Seeing any Posterization issues with A7r ll?

[Guy Mancuso]

I'll post images later but in real world shooting I shot silent shutter 12 bit and normal shutter 14 bit. I'll be damned if I can see any difference . I'll look more but first blush. Not sure I see a argument visually

 

[Knorp]

All I see is more and more posterization in this thread ...

 

[dmward]

Reading through this thread, here's what I understand:
If I shoot continuous bracket, which I do often, then I get a 12 bit file with lossy compression. The reason I shot continuous bracket is so I can combine the three frames, 2 EV apart, into a 32bit floating point file to ensure I have the highlight and shadows without clipping or blocking. That appears to make the difference between 12 and 14 bit irrelevant. What I've noticed is that most often, the normal or under exposed frame has sufficient dynamic range to cover both highlights and shadows.

For example:



Electronic front curtain shutter means 12 bit lossy file. Mostly using the electronic front curtain shutter is going to be important either on a tripod to minimize camera movement or to reduce noise. The tripod mounted situation is probably going to include bracketing. The noise concern is most likely to be in a situation indoors where either there will be really contrasty light and the shadows are going to have to block up regardless. I shoot weddings so its something to check. Luckily I have an opportunity to do that this Saturday. Two well known Chicago landmarks for testing; Old St. Pat's and Trump Tower. I'll let you know what I find.

Other people testing is interesting but without a lot more information its subject to a wide interpretation. Having to subscribe to read the details is a deterrent.

Based on my experience, there is always room for interpretation. Everything that is built to a specification includes tolerances. Its the only way to make anything affordable. I also know from experience, that one can use the tolerances to manipulate the outcome. i.e. a friend wanted maximum horse power from the engine in his acrobatic airplane so I built the engine taking care to minimize all the tolerances that related to compression. And, made sure all the bearing tolerances were a little past the midpoint to reduce friction and improve oil flow.

The real significance is that there will always be an opportunity to find fault. The reality is that 98% of the time the fault will be with my technique rather than equipment limitations. So, I focus on improving my technique.

 

[pegelli]

Electronic front curtain shutter means 12 bit lossy file.

Is this so? I understood that Silent Shutter is 12 bit and that EFCS is lossy 14 bit (or lossy 13 bit +1 as some people call it). Happy to be corrected if I'm wrong.

 

[ErikKaffehr]

Hi,

There is a thread over at DPReview discussing the issue at hand:
http://www.dpreview.com/forums/post/56349109

There are some really knowledgable guys on that thread, one of those, Iliah Borg has analysed the actual raw file and found that the main culprit is conversion from Prophoto RGB to Adobe RGB that causes extensive clipping. So I would say that it is a processing error by Diglloyd.

Personally I don't expect we can count on a firmware fix for the tone-curve compression, that is mapping the 14 bit linear space into a non-linear 11-bit space. The raeson for this is that I would suggest the Sony ASICs are relly only supporting a 12-bit data path.

Now, Sony also does a block compression that goes over 16 pixel blocks. It is actually non-lossy as long as gradients are small, but can cause artefacts with large gradients, think of sun illuminated bright hair agains black bacground.

Well, I think Lloyd makes a bit to much noise about this. Also I think he should release the raw image, albeit I understand the issues Tim Ashley rises.

I would add that Lloyd (Diglloyd) does publish a lot of good stuff. he was first to detect the shutter vibration on the A7r, for instance, but he also failed to understand the real cause of it, namely the vibration caused by the front shutter blade.

Best regards
Erik Kaffehr

 

[mark1958]

This is my understanding as well

Is this so? I understood that Silent Shutter is 12 bit and that EFCS is lossy 14 bit (or lossy 13 bit +1 as some people call it). Happy to be corrected if I'm wrong.

 

[Viramati]

Electronic front curtain shutter means 12 bit lossy file. Mostly using the electronic front curtain shutter is going to be important either on a tripod to minimize camera movement or to reduce noise. .

in my very unscientific testing I compared my A7s and A7rII both without EFCS on and you can feel in your hand that the A7rII shutter mechanism is noticeably 'damped' compared to A7s as Sony claims. I never had the A7r but I am wondering if this newly dampened shutter mechanism will go a long way to getting rid of any shutter shock at lower speeds on a tripod.

