a7RII FW 4 star-eating test results

JimKasson

Community Leader
Forum Moderator
Messages
52,256
Solutions
52
Reaction score
59,047
Location
Monterey, CA, US
I've tested the new firmware, and don't find any material differences in the camera's appetite for stars from FW 3.30.



57f4b007f07f472dac2abc5dc97daf3c.jpg.png



26d12fba40be4a9197e7635355717890.jpg.png

c5d63ea766c84b49b587a09ded102665.jpg.png

f744472321464326a1690c9717a540e4.jpg.png

Details here:


Jim

--
 
I guess I'm lucky I was too lazy to upgrade my FW 3.2 a7rii. Thanks for the info Jim. Do we all know for sure that it was version 3.3 that introduced the problem? I read in another thread somewhere or on fredmiranda that it was suspected as far back as FW 2.0
 
I guess I'm lucky I was too lazy to upgrade my FW 3.2 a7rii. Thanks for the info Jim. Do we all know for sure that it was version 3.3 that introduced the problem?
That's where I first saw it. It didn't happen with FW 2.0 at 4 seconds. Not sure exactly when that behavior was introduced. I wish Sony were less opaque about the changes effected with each FW release.
I read in another thread somewhere or on fredmiranda that it was suspected as far back as FW 2.0
Jim
 
Thanks for testing this and removing all doubts! So Sony really listened this time. For me, the biggest improvement (and unexpected surprise) with firmware 4.0 is the news "Focus Standard", makes the flexible AF spot so much more enjoyable.
 
Last edited:
Thanks for testing this and removing all doubts! So Sony really listened this time.
I don't understand. They listened (maybe), and didn't fix the star-eating. How is that good news?
For me, the biggest improvement (and unexpected surprise) with firmware 4.0 is the news "Focus Standard", makes the flexible AF spot so much more enjoyable.
Jim
 
Thanks Jim, I appreciate the effort you put into these posts. We all benefit from them.

The only real solution for Sony is to offer the ability to turn off the low pass filtering. No doubt that will make the RAWs look really noisy with coloured dots everywhere in the image but so far even LR's colour noise slider does a good job of getting rid of that.

It would be good if Sony did that as it seems like all the major camera brands cook their RAWs these days.

Greg.
 
Thanks Jim, I appreciate the effort you put into these posts. We all benefit from them.

The only real solution for Sony is to offer the ability to turn off the low pass filtering. No doubt that will make the RAWs look really noisy with coloured dots everywhere in the image but so far even LR's colour noise slider does a good job of getting rid of that.

It would be good if Sony did that as it seems like all the major camera brands cook their RAWs these days.

Greg.
Agreed.

Or better still, give us an option for "raw" that is actually RAW. ; )
 
thx a lot for the test Jim.

so Star Eater is not fixed at all, only marginal if any improvement...

i made a slide to summarize how much we have lost in resolution since the introduction of star eater

SONY MUST ALLOW USER TO DEACTIVATE STAR EATER!



468a3cd882cb4510893e0667902b869a.jpg
 
Thanks Jim, I appreciate the effort you put into these posts. We all benefit from them.

The only real solution for Sony is to offer the ability to turn off the low pass filtering. No doubt that will make the RAWs look really noisy with coloured dots everywhere in the image but so far even LR's colour noise slider does a good job of getting rid of that.

It would be good if Sony did that as it seems like all the major camera brands cook their RAWs these days.

Greg.
Agreed.

Or better still, give us an option for "raw" that is actually RAW. ; )
Be careful what you wish for. No CDS? No mapping out dead or stuck pixels? No PDAF pixel interpolation?

All that could work -- except for no CDS -- but the raw developers would have to get smarter, and there would have to be a pixel map in the metadata, like with disks.

I would like it if things moved in that direction, but going all the way would be a big change.

Jim
 
Thanks Jim, I appreciate the effort you put into these posts. We all benefit from them.

The only real solution for Sony is to offer the ability to turn off the low pass filtering. No doubt that will make the RAWs look really noisy with coloured dots everywhere in the image but so far even LR's colour noise slider does a good job of getting rid of that.

It would be good if Sony did that as it seems like all the major camera brands cook their RAWs these days.

Greg.
Agreed.

Or better still, give us an option for "raw" that is actually RAW. ; )
Be careful what you wish for. No CDS? No mapping out dead or stuck pixels? No PDAF pixel interpolation?

All that could work -- except for no CDS -- but the raw developers would have to get smarter, and there would have to be a pixel map in the metadata, like with disks.

I would like it if things moved in that direction, but going all the way would be a big change.

