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

Would you agree? A7RIII STAR EATER

  • Please log in to reply
13 replies to this topic

#1 calypsob

calypsob

    Aurora

  • *****
  • topic starter
  • Posts: 4683
  • Joined: 20 Apr 2013

Posted 22 August 2019 - 07:58 PM

I have made it a point to email Sony once a week about the ever present star eater issue in the a7riii. They never truly gave us the ability to process true un adulterated raw data. I can see star colors and cores being eaten up in long exposures but just to get a second opinion, im curious what you think. Does this cropped image appear to have been negatively impacted by star eater? Other than the poor non round star profile created by the less than perfect gm 24-105 I see faint background stars which have been desaturated and foreground higher mag stars which were for some reason able to accumulate color data. I did not buy this camera to do astro work but thought it would be a bonus to get clean milkyway shots. After seeing how spatial filters impact these images I decided to bust sonys balls about this issue. What do you think?

https://flic.kr/p/2gRx31u

https://flic.kr/p/2gRxNfh

Plus a bonus Sonys response to my star eater complaint after providing these images


Hello Wes,


Thank you for contacting Sony Digital Imaging Support Team.



I understand that you have some question about the star eater issue from the images you took using your ILCE7RM3 camera.


We are able to access the sample image you provided, the sample image of the sky is normal and star eater issue is not present. If the star eater issue is present you will not see any bright spot on that image. And also, we no longer received any star eater issue since that the new firmware update was released for our ILCE7RM2 cameras. The star eater issue happens with the ILCE7RM2 not with your ILCE7RM3 camera model.


Should you have any technical questions or concerns, please feel free to send us an email. Our support hours are Monday to Friday 10:00 AM - 9:00 PM; Saturday 10:00 AM - 7:00 PM EST; Closed on Sunday.



Sincerely,

Carl
Sony Digital Imaging Support Team


Edited by calypsob, 23 August 2019 - 03:23 AM.


#2 whwang

whwang

    Mercury-Atlas

  • *****
  • Posts: 2825
  • Joined: 20 Mar 2013

Posted 22 August 2019 - 08:10 PM

Hi Wes,

 

I think what you should really do is to stack many 3-second exposures and also stack a few longer, Bulb exposures.

By comparing these two, you can get a definite answer about whether there is some damage to stars in the longer exposures.


  • agavephoto and calypsob like this

#3 gezak22

gezak22

    Mercury-Atlas

  • *****
  • Posts: 2948
  • Joined: 15 Aug 2004
  • Loc: On far side of moon. Send help.

Posted 22 August 2019 - 10:15 PM

These data support that the A7R3 is still a star eater, and Sony are sticking their heads in the sand.



#4 whwang

whwang

    Mercury-Atlas

  • *****
  • Posts: 2825
  • Joined: 20 Mar 2013

Posted 22 August 2019 - 10:29 PM

These data support that the A7R3 is still a star eater, and Sony are sticking their heads in the sand.

 

Ouch.  That's ugly.



#5 t_image

t_image

    Gemini

  • -----
  • Posts: 3154
  • Joined: 22 Jul 2015

Posted 22 August 2019 - 11:37 PM

I have made it a point to email Sony once a week about the ever present star eater issue in the a7riii. They never truly gave us the ability to process true un adulterated raw data. I can see star colors and cores being eaten up in long exposures but just to get a second opinion, im curious what you think. Does this cropped image appear to have been negstively impacted by star eater? Other than the poor non round star profile created by the less than perfect gm 24-105 I see faint background stars which have been desaturated and foreground higher mag stars which were for some reason able to accumulate color data. I did not buy this camera to do astro work but thought it would be a bonus to get clean milkyway shots. After seeing how spatial filters impact thee images I decided to bust sonys balls about this issue. What do you think?

https://flic.kr/p/2gRx31u

https://flic.kr/p/2gRxNfh

Plus a bonus Somys response to my star eater complaint after providing these images


Hello Wes,


Thank you for contacting Sony Digital Imaging Support Team.



I understand that you have some question about the star eater issue from the images you took using your ILCE7RM3 camera.


We are able to access the sample image you provided, the sample image of the sky is normal and star eater issue is not present. If the star eater issue is present you will not see any bright spot on that image. And also, we no longer received any star eater issue since that the new firmware update was released for our ILCE7RM2 cameras. The star eater issue happens with the ILCE7RM2 not with your ILCE7RM3 camera model.


Should you have any technical questions or concerns, please feel free to send us an email. Our support hours are Monday to Friday 10:00 AM - 9:00 PM; Saturday 10:00 AM - 7:00 PM EST; Closed on Sunday.



Sincerely,

Carl
Sony Digital Imaging Support Team

 

I applaud your efforts but am confused by your presentation of your case.scratchhead2.gif

 

In the past compelling evidence was presented for offending camera models by taking images with settings known to trigger the spatial filtering,

compared with images where a known baseline reference (like with settings where the filtering was not triggered).

