95 replies to this topic

[GTom]

Found this on DPR, quite drastic. Full star formations, far from the noise level, wiped out:(

 

You have been warned...

Edited by GTom, 29 November 2016 - 04:18 PM.

[whwang]

Like someone pointed out in the DPR thread, I am also not convinced that the test there is conducted properly to show the real problem.

 

What should really be done is series of 30 sec exposures in B mode and M mode on the same target in the same night under the same temperature.  Then the difference between the stacked B and M images can tell you whether there is really a serious filtering.

 

I did the test on A7S2 using dark exposures.  (I did not have a chance to bring the camera to stars yet, because of the weather.)  I can see differences between the stacked B and M exposures (each contains 16 30-sec exposures).  So there is definitely still some filtering.  However, most of the hot pixels are not filtered.  The difference between the stacked B and M exposures is also much smaller than that from A7R.  So there is a possibility that Sony had reduced the strength of filtering (like what Nikon did a few years ago).  Nevertheless, I need to bring this camera to stars so we can say for sure what's going on.

 

Cheers,

Wei-Hao

[sharkmelley]

It looks quite likely to me.  The raw file shows evidence of the "pixel pairing" whereby a hot pixel has its intensity reduced to be identical to the warmest of its neighbours (of the same colour).  Jim Kasson has also taken noticed a difference in the Fourier transform between 3.2 and 4sec exposures similar to his analysis of the original star eater issue - evidence of spatial filtering.

 

I really hope it's not true though.

 

Mark

Edited by sharkmelley, 29 November 2016 - 07:37 PM.

[erictheastrojunkie]

So does this affect all models or just the A7s and A7r2? Really really disappointing if true. 

[sharkmelley]

So does this affect all models or just the A7s and A7r2? Really really disappointing if true. 

Originally, the star-eater only happened in bulb mode and affected all models in the A7 series and the A5100, A6000 and probably many more.

 

Having said that, Wei-Hao (above) believes something different is happening on the A7SII.

 

In any case, it now seems that the latest firmware update on the A7RII also applies star-eater to manual exposures 4sec or longer.

 

Mark

Edited by sharkmelley, 29 November 2016 - 10:38 PM.

[whwang]

It's almost for sure that Sony is still doing something bad.  The question is whether it practically affects astrophotography.  Nikon is still trying to mess around with the NEF file, but the effect is small enough in recent generations of Nikon cameras that we do not have to worry too much about it.  Is it the same for Sony?  I don't know, and I am eager to know.

 

If Sony does the same to even short M-mode exposures, then it will be much harder to verify how bad it is.

 

And allow me to constantly remind people this is how bad it used to be on A7R:

http://www.asiaa.sin.../star_eater.jpg

 

Whether the A7?2 series is equally bad remains to be seen.

 

Cheers.

Wei-Hao

[sharkmelley]

Hi Wei-Hao,

 

Is it possible for you to upload a raw bulb mode A7SII dark ?  I'm interested to take a look.

 

Mark

[whwang]

30-sec Bulb dark from A7S2: 

https://dl.dropboxus...S2/DSC00357.ARW

 

30-sec M dark from A7S2:

https://dl.dropboxus...S2/DSC00358.ARW

[sharkmelley]

30-sec Bulb dark from A7S2: 

https://dl.dropboxus...S2/DSC00357.ARW

 

30-sec M dark from A7S2:

https://dl.dropboxus...S2/DSC00358.ARW

 

Are those files TIFF or ARW?  They download as TIFF but refuse to open in PixInsight because the "ImageLength" field is missing.

Mark

Edited by sharkmelley, 29 November 2016 - 11:42 PM.

[whwang]

They are ARW, copied from camera's SD card to my HD, and then to the shared location.  Latest Photoshop CC can recognize them and covert them.

[sharkmelley]

They are ARW, copied from camera's SD card to my HD, and then to the shared location.  Latest Photoshop CC can recognize them and covert them.

Thanks - I opened them in PS CC and saved as DNG and then into PixInsight.

 

The undebayered bulb mode A7S2 file has the usual pixel pairing signature typical of the star-eater algorithm.

 

Also, I obtained a 4 second A7R2 file with the new firmware from the person who made the DPR post.  Unfortunately, this also has the pixel pairing signature of star-eater.

 

Mark

[whwang]

Did you find similar pixel pairing in M-mode exposures from older A7x cameras?  Those are supposed to free from the problem.

 

Cheers,

Wei-Hao

[sharkmelley]

Did you find similar pixel pairing in M-mode exposures from older A7x cameras?  Those are supposed to free from the problem.

 

Cheers,

Wei-Hao

No there's no pixel pairing in M mode - that's the mode I use for all my A7S astrophotography.

 

Mark

[GTom]

I have a sony a6000 ATM, I can test it hopefully tonight.