 

[davidstock]

Hi,

"There is a thread over at DPReview discussing the issue at hand:
http://www.dpreview.com/forums/post/56349109

There are some really knowledgable guys on that thread, one of those, Iliah Borg has analysed the actual raw file and found that the main culprit is conversion from Prophoto RGB to Adobe RGB that causes extensive clipping. So I would say that it is a processing error by Diglloyd."

Actually, if you read further in the thread, Iliah Borg agrees with Lloyd Chambers that there is in fact a problem with the red channel of the RAW file; it is not just a processing issue in his opinion. Still to be determined, AFAIK, is whether the use of a polarizing filter played a significant role here. And, of course, how often this issue is likely to arise.

I know nobody likes to hear anything bad about their camera, especially when it is new. But every camera has limitations and weaknesses, and I personally am interested in finding out about them. I don't take it as a personal insult when I hear that my cameras (or lenses) might have an issue or two. I want to know. Even if those issues only show up in extreme cases. Don't we want to optimize our equipment, weaknesses and all?

I suggest caution and open-mindedness.

Remember, some people said (and still say) that there is virtually no shutter shock with the A7R. People said (and still say) that there is no proof that Sony's RAW compression scheme ever leads to image degradation. People said (and still say) that the original lens mount used by the A7 cameras was plenty strong enough--or even that it was better than Sony's new (whoops!) replacement stainless steel mount. People say there are no RAW artifacts at bright/dark transitions in A7RII files. And people sometimes blame the messenger, only to realize later on that the messenger had a point. Kind of embarrassing, right?

We can disagree about technical stuff and raise all the objections we want. We can criticize perceived errors or bias. But I don't think we need to launch personal attacks.

I used to subscribe to some of Lloyd Chambers' sites. I didn't always agree with him, but I found a lot of his work very useful. He's overblown a few issues, but he's raised several that were quite legitimate. He's a skilled photographer, who takes photo gear seriously. He's been early to highlight a number of important issues with the digital imaging chain and with various cameras--issues that others jumped on board with later. If you aren't aware of those, you might want to go back and do some research. He's irreverent towards the camera and software companies, including Leica and Adobe, Canon and Nikon. And Sony. Good for him.

We'll find out soon enough if Lloyd is right, wrong, or overreacting in this case. IMO, he doesn't deserve to be disrespected. Just my 2 cents.

--d

 

[tn1krr]

Electronic front curtain shutter means 12 bit lossy file.

Wrong. EFCS does not drop bit depth. Continuous bracketing/shutter, LENR, silent shutter and bulb do.

Single bracketing is still 14 bit, so bracketing with the IR/wired remote is just fine. The only one that annoys me is bulb, 12 bit and with spatial filtering ("Star eater" as they say).

The shutter mechanism in the A7R II is more damped indeed vs. the old A7R, but there really is no reason to disable EFCS unless one encounters the exposure "evenness" problems with very high shutter speeds.

 

[tashley]

I second Davidstock above. I have also read the entire DPR thread and can't see where Iliah claims to have seen the RAW file. But he quite clearly agrees with a fair part of what Lloyd says.

And David, I was about to post a link to one of the threads in which people manned the barricades to defend the shutter in the A7r. That makes interesting historical reading. Thanks for making the point, which I think is highly valid.

 

[tn1krr]

I second Davidstock above. I have also read the entire DPR thread and can't see where Iliah claims to have seen the RAW file. But he quite clearly agrees with a fair part of what Lloyd says.

And David, I was about to post a link to one of the threads in which people manned the barricades to defend the shutter in the A7r. That makes interesting historical reading. Thanks for making the point, which I think is highly valid.

Iliah has seen and analyzed the RAW file and he still agrees the RAW format plays a real part in this.

There was indeed huge barricades build to deny existence of shutter shock and artefacts in star trails. Until the the photographer (local guy from here, a super-dedicated landscape shooter) who took the star trail photo released the RAW it was labelled as fake/fabrication, after that it was "wrong exposure, wrong raw converter..." none of which made any sense whatsoever.

As for LLoyd, his representation or choise of words of issues may not please everyone but his hit rate on real issues is very very high.

 

[tn1krr]

I second Davidstock above. I have also read the entire DPR thread and can't see where Iliah claims to have seen the RAW file. But he quite clearly agrees with a fair part of what Lloyd says.

And David, I was about to post a link to one of the threads in which people manned the barricades to defend the shutter in the A7r. That makes interesting historical reading. Thanks for making the point, which I think is highly valid.

Iliah has seen and analyzed the RAW file and he still agrees the RAW format plays a real part in this.

