Diglloyd posted two new A7rII articles:
Understanding Compression of the Sony A7R II Uncompressed ARW Raw Format
and
Reader Question: Convert to DNG for Sony’s Space-Wasting 14-bit Uncompressed RAW Files?.
And here are two key insights he shares about the new A7rII uncompressed RAW format:
1) The Sony A7R II captures 14 bits per photosite. The Sony uncompressed raw format then stores these 14 bits in 16 bits, bloating the file by 14%. In other words, Sony wastes about 9MB per image for no value or purpose at all. So Sony users lose triply: wasted storage, degraded camera responsiveness, fewer shots before the camera buffer fills.
2) A lossless-compressed format can save far more than even the bit packing mentioned above. Nikon and Canon use Huffman encoding. But it might be that the CPU in the Sony cameras is not fast enough for Huffman encoding?
In short, we would all save space, time and have a faster camera if Sony could add lossless-compressed files and not use the 16bit format. Let’s hope Sony has this on their todo list for the next firmware update!
–
Sony A7rII store links to [shopcountry 41860].
175 Responses
Lol
Some people are just half-empty type of guys.
The good news is if we know they are half-empty type of guys we can compress the hell out of them i suspect good compression results in the >80% range vs guys who are just full of it compression on them becomes messy quickly 🙂
Is this guy ever happy?
Don’t want to waste space? Then use compressed format. You won’t be able to tell 99% of the time, and let’s be blunt, Lloyd, using any lossless format isn’t going to improve your photographs above the level of distinctly average.
No he’s not, like us. 😉
Everything non-Sony makes him happy.
Who cares???
Surely smaller larger RAW whatever. May it will be nice to get some compression and lots of options. Good thing for Sony engineers to take on when nothing else is broken. This is not the case though. Uncompressed RAW and compressed RAW option is quite good if it were available on all models – A7ii, A77ii, A6000, A7r, A7s, A7…..even A5000, A5100
I would rate lack of 14bit RAW on crop cameras much higher if uncompressed RAW is released to all A7 cameras.
RX bro?
Oh man. Give us a f ing break.
The option is there for those who want it or need it.
I appreciate Lloyd’s contribution to Sony providing true 14 bit files for us to have as an option but we need not to be so pedantic, annoying and an al about this anymore.
No need.
It was already established the cRAW Sony used gave you 90+% of the perceptual IQ, so if you need to save storage, use it. If you need the other 10%, don’t b i t ch about it. It’s there. Use it wisely.
Actually, I think the point that losslessly compressed RAW should have been implemented is a legitimate one. Significant space savings can be made for free with a few lines of code and almost no processing power, which would also reduce write times. It’s not like this is an unknown technology requiring anything more than basic effort to implement. The frustration at Sony not bothering to go that extra half step and doing a bit of a half-arsed job does have merit.
I’m a software developer for embedded devices and I would store it exactly the same way. Especially in memory (Buffer). If you read out the “Pixel” the 14 Bit value is stored in a register that is 16 Bit wide. you copy that register into memory. If you want to safe some space and append it to the last value you several mcu operations which will take some time. That means not wasting the 2 Bits would make the responsiveness worse and not better.
I think now I want 15 bit because I’m just an average guy.
Actually Sony should just start writing the whole 16 bit and be over with this. Their pro level camcorders record 16 bit RAW. Why not have cameras record 16bit RAW.
I would believe they have 16 bit in stock just pruning 2 bits to be in line with industry….and may be not much of the RAW softwares can deal with 16bit.
Lol. Who didn’t expect that? We would all save time by not reading Diglloyd.
“the CPU in the Sony cameras is not fast enough for Huffman encoding”
I really doubt that. Lossless compressed RAW files are common across the industry, in cameras with older CPU’s.
Diglloyds complaints are justified but should be no surprise on a Sony camera. Sony cameras never were, and probably never will be, well thought-through, finished products. We know that, that’s the obvious backdraw if we want cutting edge specs.
If Sony would develop a camera like Leica, it wouldn’t have mind blowing specs, right? Everything is a compromise.
I think more and more, that it makes no sense to criticise Sony for the way they develop their products. Thats the way that brought them to this point, or, in other words, brought us to a camera with ISO 409.600. Sometimes, we (I’m talking especially to myself) should be happey with what we get, instead of complaining about, what we not have.
I am happy with what I have (RX1 and RX100 user since release) and while I wish Sony would have improved AF via firmware on the RX1 and release uncompressed RAW for RX1 as well, I do not become unhappy with what I have.
That being said, a company that can do these things has no excuse for not doing them. So, uncompressed RAW as a development of FW updates is a very, very welcome development.
If it can improve the product an it can be done, why not do it?
I never felt an inkling to get a7 series cameras because none can have a lens like the Sonnar nor its leaf shutter.
