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

AA filter, spatial filter and star colours

  • Please log in to reply
164 replies to this topic

#151 SteveInNZ

SteveInNZ

    Viking 1

  • *****
  • Posts: 941
  • Joined: 07 Oct 2007
  • Loc: Auckland, New Zealand

Posted 24 December 2019 - 04:06 PM

I guess those of us who use Canons should just throw them all in the trash then.

If it's any consolation, I'll be (metaphorically) throwing my A7s in the trash to fund an EOS-R. The pros vs cons just come out that way for me.

 

Steve.


  • SandyHouTex likes this

#152 Michael Covington

Michael Covington

    Author

  • *****
  • Posts: 6,868
  • Joined: 13 May 2014
  • Loc: Athens, Georgia, USA

Posted 24 December 2019 - 04:40 PM

That reminds me of what I was about to say -- Slightly obsolete cameras can be BARGAINS for beginners and experimenters.  They still work as well as they ever did. 


  • fmeschia and 2ghouls like this

#153 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,626
  • Joined: 20 Mar 2013

Posted 24 December 2019 - 10:05 PM

Since it's a known fact that Canon cameras can have equally low read noise at ISO 1600 or 3200 and higher, they should be capable of producing nice pictures at these high ISO.  Where they can't compete with Sony-inside cameras is at lower ISO, but deep-sky astrophotography is the kind that doesn't demand low ISO.

 

One thing that I don't have good answers is the Bayer filter transmission curves.  Based on some older curves I had seen in the past, Canon cameras have broader curves than other cameras.  One theory I heard on internet is that Canon is using this trick to use broad filter transmission to compensate the lower QE of its sensors.  In a superficial way, this doesn't matter.  If you get 100 photo-electrons in a G pixel, it's 100 photo-electrons.  It doesn't matter whether these 100 electrons are produced because the QE is higher or because the filter passband is broader.  However, broad RGB curves mean lower color saturation to start with.  On color images produced with broader RGB filters, you will need more saturation boost in post processing, and this amplifies noise.  For two images that have the same total amount of photo-electrons, the one that needs more saturation boost will appear noisier.  This matches exactly what I experienced on my 5D2 and D800.  My D800 images require much less saturation boost to reach the same color as my 5D2 images with strong saturation boost.  The end result is that the D800 images are much cleaner.  I think this is a subtlety that's never seriously investigated/discussed by astrophotographers, and I am not sure whether current-generation Canon still have this kind of characteristics.  Someone with a good spectrometer can try to measure the filter transmission curves of some Canon and non-Canon DSLRs to further study this.


  • SandyHouTex, sharkmelley and Michael Covington like this

#154 Michael Covington

Michael Covington

    Author

  • *****
  • Posts: 6,868
  • Joined: 13 May 2014
  • Loc: Athens, Georgia, USA

Posted 24 December 2019 - 10:17 PM

Yes, my Canon 60Da is a fine camera at ISO 800, and it didn't stop working when these Sony sensors hit the market :)



#155 sharkmelley

sharkmelley

    Aurora

  • *****
  • topic starter
  • Posts: 4,694
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 25 December 2019 - 02:27 AM

However, broad RGB curves mean lower color saturation to start with.  On color images produced with broader RGB filters, you will need more saturation boost in post processing, and this amplifies noise.  For two images that have the same total amount of photo-electrons, the one that needs more saturation boost will appear noisier.  This matches exactly what I experienced on my 5D2 and D800.  My D800 images require much less saturation boost to reach the same color as my 5D2 images with strong saturation boost.  The end result is that the D800 images are much cleaner.  I think this is a subtlety that's never seriously investigated/discussed by astrophotographers, and I am not sure whether current-generation Canon still have this kind of characteristics.  

You are correct about this and it's not something that is generally known in the astro-community.  For instance the difference is quite extraordinary between the mirrorless Nikon Z6 and the competing Canon EOS R:

 

Go to the "Color Response" tab in both cases and look at the color matrix.  The coefficients in the Canon matrix are much larger which means more colour noise is introduced during the "raw development" of a terrestrial colour image.  Astro-imagers will typically implicitly take account of the matrix differences using increased colour saturation for Canon, which again increases colour noise.

 

Mark


Edited by sharkmelley, 25 December 2019 - 02:35 AM.

  • SandyHouTex, agavephoto and whwang like this

#156 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,626
  • Joined: 20 Mar 2013

Posted 25 December 2019 - 04:13 AM

Interesting. Thank you.  I am not aware about this matrix of DxOMark.  Another proof that DxOMark is useful if one knows what to look at.



