September 17, 2014, 01:48:02 PM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - sarangiman

Pages: 1 2 [3] 4 5 ... 19
31
At the moment the big differences are just at low ISO. Low ISO is generally the realm of different kinds of photography, photography where it's difficult to impossible to actually control the lighting, and where dynamic range can easily surpass 11-12 stops...and the differences really are significant. Maybe it's just the number "two" stops of DR. Two is a small number, it doesn't seem all that meaningful. It is a factor of four difference in the range of usable tonal levels. Four is still a "small" number...so when you get right down to it, it's basically the difference between having 12-bit or a 14-bit data...now THAT number is big. That number is 12,288...which is a really big number. That number is a better indication of the differences between a Canon sensor and an Exmor. If we regularly talked about differences in dynamic range with big numbers like that, maybe the real benefit of having "two more doublings" would be more obvious.

Try 3 stops of difference between the 5D3 and the D810. More if you account for FPN.

If you need DR, then I would offer that DR is more than just a nice to have. If you NEED DR, then having more DR is essential. I don't need DR in all my photography...when I do need more DR, I already have some tools that help me resolve the issues with having less...GND filters for example. HDR is another option, albeit one that is imperfect and adds work. More DR is more DR...having two more stops is huge...it's very meaningful, and much more than just a nice to have. It reduces the amount of time I have to spend figuring out which GND filters to use, how many to stack, how to blend across the contrast divide, etc. Instead of three filters, maybe I can get away with just one. With two more stops of DR, I could get away with a heavier shadow lift instead of having to apply an LR/ACR gradient filter in post. More in-camera DR, work. Less work with literal filters, less work in post.

Don't forget that GND filters are *still* useful for Exmor, to combat shot noise. You may or may not care if somewhat noisier shadows (b/c of noise in the sampling of light itself) don't bother you - since they're generally less visible.

The other way to do this for a scene that does fit within the DR of your sensor is to shoot a number of frames where you've exposed for the highlights, and then average them. This decreases shot noise (importantly: in your shadows), and allows you to then process the single averaged image. And image averaging is much easier than HDR merging - I can't stand the results most automated HDR software produce. Hence I do it all by hand. Which can be quite painful.

What I just explained above is a significantly better way to shoot HDR scenes than what I typically had to do with Canon. And the best part is - if you don't care about some shot noise in your shadows (a +3 EV push of ISO 64 shadows on the D810 is pretty much like ISO 500 FF levels of noise, which might be perfectly acceptable), then you just use your single frame.

And btw, I have a feeling the A7s has a slightly different architecture from most other Exmor sensors - giving it slightly more downstream read noise (so a bit of a base ISO DR cost, though nowhere near a Canon DSLR) and very low upstream read noise at higher ISOs (which gives it higher ISO DR). So, in a sense, it's a nice compromise btwn low ISO DR and high ISO performance. Slightly better ISO performance at ISO 25.6k doesn't matter to me, and neither does high ISO DR (since I shoot using ISO-less techniques, which actually gives the A7r more DR at higher, lower - above base ISO - DR, if that makes sense), so I prefer the A7r.

32
Your snarky comments add nothing to the conversation.
But DRones who turn any thread to the discussion of DR, or trolls who start new topics to do the same...those are what, public service announcements?   ::)

No, but expect a certain amount of counter-argument when Canon apologists start touting broad generalizations like 'DR doesn't matter b/c Canon's doing just fine' and uncontrolled unscientific comparisons/claims like 'oh look at this shadow recovery example I did therefore Canon's just fine' or blanket statements like 'the 5D3 AF system is superior to Nikon', etc. Extending sometimes to as broad a generalization as: "Canon 5D3 is a more capable camera overall'. Or myths like 'Canon high ISO is better' or 'use DPP for more DR' (and on that latter note: those differences are simply down to the type & extent of NR).

These myths often persist b/c of erroneous information constantly propagated on forums such as these.

I'd say the reason that some feel the need to step in & make people aware of some of this stuff is b/c perhaps, many years ago, they were mislead by the same myths that keep recycling.

