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

Is Sony Really Alpha?

  • Please log in to reply
1805 replies to this topic

#301 TimN

TimN

    Gemini

  • *****
  • Posts: 3,177
  • Joined: 20 Apr 2008
  • Loc: Ontario, Canada

Posted 21 April 2015 - 10:22 AM

I don't have a Sony camera but my Nikon D5100 has a Sony IMX071 sensor. I have given up darks awhile ago - as long as I used high or aggressive dithering. Like pedxing above, I could not tell any difference after lots of testing. In fact the darks sometimes made it worse if my temperature match was off too much. 



#302 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 21 April 2015 - 11:09 AM

Thanks Kolenka/Curtis/Anak/Mark/Pedxing/Tim; your work/comments concur my assessment [i.e., folks getting away from dark-calibration...] One thing I would like to mention is that I image at ~2300mm FL refractor and see minor NO-dark-calibration flaws like daylight, and that’s something that keeps forcing me to keep doing some kind of dark calibration. NOT having mastered manual dark calibration in its peak days, I keep resorting to LENR.

 

 

I am trying to experiment FEW (3-4) manual darks to get away from LENR. Your help in sorting out my technical challenges will be appreciated, most of which has to do with my inexperience with manual dark-calibration. Will post soon what I am dealing with in this regard. Thanks


Edited by mmalik, 21 April 2015 - 11:17 AM.


#303 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 21 April 2015 - 01:02 PM

Will post soon what I am dealing with in this regard. Thanks

 

I have posted my dark calibration problem here...



#304 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 21 April 2015 - 01:04 PM

One thing I notice about Mike's uncalibrated stack is that the blue and read hot pixels appear in pairs and the green hot pixels appear in pairs or triples.  I notice the same thing on my own A7S but I haven't seen it mentioned elsewhere.

 

Since I can't think of any physical reason why hot pixels should cluster in this way, I am forced to assume it is the result of some pre-processing algorithm performed on the image before it is written out as a raw file.  Does anyone have any insight into this?

 

Mark


Edited by sharkmelley, 21 April 2015 - 01:06 PM.


#305 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 22 April 2015 - 04:32 AM

One thing I notice about Mike's uncalibrated stack is that the blue and read hot pixels appear in pairs and the green hot pixels appear in pairs or triples.  I notice the same thing on my own A7S but I haven't seen it mentioned elsewhere.

 

Since I can't think of any physical reason why hot pixels should cluster in this way, I am forced to assume it is the result of some pre-processing algorithm performed on the image before it is written out as a raw file.  Does anyone have any insight into this?

 

I would go a step further; I see, for the most part, green mottle more discrete or clustered in doublets/triplets, while blue and red mottle is much more diffuse (may be more than doublets/triplets [unless red and blue are just getting optically bloated more than green?])

 

 

Note: Following is an a7S example.

Attached Thumbnails

  • Mottle.jpg

Edited by mmalik, 22 April 2015 - 09:30 AM.


#306 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 22 April 2015 - 04:43 AM

I feel such color mottle may have more to do with underlying sensor architecture, pixel size, etc.; here are some differences in look and feel and color of such mottle in a7S and a7II. Note: a7S mottle is much more colorful than quite bland/grayish looking mottle of a7II. (Will try posting high resolutions soon) Regards
 
a7S:
 
post-205769-0-61086100-1423387896_thumb.


a7II:

 

post-205769-0-14615800-1423387679.jpg


Edited by mmalik, 22 April 2015 - 08:51 AM.


#307 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10,450
  • Joined: 03 Jul 2004

Posted 22 April 2015 - 07:25 AM

Mike,

 

Hot pixel is supposed to be an individual pixel that has elevated charge.  Since these SONY sensors are standard Bayer RGGB sensors,

the hot pixel should only come out with just one color.  A7II's are more doctored (grey color only.)

Anyway you know SONY RAW is never raw.

 

Your other observation on some hot pixels are "clustered".  I suspect the problem is related to how SONY is sacrificing some pixels just for contrast-difference AF thus these pixels are replaced by the normal same-color pixel near-by.

When the "normal, ordinary pixel" is a hot pixel, its high value gets replicated to the contrast-detect special pixel(s).

 

AFAIK, there is known method to defeat SONY cooking the sensor raw to really get to the bottom of it.

I suggest you share the RAW files in a Google Drive so people who have raw analysis software (I use Rawdigger) to peek into the images.

