Sony A7Riii Stripes

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

mjkzz
Posts: 1681
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

yes, Lou, sure, understand that, just want to point it out about it so people do not mix it with PDAF issue.

I also encountered half exposed image, quite often (so far it has occurred 5 times), by half exposed, I mean it literally have upper image exposed correctly, gray or garbage for the lower half, not under exposed, it is kind of annoying. I once had 2 of them in one deep stack (300 images). I guess it is the silent shutter problem.

Geopressure
Posts: 85
Joined: Sun Mar 26, 2017 3:18 pm
Location: Tennessee
Contact:

Post by Geopressure »

Sorry I have sent the A7Riii camera back and can't provide any new images, but here is a crop with contrast and clarity boosted. No sharpening. This is a JPEG stack taken using EFSC at 1/15 sec with LED lighting. You can see the diffuse "waffle" or striped pattern I referred to. It might help to stand back.



Image

ray_parkhurst
Posts: 3402
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Is this caused by sensitivity differences between pixels or regions of pixels? Some pixels used for AF, others not, and they have different sensitivity?

edited to add: I see a ~50% difference in RGB luminance level between the dark and light areas. That's pretty big.

Lou Jost
Posts: 5933
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Ray, there is a steadily-improving theory of why this happens: see
https://www.dpreview.com/forums/thread/4262517?page=4

Sony is no help on this. Users are developing and testing theories and solutions.

Macro_Cosmos
Posts: 1506
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

mawyatt wrote:Macro_Cosmos,

Interesting stuff about the Sony and the strips (seems the Sony Fanfolks always bashing Nikon, but you don't much in the reverse).

Is this caused by the phase detect focus sensors on the chip?
I am not familiar enough with the Sony system to verify this hypothesis one way or another. If there is a way to control for these PDAF points, then we could reach to a sound conclusion. I assume the A7r3 uses the same system, so does the a6500 and perhaps other Sony bodies. If I am able to gather some dark field raws of these bodies and compare, we may be able to see what causes those weird stripes.

As for Sony fanboys (fanboys, not users), they bash everyone from my experience, even Sony users who dare to mention some personal issues with the system. I remember a couple in my Chinese Laowa chatgroup, they pick fights with people all day and when cornered, they Ad Hom people by asking for "your portfolio" and then cherrypick. Truly sick of these people. They ended up leaving because almost everyone active muted them, and lots of newcomers were leaving at mass due to the constant flames, which is in turn bad for the brand which the group is associated with. I'm not saying such people do not exist in other camps; I just hope we can all settle down and discuss the problems, which will ultimately lead to manufactures improving these issues. It's good for all of us, some people do not get it. The a7m3 issue I mentioned above is extremely devastating. Such problems lead to higher readout noise, false colours, and could even ruin photos. The a7m3 is still an extremely capable camera, but with such a massive flaw, it warrants a recall in my opinion. If I had one (which I almost did), I will not just eat it up. It's unacceptable.
mawyatt wrote:Will be interesting to see if the new Nikon Z6 and Z7 Mirrorless cameras have any of these types of issues.
We will see. This problem is an algorithm issue if the stripes is caused by PDAF points. Assuming this hypothesis is valid, the sample I've shown illustrates inadequate algorithm processing. Luckily this is fixable via firmware (hopefully), so it's not as devastating as the other problem.
mawyatt wrote: BTW don't think you can say that one long exposure Tlong is "exactly" like N shorter exposures Tshort, where Tshort is Tlong/N. The reason being the readout noise when a pixel is digitized (readout) happens every readout. So Tshort would have the result accumulation of N readout noise effects (if non-correlated should follow rootN), where Tlong only has only one read noise effect. Depending on how much noise the sensor produces this may or may not have an effect of the final resultant image noise level.

Anyway just a thought and thanks for showing the Sony image artifacts,

Best,
Thanks, you're also absolutely right. It's not exactly the same, but rather "somewhat similar". Hot pixels however are distributed fairly identically, which is why in astroimaging, people impose dark field exposures in between their sequences for stacking.

mjkzz wrote:wow, but I can not do 10 minutes exposure, only up to 30 sec . . . well maybe it is disabled in silent mode.

Exposure for my stack usually is less than 0.5 second and interval is about 4 seconds, so the sensor should have enough time to cool down.
Yeah, mine was done with an interverlometer. At just 1/2 seconds with such a settling time, hot pixels wouldn't be a problem. What we're doing here is actually extremely similar to astroimaging. For my stuff, incorporating dark fields should be beneficial. My exposure times can be up to 6 seconds.

