111 replies to this topic

[gregbradley]

I asked the question on the DPReview Astrophotography forum if latest Sony models still have the star eater filtering that earlier models have.

 

From the responses it seems Sony has ditched the star eater algorithim in the A7R5 camera. I don't know if that applies to the A7IV camera.

 

Good news it would seem.

 

Greg

[sharkmelley]

Have you seen any evidence?  I'm very dubious because the same (incorrect) claims have been made about almost every Sony full-frame mirrorless camera since the 2nd generation and invariably the claims have been wrong:

https://www.sonyalph...he-a7riii-does/

 

Jim Kasson used to test most Sony cameras for star eater but he has moved away from Sony cameras now.  If someone has a long exposure dark frame available I can very easily put it through my usual testing for raw-data filtering:  

https://www.markshel..._filtering.html

 

Even if the raw-data filtering has now been disabled, has Sony also disabled the lens correction that leads to the concentric coloured polygons?  I think this is the biggest issue for using Sony cameras with electronic lenses for deep-sky astrophotography.

[sharkmelley]

I can now verify that Sony has definitely made huge changes to the raw-data-filtering on the Sony A7R5.  Over on DPReview is a thread where some long exposure raw darks are provided which allowed me to perform my usual analysis of raw data filtering which I repeat here.

 

The diagram below shows raw (non-bias-subtracted) pixel values plotted against the maximum of neighbouring values found in the 5x5 block surrounding them:

 

 

The important thing to note here is the strong vertical arm which represents isolated bright pixels.  We are no longer seeing the very destructive effect where pixels brighter than their neighbours all have their value capped to the maximum of their (same colour) neighbours.

 

This is very encouraging.

 

There is some residual inter-colour-channel effects in the green and blue leading to those fan type "rays" which needs a bit more investigation.  Also, in the end it's the effect on star shape and star colour that matters to astro-photographers (especially those using advanced post-processing such as Photometric Colour Calibration) and the effect cannot be easily predicted from this particular long-exposure black frame.

 

Mark

[Tapio]

Was it so that the star eater only kicked on longer than 30s exposure?

[sharkmelley]

Was it so that the star eater only kicked on longer than 30s exposure?

On the first generation of Sony full-frame mirrorless cameras the "star eater" raw data filtering kicked in on bulb-mode exposures, which effectively limited us to 30sec.  For the 2nd generation onwards, it kicked in for all exposures of 3.2sec or 4sec and longer (depending on model).  There was a workaround on the Sony A7RII where shooting in "Continuous" mode did not trigger the raw data filtering.

See https://www.markshel...#SonyA7_A7R_A7S

 

Mark

[mmalik]

Over on DPReview is a thread...

 

I thought they were closing; they still seem to be kicking. I wonder what happened?

[Alen K]

I thought they were closing; they still seem to be kicking. I wonder what happened?

https://www.theverge...ography-reviews

[gregbradley]

Mark, when you say the concentric circles using electronic leses on Sony cameras does that mean a lens like a Samyang XP14mm F2.4 lens which I believe is manual would not be affected?

 

Greg.

[sharkmelley]

Mark, when you say the concentric circles using electronic leses on Sony cameras does that mean a lens like a Samyang XP14mm F2.4 lens which I believe is manual would not be affected?

I can't say for certain because some non-Sony lenses are known to be affected.

[Digitaliz.se]

I never understood the fear of the star eating problem unless you are scientifically evaluating your images/data.

Can anyone explain if I've missed something? I shoot with Sony cameras and always loved the result it produced.

 

/Stefan

[ChristopherBeere]

I never understood the fear of the star eating problem unless you are scientifically evaluating your images/data.

Can anyone explain if I've missed something? I shoot with Sony cameras and always loved the result it produced.

 

/Stefan

 

I decided to try one for myself as i was sceptical that it was all that bad - also was very excited about the possibility of using the new Sigma DN Arts and Contemporaries.

 

I tested it properly in October this year in Crete on the dense starfields of Cygnus and was absolutely shocked at how bad the data was.

 

Modded Sony A7II + Sigma 65mm Contemporary @f/2.0 - 100% zoom center crop from a debayered raw :

 

 

 

It has completely mangled the data. Ive chucked it in a box at the back of the astro spares cupboard as i would feel guilty selling it to someone as an astro camera its so bad 

 

I own 4 modded Canon DSLRs and the data from them is immaculate in comparison.

[xonefs]

 

Even if the raw-data filtering has now been disabled, has Sony also disabled the lens correction that leads to the concentric coloured polygons?  I think this is the biggest issue for using Sony cameras with electronic lenses for deep-sky astrophotography.

 

The colored concentric polygons were only a factor on lossy compressed raw files on the A7R IV and shooting in uncompressed raw solves this, so it seems like a non-issue either way especially now that there is lossless compressed raw as an option. 