Jim
I think RAW internal processing where known "defects" like the PDAF pixels and dead and stuck pixel mapping is appropriate. It's not trying to interpret image data per exposure (no whole image filtering)

It would be nice to have 3 levels for long exposure NR (On, Off and RAW), where RAW is only available when capturing RAW and just handles the Fixed "noise" sources but nothing else.
 
Thanks Jim, I appreciate the effort you put into these posts. We all benefit from them.

The only real solution for Sony is to offer the ability to turn off the low pass filtering. No doubt that will make the RAWs look really noisy with coloured dots everywhere in the image but so far even LR's colour noise slider does a good job of getting rid of that.

It would be good if Sony did that as it seems like all the major camera brands cook their RAWs these days.

Greg.
Agreed.

Or better still, give us an option for "raw" that is actually RAW. ; )
Be careful what you wish for. No CDS? No mapping out dead or stuck pixels? No PDAF pixel interpolation?

All that could work -- except for no CDS -- but the raw developers would have to get smarter, and there would have to be a pixel map in the metadata, like with disks.

I would like it if things moved in that direction, but going all the way would be a big change.
Point taken, and agreed, I may have overstated the case.

I would, however, prefer it if *most* in-camera processing were discretionary.
 
PS: Wish DPreview would allow user (i.e. myself) delete my own stupidity..... (even after 5 min window).
My experience has been that I can revise or delete one of your posts up until someone replies directly to your post. Then it is "locked".

That's been the case for weeks / months / years after I've made a post, as long as there hasn't been a reply.
 
PS: Wish DPreview would allow user (i.e. myself) delete my own stupidity..... (even after 5 min window).
My experience has been that I can revise or delete one of your posts up until someone replies directly to your post. Then it is "locked".

That's been the case for weeks / months / years after I've made a post, as long as there hasn't been a reply.
Yup. Problem here is, Jim's correction locked BBQue's reply.

That and the post-limit at which a thread autolocks are why I moved my reverse engineering work to Dyxum. An ongoing research project REALLY needs the ability to update posts as a reference. It also makes DPreview a poor place to put reference posts - you can't collect information in one single place, which is why discussions frequently resurface in new threads here.
 
Be careful what you wish for. No CDS? No mapping out dead or stuck pixels? No PDAF pixel interpolation?
CDS cannot be removed without harm, it works on analog level before quantization (If I understood diagrams correctly), so even with "calibration information" in RAW file its accuracy can be beyond qnantization level of 14 bit raw file.
 
Be careful what you wish for. No CDS? No mapping out dead or stuck pixels? No PDAF pixel interpolation?
CDS cannot be removed without harm, it works on analog level before quantization (If I understood diagrams correctly), so even with "calibration information" in RAW file its accuracy can be beyond qnantization level of 14 bit raw file.
My point exactly.
 
thx a lot for the test Jim.

so Star Eater is not fixed at all, only marginal if any improvement...

i made a slide to summarize how much we have lost in resolution since the introduction of star eater

SONY MUST ALLOW USER TO DEACTIVATE STAR EATER!

468a3cd882cb4510893e0667902b869a.jpg
I understand the intention, but I think that this graphic is a bit... misleading:
  • The scaling is non-linear and not consistent. As it is now, the top line should be at 48 MPix.
  • I'm looking for information, but right now, I'm not sure that the A7II is affected by star eater at all. If you have any reliable info, please let me know! (I own one and would like to be sure before updating the firmware)
  • Color resolution: There never were 42 MP in the blue or red channel on the A7RII. There's about 10 MPix red and blue info and about 20 MPix green + some unused pixels at the border and maybe half a million or so PDAF-pixels (roughly - don't quote me on those numbers).
  • I'm not quite following how you arrive at those combined "polychromatic" MPix-numbers on the right. The dividers seem to be different, too (2.33 on the A7RII, 2.66 on the others). How did you calculate those?
Please don't take this as a personal attack - I'm just looking for a fact based discussion and trying to help.

--
http://vintagelensreviews.com/
Reviews of vintage Minolta SR mount lenses and more
 
Last edited:
Thank you Jim for your time and effort, you make the information digestible for someone like myself that has limited technical knowledge.

And 'booooo' to Sony, you make it hard to love you sometimes. I have a thing for long exposures, but knowing that my 42mp camera isn't really a 42mp camera for anything over a 3.2 second exposure is a real bummer.

Jim, if I may ask you to speculate, why do you think Sony chooses to use this processing scheme? Anyone else feel free to chime in, but please let's keep the replies civil. 'Because Sony's stupid!!!' is not the type of answer I'm looking for. More on the technical side, what are the trade-offs, both good and bad.

Sharif
 

Keyboard shortcuts

Back
Top