The RX1 was not for me, because of the AF, but I still consider it as one of the greatest cameras of all time. It’s just not fast enough for me. As I said several times before, the Q (although bigger) feels like a much more finished, refined product to me (I just struggle with the price).
However, we wouldn’t have the Q without Sony having the balls and bringing out the RX1. I’m very thankful for Sonys commitment to a fullframe compact. RX1 rocks, even with it’s little backdraws.
I think you do Sony a disservice!
To suggest that the A7 series cameras are not ‘thought through’ is to seriously underestimate the whole design process and to insult the team who will have had to consider every aspect.
Are they perfect? No, but neither are the Leica that you suggest are far superior. They have their issues too and some are pretty significant (IR colour shift, sensor delamination, etc.) I would suggest that Leica have far less of an excuse for any issues given the prices they charge.
Furthermore, I don’t think Diglloyd’s complaint is justified. Does he have any idea of the technicalities behind the release of the uncompressed RAW? Both he, and you with your support, imply as though some inexperienced office junior was given the task of creating the new firmware. I’m pretty sure there would have been a whole team that would have gone through every possibility, so I doubt there are ‘two wasted bits’ for no good reason.
I see differences in how Sony approaches camera design compared to Leica/Nikon/Canon. An I tried not to lower them down for that. I mentioned Leica, because it was the first coming in my mind. I could also have said Nikon. That’s my experience, my opinion.
No, I do not “imply as though some inexperienced office junior was given the task of creating the new firmware”. That’s what you want to read between the lines.
I could also have said Nikon too: Oil on sensor with the D600, AF issues
with the D800, light reflection with the D750, etc. I’m still not sure
how you judge the merits of ‘design’ and decide that Sony is somehow
slapdash about it.
As to my inference, I was mainly criticising Diglloyd. His assertion that “Sony wastes about 9MB per image for no value or purpose at all” can hardly be read any other way than they have done it through stupidity or inexperience, and you supported that position.
I simply believe that there are likely very good technical reasons why they have gone about the firmware update this way and, to repeat, it is a disservice to them to claim otherwise without very good evidence to the contrary.
It was more like a quick fix and i think with the next powerful processor we ll see lossless cRAW and 10bit video and better buffer, startup and operational speed.
As long as there is also a faster card interface!
Better bit-packing and Huffmann compression would be welcome but come for a price: CPU time, energy, heat. Before it is not analyzed in detail what lossless compression would add here, it is not justifiable to bash for “degraded camera responsiveness”.
Never ending ridiculousness. Sonys RAW is just fine. Uncompressed RAW is a waste of space. I have been printing 30×45 and 40×60 since the Nex5N, Nex7 and now the A7R. Never ever had an issue. I seriously wonder if the people are complaining are photographers at all or at least good ones. The amount of information that is the compressed RAW is un real. I cant make a fully black photo from a night shot look like day if I want. There is no reason to complain. Ill bet anything that the people that are complaining are also posting to the internet no bigger than 1200x****.
Actually… There was..but it’s not about printing big. Put a Big Stopper, and a polarizer on, and go shoot a big glass building. This is where dynamic range becomes important and why some of us are grateful they made the uncompressed. Now if c1will just update so we can use it 🙂
Can you see the difference full zoomed out or only at 1:1? Im asking because I’m seriously interested in knowing how much a difference the uncompressed RAW makes.
You could see it on the back of the camera, not even in a computer. But this is for the a7r that I’m speaking of. Haven’t done any testing with r2 as c1 can’t read it yet. One more thing. As you may or may not know, this only occurred when you switched to continuous shooting, or bulb. If your shooting single shot things are fine and using the ‘compressed’ file version works great. Personally it was about bulb mode. The studio work was and continue’s to be great. It’s just slowing it down for landscaping which hurt. Going forward I plan on using the uncompressed only when I need it, but am quite grateful to have it when necessary.
I have the A7R and 95% of the time I shoot in bulb (hundreds of photos a month). Long exposure is my specialty. Ive never noticed a problem with the blacks or with the image at all except for a few rare cases of banding. I thought the blub issue was with the A7RII where it drops it down to 10 bit (or at least thats what a friend told me). As far as Im aware there isn’t a firmware upgrade to uncompressed RAW for the A7R.
Maybe your big stopper is better than mine? May I ask what your using to slow things down?
I dont use anything in front of my lenses. All my exposure is night time.
Ahh. The issue only presented when dynamic range was challenging. The joke around the studio was if you pretended to be shooting a canon things were fine. my Big Stopper has a pretty harsh vignette, so the issue occurred for me on the regular around the edges of the frame.
I was actually going to say that. Shoot it on a 5D mkiii and see the results. Regardless of the compressed RAW its still very very good compared to the competition and price range.
