Gearóid Ó Laoi, Garry Lee: A larger sensor will be exposed to more light. Indeed yes, but not per unit area. This is NOT why the quality is better.
Gearóid Ó Laoi, Garry Lee - you've just contradicted yourself.
It's either the same light per unit area (which means more total light if you have more area), or it's the 'same total light that somehow knows about the sensor area.' Oddly enough, it's the first one.
Imagine a situation with two different sized sensors and assume these both give the same angle-of-view.Large sensor: 100mm lens, F2 (50mm aperture diameter)Smaller sensor: 50mm lens, F2 (25mm aperture diameter)
They experience the same light per unit area, but the larger sensor has four times the sensor area, so gets four times as much light. Not because the light knows, but because the aperture is bigger on the equivalent lens for the larger sensor.
KrisAK: Does Sony not provide review samples of their cameras prior to release?
"Hands-on" reports usually coincide with product announcements from Canon, Nikon, Panasonic, etc.; with Sony, we seem to be stuck waiting for anecdotal reports of dubious value. Making an informed purchase is something of a challenge.
It varies, product to product. In this instance there's been very little access given.
Light per unit area is only relevant if you're comparing images at the 1:1 pixel level. As soon as you view photos at a common size, whether as prints or on screen, then it's the total amount of light that matters (or light per portion of the image, if you prefer).
Thomas Traub: dear Mr. Butler,
were you get such informations from? How do you managed to get this infos? Very very interesting!
Many thanks for such articles.
We usually know, what the hardware of our equipment does and how it is built. But we usually do no know anything about the software inside our cams.
Thank you very muchThomas
@Thomas Traub - as I tried to make clear in my write-up, the real work was done by Alex Tutubalin and Iliah Borg. Here I'm just trying to make their finding more accessible, and relate them back to real-world shooting.
Frank_BR: What the article did not say was that the picture of the luminous sign was pushed 5 stops in post-processing! If you doubt it, compare the picture in the article with the picture taken with ISO 200 and shown here:http://www.dpreview.com/articles/7450523388/sony-alpha-7r-ii-real-world-iso-invariance-study
No surprise that some artifacts appeared after such extreme PP. But a photographer does not need to push 5 stops, unless he has committed an extremely gross exposure error.
The article seems to ignore that the compression used by Sony is very effective in reducing the size of files. For example, the RAW files shown in DPR Studio Shot Comparison for ISO 200 have the following sizes:Sony 7RII: 41.4MBNikon D810A: 74.3MB
It is clear that the Sony 7RII stores RAW images at higher efficiency than the Nikon D810A, for example.
Any compression algorithm incurs a tradeoff between efficiency and appearance of artifacts. If you look closely enough, you will always find artifacts.
I'm sorry, I wrote 4EV when I should have said 5EV.
However, the damage is there, even before you push the file. It just becomes more visible, the more you push.
notpc: Not one mention of the outlandishly outrageously offensively ridiculous prices of any camera/lens that can be considered decent. Even Canon's new Rebels are MORE expensive than an old "digital rebel" of over a decade ago. WTF? Almost all other electronics plummet in price as the technology matures and are mass produced. Back then, we were told it was because those 22mm sensors were damn expensive to produce. And maybe they were at one time. But 12 years later, they're still maintaining an artificial price level. And people aren't buying? Shocker there.
There are so many point here, so I'll address the main ones:
I've not denied that the cost of other electronics has fallen faster, but the cost of cameras has still fallen, in both money terms and real terms.
Reserving features for more expensive models is basic product differentiation. Yes, it's all the more galling when you know that it wouldn't cost anything extra to add, but the alternative is that everyone buys the cheapest model, you make no money and you go bust. That's not a business model that shareholders invest in.
The K-50 is almost certainly and out-of-production model (probably replaced by the K-S2). The price you sell things off at, to get rid of excess stock is not a great data point.
The trend in lens pricing I'm less likely to disagree with. The added R&D costs and complexity of manufacture of lenses will push prices up but it looks like some manufacturers are creating a self-fulfilling prophecy by which only a handful of high-end users buy high-end lenses.
@12345ccr - You've never been able to get so much camera for so little money and there's never been a wider choice of good cameras.