On the specific topic of DR: at least now, on these forums, there's a general awareness & acceptance of the DR differences (albeit usually accompanied with a bitter 'let's not talk about it anymore' sentiment). I remember when years ago people had to fight just to get their clear, controlled comparisons showing the vast DR differences - and implications - accepted. Over and over again, persistently, just to convince people of what would be clear if you just did a proper side-by-side.

And now you complain it's brought up too often... anytime anyone mentions it.

Talk about a Catch 22...

33
As for DxO's 'image science', hallmarks of good science include transparency about methods with disclosure sufficient for someone skilled in the field to fully reproduce the experiment/test, and attempting to avoid bias.  DxO does not disclose their formulae or weightings for determining their Scores, and what they do disclose of their methods shows that their scores have intentional bias. 

I don't disagree with the bit about disclosing methods, being a scientist myself. You saying they have 'intentional bias' requires more proof than your simple claim, though. Bias towards what? Any formula has inherent bias, as any weighting system must. It's when people imply that there's a brand-specific bias that I take issue. I'm not saying that's what you're implying, but there are certainly those who imply it.

You could argue that they weight base ISO DR too much, and if you don't care about base ISO DR, then I can see how that'd bother you. OTOH, I think it's perfectly fine to weight base ISO DR far more heavily than high ISO DR b/c: (1) not doing so runs the risk of rating all similar size sensors roughly the same, and how does that help a consumer actually understand the *differences* between sensors?; and (2) if you shoot in an 'ISO-less' manner (even partially), you can - for a sensor with high base ISO DR - retain far more DR at high ISOs than the actual measured DR for any given higher ISO. Which - for me - makes the higher ISO DR numbers meaningless, and only the base ISO DR number relevant.

The low ISO bias is one example, yes.  It's not just their DR measurement, their 'color depth' score is also taken at base ISO for calculation of the overall Score.  Color depth is basically a measure of chroma noise, something essentially not visible in practice at low ISO, yet it figures prominently in the overall score (where two of the three Subscores are at base ISO).

Their Lens Scores are another example of their intentional bias.  Does it make sense that the Canon 50mm f/1.8 II receives a higher Score than the Canon 600mm f/4L IS II?  Is the nifty fifty a better lens?  Well, if you intentionally bias your scores based on 'performance in 150 lux illumination' (light level of a dim warehouse), I suppose it does.  Similarly, comparing a Canon and Nikon lens pair (e.g. a supertele) where the lens measurements show the Canon to be sharper, have less distortion, less CA, and equal transmission, the Nikon lens receives the same (or sometimes a higher) lens Score.  Why?  Because the sensor score of the bodies on which the lenses are mounted affects the lens Scores (and ironically, a sensor score biased toward low ISO performance then artificially inflates a lens score based on a dim light, i.e. high ISO, use case).  Bias on top of bias...bad science.

You keep quoting the overall score as why DxO is bad science, never mentioning the value of the individual metrics, which they publish. Very few of us ever quote DxO scores - we quote the individual numbers. The score is very one-dimensional, as it must be b/c it's a single number. Perhaps you'd like them to create fancy 3-D plots with various things - like illumination, focal length, and aperture - changing along X different axes in X-dimensional space. Because that'd be more easily digestible.

The simple fact that they average the performance across the focal length range & aperture range already makes it useless for me. But that doesn't make their data useless. And calling the entire site biased b/c they try to distill everything down to one score based on what they think is most important (and they try & tell you what they consider 'important' for their score) really requires some perverse logic. You may as well call every site that ranks anything based on one score completely biased and therefore utterly meaningless.

Because the sensor score of the bodies on which the lenses are mounted affects the lens Scores (and ironically, a sensor score biased toward low ISO performance then artificially inflates a lens score based on a dim light, i.e. high ISO, use case)

OK, that's probably fair criticism. And yet another reason I never even look at the lens overall score.

But OTOH, yes, attaching a similar resolving power lens to a higher resolution body does, lo and behold, lead to a higher P-MPix score.

Does he shoot D800 ? Does he shoot high ISO ?

http://500px.com/JakeOlsonStudios
No.And no.

He shoots almost exclusively iso 100 w/flash (or other light additional source)...

Nice shots. However, not my style.

Flash and reflectors. He also has some wicked processing knowledge, which I think in part is where is amazing boke comes from. He teaches a class...I might take it just to learn his PP techniques.

