Here's a bit of a short stack on NGC346 in Ha (6 x 10 minutes only). They have been stacked without any darks of bias frames to see what the raw data is showing. As you can see, there is a serious artifact there. It is composed of a series of bright spots that seem to correspond to the number of subs taken. So the images, whatever they are are marching across the image. Yet the image itself is stable enough. Which I suppose means my guiding was adequate. But if that's so, why are these dots moving?
The bright dots are hot pixels, and there is a slight drift between the subs, indicating that guiding pauses during downloads/refocusing or that dithering pushes the mount to one side, suggesting that either mount is not balanced or polar alignement could be improved.
The bright dots are hot pixels, and there is a slight drift between the subs, indicating that guiding pauses during downloads/refocusing or that dithering pushes the mount to one side, suggesting that either mount is not balanced or polar alignement could be improved.
Thanks Suavi. I'll redo the PA. I have been using SharpCap for PA and found it pretty good. I'll see if that helps. And I'll reduce the length of guiding subs too. Currently I use 5-second subs but I'll cut that to 1 second and see if that helps too. Always something isn't there?
Not sure what mount you’re using Peter but when I had an EQ6 I found the 1-2s mark to be optimum.
Suavi is correct, they’re hot spots that will largely be removed when you calibrate the data. You’d be best setting up a dithering pattern for the long term though. Although what you have at the moment will statistically remove hot pixels you may end up with a form of PRNU; walking noise. You won’t notice it until you start stacking for longer integrations and it’ll really limit how much stretching you can do for the fainter signals.
Not sure what mount you’re using Peter but when I had an EQ6 I found the 1-2s mark to be optimum.
Suavi is correct, they’re hot spots that will largely be removed when you calibrate the data. You’d be best setting up a dithering pattern for the long term though. Although what you have at the moment will statistically remove hot pixels you may end up with a form of PRNU; walking noise. You won’t notice it until you start stacking for longer integrations and it’ll really limit how much stretching you can do for the fainter signals.
G'day Colin. Not sure what PRNU is. Another job for Google.
Peter
G'day Colin. Not sure what PRNU is. Another job for Google.
Peter
Come to think of it, I think PRNU is the wrong term in this situation. Walking noise is FPN, got my acronyms all mixed up FPN is Fixed Pattern Noise. CMOS sensors show it a lot more than CCDs do and that is mostly due to the lower read noise of CMOS. FPN tends to show itself in lower SNR areas (the background).
A little while ago I did a whole lot of 0.1s exposures on the star Eta Carina to capture the nebulosity surrounding it. Because I didn't dither even the "bright" Eta Carina Nebula was really faint in the image and was destroyed by FPN.
Dark and flat calibration really helps to mitigate it but ultimately you need dithering to average it out entirely.
I don't know which programme(s) you use for image processing, especially integration, but I have found a Median integration to nearly always mitigate or even remove entirely any FPN. Average integration on the other hand, seems to show it (especially on bloody SBIG cameras).
I don't know which programme(s) you use for image processing, especially integration, but I have found a Median integration to nearly always mitigate or even remove entirely any FPN. Average integration on the other hand, seems to show it (especially on bloody SBIG cameras).
G'day Lewis. I'm using a SX CCD in Maxim to image, DeepSky Stacker to stack and PS5 to process (while I try to get my head around PI).
I'm approaching the solution a step at a time. Tonight's first baby step was to redo the PA. SharpCap told me I managed an "excellent" alignment. I haven't bothered to do a SYnScan 3-star alignment and Polar align (the new version of SynScan offers a good routine for this).
If this artifact persists, I will drop my guiding subs from 5 to 2 seconds. But I'm imaging from a suburban location within a km of the sea so seeing quality is pretty bleah making it difficult to find a guide star most of the time unless you run to longer subs.
If neither of those things produces a result I'll try to find how to dither in Maxim. I can't see it on a quick skim but it's sure to be there somewhere. If not I'll switch guiding to PHD2 but that means running yet another software package and straining the old CPU a heap more than I like.
Anyway, that's the plan. It'll be interesting to whether there are incremental improvements or none at all. I'll try to come back to this thread with a report. Might be a little while though. I'm headed to get some bits of me cut off this Friday as a result of which I may not be imaging for a couple of weeks.
You need to get rid of those hot pixels prior to aligning your subs with a dark subtraction and/or a bad pixel map. Dithering alone won't be enough unless you get a lot of subs as they are pretty bright.
Peter, dithering in MaxIM is really quite easy - in the Camera Control tab, click on EXPOSE : AUTOSAVE : AUTOSAVE SETUP -> DITHER, and then select your options (either via guider or mount, and either coordinates and movement size).
That's how I do it (with a sequence), but there is likely another way.
Peter, dithering in MaxIM is really quite easy - in the Camera Control tab, click on EXPOSE : AUTOSAVE : AUTOSAVE SETUP -> DITHER, and then select your options (either via guider or mount, and either coordinates and movement size).
That's how I do it (with a sequence), but there is likely another way.
Lewis, correct me if I'm wrong, (CMIIW) please. I understand that with guided imaging, dithering should be via guider. If that's right, how does that sit with pulse-guiding vias EQMOD?ASCOM?
Also, I'm getting my money's worth here with question-time - is it necessary to do hot-pixel removal (via maps or whatever) with calibration frames or only lights?
You don't have to use via-guider, you can just get Maxim to send the movement via-mount if that is easier for you. If you tell it to via the guider then it'll just make a dither movement the same way that it would do a guider movement. It's not different to when you run your guider calibration routine.
FPN is Fixed Pattern Noise. CMOS sensors show it a lot more than CCDs do and that is mostly due to the lower read noise of CMOS. FPN tends to show itself in lower SNR areas (the background).
