The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Seeing any Posterization issues with A7r ll?

Amin

Active member
It looks like camera "sensitivity" calibration, as the gaps are at the same data numbers in all channels. May be vignetting compensation, for example, or some other compensation.
Someone said that the EXIF showed that Lloyd was using lens correction settings for color fringing. Don't know if he was also doing vignetting comp. I always turn all that stuff off and wonder if the gaps would still be present in that case, and whether in turn the "posterization" would still be present.
 

tashley

Subscriber Member
For this shot, the problem is not the compression, but missing bits in deep shadows of the red channel, that is decompression resulting in 13-bit data, and with some additional gaps (totally missing data numbers where they need to be present).
Thanks Iliah, but because I am no genius, could you spell it out for me: if the problem is in decompression but is seen in LR, Iridient and C1, then what causes the decompression problem other than the way in which the data was compressed? I am confused!
 

ErikKaffehr

Well-known member
Hi,

I don't see banding, but out of gamut colours is very common on flowers and vegetation. I don't really know how to handle them...

Best regards
Erik

Tim, my interest in all of this is mostly just intellectual curiosity. I'm far too casual a photographer to care for any practical purposes.

I'm convinced that lossy raw can cause some issues with A7RII files. I've seen a bunch of examples of this. I'm not sure if what Lloyd is pointing out with orange peel or posterization represents an issue with the Sony lossy compression, the raw support, the camera settings affecting raw, or something else. I'm interested to dig and learn more.

Again, I couldn't care less for the purposes of my own photography. I didn't care about my Fuji X10 orbs, my Nikon D600 oil spots, or my M240 banding. None of this stuff bothers me. I do have a longstanding bias towards keeping raw as raw as possible, but that's just an ideal for me.




It's very clear from the DPR thread that Iliah reproduced Lloyd's findings and found that the bulk of the posterization "damage" happened when going from Pro Photo RGB to Adobe RGB. This doesn't mean that Lloyd processed it wrong or was not cognizant of gamut issues (though he omitted discussion of their contribution until his second blog post).

There's only one way to convert to Adobe RGB in Lightroom. The methods Iliah mentions which could have prevented the bulk of the "damage" involve specific manipulations during Photoshop soft proofing, manipulations that few of us would do. So it is simultaneously predominantly a color gamut issue as well as a real issue contributed to by raw format, according to Iliah.
 

Iliah Borg

New member
Can you explain the solution to this problem?
First and foremost, try to not underexpose, that means per channel underexposure, not just overall exposure to the right. Consider this: red channel is week, and white balance multiplier for the red channel is 2.4 on this shot, that is even with green channel exposed to the right the red channel is still underexposed more than 1 stop. For daylight shooting a magenta filter (like a Lee CC40M / CC50M) on the lens will bring the channels more to the balance.

Second, do not trust numbers like 14-bit dynamic range, confine yourself to sane 8 to 9 stops, 10 max.

Third, it is not always that the cleanest shadows are at base ISO, sometimes photographically next ISO setting results in better images, test it.

Forth, exercise conversions between colour spaces with caution, check for any gamut issues before converting and address those.

That might not look like a solution, but it helps mitigate the problem.
 
V

Vivek

Guest
First and foremost, try to not underexpose, that means per channel underexposure, not just overall exposure to the right. Consider this: red channel is week, and white balance multiplier for the red channel is 2.4 on this shot, that is even with green channel exposed to the right the red channel is still underexposed more than 1 stop. For daylight shooting a magenta filter (like a Lee CC40M / CC50M) on the lens will bring the channels more to the balance.

Second, do not trust numbers like 14-bit dynamic range, confine yourself to sane 8 to 9 stops, 10 max.

Third, it is not always that the cleanest shadows are at base ISO, sometimes photographically next ISO setting results in better images, test it.

Forth, exercise conversions between colour spaces with caution, check for any gamut issues before converting and address those.

That might not look like a solution, but it helps mitigate the problem.
Thank you for that.

So, it is possible that all the folks who have not seen this posterization may be doing things right to capture photos.
 

k-hawinkler

Well-known member
Thank you for that.

So, it is possible that all the folks who have not seen this posterization may be doing things right to capture photos.
Or get lucky? :grin:

Vivek, thanks. I approach this subject from the point of view of eliminating possible sources of problems first, like for example lossy compression that should be straight forward for Sony to correct. Then focus on remaining issues. Until then workarounds are called for, like a Lee CC40M / CC50M filter as suggested by Iliah Borg.
 
V

Vivek

Guest
There are a lot of lucky people around, K-H. See post #132 from Michael. :grin:
 

davidstock

New member
I think this discussion, with Iliah's help, has sort of clarified the boundaries of the issue.

1. The artifacts Chambers showed are probably rare, but they are real.

2. There is data missing at a particular point in the red channel of the A7II RAW file. This is the root cause of the artifacts. The artifacts are not just, or even mainly, problems with RAW developers or color management.

3. The issue can be made worse by sub-optimal exposure and poor color management.

4. Rather strict measures might be necessary to mitigate the artifacts in the limited cases where they arise. Those measures might include appropriate filtration and careful analysis of red channel exposure (at a level of precision for which the in-camera histogram is inadequate). Another possible form of mitigation might be to shoot at an ISO slightly higher than base. (I don't know for sure why this would help, but I speculate that it might have to do with the dithering effect provided by the presence of faint noise.) This would certainly be easier than the other measures!