It’s just like the 5dmkiii when using bulb mode. (Shot them side by side last year and figured all this out) Which would be fine if I didn’t leave canon for Sony. Also, we liked the 7r, and have been using it on commercial work pretty much since release Billboards = beautiful. We bought the r2 on release as well and Love it. But that 11+7 stuff will burn you from time to time, so to have the option is awesome.
Yeah, thats why I stayed with the A7R and didn’t get the A7RII. I think they have fixed that with the firmware upgrade on the A7RII but Im not sure.
The a7r uses 11+7 on continuous and bulb. 14bit with single. The recent firmware giving us the uncompressed, is what brought bulb and continuous up to single shot modes quality on the r2.
That’s actually pretty funny. “Oh, this? It’s from a Canon. Yeah. I know, right? What are you going to do?” ¯_(ツ)_/¯
Lol. Ya. Some of the commercial work we did last year would have been a deesasssdur on a Canon. Well, actually it could have been ok, just would have had to lug a ton more gear in.
Unless I’m mistaken, what you see on the back of the camera is a JPG.
It is, and usually it looks a bit better than the computer(small screen and all that). The point I was going for was, it was noticeable even there.
I do not understand those people like Diglloyd. If You want Huffman compression as it’s implemented in Nikon/Canon, then shot with Nikon/Canon. I am very happy with uncompressed file format and my camera behaves more promptly than with compressed File Format.
I am very grateful for Sony to replay for User option about uncompressed RAW file. Sony BionizX processor may be faster than those of Nikon/Canon but i may be not dedicated to support Huffman, which algorithm was not implemented in processor design.
Exaclty! Non-compression means faster response because CPU computation is not needed to compress anything.
Absolutely. And SD cards are dirt cheap these days, £25 for 64Gb high speed card in the UK, and one can always keep them compressed once on their computer.
But the entire point of Diglloyd’s article was that it would require barely any CPU computation… so you’d make gains in write speed anyway, so lossless compression actually ends up faster overall. It’s win-win, which is why it’s baffling that Sony didn’t implement it.
That all depends on the write speed of the SD card. You need a very high write speed to compensate for ‘normal’ compression (unless you use a very high compression rate, which is more computation). If writing a file takes 100 msec uncompressed and the file is halved by compression, the time it takes to write the compressed file is 50 msec. If the compression is faster than 50 msec, you win
In the real world, storage is cheap.
But not free. Which is what this solution would be.
Yeah, this is just ridiculous, with the price of a decent SD card alone you can easily get 1Tb HDD.
What is a decent SD card? Of what size? Of what speed?
It has nothing to do with CPU power. The problem is that Sony’s processor is designed/optimized for the compressed algorithm from the very beginning, so now they have very low efficiency dealing with the uncompressed data.
In other words, this will be solved in the Mk III. In the meantime, the lossy compressed RAWs will do just fine
Now please announce the new lenses.
It would simply be nice if they released firmware update for other models.
Why? I don’t see many Android phones being frequently updated after a year.
Because when i pay nearly 3,000 dollars for something, i don’t expect to go obsolete in less than a years time. Thats why. Are you seriously comparing a phone to a camera? Stop trolling.
Yes I am serious on that comparison. Except Google and Apple, no manufactures would keep pushing new systems after one year or so. Instead, they would ask you to buy a new phone. Only individual developers would port it for you.
Just buy a new camera when mk3 comes out and you will have whatever features you want. Yes, that sounds stupid. So I stopped paying for Sony cameras.
A7ii is almost year old…..but not quite….
You mean the camera that trades off IQ for battery-life?
Exactly. Like, how often did they swap out the circuit boards on the space shuttles? Like, never. So shut up Kelly!
The following “To test compressibility of Sony A7R II images, I used “gzip -9 -r”. Compressibility for this bracketed series ranged from 19% space savings to 38% space savings. Those figures could be less or more with other images, but it makes the point: Sony’s uncompressed format is hugely wasteful of storage space.” demonstrate that Digiloyd don’t know what is talking about, neither sonyalpharumors…
Comparing DEFLATE VS uncompressed is stupid, and has nothing with the 2 extras bits of data.
Anyway, I don’t know why Sony is still sharing such article…. speechless
SAR admin of course doesn’t know anything, he just likes the drama and clickbaits.
On the bright side, if Diglloyd is wrong, he will be humiliated by people such as you and Heiko.
This too is entertainment to look forward to.
Best to remember that this solves a problem a lot of us have never had – however, now when shooting start trails or night scenes (neither of which require fast shooting rates and deep buffers) we can now select the new mode (well those of us who have latest models). If they do the upgrade for a99 I will install the firmware, but I won’t lose any sleep if they don’t go back as far as the a99…
Here’s one key insight into Diglloyd: he loves to f***ing whinge.
