15th Feb 2024: Astro Pixel Processor 2.0.0-beta29 released - macOS native File Chooser, macOS CMD-Q fixed, read-only Fits on network fixed and other bug fixes
7th December 2023: added payment option Alipay to purchase Astro Pixel Processor from China, Hong Kong, Macau, Taiwan, Korea, Japan and other countries where Alipay is used.
As Vincent corrected me here, darks actually do not help reduce noise so it is not a surprise to me that the noise hasn't gone down. Good reasons to create new darks are
A change in gain or temperature
Driver updates (ZWO is known to change the default offset in their drivers in at least one occasion)
An increased number of hot pixels
Darks typically have to be shot only once a year, mostly because of the third bullet.
Also be aware that noise is always part of the game. Having very good calibration data will get that to a better level, but really getting the overall noise down is only possible (if you want that to be non destructive) with more data.
i think another problem ist, that i probably not worked backgroundlimited (I dont know its the right words in english). I mean that's almost 26 hours of exposure time at Gain100 and f/5. I've seen other shots of NGC7023, with less than half the time with identical data that contains more detail. I think there is something I have not done optimally.
The pictures you are posting are taken from the raw stack right? I mean, you process the subs up to and including tab 6 INTEGRATE and that's all? As I explained before, that is only part of the process. After that you will need to post-process the image, which is an entire process in itself with many subtleties, to get the most details out of the image.
Have you checked that all light frames were calibrated as expected? By looking at the frame column in the frame list panel at the bottom ?
What kind of noise reduction do you see in the metadata/header of the stack/integration?
The noise reduction metrics will tell you and us if the integration went good or not... for instance, no dithering or too small dither steps will lead to less noise reduction as ideally possible 😉
This is an example:
HDU1 - NOISE-1 = ' 1,3622E-04' / noise level of channel 1 HDU1 - SNR-1 = ' 3,1276E+00' / Signal to Noise Ratio of channel 1 HDU1 - NOTE-2 = 'NR = Noise Reduction' HDU1 - NOTE-3 = 'medNR = noise in median frame / noise in integration' HDU1 - NOTE-4 = 'refNR = noise in reference frame / noise in integration' HDU1 - NOTE-5 = 'ideal noise reduction = square root of number of frames' HDU1 - NOTE-6 = 'the realized/ideal noise reduction ratio should approach 1 ideally' HDU1 - NOTE-7 = 'the effective noise reduction has a correction for' HDU1 - NOTE-8 = 'dispersion change between the frame and the integration' HDU1 - NOTE-9 = 'because dispersion and noise are correlated' HDU1 - medNR-1 = ' 4,6778E+00' / median noise reduction, channel 1 HDU1 - refNR-1 = ' 4,6778E+00' / reference noise reduction, channel 1 HDU1 - idNR-1 = ' 4,5826E+00' / ideal noise reduction, channel 1 HDU1 - ratNR-1 = ' 1,0208E+00' / realized/ideal noise reduction ratio, channel 1 HDU1 - medENR-1= ' 2,5241E+00' / effective median noise reduction, channel 1 HDU1 - refENR-1= ' 2,5159E+00' / effective reference noise reduction, channel 1
In this example, the ratNR-1, realized / ideal noise reduction ratio, is 1,02. Theoretically, it can't go higher than 1, practically it can. But this means that this integration is very well performed in terms of noise reduction through the integration/stacking process.
If a stack is not optimally integrated, the realized/ideal NR ratio will be clearly lower than 1. Can you check the value for your stack?
Cheers,
Mabula
This post was modified 3 years ago by Mabula-Admin
There is nothing wrong there no, it could be a bit better maybe, but to be honest I really don't see anything wrong with the data. It looks really good, I think it's more that other people with that amount of data use quite a bit of extra post processing to make it look nicer. However, I tend to like a bit more noise, as this usually means you preserve more details.
Your noise reduction realized/ideal ratio through stacking/integrating all of your data is between 0,87 - 0,92 ( 9,2000E-01 is scientific notation for 0,92 😉 ).
Ideally it would be 1. So that means that when the data was integrated ideally/perfectly, the noise in your final integration/stack would have had 1/0,92 = 1,086 -> 8,6 % less noise.
A couple of % you will not see easily with your eyes, when it becomes 10-20% you will see the difference with your eyes in terms of how noisy the result is.
Now since, you have integrated 26 hours of exposure time divided in many ! frames I assume, there is definitely room for improvement I think.
I think the most and clear improvement will be gained in terms of noise reduction if you start to improve the dithering in the capture process. So did you dither? And if so, what kind of dither steps, how large in pixels are dither steps between the frames? If you dithered, did you do it, between all exposures?
Thank you both for your feedback. Yes, I always use dithering, and that on every shot. I think the steps I use are 5 pixels.
I probably had high expectations because of the many exposures. I simply worked several times with some tools and removed the smoking as good as possible. Also I had more expectation concerning the lifting of the dark clouds to the background. I think I'll let that be, and close NGC7023 for me for the first one. I would like to plan more precious exposure time for other subjects for the time being. I think for the beginning it is already ok.
But it would be cool if there were more possibilities in APP in the future to optimize the noise already during stacking. 😉
