Dithering

[adamsp123]

What is dithering whilst imaging and how do you go about it?

[MartinB]

Between each exposure a small shift is made to the guide star position either by a mount movement or by the guider. The result is a small random shift between all the subs. This reduces noise and hot pixels. In Maxim 5 the dithering option is found in the sequence setting window. You also need to set a suitable guider delay. Maxim lets you choose either a time delay or will start the next exposure when the guide star has a low enough tracking error e.g.0.2 pixels (this is a better option). I normally use "dither via guider".

[syd_malicious]

This reduces noise and hot pixels.

I agree with most of what you said but it is not the dithering that reduces the hot pixels. You have to have sufficient subs so that your combine method can statistically weed out the hot pixels. You, therefore, have to use the correct combine algorithm too.

Now the longer answer

No detector is perfect. Most (all??) will have some pixels that are hot - ie show up white in the exposure. If you took subs with the CCD in the same place then to your processing program it may be a star or may be a hot pixel, it doesn't know.

What you want is a method of allowing the statistics of stacking to differentiate.

So if you move the CCD a bit between each sub then the target falls on a slightly area each time. This means the subject is in a slightly different place of the shot each time BUT the hot pixels stay in the same place. This allows the stacking process to statistically drop the hot pixels. There are a number of different methods of stacking, some are designed to reduce outlier pixels (median, SD Mask) but some will not (sum) so the method of stacking is critical too.

And that is just the first part of your question. As to how then it depends on the capture software you use.

Paul

[adamsp123]

Thanks for the replies, I am using PHD for guiding which AFAIK doesn't allow for dithering and stack in DSS using the Kappa Sigma process with hot pixel detection set to remove.

I use DSS for both my Canon OSC and Atik mono CCD.

[jgs001]

If you combine PHD with Nubulosity, you can enable dithering, by using the server/client communications that Craig Stark has put into the two programs. I've not worked out how to do this as yet, but it should work.

[MartinB]

I agree with most of what you said but it is not the dithering that reduces the hot pixels. You have to have sufficient subs so that your combine method can statistically weed out the hot pixels. You, therefore, have to use the correct combine algorithm too.

Paul

Sorry, I'd assumed that was obvious.

[martin_h]

If you combine PHD with Nubulosity, you can enable dithering, by using the server/client communications that Craig Stark has put into the two programs. I've not worked out how to do this as yet, but it should work.

What you need to do, is drop down the "tools" menu in PHD, and click "Enable Server" then in nebulosity drop down the "view" menu and click PHD Link. A small window will open, and there you can enable dithering between images..

Essentially, Nebulosity sends a command to PHD, and in between exposures, PHD will move the guide star in a random direction a specified amount of pixels.