5. These mitigation measures still may not eliminate the artifacts altogether in all cases, but they would optimize our chances of a good file.

--d
 

Amin

Active member
David, I don't agree that the boundaries of the issue have been defined.

We don't know to what extent (if any) Lloyd's use of lens compensation settings affected the gaps in the raw data.

We don't know to what extent (if any) another camera with uncompressed or losslessly compressed 14-bit raw would have shown similar effects with the same subject, degree of relative underexposure, and color management.
 

Slingers

Active member
Thanks for this. I find it interesting that's it's the same artefact problem as the star trail image that is used everywhere but with the inverse of sky highlights against the dark background of the tree but the same rounding errors of the compression. It's a scenario I guess we need to be mindful of until the compression is fixed.
 

Guy Mancuso

Administrator, Instructor
Hey folks first I want to apologize here on this thread. I may have been a little aggressive in some of my comments so please do not take any offense to that. I simply hate bad data and its been going around on this Sony for reasons I am not sure and people that should be doing there job are in my mind not being very Professional. As you all know I have been in this a very long time and my passion level for this industry is extremely high. I made a good career out of it and it served me well so i care a great deal. Which sometime my red hot Italian blood does get to a boil once in awhile. So for that I am sorry. I also had stuff going on with a personal level that does concern my wife and her health. We have not won the battle but today we are on at least on some neutral ground. Some good news and some stuff we need to address. I get PM's from you folks and I love you all for it. I try not to talk to much about it but holding stuff in is not good for me either and this battle has been going for almost 6 years. In closing I care dearly for our members and I love the respect you show me even though there maybe days I don't deserve it. You are my friends and I do care about you all and your love for Photography..

Okay working on my test stuff. Hope to have this up soon. Im pretty blown away by what Im seeing. Have a great weekend
 

mjm6

Member
First and foremost, try to not underexpose, that means per channel underexposure, not just overall exposure to the right. Consider this: red channel is week, and white balance multiplier for the red channel is 2.4 on this shot, that is even with green channel exposed to the right the red channel is still underexposed more than 1 stop. For daylight shooting a magenta filter (like a Lee CC40M / CC50M) on the lens will bring the channels more to the balance.

I've been working with a variation of this to improve the file for B&W conversion! Shhhh... it's a secret.

Basically, if you are shooting for B&W conversion, you are at a distinct disadvantage if you intend to 'red' or 'orange' filter the image as part of the conversion because of exactly what Iliah has indicated in this quote.

Add to that all the baked in RAW adjustments that you don't even know are occurring, and you have possibly lost a huge amount of data from the Red channel.

Now, I don't know about you folks, but using a Red or Orange filter on the image in the traditional manner with film often produced a better aesthetic for my work than shooting without, and I certainly don't want to use a Blue filter unless you were aiming for the old Orthochromatic look. A Green sometimes worked OK for me, but normally, I favored the Y, OR, or R.

So, the point is to knock down the Green and Blue channels somewhat to raise the Red channel and ensure that you have sufficient information in the Red channel for your manipulation that you will be doing later in LR.

You don't really have to fret the G channel since it is much stronger than the R.

If you've ever done a B&W conversion using the filtration approach (rather than just combining the channels in to a single channel), and then had problems with posterization in the sky when you went to do any burning, you know why this approach is will produce a better result. Actually, I've seen posterization without even doing any burning...

I suppose I should do a writeup on this and become internet famous (and not for my account at VM!). Just kidding about the VM account.


---Michael
 

Guy Mancuso

Administrator, Instructor
Wanted to go back to this for a second because I have shot many times in Yosemite national park at the 10k and above elevations and we used to use with our ND filters up at that elevation IR ND filters because of the IR bleed at those heights Color is actually diffrent at these elevations because of the IR .

For our engineers folks and scientist would this not have a color effect. Certainly seems it since a IR ND works best. Just a thought
 

jlm

Workshop Member
in my high altitude experience a UV filter was required to get rid of the excess short wavelength (blue-UV) light less absorbed by the thinner atmosphere and reduced absorption path compared to lower elevations
 

tashley

Subscriber Member
in my high altitude experience a UV filter was required to get rid of the excess short wavelength (blue-UV) light less absorbed by the thinner atmosphere and reduced absorption path compared to lower elevations
so that might mean that the relative proportions of red to blue light were low? Not sure if Lloyd's polariser was also a uv?
 

Guy Mancuso

Administrator, Instructor
That was why I brought it up. It was something that was never mentioned was the elevation where these shots are taken . The light change is pretty big

Tim I'm not so sure a polarizer has that built in to it. Never seen one where that was advertised.
 

tashley

Subscriber Member
That was why I brought it up. It was something that was never mentioned was the elevation where these shots are taken . The light change is pretty big

Tim I'm not so sure a polarizer has that built in to it. Never seen one where that was advertised.
I'm sure I've got a couple knocking about somewhere. If my filter draw didn't make Pandora's Box look like a Scandinavian design statement, I'd look.
 

Guy Mancuso

Administrator, Instructor
Lol

Well I do think this elevation has a effect on that file. Hopefully our brain trust here can figure it out.
 
Top