after a relatively long absence from AP I used one of the best nights possible to get back into it.
The RAW´s I got from my DSLR are beautiful and low noise. I took almost 30 images and stacked about 20 including darks, flats, dark flats, offsets.
The staking went without error, I followed DSS´s recommended settings but changed from Median Combination to Average.
The result is disconcerting: the stacked image looks like someone racked through it. Or as if strong winds are blowing through the (Lagoon) nebula.
I have attached a single exposure as well as the stacked image. Both have been significantly reduced in size/quality for web posting purposes. But it is imediately evident that the single exposure is much smoother and shows none of the artefacts that (I guess) were intoduced through stacking.
The "racking"-issue is not unknown to me and is more or less present in all my stacked images. So far, however, the stacking gain outweighed it. I this case I might as well not stack at all and be happy with what I got out of the box. I would like to be better with this. Does somebody here know what the culprit might be?
I would really appreciate if somebody could point me in the right direction.
Details below:
Camera: Canon 450D
Filter: Astronomik CLS
Scope: Celestron OmniXLT150 (150mm Newton)
Exposure Time: 210s
Number of Exposures in stack: 21
Percentage of Exposures stacked: 80%
Hi Max, that is noise and pixel defects from the camera that the darks haven't completely eliminated. The solution is to "dither" while taking the image, moving the telescope very slightly as often as possible. I think this post explains dithering better. http://www.iceinspace.com.au/forum/s...05&postcount=2
thank you for your response. It seems plausible but I wonder if it couldn't be caused by something else.
I read a lot about astrophotography but dithering was only ever mentioned in conjuction with retaining star colours in exposures long enough to saturate stars. These streaks aren't visible in any of the single exposures.
Couln't be a slight misalignment in the stacking process, maybe introduced by slight field rotation due to a slightly misaligned mount?
Please correct me if I am headed in the wrong direction.
You're partly correct in that the streaks are caused by star drift and correction in the stacking process, but the prime cause is noise. If alignment was perfect and there was no flex etc there wouldn't be streaking, but the stack would still be noisier than if it was dithered.
thanks for your patience. I guess I will have figure out how to dither. And how to automate it. Any chance that PhD can do that for you with a plug of some sort?
Max. Dithering with DSLRs is almost compulsory for a number of reasons, other than sub-pixel sampling. As Kevin mentioned masking calibration errors, evident in the attached images.
Thinking of consumer CMOS sensors as a whole, fixed pattern noise, banding, dead and hot pixels and so on. Dithering spreads the image over a greater sampling of good pixels, as well as improving flat fielding.
DSLR RAW data, particularly darks and lights is not particularly linear and calibration is not always accurate. Dithering evens all this out. If you have Berry and Burnell's Handbook... their advice is to dither by no less than 12 pixels. I find within reason the more the better 15+. Some software might limit this, so I guess work with what you have.
Dithering only works if it is done between frames - guiding software generally has no idea when frames start and end, so it has to be told when to dither by the acquisition software. For example phd will implement dither when told to by Nebulosity - don't know about other software.
The dithering referred to here is performed during image acquisition, not during processing. It involves moving the mount my a small amount (eg 12-15 pixels as per Rowland's advice) between each sub frame. This means that when you register the light frames, the signal is aligned but the noise is not.
Right, thanks Richard. I'm yet to do software-based image acquisition, hence my naievity (my DSLR cannot be computer controlled, alas).
But I'll soon have my first CCD operational (and Orion SS G3 Mono) -- Orion has some packaged acquisition software with it which I will try out first, so I'll seen then.
If you are not guiding, it's easy enough but tedious to dither using your hand controller buttons. My hand controller is set to the "lowest, high speed response," if that makes sense. I find this shifts the RA and DEC axii, just the right amount.
I've posted this before to give an idea of what I mean - not the only solution by any means, but easy enough. This pattern keeps the object within the FOV as you progressively dither between images. Every second or third image will probably do.
EDIT: Using the best quality sub as the stacking reference image may alleviate the problem as well. At the end of the day, temperature matching darks will improve things substantially. In the early days I would wait for a night of similar conditions and shoot a new bunch of darks.
