit more commonly used). Anyway, basically one pass will remove any linear gradient in both directions, so if you try again there will be nothing (ish) to remove.
cheers,
Robin
you are right that it will leave higher order variations behind, since it only removes the linear component. If you ran it twice then the second pass would do very close to nothing (not quite nothing, since the 'least absolute deviation' fit used to find the background is a bit more complex in its response than the 'least squares' fit more commonly used). Anyway, basically one pass will remove any linear gradient in both directions, so if you try again there will be nothing (ish) to remove.
cheers,
Robin
- admin
- Site Admin
- Posts: 14358
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
What's new in SharpCap this week... 18th April 2022
Hi folks,
Monday again - new version of SharpCap is uploaded and ready for you to play with.
Satellite Trail Removal
The big change this week is a major improvement to satellite trail removal. It's not that the old satellite trail feature didn't work... It worked too well in some cases, and was particularly keen at picking up the cores of globular clusters as multiple satellite trails and giving them a 'frosted glass' look
The latest update fixes this issue by applying a second stage processing to any potential trails (detected lines) in the image - making sure that those lines are not too 'gappy' and that fit into a sensible pattern rather than being criss-crossed over each other.
Please still remember to check your satellite trail removal settings using the 'Test' mode before turning the setting to 'Remove'. At the minimum, in 'Test' mode, make sure that none of the parts of the target you are trying to image are wrongly detected as satellite trails.
Small Improvments
The temperature readout from an ASCOM focuser will be displayed along with the other focuser controls if your focuser has a temperature sensor and is connected in SharpCap.
When saving a deep sky annotated image while live stacking is active, the saved file will be placed in the live stacking 'processed' folder (or according to the live stacking processed image template) to keep it with the other images from the same stack.
Bug Fixes
Fixes a crash in new UI if the selection box is visible, is also very small, and you have zoomed out to the extent that the calculated size of the selection box is less than one pixel wide or high.
An Issue to watch out for
If SharpCap crashes for you with an error that contains this text
the most likely cause is buggy or out-of-date graphics card drivers. If you see this problem, please download and install the latest drivers for your graphics card from your computer manufacturer's website (or the website of the manufacturer of your graphics card).
cheers,
Robin
Monday again - new version of SharpCap is uploaded and ready for you to play with.
Satellite Trail Removal
The big change this week is a major improvement to satellite trail removal. It's not that the old satellite trail feature didn't work... It worked too well in some cases, and was particularly keen at picking up the cores of globular clusters as multiple satellite trails and giving them a 'frosted glass' look
The latest update fixes this issue by applying a second stage processing to any potential trails (detected lines) in the image - making sure that those lines are not too 'gappy' and that fit into a sensible pattern rather than being criss-crossed over each other.
Please still remember to check your satellite trail removal settings using the 'Test' mode before turning the setting to 'Remove'. At the minimum, in 'Test' mode, make sure that none of the parts of the target you are trying to image are wrongly detected as satellite trails.
Small Improvments
The temperature readout from an ASCOM focuser will be displayed along with the other focuser controls if your focuser has a temperature sensor and is connected in SharpCap.
When saving a deep sky annotated image while live stacking is active, the saved file will be placed in the live stacking 'processed' folder (or according to the live stacking processed image template) to keep it with the other images from the same stack.
Bug Fixes
Fixes a crash in new UI if the selection box is visible, is also very small, and you have zoomed out to the extent that the calculated size of the selection box is less than one pixel wide or high.
An Issue to watch out for
If SharpCap crashes for you with an error that contains this text
Code: Select all
UCEERR_RENDERTHREADFAILURE
cheers,
Robin