 

This news is very promising, I may have to get an A7R V now

Edited by xonefs, 29 December 2023 - 02:41 PM.

[Alen K]

I decided to try one for myself as i was sceptical that it was all that bad - also was very excited about the possibility of using the new Sigma DN Arts and Contemporaries.

 

I tested it properly in October this year in Crete on the dense starfields of Cygnus and was absolutely shocked at how bad the data was.

 

Modded Sony A7II + Sigma 65mm Contemporary @f/2.0 - 100% zoom center crop from a debayered raw :

 

 

 

It has completely mangled the data.

Yikes. One word comes to mind to describe what all the dimmer stars look like: maggots.

[Digitaliz.se]

Ok, that looks horrible

 

/Stefan

 

I decided to try one for myself as i was sceptical that it was all that bad - also was very excited about the possibility of using the new Sigma DN Arts and Contemporaries.

 

I tested it properly in October this year in Crete on the dense starfields of Cygnus and was absolutely shocked at how bad the data was.

 

Modded Sony A7II + Sigma 65mm Contemporary @f/2.0 - 100% zoom center crop from a debayered raw :

 

 

 

 

It has completely mangled the data. Ive chucked it in a box at the back of the astro spares cupboard as i would feel guilty selling it to someone as an astro camera its so bad 

 

I own 4 modded Canon DSLRs and the data from them is immaculate in comparison.

[erictheastrojunkie]

I decided to try one for myself as i was sceptical that it was all that bad - also was very excited about the possibility of using the new Sigma DN Arts and Contemporaries.

 

I tested it properly in October this year in Crete on the dense starfields of Cygnus and was absolutely shocked at how bad the data was.

 

Modded Sony A7II + Sigma 65mm Contemporary @f/2.0 - 100% zoom center crop from a debayered raw :

 

 

 

It has completely mangled the data. Ive chucked it in a box at the back of the astro spares cupboard as i would feel guilty selling it to someone as an astro camera its so bad 

 

I own 4 modded Canon DSLRs and the data from them is immaculate in comparison.

I used to use a Sony A7s, mainly for timelapse nightscapes, but out of necessity I used it for regular tracked nightscape panos a couple times. I paired it with a Voigtlander 50mm APO and a Voigtlander 21mm Nokton, I got pretty much the same results as you did. Downsampled and wide? Sure, looks like a good picture, but once you look at the individual frames at even a modest zoom level the star mangling is obvious and disgusting. This was the best I could manage from several panos that got destroyed by the star eater, when people deny what the algorithm does because "I don't see what the problem is, I get good photos" or "there's already too many stars anyways, what's a few missing going to do?", I just laugh. 

 

 

Here's a link to a single, unedited, frame that is part of that pano, zoomed out it doesn't look terrible, at 100% it looks absolutely aweful:
https://i.imgur.com/PVrxaeK.jpg

[whwang]

I have been telling people about the terrible star eater of Sony since 2015.  So.... I told you.  

 

Well, earlier this year I had a chance to use an A7R5 for a short period.  This is one of the picture I took:

During the processing, and when I looked at the image at various stages at 100%, I did not feel it has the old terrible star eater problem. I couldn't confirm that it's completely gone, but I wasn't annoyed by anything either.  Later if I have another chance to put my hand on an A7R5, I will do more deep imaging with sharp optics.  I can offer the raw files for others to examine as well.

[xonefs]

What I am curious about now is if other recent bodies like the A7C II have the updated star eater improvements. The A7 IV still suffers from star eater, but the A7C II and A7C R are newer so maybe not. 

[sharkmelley]

Well, earlier this year I had a chance to use an A7R5 for a short period.  This is one of the picture I took:

 

 

During the processing, and when I looked at the image at various stages at 100%, I did not feel it has the old terrible star eater problem. I couldn't confirm that it's completely gone, but I wasn't annoyed by anything either.  Later if I have another chance to put my hand on an A7R5, I will do more deep imaging with sharp optics.  I can offer the raw files for others to examine as well.

Earlier this year!   

We're only a few days into this year!

 

That's a great image!

 

I took a look at one of your raw lights and agree with your verdict.  I can find no evidence of any damage to stars.

 

Generating my usual plots, I can see some evidence of raw-data processing but for some reason it's only in the blue channel.  Note the fan-type rays similar to what was seen in the plot in post #3:

 

 

So whatever processing is being performed on the raw data, it seems to be benign.  Finally we have some good news from Sony.  One caveat is that we have no information on the potential concentric coloured polygons problem because the lens used (Sigma 40mm F1.4 DG HSM Art) is not recognised by the firmware. 

 

Mark

[xonefs]

Mark, what is the concern still with concentric coloured polygons? That issue has been complete resolved in all scenarios when shooting in uncompressed RAW and only affected lossy compressed raw on older sony bodies.