Why would you pay someone to learn their techniques when he is obviously foolish enough to continue using a camera system that imposes a huge, burdensome requirement to...

...spend a lot of time and use a lot of special techniques and extra tools that add to your total cost to make the photo look good...

Neuro - you always wake up on the wrong side of the bed? jrista himself still shoots Canon; he just points out one of the things about the system that he finds limiting, and you suddenly find it necessary to mock him for wanting to take classes from someone else who shoots Canon?

Logical much?

IIRC, Marc Adamus makes his images with both Nikon and Canon cameras. Those of us who find Canon sensors limiting are - most of the time - not oblivious to the fact that you *can* make fantastic photos with almost any system. People made fantastic photos with the limited DR of slide film, for crying out loud. Your snarky comments add nothing to the conversation.

And as an aside - I can't believe I'm here defending someone who himself called my work 'bull' when I tried to point out that the 'ISO-less' nature of the D800 opened up opportunities for shooting the 5D3 didn't allow years ago. These forums are really something!

34
Reviews / Re: Tony Northrup - D810 vs. 5D Mk3
« on: September 09, 2014, 12:44:19 PM »
B/c the 5D3 simply won't track across the frame, while the D800 will do so marvelously.

User error.
+1 pretty much. RTFM.

Nope. Not user error.

The 5D3 only has the 'capability' to track across the frame using depth information from the AF sensor, which might work for subjects that don't change depth much (e.g. Birds), but doesn't even remotely work for erratically moving subjects that change significant distance from camera (the case when shooting with wide angle fast primes, for example). In other words: doesn't work AFAIC.

Let me put it this way: if the 5D3 were perfectly capable at this, why would Canon have released iTR in the 1D X? Do any of you understand the idea behind using the metering sensor for subject recognition? Or the entire principle behind Sony SLT?

You know what's better than RTFM? Using the freaking camera. It's quite clear none of you responding have actually compared Nikon's latest 3D AF tracking to Canon's, so the authoritative voices with which you speak are rather comical.

But as someone else stated, perhaps that's exactly what makes CanonRumors so entertaining.

35
Reviews / Re: Tony Northrup - D810 vs. 5D Mk3
« on: September 09, 2014, 04:24:56 AM »
Im not an AF expert as i rarely use anything other then the center AF point in one shot mode. :) and 40% of the time i focus manually.

I think the 5D MK3 AF is very capable and i don´t know if the D800 AF is worse.

But i think even when it´s worse, it doesn´t make a big difference for the majority of shooter who lust for a D800 sensor in a Canon body.

As allround camera the Canon 5D MK3 is sure more capable.
But that´s Pickup vs. Sportscar talk again.

Again, that's your opinion.

You say the 5D3's AF is very capable, and you don't know if the D800's is worse? I'm telling you that from my perspective, the D800 is not only not worse, it's better. B/c the 5D3 simply won't track across the frame with anywhere near the accuracy & ability with which the D800, for example, will.

So, why is the 5D3 all-round more capable?

From my perspective, having used the 5D3 for the past few years (and Canon all my life, so I'm very familiar with their tech), it has a worse sensor than the D800 (far worse for some of my use-cases), and its AF is not nearly as capable for my wedding/shallow DOF (fast prime) work b/c it doesn't track subjects across the frame. Meaning I have to manually select the appropriate AF point when I can't focus & recompose (b/c of the focus plane shift), and I don't have the ability to rely on the camera to stay on what I want focused (the eye of a baby, for example) as my subject moves around or I move/re-compose.

That last bit is relevant to running brides, moving babies, and sports photography where you want to decouple the composition from AF (that is, you don't want to find yourself forced to constrain your composition simply b/c you don't have time to move your AF point - forcing you to keep your framing such that the subject of interest is underneath the selected AF point).

At least the 1D X took a stab at closing the gap between Canon & Nikon in terms of this type of tracking when it introduced iTR which worked in conjunction with the metering sensor. But it's the only Canon - to date - to have this feature, and in my tests it does not keep up as well as the Nikon D800/D4/D810, etc. Which is to be expected for any 1st generation tech.