 

Clear Skies!

 

ccs_hello



#308 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 22 April 2015 - 08:56 AM

the hot pixel should only come out with just one color

 

Agree with you and that's what I implied in my response above. [FYI: Mark had presumed them to be hot pixels which they didn't seem like to me as well]

 

 

Note: If any, one in red square seems like a hot pixel? Following is an a7S example.

Attached Thumbnails

  • HotPixel.jpg

Edited by mmalik, 22 April 2015 - 09:28 AM.


#309 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 22 April 2015 - 09:03 AM

I suggest you share the RAW files in a Google Drive so people who have raw analysis software (I use Rawdigger) to peek into the images.

 

If you missed the link above..., here... are the a7S RAWs (4) for the color mottle example under discussion. Regards

 

 

Note: These are 4 standard, non-LENR RAWs of an actual astronomical target [ISO2000, 2min each]


Edited by mmalik, 22 April 2015 - 09:09 AM.


#310 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 22 April 2015 - 09:17 AM

Here is what an a7II color (RGB) mottle looks like upon standard (non-LENR) exposure; RAWs and JPGs of both standard and LENR types are here...

 

 

Note: Following is a highly zoomed (400x) JPG section; greyish look was from a distance, up close one can discern RGB pattern. Following is an a7II example.

Attached Thumbnails

  • a7II_Mottle.JPG

Edited by mmalik, 22 April 2015 - 09:29 AM.


#311 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 22 April 2015 - 12:54 PM

Looking at pixel values in the Raw files leaves me in no doubt: the Sony A7S is using a HPS (hot pixel suppression) algorithm.  This explains the pairs of pixels with identical values - just like Nikon cameras have.  The algorithm is probably similar.  I'll dig out some links later - I seem to remember quite a few threads on DPReview on the various Nikon HPS algorithms and the resulting pixel pairs.

 

Mark


Edited by sharkmelley, 22 April 2015 - 12:57 PM.


#312 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 22 April 2015 - 03:30 PM

A few Nikon HPS related posts in DPReview:

http://www.dpreview....s/post/34575627

http://www.dpreview....s/post/37071846

http://www.dpreview....s/post/50875417

 

Mark



#313 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 23 April 2015 - 04:43 PM

The problem is worse than I feared - it affects any pixels that are brighter than their surroundings and any pixels that are dimmer than their surroundings.  Everywhere I look in the image, the brightest Red pixel value in any small area always has an adjacent twin and the dimmest Red pixel in any small area also has an adjacent twin.   The same happens in the Green and Blue channels.

 

I fear that the effect of this will be similar to the infamous Nikon star-eater algorithm, erasing or severely attenuating undersampled stars.  I'll start a new thread to discuss this issue. 

 

Mark


Edited by sharkmelley, 23 April 2015 - 04:45 PM.


#314 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 23 April 2015 - 05:45 PM

I have some Canon images to compare against (Crescent nebula); will post some comparative analysis soon. Regards



#315 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 24 April 2015 - 09:35 AM

Some 60Da, 6D, and a7S star comparisons of Single RAW Exposures:

 

 

Note: 60Da and 6D are 100% zoom of RAW in Canon Digital Photo Professional; a7S is 6D equivalent zoom of RAW in PixInsight with STF

 

 

Note: Besides acquisition differences (ISO/Duration/Filter, etc.), same scope/hardware was used for all the images

 

 

Attached Thumbnails

  • 60Da_ISO400_10min_LENR_LPS-P2.JPG
  • 6D_ISO1600_5min_LENR_LPS-V4.JPG
  • a7S_ISO2000_3min_LENR-LPS-D1.JPG


#316 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 24 April 2015 - 09:37 AM

Here is a side-by-side comparison of integrations:

 

 

Note: Besides acquisition differences (ISO/Duration/Filter, etc.), same scope/hardware was used for the images

Attached Thumbnails

  • SideBySide2.JPG

Edited by mmalik, 24 April 2015 - 09:53 AM.


#317 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 24 April 2015 - 09:52 AM

A high resolution of a7S ISO2000/42mn integration here... if folks would like to compare stars against their own crescents. Regards


Edited by mmalik, 24 April 2015 - 09:56 AM.


#318 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 24 April 2015 - 01:06 PM

That's a beautiful image of the Crescent!