I didn't use this feature, yet. Astrojantools has this function with PHD. PHD has a server and connect with Astrojantools.
I am liking very much of Astrojantools to control my DSLR - Canon T3. But it is very important in the configurantion tab let Astrojontools measure the time of transference and processing, that depends of USB speed and computer used.
In my set it is 9 - 10 seconds. To use automatic and programable sequence of photos this time is very important to don't cause interuption of communication with the camera.
as others have suggested, dither will definitely help, but I don't think that's the issue of your artefact.
I've done a fair bit of stacking in DSS without dithering but haven't seen artefacts like yours. its likely some setting in DSS.
Does the same thing happen with Median combine?
try a few different settings and methods and if you're not getting these artefacts in one or more of those methods, then its the stacking settings.
you can stack without dither but noise removal won't be effective. But doesn't look like you've dark subtracted anyway, so shouldn't be an issue.
are you able to upload a few raw subs in dropbox or other so we can try and stack as well?
btw, BackyardEOS 3.0.3 and above can dither with phd.
That's my personal preference for imaging with a dslr.
But doesn't look like you've dark subtracted anyway, so shouldn't be an issue.
I'm pretty sure he did, but not effectively. This is typical DSLR teething troubles.
I agree try other combines, but reference to a good sub and maybe retake the darks, plus some - it's a nice image and well worth the effort. Dithering is a bonus.
I've done a fair bit of stacking in DSS without dithering but haven't seen artefacts like yours. its likely some setting in DSS.
Does the same thing happen with Median combine?
try a few different settings and methods and if you're not getting these artefacts in one or more of those methods, then its the stacking settings.
I have seen this before a lot in my own stacks. It's not a setting in DSS. I've sat there for ages and tried all the stack settings, average, median, maximum, Kappa Sigma clipping etc etc with the same results. It's noise.
I discovered by accident that stacking two different sessions with slightly different alignment helped. What I accidentally did was dither the two sessions. If I did multiple sessions with the same alignment, the noise just got worse.
Of course the best cure for this is to have no noise in the first place. Otherwise dithering helps.
I discovered by accident that stacking two different sessions with slightly different alignment helped. What I accidentally did was dither the two sessions. If I did multiple sessions with the same alignment, the noise just got worse.
I think if you throw all the feedback together and pick the eyes out of it, you get a pretty good strategy combining a number of key processes, from acquisition to preprocessing. DSLRs are great providing you understand their limitations and how to manage those best. It's never just one thing? Often carrying one error forward until you have a processing nightmare.
I have seen this before a lot in my own stacks. It's not a setting in DSS. I've sat there for ages and tried all the stack settings, average, median, maximum, Kappa Sigma clipping etc etc with the same results. It's noise.
I discovered by accident that stacking two different sessions with slightly different alignment helped. What I accidentally did was dither the two sessions. If I did multiple sessions with the same alignment, the noise just got worse.
Of course the best cure for this is to have no noise in the first place. Otherwise dithering helps.
My experience has been different.
I started off with my 1000d and haven't had the chance to look at dither with BEOS and phd, so all my dslr images are stacked without dither.
of course, noise removal isn't effective, but I don't see any of the artefacts mentioned here when I stack 15 or more subs.
here is a stack of 16x5min subs, no processing, just stacked. no dither either. I've blinked through all 16, and they barely move as I use an OAG.
taken with an unmodded stock 1000d. https://www.dropbox.com/s/8u2uaw6m9r6y697/Trifid.png
Maxim does a very good job at stacking as well and if Max can post say 10 raw subs, I think its worth a stack attempt.
I'm also not sure what Max's mount is and how accurate the tracking was. if it wasn't too accurate, a bit of field rotation or drift does offset the series by a few pixels.
I'm fairly certain that the artefacts can be eliminated, depending on the raw subs of course.
if its noise, well he could dither next time, but if not, why waste that data?
I'm pretty sure he did, but not effectively. This is typical DSLR teething troubles.
you're right, he has used calibration frames.
it could be an issue with calibration as well. worth trying a stack without darks, but just flats, and then try darks and flats, and finally try darks, flats, bias.
if you have an issue with any of the calibration frames, you could identify that by stacking in stages.