And for those who talk about high ISO DR, the D800 still doesn't fare *worse*; its DR just falls back down to Canon 5D3 levels. And if you know how to, you can actually have the D800 retain far more DR at higher ISOs by simply underexposing by lowering your ISO, then boosting your exposure selectively in post to get your brightness back up to where it would've been with the proper 'hardware-level' ISO. So even there, it has the potential to have more DR under situations requiring higher ISO shooting - if you shoot Raw.

In fact, really the only thing I miss about the 5D3 are its excellent cross-type AF points all over the frame - which are far less fallible to hunting/failing than non-cross-type.

So before making blanket statements like 'the Canon 5D3 is sure more capable'... perhaps think twice. I guess none of my AF talk matters to you if you focus manually 40% of the time & use the center point the rest of the time. But many professionals stress the camera far more than that, and for them it's going to come down to which camera has the right technologies that allows the camera to 'get out of the way' the most.

And for them, it's not as clear cut as 'the 5D3 is more capable'. I assure you, for my work, it's exactly the opposite. And I didn't even know until I spent enough time with Nikon.

And I'm not blind to Nikon's issues either. Lack of more cross-type AF points and radio-controlled flash are rather egregious, and I'll inevitably have some glass envy. But, in the end, the Nikon D810 'gets out of the way' the most. And its grip is finally beefy enough that I can hold the camera properly :)

36
Reviews / Re: Tony Northrup - D810 vs. 5D Mk3
« on: September 09, 2014, 03:55:00 AM »
The wrong cameras are being compared.

When Nikon's D750 is released, that will be the camera to compare with the 5D3, not the D810.

Why? The 5D3 and D810 are similarly priced. Just curious as to your logic.

37
Just to be clear, I am only disappointed with the 5D III for landscapes and low ISO work.

Here's a link to a guy who's a hot topic on 500px at the moment, that CR members may enjoy seeing. Does he shoot D800 ? Does he shoot high ISO ?

http://500px.com/JakeOlsonStudios

The only noise comes from people complimenting his work.


When I see images like that, I think "Wow, that's one talented MFer."

"I wonder what gear he shoots with" is the last thing on my mind. I suppose that makes me a freak :o

I'm actually with you there, V8Beast :)

And everyone else makes fair points here.

sdsr - yup, how much it matters will totally vary based on some of the things you mentioned. And good point about adapting Canon lenses on the Sony bodies. Much easier to do than adapting Nikon lenses, which still have that mechanical aperture. Canon lenses are much easier to use on the A7 cameras. Definitely holding on to my Canon 1-5X 65mm lens. That thing's a beauty!

LostBoyzNZ - A7r is nice, for sure, but sometimes adapting wide-angles gives me a lot of trouble in terms of edge-to-edge and corner performance. Not surprising once you consider the effects of adapter mount variations, and the variations in the thicknesses of cover glass on the sensor (http://www.lensrentals.com/blog/2014/06/the-glass-in-the-path-sensor-stacks-and-adapted-lenses)

Also, A7r shutter shock is a real problem for longer focal lengths. Electronic first curtain on the D810 is incredibly welcome.

38
As a working pro stated earlier, shadows are important for the art...and shadows are supposed to be dark. 

This is a loaded topic. You could argue that *because* today's monitors/output devices have pretty small DR (10 stops for the better monitors today), that you shouldn't try to pack more than 10 stops worth of DR into an image. This presupposes you want to maintain a linear relationship between tones - the same relationship you would've had in the real world. So is that what you want to do, or do you want to maintain the global DR your eye-brain experienced in the real-world for a particular scene? Or something in between?

It all depends on what you want to emphasize. If you want to emphasize light vs. shadow, you might even decrease the DR of what you captured by darkening darks and brightening brights past the relationship they had in the real-world. Or you might do that for some scene elements while retaining more global DR.

The point is that when your sensor introduces little to no noise over your image data, you have the freedom to do whatever you want. You even have the option of - in the future - one day going back to your high DR Raw files so you can reprocess them for that new HDR display that actually displays 18 EV of DR (the motion picture industry is very interested in HDR displays, e.g.). And a HDR display that doesn't just give you blacker blacks, but actually gives you brighter whites as well. Ever wonder why Velvia looked so beautiful on a lightbox? B/c it had 11-13 EV of output DR (though only ~5-6 EV scene DR), with a white point on your typical lightbox 5-6x brighter than your digital LCD monitor with its brightness maxed out. It actually expanded contrast - possibly getting closer to maintaining absolute brightness differences between objects in the real-world in the process... but I digress.

