This is supposed to be better than GraXpert, which is better than DBE by miles. My experience showed that difference between Gradient Corrector and GraExpert is target dependent and picture quality dependent. Importantly, if flats are not good and speckles/rings appeared on integrated picture, then neither will correct.
I think will a sufficiently small scale and smoothness value and a lower tolerance setting, it can do a good job of correcting the rings. I will be experimenting some more on that with some really old data I have. However I always recommend trying to get the best quality data and getting good quality calibration frames to minimize the amount of processing required later.
DBE is one of those tools that can perform miracles if the user knows how to use it but most people have misconceptions about how to use it and throw their hands in the air and say it is not that good.. Yes, as a default setting, GC works ok but it could be a lot better with a real time preview. With complex gradients not on the edges, it starts to introduce strange rings and the PI team are actively working on fixing that. I am sure that sooner rather than later, GC will evolve into a great tool.
I have been testing GC more and you are correct. With complex gradients I sometime end up getting strange rings. GraXpert has performed better overall as I found in my latest video. DBE is great but can take a significant amount of time to correct really complex gradients.
Hey Abdur, great tests in this video. I ran a test on the subs collected on M101 with my new refractor and I was impressed. Though it was not difficult to achieve, because my subs were shot at a dark site before the moonrise on Thursday and Friday this last week. My ultimate test would be to re-edit my M45 project, which lasted from October to February. There I would see the real potential. Just out of curiosity, were you using GraXpert? I was always testing both DBE and GraX and picking the best result of the two. But this was time consuming though, so I’m hoping Grandient Correction + the upcoming MARS project will provide better results in less time wasted testing different parameters.
Hi Patrick. I look forward to hearing about how it performs with your M45 project. I don't use GraX as DBE was able to do almost everything I needed (with a lot of manual work in some cases) but I will process some images with both GraX and Gradient Correction and see how they compare. From everything I have seen so far Gradient Correction does a better job than GraX so I am not expecting any surprises.
Thank you. This video shows some of my processing workflow: ua-cam.com/video/Eom27tUsCbw/v-deo.htmlsi=QWRFACRWxDe3OyJz Of course this is for simple cases. For monochrome images or images that require more advanced processing, my workflow was more complex. I will make some more videos demonstrating the different processing techniques soon :)
like #20
Thank you, very clear and useful, I now have to re-process old images using that tool and the AI helpers !
Thank you. Me too! Can't wait to see what results we can get with all these new tools compared to a few years ago :)
I think its great! But I wish they could add an option for the user to enable a checkbox that let us see the extracted gradient.
Hopefully that's something that will be added later. We might be able to use Pixel Math in the meantime to generate the gradient.
This is supposed to be better than GraXpert, which is better than DBE by miles. My experience showed that difference between Gradient Corrector and GraExpert is target dependent and picture quality dependent.
Importantly, if flats are not good and speckles/rings appeared on integrated picture, then neither will correct.
I think will a sufficiently small scale and smoothness value and a lower tolerance setting, it can do a good job of correcting the rings. I will be experimenting some more on that with some really old data I have. However I always recommend trying to get the best quality data and getting good quality calibration frames to minimize the amount of processing required later.
DBE is one of those tools that can perform miracles if the user knows how to use it but most people have misconceptions about how to use it and throw their hands in the air and say it is not that good.. Yes, as a default setting, GC works ok but it could be a lot better with a real time preview. With complex gradients not on the edges, it starts to introduce strange rings and the PI team are actively working on fixing that. I am sure that sooner rather than later, GC will evolve into a great tool.
I have been testing GC more and you are correct. With complex gradients I sometime end up getting strange rings. GraXpert has performed better overall as I found in my latest video. DBE is great but can take a significant amount of time to correct really complex gradients.
Hey Abdur, great tests in this video. I ran a test on the subs collected on M101 with my new refractor and I was impressed. Though it was not difficult to achieve, because my subs were shot at a dark site before the moonrise on Thursday and Friday this last week. My ultimate test would be to re-edit my M45 project, which lasted from October to February. There I would see the real potential. Just out of curiosity, were you using GraXpert? I was always testing both DBE and GraX and picking the best result of the two. But this was time consuming though, so I’m hoping Grandient Correction + the upcoming MARS project will provide better results in less time wasted testing different parameters.
Hi Patrick. I look forward to hearing about how it performs with your M45 project. I don't use GraX as DBE was able to do almost everything I needed (with a lot of manual work in some cases) but I will process some images with both GraX and Gradient Correction and see how they compare. From everything I have seen so far Gradient Correction does a better job than GraX so I am not expecting any surprises.
I would like to see a comparison between GraXpert and gradient correction.
I just finished doing one :)
ua-cam.com/video/oeLAyyjoS0Y/v-deo.html
Thanks for a good video. I’d like to know how were your images processed prior to using the the new tool? (I mean after stacking).
Thank you. This video shows some of my processing workflow:
ua-cam.com/video/Eom27tUsCbw/v-deo.htmlsi=QWRFACRWxDe3OyJz
Of course this is for simple cases. For monochrome images or images that require more advanced processing, my workflow was more complex. I will make some more videos demonstrating the different processing techniques soon :)
Thanks for the tutorial Abdur. I just updated Pixinsight, but it doesn't show GradientCorrection. Any idea what went wrong?
My bad, had to download an entire new version. Not just update
Glad to hear it worked out :)
Thanks will try it
Let me know if you like it :)
like # 19