#157 netwolf

netwolf

    Mariner 2

  • *****
  • Posts: 280
  • Joined: 03 Sep 2005
  • Loc: NSW, Australia

Posted 13 July 2020 - 12:16 PM

Just wondering if the Nikon Z50 has this spatial filtering? 



#158 LCWASTRO

LCWASTRO

    Viking 1

  • *****
  • Posts: 766
  • Joined: 21 May 2020
  • Loc: Columbus, Ohio United States

Posted 14 September 2020 - 07:36 PM

yeah im considering the z50 would love to know



#159 sharkmelley

sharkmelley

    Aurora

  • *****
  • topic starter
  • Posts: 4,694
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 15 September 2020 - 01:13 AM

Just wondering if the Nikon Z50 has this spatial filtering? 

As far as we know, Nikon uses spatial filtering on every camera, so we can be pretty confident it is being used on the Z50.  The question is which variation of the algorithm is being used.

 

All I need is a long exposure dark (with all noise reduction switched off) and I can analyse it.

 

Mark



#160 LCWASTRO

LCWASTRO

    Viking 1

  • *****
  • Posts: 766
  • Joined: 21 May 2020
  • Loc: Columbus, Ohio United States

Posted 15 September 2020 - 08:32 AM

sorry i meant to say M50 as in the cannon



#161 sharkmelley

sharkmelley

    Aurora

  • *****
  • topic starter
  • Posts: 4,694
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 15 September 2020 - 11:42 AM

sorry i meant to say M50 as in the cannon

In general, Canon cameras don't apply any hidden raw data filtering. 

 

On the other hand there is evidence that there is some kind of noise reduction or raw data filtering being applied to the Canon EOS R5/R6.  The evidence is seen in the Fourier transforms but any I have not been given any files to perform my own analysis.

 

The Canon EOS M50 also shows evidence of some kind of data filtering:

https://www.dpreview...s/post/60976788

Again, I have not been given any files to perform my own analysis so I don't know what, if any, effect this has on stars.

 

Mark



#162 cabfl

cabfl

    Explorer 1

  • -----
  • Posts: 68
  • Joined: 21 Oct 2018

Posted 16 November 2020 - 02:31 PM

The Sony A7III has an option in the menu to enable or disable the long exposure noise reduction filter. It seems to not completely disable.

 

https://helpguide.so...0001629724.html

 

 

When you set the shutter speed to 1 second(s) or longer (long exposure shooting), noise reduction is turned on for the duration that the shutter is open. With the function turned on, the grainy noise typical of long exposures is reduced.

On:
Activates noise reduction for the same duration that the shutter is open. When noise reduction is in progress, a message appears and you cannot take another picture. Select this to prioritize the image quality.
Off:
Does not activate noise reduction. Select this to prioritize the timing of shooting.



#163 erictheastrojunkie

erictheastrojunkie

    Apollo

  • *****
  • Posts: 1,027
  • Joined: 15 Jul 2016
  • Loc: Salt Lake City

Posted 16 November 2020 - 02:42 PM

The Sony A7III has an option in the menu to enable or disable the long exposure noise reduction filter. It seems to not completely disable.

 

https://helpguide.so...0001629724.html

Ignore anything Sony says with regards to their noise reduction and spatial filtering, they aren't telling you the whole story. Listen and read to what Mark says/has done with his testing, it's exposes the reality. 



#164 sharkmelley

sharkmelley

    Aurora

  • *****
  • topic starter
  • Posts: 4,694
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 16 November 2020 - 04:45 PM

The Sony A7III has an option in the menu to enable or disable the long exposure noise reduction filter. It seems to not completely disable.

 

https://helpguide.so...0001629724.html

Long exposure noise reduction (LENR) is something unrelated to the raw filtering ("star eater") discussed in this thread.  LENR is a kind of dark subtraction.  For instance if you take a 20sec exposure with LENR switched on, the camera takes a 20sec dark after the exposure ends and subtracts it in order to cancel out the fixed pattern noise of the long exposure.  It means that a 20sec exposure then takes 40sec in total.  The LENR can be enabled or completely disabled in the camera menu.

 

Unfortunately there is no way to disable the (undocumented) raw data filtering that Sony applies to all A7III exposures of 4sec and longer.

 

Mark


Edited by sharkmelley, 17 November 2020 - 05:40 AM.

  • Michael Covington likes this

#165 sharkmelley

sharkmelley

    Aurora

  • *****
  • topic starter
  • Posts: 4,694
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 30 December 2020 - 04:08 AM

An analysis of the Sony A7SIII spatial filtering can be found here:

https://www.cloudyni...pha/?p=10766836

 

Mark


Edited by sharkmelley, 30 December 2020 - 04:08 AM.



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