So when you say that shadows are supposed to be dark... that can really open up a whole can of worms. For example, if you shoot a sunset where you've shot to preserve the orange/red tones in the sky, the cityscape buildings in the shot might be completely black when you view it on your monitor - but they were perfectly visible when your eyes saw the sunset. These 'shadows' weren't exactly 'shadows' in the real-world, yet they're 'shadows' now on your monitor b/c your entire monitor's brightness scale is much smaller than, and on the lower end of, the brightness range we experience in the real world. So are they really shadows, or are they just shadows b/c of your exposure decision & your imaging hardware? You might decide they're not 'shadows' at all; that they should be darker midtones, say. Low read noise will enable you to pull those 'shadows' up to darker midtones, then assign some other even darker tones to 'shadows' so that your final image does have good output DR/contrast.

Look at Ryan Dyar's or Marc Adamus' landscapes for gorgeous examples of capturing a ton of global DR while still having shadows/dark tones in the image that give the impression of high contrast with low global contrast (high DR). That sounds counterintuitive, but they pull it off beautifully. And they effectively have to 'tone-map' for our LDR output devices & prints. I can only imagine how much more stunning they'd be on higher DR monitors - and of course that'd likely require entire re-processing from the Raw(b/c the tone-mapping would have to be different). Or not - maybe our eye-brain system does enough 'filling in the gaps' that we'd only really appreciate HDR display of such content in a side-by-side next to a LDR monitor.

Anyway, my point here is that how you define 'shadows' itself is flexible. I can tell you one thing shadows *aren't* supposed to have: FPN & read noise. :)

39
So, if you were referring to me, it's you who conflated two independent concepts which I discussed. 
I wasn't referring to you. The particular comment I was referring to was as follows:

Quote
Besides, DxO mark isn't relevant.  Despite them scoring Nikon/Sony higher and higher against Canon product head to head, Canon still went from a 4% market share lead 4 years ago to a now 20% market share lead.  Nobody cares or nobody believes because of just that:  The garbage "science" they are doing.

Now that's garbage. This person clearly correlated the two, saying that nobody believes DxO's 'garbage science' because despite DxO's measurements, Canon's lead strengthened. DxO's raw measurements are perfectly fine, but they have - as you yourself say - little to no bearing on aggregate purchasing decisions (nor vice versa), and so correlating the two is utter nonsense.

DxO has scored Canon sensors lower than their competitors for several years, during which time Canon did not lose, but rather gained market share.  Therefore, as far as influencing the aggregate buying decisions of consumers, DxO's Scores are meaningless.
I don't disagree, as long as you have your qualifier in there. Anyway, who here was trying to suggest DxO scores influence the aggregate buying decisions of consumers? What their sensor *measurements* do do, OTOH, is inform a savvy consumer which camera will perform well in terms of image quality for landscapes (DR and SNR data), or light-limited applications (SNR data). Years ago it certainly turned me towards which cameras I could turn to for my landscape work, and I appreciate that they provide that data. Their data also allows one to look for interesting trends in technologies (which sensorgen also helps to distill, from DxO data). While we're on the subject -- what they're trying to do with P-MPix is quite clever, I think - currently it's so daunting to figure out what you can expect from different lenses across different systems of different sensor sizes, etc. That said, lens data will only be rigorously valuable once many copies are tested (so as to not run into issues like the 70-200 fiasco that keeps being brought up over and over again to beat DxO over the head with).

As for DxO's 'image science', hallmarks of good science include transparency about methods with disclosure sufficient for someone skilled in the field to fully reproduce the experiment/test, and attempting to avoid bias.  DxO does not disclose their formulae or weightings for determining their Scores, and what they do disclose of their methods shows that their scores have intentional bias. 

I don't disagree with the bit about disclosing methods, being a scientist myself. You saying they have 'intentional bias' requires more proof than your simple claim, though. Bias towards what? Any formula has inherent bias, as any weighting system must. It's when people imply that there's a brand-specific bias that I take issue. I'm not saying that's what you're implying, but there are certainly those who imply it.