Or is paid to whine. I know I’m a cynic, but I sense that he’s put up to this by the marketing divisions of Canikon. Really doesn’t make much sense otherwise–his whining. That, or he’s taking too many scripts from today’s political arena.
Well, paid to whine or not, let’s just agree his website looks faaaabulous! I wonder if he and that other pri*k, what’s his name? That frowny, no-talent hack that’s featured here every now and then? Anyway… they must have the same designer… ’cause talk about killing it in 2015, girl!
I think that’s funny. We had 12bits, some asked for 14 and got 16 and now they are unhappy.
You are missing the point. People want 14 bit since it’s valid for current sensor technology. Of course you could also make it 128 bit files, but you are storing a bunch of zero. On the other hand, lossless files, aka lossless compression, was asked for. Not uncompressed.
No matter what reason it is, Sony decided to leave room for mk3 models.
I never asked for either. What I received was lossless. Whether it’s
compressed or not is a minor technicality. The image quality surely
trumps all else for people who wish this mode.
Either way, I’d like to bring to your mind the consideration that you and Digilloyd might be wrong. Read Heiko and Josselin Cornou’s posts further below.
What Sony did is the best they could do, Sony saved CPU cycles, Diglloyd wants Sony to waste CPU cycles and make the camera slower.
Lossless, on most scenarios, is used to describe compression. Uncompressed is just uncompressed. Of course you could have lossy uncompressed, say 12 bit uncompressed files compared with 14 bit uncompressed.
When comes to IQ, I still recall many great minds trying to tell people there was no difference between 12 bit compressed file and 14 bit lossless files, or at least you can’t see it. By that standard, you are trolling by saying uncompressed brings better IQ.
“Of course you could have lossy uncompressed, say 12 bit uncompressed files compared with 14 bit uncompressed.”
Both these will come at drawbacks. 14 bit uncompressed comes at a drawback which you refuse to understand. You probably have some fairytale imagination that believes everything would be perfect, if only Sony would do as Digllous says and set their cameras to waste CPU cycles here and there.
Oh well, that 12/14 bit discussion had been there for many years. I don’t plan to go over it again. Regardless, other cameras from Canikon solved it very well. So there is a mature solution from “dying companies” but it takes some “innovation” from Sony,
Do you at least acknowledge the positive advantages to Sony’s current method, or are you just a Diglloud sheep?
From what I understand, their current method is the direct result of making compromises to their (intentionally) “crippled” design in the first place. As I said, solution have been around for many years. You don’t need to pay $3000 to let a camera output lossless compressed raw for a higher rate.
So you’re completely unwilling to acknowledge what Sony currently does has certain advantages. That’s really all I needed to know about you.
Based on their current hardware, lossless compression is going to have down-sides as you said. The root of their issue is that they adopted this design while they have an option that they don’t need to face it, which is a result of some business decision.
Everything has a circumstance. For all you know, they might have needed to make their cameras as large as the A-mount cameras to accommodate for additional microchips to do exactly as you wanted.
At that point, you bring a whole load other disadvantages to the table if you wanted the E-mount cameras to behave more like the A-mount cameras.
Is it worth it to wreck your head over these baseless speculations?
From what I know, they want you to buy a new camera every 15-18 months. And this leaves room for new models to come out quick and with low development cost. A7 is the only camera I see that can run an multi-language keyboard App. They apparently used that processing power for something other than taking photos.
Of course, I always have to keep in mind that you are a Diglloud sheep who can never see advantages to Sony’s methods, and never see disadvantages to Diglloud’s suggestions.
There is only 8 bit, 16 bit, 32 bit TIFF or PNG, which means uncompressed files should be in multiple of 8 bit. So I think Diglloyd is making a mistake here.
And yes, Sony done something wrong in the first place. What they do now doesn’t cancel it or justify it since they build everything on that basis. Just like Fuji putting an old apsc sensor in their camera. No matter how much they improve it, it’s still out-dated in performance.
Be clear, I don’t question Sony’s technical ability. I question the way they do business.
I don’t care what technical reason is behind this. What I concern is, other companies have it and Sony used to have it in the past. Then it must be a business decision that you, as a user, couldn’t have it today.
Don’t forget lossless compression and uncompressed raw are not new technology. They didn’t even let you enjoy the full speed of UHS-1 now.
Sony have never had lossless compressed. They used to have uncompressed RAW and CRAW, which became the standard compression we see now.
Okay, I am not familiar with cameras older than A57. That sounds like they decided not to address this very basic issue after so many years.
If that the case then easy: hand out a batch processing tool that could convert the half-baked 16 bit files into proper 14 bit on your computer! Right now they’re really wasting disk space for nothing. It adds up to huge waste when you store thousands of images.
