Dithering

[rcheshire]

Here is a nice little 'beginners' article on dithering (http://www.astro.uni-bonn.de/~mischa/dithering.html) :D - the last two paragraphs sum it all up - still very interesting. So perfect polar alignment is not always a good thing if you don't have dithering software, like me.

Below is an example of what I have been told is associated with not dithering, called differential flexure. Polar alignment was particularly good on this occasion. So even DSLR users need to dither - and I thought that came with age.

[OzRob]

Something that I tried while I was using a DSLR was manual dithering. What I did was take perhaps three photos without touching anything then move the mount before taking another three images. I took at least three sets like this. I then stacked each set and then the results of each stack to produce the final image. I didn't really try it long enough to see if it was worth the effort but maybe it might work for someone.

[irwjager]

Funny - could Okiestar's problems maybe be resolved by dithering? :)

http://www.iceinspace.com.au/forum/showthread.php?t=71637

[Octane]

Yep, dithered from day dot with the DSLR and now also dither with the CCD.

It makes a world of difference. : )

H

[ericwbenson]

Wait a sec, differential flexure is a problem during guiding (you get trailed stars even if the guiding is perfect) and dithering is what you or software does between exposures, while the guider is stopped. You dither so that sensor artifacts can get rejected during image stacking. Am I missing a turn of phrase above?

EB

[rcheshire]

Dithering is also used to prevent stacking pixels one on the other which reduces SNR - according to the tutorial posted above. Differential flexure in the absence of guiding software was a little difficult to understand. And I'm not completely sure still.

[RobF]

If you animate your aligned subs (still uncal'd) in something like CCDStack or PixInsight, then zoom right in, its quite a surprise to see what's happening with dithering. At such small scales you see a lot of the noise from the camera jump back and forth with each image. By dithering its a bit like looking through dirty glasses at an object - if you smoothly keep moving your head around you still get to create a complete composite image of what you want to see without any of the splotches permanently blacking out any of the view.

Oh heck, I'm resorting to analogies again....:rolleyes:

[Octane]

Good analogy, Rob. : )

H

[rcheshire]

I like it too. Analogies are a good educational tool. I get it!

As far as dithering is concerned on an unguided setup (given that polar alignment is getting easier and better with experience, but I'm stacking pixels one on the other), I plan on using DEC with the paddle control to shift the image in the finder. Alternatively, should I use RA to make the adjustment - keeping the image in the same rotational plane. Does it matter?

I can use a DSLR program to view the image on-screen and use that for more accurate adjustments, rather than peering through the viewfinder. While I think of it, by zooming in it should be possible to make fine adjustments - say 2 pixels at a time:question:

[DavidTrap]

H,

So in Maxim, how many pixels do you tell it to dither by?

DT

[rcheshire]

I was planning on eyeballing it.

[Octane]

David,

I just chose a random number -- 5. There was a thing I read, must have been the autodither plugin for CCDSoft, which goes into great detail about how much you should move the scope between exposures. Will have to dig it out and re-read it and implement the changes.

Rowland, you want to move the scope in a random direction after each exposure. Moving the scope in both RA and DEC is the best way to do this.

The plugin documentation mentioned above, recommended the spiral dither.

H

[rcheshire]

Thanks H. That makes better sense.

[DavidTrap]

You are living life on the edge with choices like that H!

Does Maxim do a "spiral dither"? Do you choose to dither by guider or mount? I picked by guider last night.

DT

[RobF]

I think it must be guider pixels David. It probably only matters if you're doing mosaics if you bump it too much - you might wander outside where you need to be aimed. I have a settle value programmed in too to ensure everything is steady before the next sub starts. I've currently got 0.25 in there, and introduces I nice small jump between subs without too much moving about. Mind you, my guider is about 5 arcsec/pixel I think, while the QHY9 on the Newt is more like 1.1 arcs/px, so a small guider movement gives me a decent shift between shots.