You could argue that they weight base ISO DR too much, and if you don't care about base ISO DR, then I can see how that'd bother you. OTOH, I think it's perfectly fine to weight base ISO DR far more heavily than high ISO DR b/c: (1) not doing so runs the risk of rating all similar size sensors roughly the same, and how does that help a consumer actually understand the *differences* between sensors?; and (2) if you shoot in an 'ISO-less' manner (even partially), you can - for a sensor with high base ISO DR - retain far more DR at high ISOs than the actual measured DR for any given higher ISO. Which - for me - makes the higher ISO DR numbers meaningless, and only the base ISO DR number relevant.


40
Reviews / Re: Tony Northrup - D810 vs. 5D Mk3
« on: September 07, 2014, 05:07:20 AM »
It's very difficult to compare total different AF systems with each other. The Canon AF system of the 5D3 and the 1Dx is even now, much more advanced then the Nikon D4s or the D810. What Northrup want to tell us that the D810 has a higher keeper rate compared to the 5D3 is his opinion, but I'm sure he is not really aware then of all the possibilities of the much more advanced AF of the 5D3.

It's your opinion that the AF system of the 5D3 is more advanced than what the Nikon D4s & D810 offer.

It's my opinion that Nikon's is far more more advanced in practice. Simply b/c of the 91k pixel RGB metering sensor for subject tracking.

Our difference in opinion probably boils down to you not caring about iTR on Canon, or Nikon's analog: 3D tracking. (considering you said: "So even the iTr on a 1Dx is hardly used in my case because of the lower AF speeds of the zone and 61 points AF compared to the other methods.")

Nikon's 3D AF tracking enables me to capture shots I wouldn't have dreamt of catching with my 5D3. Using fast f/1.4 wider primes (24mm and 35mm), I can always use the center point to initiate tracking, and then keep that subject in focus even after recomposition or the subject moving (which babies & running brides tend to do a lot). My hit rate for this type of photography shot up dramatically going from a 5D3 to a D810. And I wonder how many people actually realize that 3D AF tracking can be used for this to such great benefit. The 91k-pixel sensor does remarkably well in this regard of recognizing, say, an eye & sticking with it - it's so good sometimes, and so fast at shifting the AF point, that sometimes it feels like the camera is using the accelerometer data to measure my hand movements in order to shift the AF point (obviously, this isn't actually the case)!

It's funny - the term '3D' here is almost misleading, as it implies depth. Well, AF in general (in AF-C or Servo modes) tracks depth by default. The point of '3D' here is that it also tracks across the two dimensions of the frame; hence 2D + depth = 3D. But, at first, it seems counterintuitive, if you see what I mean - it's the 3D mode that tracks across the 2D plane, and all the other modes that automatically track depth (in AF-C/Servo).

Now, there are some other advanced features Canon offers - 5 high sensitivity dual cross-type points in the center with wider baselines, more cross-type points on the sides, and spot AF. Of these, I miss the cross-type points on the sides the most. I have yet to see benefits of the high sensitivity points, and spot AF, for my work. I'd love to quantify how useful these can be, especially the high sensitive wider baseline points for low light work. Roger Cicala's work showed no real difference in precision between the modern Nikon and Canon AF systems, but I believe that was in good light.

The 1Dx offered *the potential* to combine all those pluses of the Canon AF system with the '3D' tracking capability Nikon's offered for a long time (and Canon did as well, just using only depth information from the AF system - in other words, it didn't do it very well). Unfortunately, I never found it able to keep up, and stick as well to the initial subject, as Nikon's D810 or D4s. No matter what combination of use-cases/settings within those use-cases I tried on the 1Dx. The 1Dx is, however, much more capable than the 5D3 at this type of AF. It approaches the Nikons in this regard, but still has some catching up to do - as you'd expect for a 1st generation tech.

So, for my type of shooting, the AF system on the D810 far outperforms the 5D3 I used extensively for 3 years. And that was like icing on the cake, since I'd been wanting a better sensor for a long time.

YMMV for the type of photography you do, of course. If 3D AF tracking doesn't matter to you, the 5D3/1Dx offer very compelling AF systems.

It's just that when you make blanket systems like 'the much more advanced AF system of the 5D3', I really can't sit silent...