My experience of your links doesn't do this. I am missing something? Is it an animated GIF that isn't playing for me?????

I see a widefield, and a crop 'untitled.'

It also seems clear "Carl" doesn't understand the concepts involved, and wasn't visual persuaded by your images (or at least is parroting someone else who wasn't persuaded by your images)....

 

Seems you got some kind of audience with someone there at Sony, with an invite to continue to email.

 

If your links are your best case, I'd consider trying with a better presentation as if you were explaining it to a 5th grader....

 

There are many ways that are obvious that Sony hasn't eliminated the secondary unintended consequences of their spatial filtering algorithm,

but since you have access to the a7riii and want your point two be heard,

please work on the delivery. Take for starter's Wei Ho's suggestion......

 

I'd also advise word careful word choice. Sony and others are influencers are now using "star-eater"----but unfortunately the term is very results-oriented.

 

As Mark Shelley and Jim Kasson have made efforts to define it as a spatial filtering process,

IMO we need to help Sony admit what is going on is caused by an intentional data cooking process.

If say, Carl for example (imagine he wasn't just a parrot of scripted information posted on his computer screen),

had to address whether or not Sony in the a7rIII with the recent release was still intentionally cooking the data (process),

instead of answering about very vague or pixel-peeping dots results,

than he would have been less able to state "If the star eater issue is present you will not see any bright spot on that image."

 

Using the term "star-eater" has allowed at least Carl to frame the debate to a very specific set of results,

rather that the whole issue that

an intentional data-cooking is producing secondary unintended detrimental results in imaging stars that could otherwise with the same gear be imaged better....

 

 

I think those that want Sony to completely fix any negative effects are not being realistic.

 

Please can we just get Sony to see that there is a population of users that want Sony to give them the option

of whether Sony's latest spatial filtering data-cooking will apply to the desired images......

 

How about just saying "Hey Sony, I get you haven't been able to completely fix it.

Just give us a menu option so whatever current spatial filtering doesn't  have to be applied to our photos of the stars........"



#6 agavephoto

agavephoto

    Explorer 1

  • -----
  • Posts: 89
  • Joined: 14 Jul 2009

Posted 22 August 2019 - 11:43 PM

I applaud your efforts to tirelessly send in problem reports to the manufacturer ... but, it seems this isn't getting to a group who really understands the issue. Isn't this where you sell the camera and buy one that doesn't do this?



#7 sharkmelley

sharkmelley

    Vanguard

  • *****
  • Posts: 2302
  • Joined: 19 Feb 2013

Posted 23 August 2019 - 12:58 AM

The Sony Digital Imaging Support Team are sticking to what they've been told.  They use a narrow definition of "star-eater" (i.e. the original crude algorithm) and using that definition they are able to claim to have fixed the issue.  For anyone new to this, you can find my analysis of the "fixed" algorithm here: http://www.markshell...areater_v2.html

 

Possibly the way to approach this is not to complain about "star-eater" because they have been told it is fixed.  Instead show evidence that there is an important difference between a star image at 3.2sec and 4sec (which is where the raw data filtering kicks in).

 

To be honest, I think the reason for the desaturation of the small stars in your cropped image is that noise reduction has been applied.  The effect of the "fixed" algorithm is to make tightly focused stars less bright and turn them green, not desaturate them.

 

Mark


Edited by sharkmelley, 23 August 2019 - 01:02 AM.


#8 calypsob

calypsob

    Aurora

  • *****
  • topic starter
  • Posts: 4683
  • Joined: 20 Apr 2013

Posted 23 August 2019 - 03:20 AM

Hi Wes,

 

I think what you should really do is to stack many 3-second exposures and also stack a few longer, Bulb exposures.

By comparing these two, you can get a definite answer about whether there is some damage to stars in the longer exposures.

That is a great idea, I will have to put together a test later this month.



#9 calypsob

calypsob

    Aurora

  • *****
  • topic starter
  • Posts: 4683
  • Joined: 20 Apr 2013

Posted 23 August 2019 - 03:22 AM

These data support that the A7R3 is still a star eater, and Sony are sticking their heads in the sand.

And this https://blog.kasson....-exposure-time/



#10 calypsob

calypsob

    Aurora

  • *****
  • topic starter
  • Posts: 4683
  • Joined: 20 Apr 2013

Posted 23 August 2019 - 03:29 AM

I applaud your efforts but am confused by your presentation of your case.scratchhead2.gif

 

In the past compelling evidence was presented for offending camera models by taking images with settings known to trigger the spatial filtering,

compared with images where a known baseline reference (like with settings where the filtering was not triggered).

My experience of your links doesn't do this. I am missing something? Is it an animated GIF that isn't playing for me?????

I see a widefield, and a crop 'untitled.'

It also seems clear "Carl" doesn't understand the concepts involved, and wasn't visual persuaded by your images (or at least is parroting someone else who wasn't persuaded by your images)....

