p.5 #1 · Sony compression creates artifacts in A7r ?
Snapsy gave some good advice that I should back off and let the thread go organic. I am going out to replicate the shot with my NEX7 and the A7r.
I also asked the Nikon 800/e guys for their input hoping they will add to the number of views in this forum.
I am taking my two best lenses out to test this - the 85mm 1.2 and the 135mm APO.
As well, I will try some other compositions - I think a pure sky gradient will be easy for the algorithm as will a complicated (leafy trees scene) the 'problem areas' are only in these middle ground tonal transitions (sky reflected into muddy green shaded water)
Maybe it's 1 in a 1000 issue for a photo but like many I want a camera I can trust. I can't afford to go back to Switzerland or Istanbul or New Zealand to reshoot what my camera failed to capture correctly.
It's much cheaper to buy a better camera/monitor/capture software or learn someone's better process.
The RAW file is here for anyone that wants to give it a try.
p.5 #2 · Sony compression creates artifacts in A7r ?
MaxBerlin wrote:
I am going to look at the images on a smaller - lower res monitor and lesser video card.
Maybe I will finally be able to not see what everyone else is not seeing.
I have updated my post with smaller jpgs of the images I was using - you can see how the shading of the sky could give problems but - until I have a new monitor - I think I saw a monitor related issue.
The shots I put on Flickr are 4 from A7r and one from NEX-7. I can't do an exact comparison as I sold the NEX but I am happier with the tree shots from the A7r due to better gradation of the foliage on the trees in the background. Obviously (and unfortunately) the NEX and A7r shots were on different days but lighting was similar.
I can appreciate you want certain things from the camera and you're not getting them but at the moment I don't see a problem. However I do have access to an A0 printer so who knows...........
p.5 #3 · Sony compression creates artifacts in A7r ?
I think max is too far in the weeds on this, but the artifacts remind me some of the native ISO and multiples issue/artifact that was observed with the a900. Anyone remember that?
As it was, you could see a slightly more noisy (magenta noise mostly) with the camera when used at ISO 100, but at 160, it was cleaner.
You really had to peep to see the differences... Slightly elevated noise in the pixel grain, and a slight magenta shift in that grain. Once documented, I got in the habit of using the 160 base for my shooting. It meant very slightly cleaner files. No big deal either way, unless you are going to really be pushing the file curves.
As I recall, it was all documented in FM.
I suggest max consider testing different ISOs to see if there is a difference in the performance.
p.5 #4 · Sony compression creates artifacts in A7r ?
mjm6 wrote:
I think max is too far in the weeds on this, but the artifacts remind me some of the native ISO and multiples issue/artifact that was observed with the a900. Anyone remember that?
As it was, you could see a slightly more noisy (magenta noise mostly) with the camera when used at ISO 100, but at 160, it was cleaner.
You really had to peep to see the differences... Slightly elevated noise in the pixel grain, and a slight magenta shift in that grain. Once documented, I got in the habit of using the 160 base for my shooting. It meant very slightly cleaner files. No big deal either way, unless you are going to really be pushing the file curves.
As I recall, it was all documented in FM.
I suggest max consider testing different ISOs to see if there is a difference in the performance.
I was reminded of the same, but I was afraid to open that can of worms. I agree that ISO 160 did clean up the A900's shadows, but it also did add a bit of grain to mid tones and slightly decrease DR, so it was a trade off.
p.5 #6 · Sony compression creates artifacts in A7r ?
I sincerely appreciate the help being offered. I want to be able to keep and use this camera. I would try the A7 if not, then back to my NEX7 and perhaps the leap to an 800e if I have to (which there is some form of regret today for having waited)
p.5 #7 · Sony compression creates artifacts in A7r ?
I think this thread may be on to something.
I didn't have a Coke bottle, so I was forced to take a picture of my cat. I converted it to a DNG and back and forth from SRGB to ProPhoto a few times as that is my workflow.
Looking at his whiskers at 800% it's clear that there are some strange pixels there. I've looked at it on my 30 inch calibrated Apple Cinema Display, as well my Google Glass and a Zenith TV and even when I turn my head sideways the pixels are still there.
Regretfully, I'll be returning the A7R not only due to this defect, but also because of the "crap" implementation of exposure bracketing. Not to mention that the shutter shock is giving me carpal tunnel syndrome.
On top of all this, the camera is an ugly shade of black. My colorimeter measurements of the battery cover show RGB values of 1,2,1 not 0,0,0 as you would expect if Sony had any idea how to dye plastic. Of course I made those measurements in the Northern Hemisphere, so I plan to redo them in the Southern Hemisphere later today to double check.
Anyway ... see for yourselves .... WHO'S CRAZY NOW, HUH?
p.5 #8 · Sony compression creates artifacts in A7r ?
mjm6 wrote:
I think max is too far in the weeds on this, but the artifacts remind me some of the native ISO and multiples issue/artifact that was observed with the a900. Anyone remember that?