We did not get 16 bits, we got 14 bits, plus 2 bits that are always 0. That said, I cannot really be angry at Sony at this point.
“But it might be that the CPU in the Sony cameras is not fast enough for Huffman encoding?”
Then don’t use Huffman. I mean: it can compress to JPG, so compress it to lossless JPG. Should be possible and it can skip a step, creating the lossy bitstream to compress. Even if you keep the initial 16 bits (14 bit signal plus 2 dummy bits) it will compress to something between 10 to 12 bits per pixel (http://imagecompression.info/lossless/)
JPEG uses Huffman compression anyway, but probably with smaller dictionary
Lossless JPEG uses Huffman.
Huffman encoding is used in almost any compressed file format in ADDITION to other more sophisticated and mostly lossy algorithms. Huffman coding is widely used due to its simplicity, high speed, and lack of patent coverage. As I said earlier, I highly doubt the CPU is too slow.
Bottom line: They could give you the compression option and tell you it’s slow so use at your own risk.
I meant the opposite. I was trying to say, that I doubt that the Sony cameras are lacking circuits optimized for fast Huffman encoding, because Huffman encoding is used in JPEG too and the algorithms are fast to compute.
Just because they are both called JPEG doesn’t mean that they use anything like the same processing. Lossy JPEGs in cameras use special circuits to take the load off the CPU (see if your laptop can render 5 x 42MP JPEGs a second, and then imagine doing it without less than 5% of the electrical power available).
A few issues: JPEG doesn’t support more than 8-bits* and the RAW files are stored at a mosaiced, non-rastered stage.
* The common JPEG format at least. JPEG2000 and JPEG XR do, however, but those are unlikely to be included in the camera. Anyway
Thank you Captain Obvious!
BS.
2 bits are left around for the future, so Sony doesnt have to redo file format each time sensor DR increases,
Nikon’s uncompressed RAW format uses 16 bits as well. Processing 14 bit aligned data is computationally horrible. It wastes a lot of CPU and, I’ve no doubt, that LR and other RAW converters probably expect it that way.
Using 14 bit alignment would, in effect, be a type of file compression. Really it would be better to do the job properly if there’s sufficient cpu power available.
Unfortunately Sony SD card interface is very obsolete. It can barely hit 40MB/s! For a camera with files this big that’s ridiculous.
Oh I never knew that Lloyd is a computer engineer who knows more about algorithms then Sony engineers. Learned something new today.
Really……he’s a visionary in his own 11+7-bit mind, LOL. Actually, and its far more illuminating, is Lloyd’s rant about fast glass. Here’s a man who oblivious to “the look” generated from fast glass, he’s so wrapped up in technical perfection he cannot understand how perfectly boring his technically perfect images are.
And the architecture of the processor, which is also key in the “lossless compresion” is possible with minimal processing power.
Seems people forget that any form of compression needs computing power.
His disciples harass me below!
how much processing power do you need to remove extra zeroes?
You would need to test that to know.
Maybe Sony did the cost/benefit analysis and determined this was best.
Maybe they favoured the consideration for the already limited battery power.
Something tells me that the marketing division of Canikon whispers sweet nothings into his ear. Unfortunately, he then grabs a megaphone to blast the comments. Sigh…take his comments for what they are worth and you can decide what that is….
Facts is cheating!
I won’t bother to read the article, but I have to admit that I’m curious if the the two bits are on the LSB or MSB side….
Seriously, seems that not one take notice that sony never allowed and never will allow the true raw because they want to make their lenses look better than they really are. Someone will find it, like Leica T i guess that has very bad kit lenses and rely on software correction. I guess all sony lenses rely on software correction some lenses more some lenses less, just waiting for the rogers cicala discoveries on the sony lenses.
In a digital world it’s crazy to not use it for everyones benefit. Smaller cheaper lenses that put out the same look as bigger more expensive lenses.
Crazy? This is omiting information, kind of lying… unaceptable, and the worst part are people who think like you… At least it should be warned not expected. some of their lenses are incredibly expensive but with no real reason.
I take it you own a Samsung phone because Apple products are too expensive; therefore, morally repugnant? The market will set prices and demand–always does and no amount of calling ethics and morality will change that.
You buy a lens for a price, this price is achieved by cost of materials, cost of production, cost of design, cost of transportation, performance of the lens, etc… and at last profit.
You pay premium for the lens that only performs well based on a software correction… Doesn’t feel right at all, this omission can omit bad quality production sample, true resolution, vignetting, distortion, etc… I would feel robbed, fooled, could have bought another lens for a better price…
I have a samsung btw, but it just doesn’t matter i don’t have brand loyalty.
You buy an iphone its a one piece hardware.