 

It's difficult to tell for sure but I do get the impression that stars on the A7S version are dimmer.  I haven't looked at a LENR frame yet to see if it shows the tell tale pixel pairing.  I hope to do some tests over the weekend - some with an artificial star to get a better idea of what is going on and under what circumstances.  My flat frames (taken with a sub second exposure time) do not show the issue.  So I suspect that, just like Nikon, this algorithm is only switched on at longer exposures.

 

Mark



#319 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 24 April 2015 - 06:44 PM

I've started a new thread for the discussion on the Sony A7S hot pixel suppression algorithm:

http://www.cloudynig...ater-algorithm/

 

Mark



#320 bwallan

bwallan

    Viking 1

  • -----
  • Posts: 563
  • Joined: 08 Jul 2010
  • Loc: Alberta, Canada

Posted 24 April 2015 - 11:59 PM

Here is a side-by-side comparison of integrations:

 

 

Note: Besides acquisition differences (ISO/Duration/Filter, etc.), same scope/hardware was used for the images

The A7S appears to had some significant noise reduction?  The nebulosity looks plasticy?

 

bwa



#321 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 25 April 2015 - 06:18 AM

The A7S appears to had some significant noise reduction?  The nebulosity looks plasticy?

 
Check out high my ISO2000/105min high resolution integration of Pelican Nebula here... and here...; this was my first decent integration with a7S. Acquisition details here.... Your thoughts on this? Regards


Edited by mmalik, 25 April 2015 - 06:24 AM.


#322 bwallan

bwallan

    Viking 1

  • -----
  • Posts: 563
  • Joined: 08 Jul 2010
  • Loc: Alberta, Canada

Posted 25 April 2015 - 01:20 PM

A high resolution of a7S ISO2000/42mn integration here... if folks would like to compare stars against their own crescents. Regards

I certainly don't see a problem with this image!

 

bwa



#323 bwallan

bwallan

    Viking 1

  • -----
  • Posts: 563
  • Joined: 08 Jul 2010
  • Loc: Alberta, Canada

Posted 25 April 2015 - 01:31 PM

 

The A7S appears to had some significant noise reduction?  The nebulosity looks plasticy?

 
Check out high my ISO2000/105min high resolution integration of Pelican Nebula here... and here...; this was my first decent integration with a7S. Acquisition details here.... Your thoughts on this? Regards

 

Why do you use Long Exposure Noise Reduction (LENR)?  I find this in-camera noise reduction (or any in-camera noise reduction) eliminates nebulosity, faint stars and other worthwhile details.  It can also give a plasticy look to images.  I much prefer to do any necessary noise reduction in postprocessing.

 

I don't, however, see any indication of the hot pixel doubling being discussed?  Wondering if this problem is the result of certain camera settings (or firmware) as opposed to an inherent camera glitch?  I've gone back and pixel peeped a number of my images and can't find any hint of the problem.

 

bwa



#324 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 25 April 2015 - 02:37 PM

Bwa, if anyone took a standard dark (without LENR) on any alpha (I tested a7S and a7II), one will see the following pattern and folks are starting to call it 'pixel doubling' [I am not sure what to call it just yet or if it is even a problem?]. Go ahead and do it on your own alpha and you'll see it.

 

 

Note: If one were just stacking lights (the modern trend) this pattern will be found in the stacked image as well.

 

 

With that said, I am not sure where it leaves things?

 

 

Plan: This is what I am going to do; I am going to take similar darks with 6D and 7DII and post results for comparison sake and take the discussion from there. At this point, I see it no more than inherent characteristics of a given sensor and its design (pixel size, etc.). Regards

 

 

Note: Following are about 600% zoom of in-camera JPGs.

Attached Thumbnails

  • a7s_a7II-DarkMottle.jpg

Edited by mmalik, 25 April 2015 - 04:25 PM.


#325 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • topic starter
  • Posts: 10,148
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 25 April 2015 - 05:21 PM

Following are standard (without LENR) darks from 6D and 7DII; I am seeing somewhat equivalent color mottle in Canon especially in 7DII.

 

 

My take is this so called 'pixel doubling' theory is a non-issue; it is no more than inherent characteristics of a given sensor and its design. Regards

 

 

7DII Dark RAW/JPG here...

6D Dark RAW/JPG here...

 

 

Note: Following are about 600% zoom of in-camera JPGs.

 

Attached Thumbnails

  • 6D_7DII-DarkMottle.jpg



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