 

However, from your feedback, it is clear to me, that I'll avoid the a7s as an astrophoto-only camera. Limiting to 30s even at ISO6400 is a compromise putting the camera behind the (actually cheaper) asi1600mm-cool on my list.

[sharkmelley]

I have a sony a6000 ATM, I can test it hopefully tonight.

 

 

It's already known that the A6000 is a star eater in bulb mode.  To see if that is a concern to you, take a shot in manual mode and compare it with an identical length shot in bulb mode. 

 

Make sure there is absolutely no star trailing in both images - not even by a single pixel - otherwise the test will be invalid.   Star trailing protects the stars from being eaten.

 

Mark

Edited by sharkmelley, 30 November 2016 - 06:46 AM.

[GTom]

It's already known that the A6000 is a star eater in bulb mode.  To see if that is a concern to you, take a shot in manual mode and compare it with an identical length shot in bulb mode. 

 

Make sure there is absolutely no star trailing in both images - not even by a single pixel - otherwise the test will be invalid.   Star trailing protects the stars from being eaten.

Pure curiosity is it now. My a6000 is not converted and apparently it stays so. I'll keep my converted+FW-hacked D5100 until an APSc or larger monochrome CMOS camera comes to market at a reasonable price.

[sharkmelley]

Here is the first info I have seen that the issue now affects the MkII version of the Sony A7S for exposures 4sec and greater:

http://www.sonyalpha...hy-fix-request/

 

Sony is clearly not listening

 

Mark

[whwang]

At this very moment, my computer is processing my images taken with my modified A7R. I have to stack 100 frames of 30s each, in order to avoid the star eater problem.  The price to pay?  Lot of processing power and processing time, and thicker electricity bills needed for running both the computer and A/C.  I can feel the heat coming from my computer right now.  Sorry, trees.  Sorry, polar bears.  It's my fault picking Sony.  I won't make the same mistake again.

[garret]

 

I can feel the heat coming from my computer right now.

If your PC is water-cooled you have a boiler

 

I hope the up coming Canon 6 D mk2 is as good as the Canon 5 D mk4, My current Canon 5 D mk2 is getting old and noisy.

 

Garret

[TareqPhoto]

Not sure if i will modify my A7r or not because i want to sell it and upgrade to A7rII, i use this camera mainly for landcapes, nightscapes and portraits, so i don't want to modded it and then i can't sell it to upgrade.

 

But if i can modify my a7r without opening it or sending it for that then i will be interested in.

[sharkmelley]

At this very moment, my computer is processing my images taken with my modified A7R. I have to stack 100 frames of 30s each, in order to avoid the star eater problem.  The price to pay?  Lot of processing power and processing time, and thicker electricity bills needed for running both the computer and A/C.  I can feel the heat coming from my computer right now.  Sorry, trees.  Sorry, polar bears.  It's my fault picking Sony.  I won't make the same mistake again.

I know the feeling - I often have hundreds of 30sec Sony A7S exposures for a single image.

 

Mark

[calypsob]

 

At this very moment, my computer is processing my images taken with my modified A7R. I have to stack 100 frames of 30s each, in order to avoid the star eater problem.  The price to pay?  Lot of processing power and processing time, and thicker electricity bills needed for running both the computer and A/C.  I can feel the heat coming from my computer right now.  Sorry, trees.  Sorry, polar bears.  It's my fault picking Sony.  I won't make the same mistake again.

I know the feeling - I often have hundreds of 30sec Sony A7S exposures for a single image.

 

Mark

 

what iso are you two shooting at, just curious.  

[whwang]

what iso are you two shooting at, just curious.  

1600.  to minimize the impact of read noise in short exposures.  

[sharkmelley]

 

 

At this very moment, my computer is processing my images taken with my modified A7R. I have to stack 100 frames of 30s each, in order to avoid the star eater problem.  The price to pay?  Lot of processing power and processing time, and thicker electricity bills needed for running both the computer and A/C.  I can feel the heat coming from my computer right now.  Sorry, trees.  Sorry, polar bears.  It's my fault picking Sony.  I won't make the same mistake again.

I know the feeling - I often have hundreds of 30sec Sony A7S exposures for a single image.

 

Mark

 

what iso are you two shooting at, just curious.  

 

I'm using ISO 10,000 which for my sky and my optics puts my back of camera histogram peaks just to the right of centre.  I find this is necessary to overcome coloured concentric banding effects and the split sensor left-hand-side horizontal coloured banding effects.

 

Mark

[navmannz]

I'm a newbie here, and have only recently begun to more systematically explore astrophotography with my A7RII - but I was so discouraged when I found out about the effects of the star-eater algorithm that I've written a blog describing my experience of its impacts here https://johnleathwic...trophotography/. I wish I could find a way to get Sony to engage on this, but never seem to get anything beyond a brief and bland assurance in classic corporate-speak.

 

Does anyone have any suggestions on how to establish a more effective line of communication with them?

 

-John