If the lenses are or not software corrected, doesn’t matter too much, because you had no choice… i don’t think any sane people buy a phone based on its lens image quality only. Its composite you have to think about every aspect or at least the most important aspects for you.
It performs as it should.
It would be better if they shown the true performance of each component? Yes!
A lens is a diferent case, because you have the choice, if you are purist like many are, you feel like it could have invested in another lens or maybe even another system, because a lens is determining factor on photography.
Well, the worst thing is because of this cooked raws, the camera is probably eating more battery, because it has to process one more step and its processing it in realtime when you look thru the EVF or the back Screen, its slow down the lag from shot to shot, to even turning the camera on until appears the image, so more battery every single time is consumed, probably slowdown the max fps, worst case in video, its applying it on video in realtime, so more things to process can be a cause of overheating.
Even if it a direct hardware feature it causes lag.
So we get a worse final image quality because Sony wants to hide their faults.
Maybe thats why Nikon can achieve higher dynamic range with the same sensor, i really don’t know, but makes sense when hiding vignetting for example.
First it was raw file concern trolls and now it’s lens concern trolls.
Science is repeatable, you are proving it yourself, want to discuss, im open.
Trolls don’t have reasons just talk like you did twice.
Ever listen to two politicians debating each other, they both skirt the truth to make their points, but never quite dress their own issues. Well, that’s what I hear from trolling commentators.
You, Marcelo, need to log off the Internet and go somewhere peaceful to shoot on film.
Larry, fuc.off…
Hey man, that wasn’t a sarcastic comment– and you’ve only proven the validity of my advice. I’ve recently been going over some film shot more than a decade ago, and it’s amazing how looking at it and appreciating it can pull your mind out of the technological gutter. Anything taking on a recent Sony APS-C sensor (or larger) with a decent-ish prime lens produces absolutely stunning results. Stunning.
Yep, you’re trolling and talking fluent gibberish.
You need to calm down and learn to control your breathing, this will help you with your panic attacks.
Yeah, ritalin with coffee is a very bad combination in the morning…
Eh? What is “true” RAW? Any Sony RAW file can be viewed with or without corrections.
Nope look after lensrental last article about FE lenses. But a quick answer, how could you know if the only image that you can see is from a screen or evf that can be already modified when it gets to your eyes?
You’re not making any sense. It is perfectly possible to view an untouched RAW file.
Case in point, look at an uncorrected file from the 16-50 PZ. It a mess of distortion and vignetting. Saying that, its corrected results are perfectly acceptable for its intended market.
Wise men speak because they have something to say. Fools speak because they have to say something.
Wiser men read the comments sections of articles.
Hahaha!
Nice
Wow, people are easily fooled if you wanted to fool someone to think that one think is untouched what would you do, just don’t fully correct the problems…
I don’t care if its perfectly acceptable i care about morality and ethics that seems to not exist anymore, if everybody is unreliable we tend to caos…
Industry is full of shit and it has to end, if we accept to be fools they will make fool of us.
That much I agree with, people are easily fooled. You’re a prime example.
Yeah, you should note why, otherwise you are the one who is making fool of yourself.
“I don’t care if its perfectly acceptable i care about morality and
ethics that seems to not exist anymore, if everybody is unreliable we
tend to caos… Industry is full of **** and it has to end”
Good luck with that buddy. You’ve chosen a noble fight.
Buy an OTUS if you want a lens that’s designed to perform optimally without software corrections. Note its SIZE and note its WEIGHT and finally, note its COST. I’ve read your comments here before, you appear to be ill-informed or you like to troll.
What is an untouched RAW?
Only the RAW-values without preview JPEG?
Only if meta-data about ISO, aperture, focal-length, WB etc. is missing?
Only the electron-states on the sensor before reading it out to a chip and writing this to RAM?
Is really a debatable question.
I was searching right now and seems that nowadays even some sensors have some kind of advanced internal processing.
My concept of untouched raw would be a file that does not receive any kind of post processing that alters the image quality. Meta-data does not alter image quality.
My point is, even if you have a plain simple .bmp file on your computer, with each pixel consists of 8+8+8 bit as RGB value, there is a ~54 byte header that specifies stuff like compression-mode=off, number of lines, number of columns, bits per pixel, etc.
So this 54 byte header belongs not to the actual bitmap, but it is part of the bitmap-file.
Same with RAW, but each company has a different method of defining how the RAW-file has to be stored as a file. Header, embedded JPEG, exif + meta-data.
Then: are there first all red-pixels stored, than all green ones. Or is it alternating… Who knows.
And when it is compressed, which scheme?
So I think, the Sony ARW is the real-camera-sensor readout, stored unmodified in any file-format. And 2 bits wasted when stored as uncompressed ARW because the algorithm is more simple 🙁
I don’t think there is any “correction” applied within the RAW file. But in the metadata there is maybe a profile stored for a lens-correction that has to be used by the viewing-software. But I am quite sure the RAW-pixels itself are unmodified.
