Cookie consent

AstroBin saves small pieces of text information (cookies) on your device in order to deliver better content and for statistical purposes. You can disable the usage of cookies by changing the settings of your browser. By browsing AstroBin without changing the browser settings, you grant us permission to store that information on your device.

I agree

walking noise

03 Jan, 2020 15:06

Streaks appear on my photo after stacking my images, all oriented in the same direction. This does not happen every time but, I have the impression that when the signal / noise ratio is lower the final image are more prone to this problem.According to my calibration of my mount on the celestial pole I am a little less than 2 sec of arc from the ideal position. Each 3 min photo seems ok, no streak. I use pixinsight, imagesplus and registar software.So how do you get rid of these streaks with pixinsight software? I tried to increase the number of darks composing the master dark but that does not change anything on the final appearance of the stack.full spectrum Canon T3i cameraAny suggestions? Thank you.

03 Jan, 2020 17:18
It's very hard to get rid of walking noise after capture. The two best methods for eliminating walking noise are to get your polar alignment 100% right and to use dithering. I don't dither yet, so in cases where I've had walking noise (thanks to bumped mount, for example), I simply gather more data and use kappa sigma clipping when stacking. In the past, that's kept me from having to trash data with walking noise.
03 Jan, 2020 17:35
What Chris said.Any kind of noise which has a pattern is hard to get rid of. DSLRs are particularly bad. Dithering is by far the best solution, even with good polar alignment. Programming an aggressive random dither every 2-3 frames will make a major improvement. Both SGP and Backyard EOS support dithering.
03 Jan, 2020 17:39
Hello Chris,

i've tried dithering once with ByEOS software but it was very time consuming. But perhaps it is a good idea to use it anyway.

I will check my polar alignment as you suggest and thanks for the advice for the  kappa sigma clipping i will try it.

03 Jan, 2020 17:43
Hello Arun,

Ok so dithering is the solution whatever the polar alignment quality. So I will do it !


03 Jan, 2020 17:44
Yeah, that's the main reason I don't dither. Ever since I got a PoleMaster, I've only experienced it once. I knew I bumped my mount but underestimated how badly I knocked things off. Adding another 4 hours of data got rid of it.
03 Jan, 2020 20:44

The root cause is pattern noise on the sensor, and the slightest inaccuracy in polar alignment will add the patterns during integration, resulting in the dreaded streaks.

I have found that this kind of noise is reduced significantly with good darks (which is not the same as *many* darks, essentially you need to match your sensor temperature while taking the lights). It disappears almost completely (at least in subs taken with my D7500) if I use a Bad Pixel map: (scroll down to the section "Bad Pixels"smile

I am not sure why. Most likely explanation is BPM interpolation "destroys" the noise pattern, so there is nothing to cause streaks. You still get other artifacts of problematic tracking, such as more elongated stars.

03 Jan, 2020 22:18
Hi D.,

ok, interesting i will check the quality of my darks… and the info on web sites you suggest too !


28 Jan, 2020 04:59
If by streaks you mean vertical or horizontal bands, CanonBandingReduction script works very well, just make sure you apply to to linear image before anything else (before DBE etc) and rotate the image so that your bands are horizontal. Ideally this should be applied to each individual frame, but I found it in my case it works very well on integrated image. Hope this helps

28 Jan, 2020 12:18
Hello Saulius,

this is a different problem than the horizontal or vertical bands. However, I  frequently use CanonBandingReduction but after DBE  and without proper rotation so,  I take note of your suggestions which will be very useful

thank you

Edited 28 Jan, 2020 12:19
Register or login to create to post a reply.