 

Seems you got some kind of audience with someone there at Sony, with an invite to continue to email.

 

If your links are your best case, I'd consider trying with a better presentation as if you were explaining it to a 5th grader....

 

There are many ways that are obvious that Sony hasn't eliminated the secondary unintended consequences of their spatial filtering algorithm,

but since you have access to the a7riii and want your point two be heard,

please work on the delivery. Take for starter's Wei Ho's suggestion......

 

I'd also advise word careful word choice. Sony and others are influencers are now using "star-eater"----but unfortunately the term is very results-oriented.

 

As Mark Shelley and Jim Kasson have made efforts to define it as a spatial filtering process,

IMO we need to help Sony admit what is going on is caused by an intentional data cooking process.

If say, Carl for example (imagine he wasn't just a parrot of scripted information posted on his computer screen),

had to address whether or not Sony in the a7rIII with the recent release was still intentionally cooking the data (process),

instead of answering about very vague or pixel-peeping dots results,

than he would have been less able to state "If the star eater issue is present you will not see any bright spot on that image."

 

Using the term "star-eater" has allowed at least Carl to frame the debate to a very specific set of results,

rather that the whole issue that

an intentional data-cooking is producing secondary unintended detrimental results in imaging stars that could otherwise with the same gear be imaged better....

 

 

I think those that want Sony to completely fix any negative effects are not being realistic.

 

Please can we just get Sony to see that there is a population of users that want Sony to give them the option

of whether Sony's latest spatial filtering data-cooking will apply to the desired images......

 

How about just saying "Hey Sony, I get you haven't been able to completely fix it.

Just give us a menu option so whatever current spatial filtering doesn't  have to be applied to our photos of the stars........"

In my email I explained the links very clearly, referred to the issue as spacial filtering, and simply requested access to true raw data with no filtering applied whatsoever. I will make another attempt by incorporating the 3s exposures as wei hao suggested.



#11 calypsob

calypsob

    Aurora

  • *****
  • topic starter
  • Posts: 4683
  • Joined: 20 Apr 2013

Posted 23 August 2019 - 03:38 AM

I applaud your efforts to tirelessly send in problem reports to the manufacturer ... but, it seems this isn't getting to a group who really understands the issue. Isn't this where you sell the camera and buy one that doesn't do this?

The camera is not for astrowork, it performs great otherwise so theres no reason to sell. It seems to me though that sony has gone out of its way to modify the data coming out of this camera, I think the end user should be the one to decide if they want that sort of processing applied.

 

 

The Sony Digital Imaging Support Team are sticking to what they've been told.  They use a narrow definition of "star-eater" (i.e. the original crude algorithm) and using that definition they are able to claim to have fixed the issue.  For anyone new to this, you can find my analysis of the "fixed" algorithm here: http://www.markshell...areater_v2.html

 

Possibly the way to approach this is not to complain about "star-eater" because they have been told it is fixed.  Instead show evidence that there is an important difference between a star image at 3.2sec and 4sec (which is where the raw data filtering kicks in).

 

To be honest, I think the reason for the desaturation of the small stars in your cropped image is that noise reduction has been applied.  The effect of the "fixed" algorithm is to make tightly focused stars less bright and turn them green, not desaturate them.

 

Mark

Mark thanks, the demo test 3s then 4s and longer stacked subs will certainly be the next test. I did not apply noise reduction on the image, just chromatic abberation correction. I need to go back and be sure camera raw did not have a nr preset activated.



#12 Coconuts

Coconuts

    Vostok 1

  • *****
  • Posts: 194
  • Joined: 23 Sep 2012

Posted 23 August 2019 - 06:13 AM

This thread is also very relevant for the upcoming (September 12th) Sony A7R IV, which uses a variant of their IMX455 sensor, with a lot of phase detect autofocus pixels.  The same goes for the already shipping Fuji GFX100, which uses a variant of the 55 mm diagonal Sony IMX461 sensor.  Both are very high pixel count, back illuminated CMOS, with the potential to be great astrocameras.  That said, I'll wait for the photographic version of the QHY600, which will have no autofocus pixels, includes deep cooling, and for sure won't be doing funny stuff with the data.

 

All the best,

 

Kevin



#13 calypsob

calypsob

    Aurora

  • *****
  • topic starter
  • Posts: 4683
  • Joined: 20 Apr 2013

Posted 23 August 2019 - 10:33 AM

My request was escalated internally at sony, to be continued.

#14 sharkmelley

sharkmelley

    Vanguard

  • *****
  • Posts: 2302
  • Joined: 19 Feb 2013

Posted 23 August 2019 - 04:45 PM

My request was escalated internally at sony, to be continued.

I've heard that before, back in the early days of the Sony A7S. It simply means you have taken the Support department beyond their standard answer sheet.

 

Mark


  • SandyHouTex and 17.5Dob like this


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