Also, would it be possible to send me the 10-minute dark field .arw file? My email is macrocosmos1949@gmail.com
Thank you!

Macro_Cosmos
Posts: 1506
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

Lou Jost wrote:The classic stripe problem is different from this, and depends critically on how a scene is lit, and also on the lens; it would not show up in long-exposure dark frames. It is almost certainly due to the PDAF pixels. It occurs in normal short exposures. As entry points to the literature on this subject, see:
http://blog.kasson.com/a9/pdaf-striping-on-the-a9/
https://blog.kasson.com/a7riii/sony-a7i ... ping-faqs/
https://www.dpreview.com/forums/post/61073358
Note that Nikon cameras also can show this (Edit:: IF I RECALL CORRECTLY).
YES! Exactly what I was thinking after the response above. Attaching a lens to the camera could cause the activation of PDAF, which could lead to these odd stripes.

I did a dark field with my D810 attached to a 70-200. In LR, it's pure black besides some of those annoying hot pixels and a couple dead ones. Loading into Maxim DL leads to chaos. The raw is full of weird mosaic patterns.

Now I just need to get home and run the test for my D810 with no lens, and then for my D7100 with and without a lens.

~ MC

Lou Jost
Posts: 5933
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

"If I am able to gather some dark field raws of these bodies and compare, we may be able to see what causes those weird stripes."
As I mentioned above, these PDAF stripes are fairly well understood and would not be visible in dark frames. It is known that the stripes appear only when there is a bit of flare in the lens.
"Attaching a lens to the camera could cause the activation of PDAF, which could lead to these odd stripes."
There is a lens dependence but it is complex. It has to do with location of exit pupil, the way the lens flares, and other subtle features. Some lenses are known to be especially good at making these stripes, others are known to be nearly immune to them. It is not a simple matter of turning the PDAF sensors on or off. In cameras where autofocusing is turned off, the stripes are unchanged. It is a passive effect of the PDAF sensels treating light slightly differently than regular sensels. That's also why it won't appear in dark frames.
"they bash everyone from my experience, even Sony users who dare to mention some personal issues with the system."
This sounds exactly like the Pentax forum that I studied to research my last camera purchase. I am glad our forum here is not like that!

Macro_Cosmos
Posts: 1506
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

An update, the stripes I have observed are very likely FPN (Fixed Pattern Noise). It's typically found on earlier CMOS sensors. There's many factors that can contribute to this disparity.

It's usually suppressed via flat field correction, an interpolation method which is cheap to implement. Long exposures exacerbates this problem. There's many better methods as well.

Those ARE NOT PDAF stripes simply because:
1. It's across the entire frame, ew. PDAF striping isn't.
2. It's a dark field exposure, meaning no light reaches the sensor [1]
3. PDAF striping is caused by lens flare, and some lenses are said to be extremely prone to it.

[1] Theoretically-- some lenses might have light leaks, the Nikon 105mm f/1.4 a friend tried out had this issue. I'm not saying every single 105/1.4 has this problem, just the one my friend got to play with. Also for DSLRs, the viewfinder window must be blocked for long exposures, that thing also gathers some light.

I am too busy with 4 subjects and a ton of assignments to actually investigate and research this problem. I will post updates once I get enough time.

Once again, feel free to email me dark field raws of 10 minutes of any kind of camera. I am especially interested in a7r3, a7r2, a7m2, Fuji x-H1, Nikon D850, Nikon D500. My email is macrocosmos1949@gmail.com.
Here are the requirements:
1. ALL in camera processing off. Long exposure noise reduction and etc.
2. Best raw format
3. Lowest base ISO
4. DSLR? MUST block the viewfinder out. If it's not blocked out, it will not be usable for analysis.

A dark field raw is easy. Just put the lens cap on, and using an interverlometer, set at 10 minutes. Don't have a shutter release? No problem, just use a phone alarm. The several seconds doesn't matter much.


Also thanks to Mjkzz for providing me with some a7m3 raws. I didn't respond to the email, sorry.

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

The Nikon D500, D800 & D850 have a built-in viewfinder shutter that closes off the viewfinder for long exposures. It's a small lever on the side of the viewfinder.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Lou Jost
Posts: 5933
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

"Those ARE NOT PDAF stripes simply because:"

Yes, that's what I've been trying to tell you. Your dark-frame pattern has nothing to do with the striping pattern that is usually discussed under that name with respect to Sony and other sensors.

Post Reply Previous topicNext topic