As it was, you could see a slightly more noisy (magenta noise mostly) with the camera when used at ISO 100, but at 160, it was cleaner.
You really had to peep to see the differences... Slightly elevated noise in the pixel grain, and a slight magenta shift in that grain. Once documented, I got in the habit of using the 160 base for my shooting. It meant very slightly cleaner files. No big deal either way, unless you are going to really be pushing the file curves.
As I recall, it was all documented in FM.
I suggest max consider testing different ISOs to see if there is a difference in the performance.
I never saw it with my a900 but that could be because I used ISO 160 after doing extensive testing for best ISO on that camera (a controversial subject I know).
p.5 #10 · Sony compression creates artifacts in A7r ?
ebrandon wrote:
I think this thread may be on to something. I didn't have a Coke bottle, so I was forced to take a picture of my cat. I converted it to a DNG and back and forth from SRGB to ProPhoto a few times as that is my workflow. [...]
Anyway ... see for yourselves .... WHO'S CRAZY NOW, HUH?
p.5 #11 · Sony compression creates artifacts in A7r ?
Well, all kidding aside. Could one of our resident technical experts look at this comparative raw histogram out of RawDigger and explain what they see (or what it means) with regard to the dramatic difference in number of "values" for the Sony RX1 vs Pentax K3? Same scene composed the same taken within minutes of one another (as the shape of the histogram also indicates).
p.5 #12 · Sony compression creates artifacts in A7r ?
I just gotta let you guys know that even today's modern digital medium format cameras can produce some nasty looking images when viewed at 100% and higher. Yet, they are the standard for important portraitures and fashion. By comparison, the modern Sony 36mp sensor is way cleaner than any of the medium format sensors. If this sensor is not good enough, then it's time to put away all the photographic equipment and wait for a future sensor.
While I have yet seen a problem with the A7R images, I do know it is Sony's philosophy to not clean up stray noise. Other manufacturers do some light noise cleaning in their RAW files. I definitely prefer Sony's approach as I have better noise suppression software on my computer and can control how much I want to clean, depending on the image's subject.
Also, I have respect for Lloyd. I have learned a lot from him and he has helped me with an issue in the past. However, he does sometimes go too far into the deep end, in pursuit for the cleanest image possible. I couldn't believe it when he trashed Hasselblad's digitial medium format cameras because it didn't reach the level of perfection he expected from the lens and files. In the end, we look at the final product, not at 100%, and to me, I wish 35mm FF could produce the colors I saw from the Hasselblad's files.
p.5 #13 · Sony compression creates artifacts in A7r ?
ebrandon wrote:
I think this thread may be on to something.
I didn't have a Coke bottle, so I was forced to take a picture of my cat. I converted it to a DNG and back and forth from SRGB to ProPhoto a few times as that is my workflow.
Looking at his whiskers at 800% it's clear that there are some strange pixels there. I've looked at it on my 30 inch calibrated Apple Cinema Display, as well my Google Glass and a Zenith TV and even when I turn my head sideways the pixels are still there.
Did you look at your cat's whiskers under magnification to ensure there weren't any strange pixels on them to begin with? Also, we can't really make a determination based on a cat photo, we need a goose photo. I assume your cat isn't losslessly compressed?
p.5 #14 · Sony compression creates artifacts in A7r ?
sflxn wrote:
I just gotta let you guys know that even today's modern digital medium format cameras can produce some nasty looking images when viewed at 100% and higher. Yet, they are the standard for important portraitures and fashion. By comparison, the modern Sony 36mp sensor is way cleaner than any of the medium format sensors. If this sensor is not good enough, then it's time to put away all the photographic equipment and wait for a future sensor.
While I have yet seen a problem with the A7R images, I do know it is Sony's philosophy to not clean up stray noise. Other manufacturers do some light noise cleaning in their RAW files. I definitely prefer Sony's approach as I have better noise suppression software on my computer and can control how much I want to clean, depending on the image's subject.
Also, I have respect for Lloyd. I have learned a lot from him and he has helped me with an issue in the past. However, he does sometimes go too far into the deep end, in pursuit for the cleanest image possible. I couldn't believe it when he trashed Hasselblad's digitial medium format cameras because it didn't reach the level of perfection he expected from the lens and files. In the end, we look at the final product, not at 100%, and to me, I wish 35mm FF could produce the colors I saw from the Hasselblad's files....Show more →
I think it's worth noting that at least some of the discussion here is about lossy compressed raws (not "Lossless" as the thread title incorrectly states) rather than the quality of Sony sensors. I know a huge amount of content within this thread (as well as reaction) seems to wondering into DpReview type forum territory.
p.5 #15 · Sony compression creates artifacts in A7r ?