41
let's say for the sake of argument that Sony sensors really are better at this point in time (a judgment that is highly subjective and very suspect, since it hinges on tiny, tiny differences in just one subset of a sensor's overall performance

There's nothing 'highly subjective' here at all. It's a quantifiable, demonstrable fact. And it's not a tiny difference in just one aspect of overall performance. Low downstream read noise not only increases base ISO dynamic range, but can allow you to maintain high dynamic range at all ISOs if you know how to take advantage of 'ISO-less' sensors.

Furthermore, to think that Canon shouldn't be interested in dynamic range performance of their sensors is pretty short-sighted. Canon's trying to make a dent in the motion picture industry, and what's arguably the one thing expert cinematographers constantly go on about?

DR, DR, and more DR. Straight from the horse's mouth (Emmanuel Lubezki, DP on 'Gravity', 'The Tree of Life', etc.).

42
jrista: I wouldn't bother. You're not going to convince someone that sensor performance matters if that person - no matter how smart he is - appears to not have a single photograph taken in challenging light (in his shared collection anyway).

Those who care are those who've struggled out in the field, time and again, and are tired of fighting their equipment at the cost of the art. Especially when there are much better alternatives out there... that have existed for years, no less.

I think confirmation bias really runs rampant in some of these threads. My favorite was a previous comment by someone that basically tried to say that since Canon is still leading the market, DxO's image science must be wrong. Yes, let's correlate two entirely uncorrelated things.

43
Aglet: all good points. I liked Canon ergonomics too, but I'm faring just fine with Nikon too, once I switched those dials and indicators. The more sensitive D-pad on the D810 is actually a pleasure to use now. I hear this argument about Canon ergonomics being better all the time and, in the end, I really don't get it. Both systems are perfectly usable and don't get in the way of my photography as much as, say, Canon's limited DR or the 5DIII's complete inability to track subjects across the frame. My point being: there are things you can easily work around, and other things that prove much more difficult to work around. So Canon ergonomics is a tenuous reason to stick with Canon.

Plus, it's easy to find something to pick on from either company. For example, Nikon has the best implementation of programmable auto ISO, with easy access to exposure compensation in M mode with auto ISO. Canon finally put EC in M mode into the 1DX, but did so in one of the most unusable ways imaginable: either via the Q menu (requiring you to take your eye away from the OVF), or via the Set button (which we all like to assign to magnified playback view). Meanwhile, there's a dedicated EC button on the 1DX at the top of the camera... but, no, why assign EC to that?

So much for 'intuitive interfaces'... :)

44
EOS Bodies / Re: EOS 7D Mark II & Photokina
« on: September 04, 2014, 05:49:26 AM »

I agree with this. In this case, this is where Canon cameras are currently superior (at least, to most...I think the D810 is starting to change this). At higher ISO, most Exmor cameras still have ~3e- RN, even more. Canon sensors, on the other hand, drop to 2e- and below at the higher ISO settings, dropping to as little as ~1.3e- on some cameras at very high ISO. Those savings in RN are what give Canon's 1D X, 6D (and even to a small degree, the 5D III) the dynamic range/SNR edge at high ISO.


No. And I wish these myths weren't propagated with a resilient persistence as readily as they are on forums. First of all, if you want to talk about 'superior', talk about the Sony A7S and its sub 1 electron read noise.

More importantly, I'd treat those read noise numbers on sensorgen a little more softly. Don't get me wrong - I do value sensorgen - but do consider there are no error bars or stdev quoted. Second, it's not just about (upstream) read noise here. You're quoting fractions of electrons differences, yet ignoring QE differences?

The reality is: there are differences, but they're more in-line with generations of tech; beyond that you're splitting hairs at best. Furthermore, simply quoting the min read noise ignores # of pixels, which matters for sensible, normalized comparisons.

Fortunately, we can look at some test shots taken at the same focal plane exposures:

D810 vs 6D vs 1Dx vs D800E
http://tinyurl.com/n3cke3v

Are those differences - at ISO 12,800 no less - really worth talking about? I guess that's up to you, and if it were me I'd rank performance D810 = 1Dx > 6D > D800E*. I'd also remind you that the D810 is performing on par with the 1Dx despite having twice as many pixels! Yes the D800E is performing slightly worse, but at more reasonable ISOs of 3200 and below, it's a wash.

