Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

ASI533mc Pro users do you dither/drizzle

  • Please log in to reply
16 replies to this topic

#1 WhitenerJ

WhitenerJ

    Viking 1

  • *****
  • topic starter
  • Posts: 704
  • Joined: 06 Feb 2012
  • Loc: Waxahachie, TX

Posted 04 March 2024 - 01:40 PM

I’m getting a 533 soon and was curious if it needs to be dithered to avoid walking noise? I’ll be using it with a RedCat 51. If you do dither do you also drizzle?

Thanks guys.

#2 unimatrix0

unimatrix0

    Fly Me to the Moon

  • *****
  • Posts: 6,845
  • Joined: 03 Jan 2021

Posted 04 March 2024 - 01:45 PM

Dither/drizzle achives 2 different things. 

 

Dither to get rid of walking noise

 

Drizzle to upscale an image, that is otherwise small

 

And yes you gonna need to dither anyway.  It's in the AA app, under the guide options on the bottom-  how many times (every 1  2 or 3rd subs and how many pixels) .  

I dither everything around 5 minutes for each session. So if I'm taking 1 minute subs, then every 5th dithered 2 pixel. If 5 minute subs, then I dither every sub.  I found 2 pixel work just fine, larger amount just makes my mount settle a lot longer. I have no walking noise with 2 pixel dithering and 5 minutes interval. 

 

Drizzling is optional, and done in the post processing software. With the 533, it's a good idea if you ask me, it's 3008x3008 pixel image, so if I select drizzle (2x)  integration in Pixinsight, then I get a 6016x6016, but I usually cropping anyway, so it gets near an APS-C size at the end (6000x4000). 

 

You can also not drizzle, they are still fine, easier/faster to process.  For galaxies that are in the center of the image, I don't do drizzle, size is fine at 8megapixel. 


Edited by unimatrix0, 04 March 2024 - 01:50 PM.


#3 idclimber

idclimber

    Fly Me to the Moon

  • *****
  • Posts: 6,804
  • Joined: 08 Apr 2016
  • Loc: McCall Idaho

Posted 04 March 2024 - 01:58 PM

There is no camera/scope that will not benefit from dithering. 

 

Drizzle depends on quality and quantity of the data. The cool part is I can skip it and add it later at 2x. Then see if there was any benefit by comparing the results. On one of my more recent projects I got better detail doing 2x drizzle and then immediately down sample using integer resample. The only thing it cost me was a few hours of time letting WBPP run again. 

 

I did not get the same benefit drizzle 1x. It made no noticeable difference. 


  • Juno18 likes this

#4 JF1960

JF1960

    Viking 1

  • *****
  • Posts: 732
  • Joined: 16 Jan 2023
  • Loc: SoCal

Posted 04 March 2024 - 01:59 PM

Under sampling means your smaller stars can look blocky.  It’s hard to see unless you zoom in.  I haven’t seen much issue with walking noise using the 533.  But I do dither every frame and apply a 2x drizzle to get nicer small stars. 2x drizzle can increase your noise.  But I don’t find that ever bad enough and noisexterminator deals with it fine.



#5 WhitenerJ

WhitenerJ

    Viking 1

  • *****
  • topic starter
  • Posts: 704
  • Joined: 06 Feb 2012
  • Loc: Waxahachie, TX

Posted 04 March 2024 - 02:04 PM

Under sampling means your smaller stars can look blocky.  It’s hard to see unless you zoom in.  I haven’t seen much issue with walking noise using the 533.  But I do dither every frame and apply a 2x drizzle to get nicer small stars. 2x drizzle can increase your noise.  But I don’t find that ever bad enough and noisexterminator deals with it fine.

Thanks for the reply, you havre the same setup. Can you share your dither settings exactly?



#6 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 34,964
  • Joined: 27 Oct 2014

Posted 04 March 2024 - 02:23 PM



Under sampling means your smaller stars can look blocky.  It’s hard to see unless you zoom in.  I haven’t seen much issue with walking noise using the 533.  But I do dither every frame and apply a 2x drizzle to get nicer small stars. 2x drizzle can increase your noise.  But I don’t find that ever bad enough and noisexterminator deals with it fine.

People worry WAY too much about undersampling.  There's an upside to it, better signal to noise ratio.

 

This image has a scale of (oh the horror!) 2.8 arc sec per pixel.  To get there with a 533 you'd need a focal length of 277mm or less.  Click on the crummy CN jpg for a good version and capture details.  Stars look blocky to you?

 

get.jpg?insecure


Edited by bobzeq25, 04 March 2024 - 02:26 PM.

  • WhitenerJ, DutchTerror and unimatrix0 like this

#7 unimatrix0

unimatrix0

    Fly Me to the Moon

  • *****
  • Posts: 6,845
  • Joined: 03 Jan 2021

Posted 04 March 2024 - 02:39 PM

The other end of the specturm.

 

Oversampled image.  Pixel scale is at 0.7 or less, can't remember what focal length my reducer was set to. 

 

qhy183C + RC6 

 

If I won't mention it's oversampled, would you even notice? 

 

gallery_355785_19233_1249473.jpg


Edited by unimatrix0, 04 March 2024 - 02:40 PM.

  • bobzeq25 likes this

#8 imtl

imtl

    Cosmos

  • *****
  • Posts: 8,604
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 04 March 2024 - 02:44 PM

The other end of the specturm.