Few things come to mind for me on this ...
1) 8X or 11X viewing ... I appreciate the scrutiny of microscopy (applaud the effort, I think) as much as the next guy, but I wouldn't kick Marilyn Monroe off the set for having a freckle smaller than a mustard seed between her toes.
2) Does Sony have OEM native software for processing rather than relying on Adobe's algorithm's?
3) The profile differences between sRGB, aRGB and ProPhotoRGB progressively "stretch/push" things farther mathematically with bigger "steps" for a given input value ... the more you push, the more artifacts you are prone to.
4) If you haven't been processing non-AA filtered images, and think you should process them the same as AA filtered images ... you're going to have a tendency to be over processing if you use the same approach without first learning the diffs. Learning to process Sony's Lossy/Lossless/ files might have a learning curve to them that don't follow the same paradigm as before.
When I switched to Canon files ... new learning curve, picked up my SLR/C ... new learning curve @ A7R not unreasonable to anticipate a new learning curve for dealing with lossy/lossless/ files.
That being said ... I like my RAW's to be uncooked and if Sony is taking some liberties, I appreciate being made aware of those liberties. Whether or not they make significant diff or are a "smart" practicality will be highly debatable for some. I'd like to take a stab at the RAW but don't feel much like working a 36MP file atm. Maybe a couple manageable crop sections of the raw would be appreciated to play with.
There seems to be some "yea", some "nay" at the orange peel (here and elsewhere). I'm inclined to think that this will be realized as relegated to algorithm, processing, profile approaches and a need for some old dogs to learn a new trick (or two).
Worth keeping an eye to as we continue to learn more about the A7R ... which raises the question @ did we ever confirm the A7R no-AA filter as truly non-existent, or is it similar to the D800E approach?
Now, where was that freckle again ... or were you too busy looking elsewhere to notice?
p.5 #16 · Sony compression creates artifacts in A7r ?
Thanks Tariq. It's easy territory to go stand up and my points could be better phrased but I see an issue and am willing to accept any help on getting these files to look better.
The camera is a dream in the field, I just shot a number of test shots and feel great about what I saw in the EVF. But a little hesitation in to how the outputs will look.
p.5 #17 · Sony compression creates artifacts in A7r ?
douglasf13 wrote:
I was reminded of the same, but I was afraid to open that can of worms. I agree that ISO 160 did clean up the A900's shadows, but it also did add a bit of grain to mid tones and slightly decrease DR, so it was a trade off.
Hahaha... Yes, I thought of not mentioning it as well... Regardless, 160 did clean up the shadows, and the grain overall was a reasonable compromise in my experience.
I have no doubt that the a7r will smoke that older sensor, and I've been very pleased with it since the beginning, so I am happily expectant of the results once i have an a7r in my hands (this week).
I still shoot the a900 and will be doing extensive comparisons with the a900, the a7r, an M9 and both Leica M and Leica R lenses that I have been using on the a900.
p.5 #18 · Sony compression creates artifacts in A7r ?
'I can't afford to go back to Switzerland or Istanbul or New Zealand to reshoot what my camera failed to capture correctly.'
Max, you're turning me into a quivering mass of jelly with statements like that.
ebrandon, beautifully done, just enough bait to hook the unwary, lol.
'the Nikon 800/e guys for their input'
I call these guys 'the usual suspects', a reference to Casablanca for non movie buffs. They appear to be the current space invaders of the Sony part of the Alt forum, the Canon users have too much decorum.
There has never been a perfect camera made, and it just goes to show if you want to look past actual photography into the dark tunnel of signal processing and matters QA in minute detail, you never know what horrors await! Lloyd will be there to hold your hand however, a fellow traveller down that dark road.
Max, I could point you to the travails of Joseph Holmes - a near genius in my view - who had huge problems with hugely expensive MF digital systems and lenses, matter of fact I will do just that:
p.5 #19 · Sony compression creates artifacts in A7r ?
Excuse me, I've read through the first two pages twice and still dont grasp the problem. Could somebody explain it in laymen terms? I AM interested in photography so no phun or irony intended, for real.
p.5 #20 · Sony compression creates artifacts in A7r ?
philip_pj wrote:
I call these guys 'the usual suspects', a reference to Casablanca for non movie buffs. They appear to be the current space invaders of the Sony part of the Alt forum, the Canon users have too much decorum.
"I'm shocked, shocked to find a lossy algorithm in my camera."
Max, I could point you to the travails of Joseph Holmes - a near genius in my view - who had huge problems with hugely expensive MF digital systems and lenses, matter of fact I will do just that:
Archive 2013 · Sony compression creates artifacts in A7r ?
I agree that ISO 160 did clean up the A900's shadows, but it also did add a bit of grain to mid tones and slightly decrease DR, so it was a trade off.
files might have a learning curve to them that don't follow the same paradigm as before. 