And this whole 'Canon cameras are superior' attitude... it's like brand religion. And it's demonstrably false. Now let's take a look at some other cameras, including the 5DIII which you think is somehow better than Exmor:

D4s vs A7R vs 5DIII vs D610
http://tinyurl.com/kd7vze3

Pretty marginal differences between the 5DIII, the A7R (Exmor), and the D610. Nikon D4s beats all of them, though. And, actually, at higher ISOs the A7R - after removal of the magenta tint in shadows - perhaps even slightly outperforms the 5DIII in normalized comparisons: http://tinyurl.com/k2dam52. At any rate, it certainly doesn't underperform compared to the 5DIII.

So where is this Canon 'superiority' and 5DIII's 'small degree [of an]... edge' at high ISO?

While we're at it, let's look at some APS-C cameras:

Canon 70D vs Nikon D3300 vs Nikon D7100 vs. Canon Rebel T5i
http://tinyurl.com/orncyb8

The Nikon D7100 (Toshiba sensor) outperforms all of those cameras, and the Nikon D3300 performs better than both the Canon cameras despite 1/3 EV less focal plane exposure (all other cameras received the same focal plane exposure).

I could go on, but I think you get my point...

The D810 ultimately drops to 1.3e- at it's highest ISO setting, but it still doesn't manage to eek out more DR than the 5D III, and certainly not as much as the 6D and 1D X.

Why are we still talking about high ISO DR disadvantage for the low (downstream) read noise camera (D810) when I clearly demonstrated in one of my previous posts that you can get demonstrably more DR at high ISO by underexposing & boosting exposure selectively - something you're less able to do with a Canon DSLR?

Furthermore, according to that site-no-one-speaks-of-here-but-who's-data-the-site-you-quoted-numbers-from-derives-its-data-from, there's no difference in normalized DR between the D810 and the 5DIII at high ISOs.

However, with the D810, Nikon seems to have introduced a bias offset, and they now have a roughly linear falloff in read noise compared to the flat RN curve of the D810. I think that further enhanced the value of using higher ISO settings, vs. using the camera "ISO-less", for anything above ISO 400.

Since you brought it up: I'm a little confused by sensorgen's numbers for the D810, as they disagree with DxO's. Sensorgen gives the D810 lower pixel-level DR than the D800/E, whereas DxO clearly shows the opposite. 

As for the value of higher ISO vs. shooting 'ISO-less' (to some degree anyway), that depends on what you care about. Comparing ISO 1600 vs ISO 400 + 2EV (same shutter speed/aperture), the tradeoff is between (1) a couple electrons more noise per-pixel, and (2) two whole stops of DR.

I personally, generally, prefer the flexibility of option (2). But now we're veering off the main point of our discussion, which is that the absence of downstream read noise (high base ISO DR) opens up possibilities otherwise impossible to achieve. And this doesn't come at some high ISO performance cost (compared to Canon). It didn't back when the D800 and 5DIII were introduced, and it doesn't now.

And it's a tenuous thread - if not utterly wrong in some cases - to claim that somehow Canon sensors in general are 'superior' to all others with respect to ISO performance. Compared to Exmor or no Exmor.

*The Nikon files have a bit more magenta in blacks at the highest ISOs, though this could possibly be from ACR's 'Shadows Tint' calibration.

45
EOS Bodies / Re: EOS 7D Mark II & Photokina
« on: September 03, 2014, 06:12:15 AM »
Let me reiterate:

There might still be *some* advantages even for the most 'ISO-less' cameras today to using, say, ISO 400 amplification vs. ISO 100 for the darkest of signals. Because there's still some finite read noise in the signal chain. However, it may be academic for many.

And for very very high ISO applications, where every fraction of an electron in noise counts, you'll see benefits to hardware-level ISO amplification. Again, b/c of that finite read noise that is still introduced by the signal processing. Or b/c of other things like conversion gain optimizations (e.g. what's suspected to be the case in the A7S). And when you're trying to make an image with 40-50 photoelectrons of less per pixel... every electron of noise saved counts.

Pages: 1 2 [3] 4 5 ... 19