Oversampled image. Pixel scale is at 0.7 or less, can't remember what focal length my reducer was set to.

qhy183C + RC6

If I won't mention it's oversampled, would you even notice?

gallery_355785_19233_1249473.jpg


Was your seeing worse than 2" when acquiring this data? How do you know you were oversampled at 0.7"/pixel?

#9 unimatrix0

unimatrix0

    Fly Me to the Moon

  • *****
  • Posts: 6,845
  • Joined: 03 Jan 2021

Posted 04 March 2024 - 02:51 PM

Was your seeing worse than 2" when acquiring this data? How do you know you were oversampled at 0.7"/pixel?

It's almost always poor to average seeing in CT. 

Average seeing is rare. 

 

I was wrong with the numbers. it's oversampled at 0.38 pixel resolution, not 0.7

 

Try 1300mm focal length and 2.4 micron (183) for numbers and add poor seeing

 

https://astronomy.to...ccd_suitability


Edited by unimatrix0, 04 March 2024 - 02:53 PM.


#10 imtl

imtl

    Cosmos

  • *****
  • Posts: 8,604
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 04 March 2024 - 02:52 PM

It's almost always poor to average seeing in CT.
Average seeing is rare.

I was wrong with the numbers. it's oversampled at 0.38 pixels

Try 1300mm focal length and 2.4 micron (183) for numbers and add poor seeing

https://astronomy.to...ccd_suitability


Make sense!!

#11 arbit

arbit

    Apollo

  • -----
  • Posts: 1,401
  • Joined: 19 Feb 2012

Posted 04 March 2024 - 08:24 PM

To add a couple of points.

Dithering distance is a matter of preference. Generally, somewhere between 2 and 5 pixels. The thing to take care of is that this is at the imaging camera, but dither is done at guide camera. So, the ratio of guide and imaging pixel scales is important. If you want to (say) dither 3 pixels, the imaging scale is 3as/px, and the guide cam is 6as/px, then the dither setting in PHD2 would be 1.5px. If you use NINA, it will work this out for you.

Drizzle is a matter of preference. I often use the Redcat51 / 533MC combo. Theoretically at an image scale of 3.14as/px, its way undersampled. But whether stars are blocky or not depends on the zoom level :-)

However, depends on what the DSO is. For widefield nebula, I dont really bother. If I want slightly better stars on zooming in upscaling works fine.

For a larger galaxy like M31 or M33, I prefer drizzle - it does seem to recover some detail, in the dust arms especially. Decon also works better. Same if imaging a section of a nebula where you want to focus on the gas cloud details.

 

Edit: The above is for a 2x (or if you have the processing power and subs, 3x drizzle). Pixinsight does recommend a 1x CFA drizzle for OSC cameras like 533MC for the best extraction of colour information, but I've really noticed any benefit.

Sent from my SM-S908E using Tapatalk


Edited by arbit, 04 March 2024 - 08:37 PM.


#12 JF1960

JF1960

    Viking 1

  • *****
  • Posts: 732
  • Joined: 16 Jan 2023
  • Loc: SoCal

Posted 04 March 2024 - 08:43 PM

Thanks for the reply, you havre the same setup. Can you share your dither settings exactly?

My OTA is packed up right now so I can’t check. I used this excellent thread to figure out how far to dither:

 

https://www.cloudyni...g-setting-help/



#13 AstroVagabond

AstroVagabond

    Viking 1

  • *****
  • Posts: 868
  • Joined: 28 Nov 2020
  • Loc: San Mateo, CA

Posted 05 March 2024 - 12:16 AM

I dither my ASI533MM up to every three exposures using Tom Palmer's Target Scheduler pluging as a N.I.N.A. user therefore I do not drizzle.

 

It is my understanding you may have to dither after each exposure if you want to drizzle. I may be 100% wrong as I often am in this hobby. 

 

~ Bill



#14 sharkmelley

sharkmelley

    Cosmos

  • *****
  • Posts: 7,599
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 05 March 2024 - 12:49 AM

CFA Drizzle (a.k.a Bayer Drizzle) is recommended standard technique for all one-shot-colour cameras, including DSLR/Mirrorless.  It avoids the interpolation step inherent in the debayering process and gives an improved (i.e. more random, less blocky) noise characteristic.



#15 imtl

imtl

    Cosmos

  • *****
  • Posts: 8,604
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 05 March 2024 - 01:27 AM

To add a couple of points.

Dithering distance is a matter of preference.


Dithering distance is not a matter of preference. It is determined by the typical FPN scale you have on your sensor. You need a minimum step size to negate that.

#16 phsampaio

phsampaio

    Mariner 2

  • -----
  • Posts: 230
  • Joined: 06 Aug 2020

Posted 05 March 2024 - 08:40 AM

Always dither.

Drizzle is optional, but at the same time, there's no harm in trying it depending on your setup (since you already dither...).

For DSLR and OSC you should at least do CFA Drizzle at 1x in Pixinsight. As others have noted, it skips the debayering interpolation, and with proper dithering, can recover much of the resolution and color fidelity that is lost because of the Bayer matrix.

#17 Forward Scatter

Forward Scatter

    Vanguard

  • -----
  • Posts: 2,109
  • Joined: 22 Jul 2018
  • Loc: Wandering the PNW

Posted 05 March 2024 - 10:40 AM

I also drizzle/dither on my 533 as well. Dizz 1X with my newt, 2X with my 72ED and Sammy 135mm.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics