Dithering (How Many Pixels)?

Just wondering on what people consider to be the minimum number of pixels to dither a frame by? I’m currently using 5 px - but maybe I should use more?

1 Like

What is your imaging scale? Think about it. You are trying to move the camera sufficiently so that pixel defects/errors do not autocorrelate. 5 pixels may be too much unless the detail you are aiming for is around that size.

Roberto

Hi Roberto,

The image scale is 1.2 - so maybe I should use a lower amount of dither (3)?

Yes, up to 3 pixels should be more than enough. Good luck! :sunglasses:

Roberto

Many Thanks…now waiting for the wx :slight_smile:

My image scale is 0.84"/px and guider scale is 1.19"/px and I set the dithering to 5 pixels without issues.

Peter

Ok, then you move the imager for 7 pixels.
What I dont know which is a usefull amount of shifting, 7 pixels or 10 or 20?
Ed

Dithering needed to avoid pattern noise and bad pixel (hot pixel, cold pixel, different photon capture capacity) and gain SNR and uniformity … so never go under 3 pixel usually its a good thing. But talking about pixel mean to understand if you are doing sequence guided or not. If you do guided dithering pixel is referred to guide sensor, so you must understand difference between main camera sensor and guide sensor, you must rescale for example 3 pixel in guide camera how is on main camera, just do simple proportional equation. If you do undithering guiding pixel are calculated directly on main camera resolution, so no calculation to do.

Nyquist theorem for sampling say about 2 times the resolution you have , in a 2-dimensional environment like an image this value become 3.83. So one rule must be to stay not under 3 or 4 pixel in dithering. But if you have a vertical band of 15 pixel where the signal is different from the adiacent 15 pixel you must use a dithering bigger than 15 … so really difficult to really do the right thing (but this really a worst case).

If you have a bayer matrix, 2x2 pixel compose the data so multiply by 2 the dithering amount you choose.

Final remember, more bigger is the dithering in pixel more time needed to the sytem to come back to quiscence.

All the best
LO

4 Likes

Welcome Lewis,

depends in your goals. To remove bad pixel 3/4 pixel are good.

All the best Leo

Nothing wrong with dithering “too much” and that’s a good thing.

Peter

Hello Leo
Many thanks for this information, now is all clear :grinning:
A little addon information wich was helpfull for me concerning the 2-dimensional Nyquist criterion…
Link: [http://www.stanmooreastro.com/pixel_size.htm]

Ed

Hi Leo,
I drive on a 60mm f/4 240mm with asi178mm. 1 pixel corresponds to about 2.05".
Among the various configurations I use I also have a 6d on a 135mm, to have 3 dithering pixels on the 6d I have to go to about 14 dithering pixels on the guide, but the maximum parameter that can be set is 9.9.
If what I calculated is correct I am asking you, if it is possible when you have time, to increase the maximum value of the settable dithering.
Thank you
Marzio

Sorry i will increase the maximum.

All the best
LO

1 Like

Thank you so much Leo
CS
Marzio