Cameras such as the RX100 series offer significantly better IQ and flexibility than a smartphone, small form mirrorless (Oly E-PL7, Sony a6000) offer greater stills and video capability in package you can easily carry around. The enthusiast camera and lens offerings from the likes of Fujifilm open up new choices to enthusiasts and sub $2000 full frame cameras open up an area of photography that was much more exclusive until recently.
Whatever your level of commitment to photography, there are some astonishingly capable cameras out there. There are ways they could all be better (a ground-up redesign, both at the entry-level and enthuiast class, WiFi connections that work quickly and reliably...), but smartphones, for all their immediacy, are pretty limited when not shot in good light.
talico: So... When DxOMark rates the sensor, do they take this into account? Does the cooking of the RAW file have any benefit to Sony that make it worth doing? Lower noise? Sony's sensors are supposed to be the best you can buy. Why do they need to do this?
The compression curve may be the cause of the slightly lower DR ratings that Sony cameras tend to get, compared with Nikons using similar Sony sensors.
As I understand it, DxO's testing (and I may well be wrong on this), wouldn't look for or detect the high-contrast edge errors stemming from the localized compression.
PVCdroid: Why can't DPR take a shot of the same scene with a Nikon and either the A7R or A7Rii? What would be so difficult about that where we could see both images put through this underexposure and then pushed so far to show artifacts? I'm hearing myself and everyone else that is content with Sony RAW being called apologists. I'm not but if DPR wants to write an article on the debilitating effects of Sony's compression scheme then let's visually see the difference.
Our a7R II review will include side-by-side results against the D810 and will show ISO Invariance data of 14-bit vs 12-bit vs 12-bit E-shutter (I'm working on that at the moment).
I spent some time working on an a7R II vs D810 shot for this article but we felt there were too many variables (not least the lower minimum ISO on the D810) that would be argued over.
@steelhead3 - that would be the opposite of the truth.
We've demonstrated that the second step of the process, which most other cameras don't do, can have an impact in out-of-camera JPEGs - which should show that you don't need to particularly push files to see that difference.
As we've made clear, the differences in DR are extremely subtle but if you look at our ISO Invariance test, DxO's data or the work of people like Bill Claff, you'll see that Raw files from Sonys tend to slightly under-perform the results from cameras using Sony sensors of the same pixel count.
Henry Richardson: Panasonic also uses lossy compression for its raw files. Here is just one place to get info about that:
That's very interesting. It sounds like a fairly similar strategy to the more destructive part of the process described here.
I'll have a look into that as soon as I get a moment.
Also, just for clarity:
MSRP for Rebel XT (350D) with 18-55mm zoom: $999MSRP for Rebel T6i (750D) with 18-55mm zoom: $899
If you then take ten year's worth of inflation (CPI) into account, the XT would cost $1220 in 2015 dollars.
To me, that looks like the T6i is around 26% less expensive than the XT was, ten years ago.
The camera market is highly competitive. Two big players dominate the market, everyone else has a huge incentive to under-cut them to build (or even hang on to) market share. The idea that everyone is *artificially* inflating prices, rather than competing, seems questionable.
There are several manufacturers whose very survival in the camera market looks suspect and there are *at least* two big companies with the financial and marketing clout to aggressively expand their market share. Does it really seem likely that all of them would decide not to compete on price?
JOrmsby: Kudos to you guys for looking into this. And I agree, uncompressed RAW should certainly be included on a $3500 high MP camera aimed at professionals.
@ KAMBIC right back to Phil's first reviews on the site, we've tried to highlight problems with cameras in the hope that this would be at least considered by the manufacturer.
The two things that make this different are that:a) We're trying to pull the relevant pieces out of reviews as stand-alone articles, to prevent them being over-looked.b) This issue applies to many cameras, not just one model, so it doesn't make sense to bury it in one single model review.
Neither of these aspects has anything to do with brand. In this instance these two factors coincided regarding a Sony product but we're not writing about it *because* it's a Sony product.
alpha dog: It would be really interesting to see a comparison of the same shot on comparable non Sony kit (using similar generation/same resolution sensors). eg Compare Nikon D750 (uncompressed and compressed) v A7ii ('14 bit' compressed and '12 bit' compressed) to see how big a difference is visible - both in the high contrast edges and in pushed shadows.