There was indeed huge barricades build to deny existence of shutter shock and artefacts in star trails. For example, until the the photographer (local guy from here, a super-dedicated landscape shooter) who took the infamous star trail photo released the RAW it was labelled as fake/fabrication, after that it was "wrong exposure, wrong raw converter..." none of which made any sense whatsoever.

As for LLoyd, his representation or choice of words of issues may not please everyone but his hit rate on real issues is very very high. He is a reviewer, not a cheerleader. All fanboys hate him, because he exposes/highlight issues on all camera brands.

 

[pegelli]

I agree with the point that Lloyd deserves respect. It's more the hate boys that immediately react to the issue he reported that could use some lessons in moderation, as well as the ones now putting all kinds of **** over what he reported and how he reported it. Immediately jumping on the lossy compression bandwagon and beating that horse to its ump-teens death is what doesn't add any value (and was plain wrong).

It may not be a camera issue but what we learned is that underexposing deep blues can create havoc in Adobe RGB (and sRGB). It's something we probably subconsciously know but bringing it out specifically like this is something to benefit from and not to criticize too much. We all make mistakes, reviewers and non-reviewers, just keep an open mind and move on is usually the best course of action.

 

[waardij]

My understanding is that the question is whether what is seen in the file form Lloyd is due to the Sony compression or not (or maybe it is not, and I just misunderstood). I am sure the red channel really is without information in the RAW file. The problem in blaming this on the compression, is that there is no compression in low contrast shadows. anyone doubting this can search back to the explanation the rawdigger site has published on the Sony compression (or maybe it was Lloyd publishing there findings on his site). and concerning the Sony files being less up to extreme post processing; this could be the case, I do not know, but I do know the Sony a900 files where more 'brittle' compared to those out of a D3x, while the a900 did have a uncompressed RAW mode. and turning this on or off did little to change the difference with the D3x.
I am sure there are differences between the d810 and the a7rii, but one might be careful in claiming to understand the underlying mechanism.

 

[Amin]

I have also read the entire DPR thread and can't see where Iliah claims to have seen the RAW file.

Right here: Re: A7RII posterization?: Sony Alpha Full Frame E-mount Talk Forum: Digital Photography Review

mdcromer: "I take it you have looked at the RAW file and analyzed it?"

Iliah: "Yes, I did."

... But he quite clearly agrees with a fair part of what Lloyd says.

There are two reasons I quoted Iliah en bloc in this post. First, because there is a high likelihood that DPR will delete that thread like they deleted the first one. Second, because those quotes clearly show that while Iliah feels that the Sony RAW format plays a role, he thinks the largest contributor to what we're seeing is color management.

I quoted Iliah en bloc, ie no cherry picking. For clarity, I've now gone back to my post and highlighted the parts about Sony RAW format in cyan and parts about colour management in violet.

Keep in mind that Iliah came to these conclusion even before being told that a polarizer was used, which he subsequently said could also be contributory (added that quote to my post as well).

 

[Guy Mancuso]

I have to disagree here. I get paid to shoot by clients and the old saying is your only as good as your last shoot. Well the missing part of that is you lost the client. Lloyd gets paid to review and that same scenerio goes with him and you get paid to review. He is no diffrent and gets respect by earning it, it's not a given and your not allowed to make mistakes. I'm sorry if a photographer is held up to that standard and loses a client, the same goes for him or any other reviewer. You simple don't post anything until it is figured out and need to take responsibility for yourself just like a Pro shooter does. There is no one to blame , no situation to blame, no raw file or anything like that to blame. You put up bad info , you will get your *** burnt off. Welcome to the real world. If I can lose a paying client than I'm accountable for that. Same with him or anyone else. Reputations are earned and lost by your integrity.

I'm taking my wife in for MRI's this morning just imagine we get bad data. I know it's a different league but it's really not. You put your name on it, you own it good or bad.

Sorry there is no hall pass. I have no bitch with anyone but I do have a serious issue with bad data. I had a real issue with this whole Sony release and a ton of bad data. I'm not paying anyone for bad data. the word Professional is not just a word, it's a responsibility .

 

[Vivek]

the word Professional is not just a word, it's a responsibility .

:clap:

 

[tashley]

I've had another look at the RAW file this morning in both C1 and LR with a variety of different settings from default in both to various tweaks in both. All this on a monitor that is both calibrated and has approx. a full Adobe RGB gamut, so one of the better ones out there. I have also had a look at the file in RawDigger and it is very gappy. Please note that RawDigger is not gamut-dependent, it is merely looking at all the information in the file and showing us where there is and where there is not data.