Well, i will wait for more results, but i have read about some lens “behaviour” diferences on A7R and A7R2.
Search for the Lensrentals article about the recent test Roger did on the FE mount lenses, is on the sonyalpharumors btw.
He says “All that being said, these lenses may look better on-camera than on the optical bench because there seems to be at least some modification of files as the raw is written. We’ll get a better handle on that as we test more lenses.”
I don’t have A7R neither A7R2 in my hands but one “easy way to discover that would be hacking the camera or changing the firmware from the lenses.
http://diglloyd.com/blog/2015/20150807_0912-SonyA7R_II-raw-file-cooking.html
Take a read on that…
I don’t know if this is really true…
But if it is true – well. I don’t care.
As long as DxOmark measures Sony snesor (AND the baked in correction) this good, the RAW correction can not be this bad.
Especially for vignetting it can be a big plus to have a consistent lighting in the complete image, stored in 14 bit. This can give a higher dynamic range, overall.
DxOmark cannot measure sensors, they can only measure RAW files. If the RAW file has been altered, the sensor may look better than it is, and the blame for a less detailed image may be put on all the lenses tested on that camera.
http://diglloyd.com/blog/2015/20150527_0544-Sony-lens-corrections.html
Anyway i’m quite sure that no brand delivers the exact untouched raw as we would like to. But my real concern is more about cooked raws that go beyond image processing and adds profiles that aid lenses or adapters to look better than their real performance value. This is omitting data to make you pay premium for mainstream products that are sold separatedely. This could be called unfair advantage when comparing to Canikon, but its bad for the consumer too.
At least we all should know about that in manuals when we buy the product.
You read, you know what you are buying and thats fair because you made a choice based on the truth.
I could say that probably all sensors has some direct hardware function that uses some algorithm processing to reduce noise or to apply sharpening or to change colors or so, i can understand that because you are buying a product you know whats is offering and you don’t have another choice of sensor to buy for that camera, because the sensor is not interchangeable.
To be true, even this is not really fair with the consumer, because if you have the knowledge of how this is done and you could do better on your own its a limiting factor.
This is not debatable at all. “Untouched” means that you get the sensor read-out unaltered, i.e. the values from the ADCs of each photo site. Loss-less compression is okay, as well as metadata, because you still get the original data.
Software corrections are way cheaper than hardware corrections (read: perfect lens in distortion, coma, etc.).
Definitly is cheaper, but should be advised.
Nobody corrects distortion or coma in-RAW.
I didn’t say anything about RAW, just software correction.
This is the future– especially if you ever want to stop paying over a thousand dollars for quality normal prime (that is indistinguishable from the “corrected” version). There will always be the purists, though, to nag and remind us of when you could still spend an entire month’s salary on an optically really-really-darn-near-perfect lens.
I love Roger’s blog but maybe he should have kept this investigation to himself. The general photographic population isn’t ready for this particular curtain to come down yet.
If this person was dying of thirst in the desert and you handed him a glass of water, he would say it was too cold.
You have to be pretty coy to buy slow SD cards today when you buy photographic gear. Sandisk Ultras are fast enough for cRAW and Extreme Pros are not very expensive given their capacities and write speeds.
If you plan to shoot big files, plan to also invest in speedy cards.
This tells me you just don’t know what you’re talking about. The write speed on the a7rii is bottlenecked by the camera’s write speeds and outdated UHS interface that Sony implemented; buying an expensive high speed card beyond a relatively low threshold will make little to no difference on this camera. You’ll only ever get around 35mb/s which means compression quickly becomes very important for buffering and write times.
Why do you continue with your blind defence of Sony’s obvious failings? Do you have stock in them or something? Or do you not own a Sony camera and hence don’t care if they don’t fix things?
Our storage technology is always the bottleneck. The CPU is orders of magnitude faster than our external storage, regardless of how much money you have to spend.
No. Sony owes you nothing. The fact they did this is nice, but it was not a hindrance to all photographers. Actually, cRAW would be a non issue to 99.9% of Sony customers, but they did this for the internet reputation.
This sense of entitlement in the internet is really messed up.
I have no idea what you’re talking about. You’re saying “no it’s not free to implement a basic software change”? Because if you are, that’s extremely pedantic – if you put the cost of even minimal lossless compression against the cost they’d already invested in implementing lossless RAW, it’s completely negligible. Sony are doing this because they want to improve their reputation so they can sell more cameras. If you factor in the bad press they’ve had from half-arsing the job twice (no 14-bit RAW, then no lossless compressed RAW), then it appears completely nonsensical.