The problem is that the shots that would be ideal for this example (A7R II vs D800) are not the same as the ones we need for a a7R II (and D810 review), which are a7R II vs D810.
I didn't want to fill this article with studio shots and risk it being dismissed as irrelevant to 'real world' shooting, so I didn't use those (though they may appear in our Raw DR pages of the review).
Mrrowe8: If I understand you correctly if I used the higher mega pixel count cameras the amount of lost information if not shot rapidly is minimal or if I shoot in bulb the higher the pixel count again the information lost should be again minimal and wouldn't it also impact what size of a print would I notice this and at what distance i would be observing this ?
The most visibly destructive aspect of the compression stretches across 32 pixel strips (Interleaved strips of 16 red, 16 green or 16 blue pixels), so these will make up a smaller proportion of high pixel-count images.
Everything else is essentially a reduction in dynamic range: reducing the processing flexibility (meaning you'll hit whatever you consider to be an unacceptable noise level, slightly earlier).
naththo: Richard Butler, I have just had a thought came up. If you are on RAW mode in A7RII or any other Sony camera, did you turn off the lens correction option including shading correction, Chroma Aberrations correction and Distortion Correction before shooting those example? I just hoping if you did.
I didn't shoot either of these examples so can't be 100% sure what settings were used.
You're right that shading correction is 'baked' into the Raw file, so could affect the corners/edges of the frame. CA and Distortion correction are only metadata tags, I believe, which can be read (or ignored) by Raw conversion software, so that makes no difference.
frosti7: does the old GH4 is limited to 30 mins in 1080p as well?
Earlier, incorrect comment deleted as a result of estarkey's correction.
@estarkey: I don't remember Panasonic US ever making that clear (though yes, I'm aware that the duty restriction is an EU issue).
[Page 392 of this manual](ftp://ftp.panasonic.com/camera/dmcgh4/dmc-gh4_en_advanced_om.pdf) suggests that US models are restricted to 29:59 for Full HD, HD and VGA recordings. Thanks for clearing things up about 4K.
Caerolle: I don't get it. If simplicity and user-friendliness are the most important things, why are P&S cameras the ones that are dying? They are small, most likely have wifi, they have automated scene modes, and all that? Yes this is the class of camera that is disappearing the most quickly?
Maybe the sensors are too small to give much better pics, even though you have more control than with a phone? Then, what about the RX100? Is it just not automated enough? (I don't know anything about them, really, other than they are a compact with a really nice sensor.).
I just don't understand all this 'they need to be simpler' stuff. Even my 6D has auto modes, and wifi. Certainly I would think that the entry-level dSLRs and mirrorless can be operated much like a phone, if they have a touch screen (and it blows my mind in 2015 that all cameras don't have them).
And if you want to do serious photography, where you are taking control of the settings, I just don't know how you make it simpler?
If you watch Hilbig's video, the sales data he shows makes it appear that ILC (DSLR and Mirrorless) sales are following a similar trend to compacts, just with a 2-3year offset. In other words, something is killing *all* camera sales.
If that's the case and we go back to only enthusiast photographers buying cameras, then we're likely to see more than one camera maker leave the industry.
larrytusaz: The one poster at that page (Andrew Freeman) is spot on. It's laziness, and people not appreciating fine craftsmanship.
WHO CARES if a Kodak Instamatic soccer mom finds a camera like my Sony A6000 "complex?" WHO CARES if they don't know what all the buttons are for? I do. Let them have their smartphones or Kodaks, but cameras like the A6000 aren't meant for them. They're meant for people who know what they're doing.
It's like the one person said, people are getting lazy with everything. You see it in grammar too, people mixing up your vs you're and it's vs its, and them saying "who cares, it doesn't matter" when you point out the proper spelling etc. That's lazy. It DOES matter. Get it right.
"The best camera is what's with you"--more excuses. Pop a Sigma 30mm 2.8 on my A5000 and it's tiny. They also make the Sony RX series, and they can upload to a smartphone for instant sharing. Soccer moms are one thing, but REAL photographers take a REAL camera with them. To not do so is lazy.
larrytusaz - it might be worth caring if your camera maker of choice is one of the ones that withdraws from the business (which is likely to happen if sales don't pick up soon).