 

Solution: don't shoot in lossy compressed raw and shoot uncompressed.

 

Now with lossless compressed raw I expect it to also be a non-issue there as well and also solve any file size concerns.

 

So what is the concern still? I don't see why anyone would want to be shooting in lossy compressed raw in the first place for astrophotography, and now with lossless compressed raw available why anyone would ever choose lossy compressed. 

Edited by xonefs, 04 January 2024 - 04:05 PM.

[sharkmelley]

Mark, what is the concern still with concentric coloured polygons? That issue has been complete resolved in all scenarios when shooting in uncompressed RAW and only affected lossy compressed raw on older sony bodies.

The concentric coloured polygons are an artefact of a lens correction that is applied whenever the lens is recognised by the firmware.  They are more apparent when lens corrections are switched on but they cannot be completely disabled within the camera's menus.  The lens correction is applied whether or not raw compression is switched on, so the two effects are unrelated.  However, it is true that raw compression makes the polygons more visible, which may be because raw compression also reduces the camera's effective bit depth to 13bits instead of 14bits. 

 

Edit:  I recently found Sony's patent for lens corrections using polygons (instead of circles):  https://patents.goog...atent/US7317482

Edited by sharkmelley, 04 January 2024 - 05:04 PM.

[xonefs]

Ok, can you give an example of the concentric polygons appearing in practice when there is no raw compression? I have yet to be able to have it happen to me or be able to reproduce it. I did have it happen to me shooting in compressed raw so I know what the issue looks like and that was frustrating, but I have never been able to get it to appear again after switching to uncompressed raw in the Sony A7R IV and using a variety of recognized lenses. 

Edited by xonefs, 04 January 2024 - 06:04 PM.

[sharkmelley]

It depends on what you mean by "appearing in practice".  In the coloured polygons article on my website, the Sony A7III example in the section "Can the lens correction be switched off?" was shot with compression off.  It clearly demonstrates that the polygons exist with compression off and from there it is simply a question of determining under what circumstances they "appear in practice".

 

Typically it is not so much of a problem for everyday photography but unfortunately, the astro-processing workflow designed to reveal faint structures and nebulosity in deep-sky images (i.e. calibration with master flats followed by stacking, background subtraction and stretching) will also reveal faint artefacts embedded in the raw data by crude in-camera processing such as the Sony polygon lens correction.

 

I don't actually have any data from a Sony A7RIV but others have tested it and claim the problem is still there but I'd love to verify this myself.  If you are willing then please create a sequence of ISO 100 flat frames according to the test protocol in this article and I'll take a look.  Shoot with a recognised lens wide-open.

 

Mark

[xonefs]

Ok I will send them over. 

 

From what you are saying I will not be surprised if it shows up since I saw it in the lossy compressed files, just haven't seen it since. It would be helpful if someone with an A7R5 can test this as well. 

 

I also have a Sony ZVE1 sitting here I am about to return (a7s3 sensor)- is there anything quick I can get off that for you to test in regards to star eater or anything? I just realized it is even newer and came out after the a7r5. 

Edited by xonefs, 05 January 2024 - 11:04 AM.

[sharkmelley]

Ok I will send them over. 

 

From what you are saying I will not be surprised if it shows up since I saw it in the lossy compressed files, just haven't seen it since.

Here are the results from a sequence of uncompressed raws taken using a Sony A7R4 with Tamron 28-75mm F2.8 Di III RXD at 75mm f/2.8 ISO 100 (lens corrections switched off):

 

 

The files were bias-subtracted, debayered, binned and then colour saturated.  Some rings are appearing at the edges of some of the shots but they are so faint that I can't even determine whether or not they are polygons.  They could be caused by something entirely different. With other camera and lens combinations, the coloured polygon effect was very obvious, even with uncompressed raws.

 

You will almost certainly not hit a problem with this camera/lens combination.  Maybe with another lens it would be more obvious - who knows. 

Edited by sharkmelley, 06 January 2024 - 06:40 AM.

[sharkmelley]

I also have a Sony ZVE1 sitting here I am about to return (a7s3 sensor)- is there anything quick I can get off that for you to test in regards to star eater or anything? I just realized it is even newer and came out after the a7r5. 

Here are the results of the dark test for raw-data filtering on the Sony ZV-E1:

 

 

It looks very similar to the that for the Sony A7R5 in post #3.  Almost certainly it is the same algorithm and star-eater will not be a problem.

 

Here are the results from a sequence of losslessly compressed raws taken using the Sony ZV-E1 with Tamron 28-75mm F2.8 Di III RXD at 75mm f/2.8 ISO 100 (lens corrections switched off):

 

 

I can't find any evidence of the coloured concentric polygons.