I didn’t say it wasn’t free to implement it. I said Sony owes you nothing. I bet the bad press isn’t catching up with the sales.
“Catching up with sales” is meaningless. However, the fact that it affects sales to some degree appears almost considering the attention it’s had. Weighing up even a handful of sales against the minute cost of implementation shows that Sony’s decision makes zero financial sense.
Because Diglloyd knows CPU architecture well enough. If it was win-win, why wouldn’t Sony do it anyway? Oh wait, the illuminati have it in for you, that’s why.
No, because Sony is a fallible company with fallible engineers who need people to tell them when something needs fixing. Do you just assume that all big corporations are untouchable, and every decision they make must have perfect reasoning behind it?
Stop being emotionally engaged. Have a drink, a cig and some chill pills.
As proven by the GH2 hacks and patches, the limit of throughoutput may only be firmware and not hardware.
Mike Kobal has RX1rII full res image samples.
Jpegs, though…
Interesting my response is pending admin to allow.
Someone please correct me if I am wrong. I updated my A7RII on Monday (have had it a week btw) and went to shoot an HDR sunset sequence. I thought natively it offered a 5 shot option through the drive selection. After the firmware update, I only have 3 shot options. Now hoping that Promote will find a way to build Sony capability into their remote to give me larger latitude HDR options. FYI, I moved to Sony from a Canon 5D series rather than the 5DS due to the technical capabilities of this beast. Thanks to all the contributors on this site as I have learned a ton of knowledge through you all.
You really need 5 shot HDR on an a7RII…?
My camera still brackets 3, 5, 7, or 9 shots in uncompressed mode just like before in compressed mode with the previous firmware. If you are talking about the built in HDR function that results in a processed HDR JPEG, the function is also unchanged from before.
Maybe Sony is thinking ahead of the Diglloyd guy, it’s using 16bit to make the format future-compatible…
haha – good one
LOL…. there’s little reason to add digits to ADC, it’s more reasonable to implement smaller pixels, which due to oversampling will have same effect + more resolution.
I thought it was 16bits to 14bits then compressed to 11+7bits?
No, it’s 14+4+2-8+9-5+6-8.
I keep forgetting. That looks about right, though.
IIRC
Captured in 14bit, processed in a 16bit container, then converted to 11+7.
I would think Sony is just saving the 16bit file after the processing(NR for example) to get the uncompressed 14 bit data in a 16bit format.
Give up guys, this will never end, even if they make 24bit that captures every color in the universe this guy will keep talking shit.
DigLloyd has fulfilled the precise caricature of the idiot reviewer (that many of us predicted), who would come along and criticize the very feature he demanded from Sony.
Reviewers that allow for zero margin of error are the exact people that shouldn’t be employed as reviewers. Life, and by extension, photography, is messy and imperfect, and it’s only what you make of it. Get down from off your plastic highchair, you’re a man acting like a toddler, Digi-luh-luh-Lloyd. XD
Not so fast there. Sony did provide non-destructive RAW format but without compression. What Sony should have provided is lossless compression where most camera makers provide with their cameras.
He did not demand uncompressed RAW. He, like everybody else that actually knows what they are talking about, asked for losslessly compressed RAW. Sony did not deliver this and he is pointing it out. Nothing wrong with that.
Actually, Diglloud was crying about image quality. That was the basis of everything in his complaints, and Sony addressed this.
Now he is crying because he is unable to see the advantages to how Sony offers uncompressed raw. This method is processed faster, it simpler to copy from memory, and wastes less CPU time.
But he’s trying to paint the picture that it’s all bad, and the problem is ignorants will take his word as gospel.
” Life, and by extension, photography, is messy and imperfect, and it’s
only what you make of it. Get down from off your plastic highchair,”
Yes, thank you, this is the most intelligent thing I have read in a long time in a photo forum. As for DigLloyd, his way to review is indeed close to absurd. He will go on ranting about anything that is not Nikon forever, pushing Canon files by 4 stops in order to show shadow noise, lossy compression Raw artefacts and such. Unfortunately the rest of the camera (A7rII) is so good that there is not much to find to critisize, so now as there is uncompressed Raw, he will go on complaining about the lack of compression in the Raws, or the Nikon 810 having 2% less noise at ISO12800 etc etc. This guy (who publishes more selfies than probably all other reviewers together) will never be happy with anything not Nikon.
He would point out that it’s too cold, and drink it. I think finding out everything there is to it is the point of his site.
Every camera on this planet capable of snapping JPGs uses Huffman encoding plus a whole lot of other algorithms specified by JFIF. I find it hard to believe that Sony isnt capable of implementing it. Even more so because it doesn’t need to be in real time. The camera could easily add the compression method in the background when it’s not busy snapping pictures.
#2 is most definitely BS