We will never know what effect the polariser had on this but I am tempted to agree with both Iliah and with Lloyd and with Alex Tutabalin of RAWDigger that the file format plays a significant part in the effect.

The file is pretty much unusable though to be fair I haven't tried gamut re-mapping, because it seems to me from looking at the file that there's not enough variation of tones there to re-map and because I don't, frankly, want to spend my time gamut remapping.

There is an unquantifiable chance that shooting this without a polariser and with more exposure would ameliorate this to a degree, but I can't guess how much of a degree. My guess is that the posterisation would be there to some extent.

From this I take away the following:

The lossy file format is just what it says. Generally it won't matter, sometimes, probably rarely, it might and I have little doubt that if data is thrown away, however skilfully, then the files will contain less information, under certain shooting situations, than those from a D810 (resolution aside). In other words the D810 will have a slightly wider shooting envelope in some respects, though the possibility is that a slightly wider DR for the RII might give it a wider shooting envelope in other respects. Horses for courses.

Put it this way, if I were in front of a tricky scene with both cameras and a wide choice of glass, I'd reach for the D810 unless I really thought the extra tad of DR were that vital.

As it is, I have both cameras. My initial view was that I'd spend some time with the MkII and then decide if I was going to sell the D810.

The D810 stays. I won't travel with it if the travel is casual but if I were off somewhere that mattered photographically I would.

We all see differently (otherwise there would only be need for one photographer, though (s)he'd be pretty busy!) and we all respond to different systems and their files in different ways. This is my way, and I don't expect it to suit everyone or indeed anyone else.

 

[GrahamWelland]

I used to subscribe to some of Lloyd Chambers' sites. I didn't always agree with him, but I found a lot of his work very useful. He's overblown a few issues, but he's raised several that were quite legitimate. He's a skilled photographer, who takes photo gear seriously. He's been early to highlight a number of important issues with the digital imaging chain and with various cameras--issues that others jumped on board with later. If you aren't aware of those, you might want to go back and do some research. He's irreverent towards the camera and software companies, including Leica and Adobe, Canon and Nikon. And Sony. Good for him.

We'll find out soon enough if Lloyd is right, wrong, or overreacting in this case. IMO, he doesn't deserve to be disrespected. Just my 2 cents.

--d

Me too.

I don't disagree with Lloyd's rigor but I do think that he gets over zealous about issues. They may be real, as they are with ALL systems, but to read Lloyd's normal synopsis you'd think that the cameras/lenses are all crap and they evidently are not. To be fair to him at least this time he's accepting that the problems are only evident in pretty much an outlier situation, albeit a random one which is probably what is making people lose their minds over this.

when it comes to the perfect camera / lens, here's the reality - it doesn't exist (well other than my Alpa/tech glass :ROTFL: Just kidding! Really.)

 

[Amin]

Lloyd initially posted this example of posterization without mentioning use of the polarizer, without recognizing the role of color management, and prior to trying in other raw converters (only afterwards did he try Iridient).

If anyone doubts that trying a different RAW converter prior to publishing would have been a good idea from such an experienced reviewer, see this: https://helpx.adobe.com/lightroom/kb/camera-standard-profile-displays-posterized.html

And now in retrospect, he says in his new blog post, "I knew the blue was intense and had I stopped to consider, I’d have known that this could be problematic for the red channel". Would have been nice to include this in the first blog post.

So what began as being "flabbergasted" at results which he suggested were due to Sony's RAW format now turns into "the Sony A7R II offers a coarse histogram unsuitable for discerning issues in small areas of the frame, and it offers no RGB histogram just for the zoomed-in area... I have also observed misleading errors in the Sony RGB histogram when comparing to the actual raw data in RawDigger..."

I for one am not "disrespecting" Lloyd. Owning multiple systems, I have no reason to defend one of them. I am a long time Leica user, and neither Lloyd's criticisms of Leica gear nor this Sony stuff bothers my in any personal way. I think that for the most part he does a good job. However, I think we can see that Lloyd is simply less rigorous in his approach than he is being given credit for by some here. At least on some occasions, he publishes findings quickly without details of methodology (eg, polarizer use), source files (raw availability for download), or serious investigation. Then he goes back and revises/updates afterwards. We saw this when he recently posted about abject Leica Q failure that turned out to be something other than that.