Messages

16

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 24, 2013, 10:33:35 AM »

Interesting question, but I'm 90% sure that the rich ladies who want a "nice" camera make up less than 2% of the market for 5D and 1D buyers.  I'd guess that 70% are professionals or are used in a professional capacity with the remaining 25% being hobbyists.

I don't follow what you mean here. Would you elaborate ? Your numbers don't add up and neither does your logic.

My point is that I believe that many of those 25% started out as Rebel users (I notice you too started out as a Rebel user ... are you a "rich lady who wants a nice camera" too ?). Also, I wouldn't completely dismiss the 70% (who may also have started off as Rebel users) -- many pro photographers are in a position of being able to choose their equipment so if they have a strong affinity with a given brand when they are hobbyists / students, they are likely to stick with it as professionals.

Whatever percentage of the potential 5 series user base start out as Rebel users, they aren't going to make up a large percentage of the Rebel user base but they are still strategically important to Canon (and other manufacturers who might want to lure them away)

17

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 24, 2013, 09:26:44 AM »

Also, whatever you might say about the "majority of rebel users" -- I wonder how many 5-series and 1-series users start out as Rebel users ? (I did) There is a delicate balance here -- they do want to be able to encourage these people to ultimately upgrade to a 5 or 1 series body but they don't want to make the entry level body so awful that these users start out with a different manufacturer.

As did I.  So for us at least, and I suspect there are lots of others like us, it seems their 'cripple the low end to drive up-marketing' strategy is working rather well. 

I suspect that there are many more who are not "like us" -- we have more disposable income than the typical hobbyists.

18

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 24, 2013, 08:26:08 AM »

My point is that when people on gear forums say there is "no reason" not to include some feature, in fact there may be a very good reason from the manufacturer's perspective.  There may be a 100,000 good reasons ...

Yet you haven't stated any for the feature that we are discussing (AFMA). Instead, you have brought up two features that are not analogous -- art filters and the print button (my comments about the art filters apply to the print button -- it's at the top level)

Also you mention it costs something to add a feature in software but gross costs are not the same thing as marginal costs. The marginal cost of putting AFMA in rebel bodies would not be very substantial (they already have the software for this)

But that's not a convincing reason for a car company to design their sports car with a towing hitch, and that towing hitch may well turn away a lot of their potential buyers.

You're doing a great job of coming up with analogies that are not relevant, but not such a great job of addressing the issue at hand (which is again, AFMA)

And if the point is to send a message, as you say, to advanced users that they really should step up, well that is a perfectly valid business decision.

Well, you are welcome to defend it as a business decision. However, Canon's policy of going out of their way to cripple their lower end products does make them less attractive (and difficult to recommend to someone on a tight budget!)

Also, whatever you might say about the "majority of rebel users" -- I wonder how many 5-series and 1-series users start out as Rebel users ? (I did) There is a delicate balance here -- they do want to be able to encourage these people to ultimately upgrade to a 5 or 1 series body but they don't want to make the entry level body so awful that these users start out with a different manufacturer.

19

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 24, 2013, 12:10:03 AM »

So why the heck did Canon "cripple" the 5DIII by omitting so many features?  Why doesn't the 5DIII have fabulous built-in creative filters like Toy Camera and Grainy Black & White?  Is Canon just being greedy?

This is the first time I've heard anyone complain about this feature, and it is in jest. I agree that it is possible for a feature to have zero or even negative marginal value, and that in-camera art filters on a pro body  would have no or negative marginal value (if they were high enough in the decision tree to be accessible they would have a negative marginal value because of the resulting clutter, otherwise it would be about 0)

I don't agree that AFMA on rebel bodies is such a feature.

For one, AFMA is completely usable even if it's buried under layers of menus, because it doesn't need to be manipulated during day to day shooting -- it is only adjusted in controlled (e.g. non time critical) settings. Even on the 5DII (for example) it's buried in a 3 layer menu (go to the right tab, then select AF/other, then select micro adjust).

I don't understand why putting that in the Rebel would overly complicate things. They do already have equally confusing items (e.g. such as a two axis white balance adjustment which would have even many advanced users scratching their heads)

Art filters by contrast in cameras that have them tend to occupy spots close to the top of the decision heirarchy (sometimes they are actually hardwired into the manual controls which would be incredibly annoying for most 1D and 5D series users). They need to be done this way to be quickly accessible.

Of course, different users have different expectations.

Of course, but my point is that one of the reasons people buy inexpensive cameras is because of budgetery constraints, not because they are unsophisticated users.

I do agree with you that Canon's point of view seems to be that the Rebel line is unsuitable for the more sophisticated user. A savvy user on a budget probably shouldn't buy a rebel.

I simply don't believe your contention that Canon's reason for omitting this feature from their Rebel (or there XXD bodies from which it was removed) is to avoid adding "unwanted features". In fact on the contrary it looks like the point is to send a message to advanced users who might otherwise purchase rebel series bodies that they really should step up and buy one of their more expensive offerings.

20

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 23, 2013, 10:03:26 PM »

Hey, elford,
I see you got the 5D mark III.
Is the AF on this model really that much better than on previous models? What is the st. dev for it in Canon's units (if you know of course)?

Sorry it's confusing but the camera model that shows up beneath our names has nothing to do with what we own. I own a 5DII and a micro 4/3 camera. Before that I had an old Rebel (XS).

The standard deviation and the "tolerance range" in number of  AFMA units are not fixed quantities -- variability of the AF will depend on things like subject distance, and available lighting (either ambient or flash/ AF assist). Then the range for which an object appears to be in focus could be 2 or 3 or 5 standard deviations depending on depth of field.

Roger at Lens Rentals had some tests where he looked at the variability of autofocus of some different bodies (including the 5DIII) with some different lenses. My recollection was that with many tested lenses, phase detection on the 5DIII is more consistent than older bodies. Contrast detect (used in live view and on mirrorless bodies) is more accurate but slower. I'd suggest searching lensrentals for the article if you are intersted, it was a good read.

21

Third Party Manufacturers / Re: DxOMark trashes the Leica M9 sensor

« on: March 23, 2013, 09:55:51 PM »

I think you misunderstood my comment about "questionable test results" to mean "biased", I am not accusing DxO of being biased ... my point is that DxO does their testing in a certain controlled environment, lab or whatever but when we make photos in real world scenarios there are far too many variables that are simply not in the realm of any one company to accurately predict and test

Well, yes, of course -- they are in the business of measuring sensor performance, no more, no less.

This -- sensor performance -- is a really big deal to some people. I agree that for some shooting scenarios it may not be terribly important. (Actually, for what you do, some of digitalrev's reviews are probably more relevant!)

But we get some DxO fans who, without knowing my needs, (and the needs of other photographers) comment saying Nikon D5200 sensor has scored higher than 5D MK III, therefore it is a better camera and that I must accept their "facts"

Well, that's horribly ill-informed on their part. But I haven't seen these fans (are you sure they're not Nikon "fans" who would drop DxO like a bad habit if Canon started scoring higher than Nikon ?)

Anyway, the DxO score only tells us that the D5200 scored higher on the aggregate score, which in this case means that the sensor has less noise at low ISO. It says very little else about the relative merits of the two cameras.

22

Reviews / Re: Most Objective and Less Objective REVIEWER?

« on: March 23, 2013, 09:37:53 PM »

PZ, for example would criticize fast wide primes at f/1.4 for corner performance, comparing them, implicitly, to slower or longer  primes. Really, Klauss?

I don't think that's an accurate criticism. When he reviews test results (like soft corners or conspicuous barrel distortion or absence of distortions), he makes a point of mentioning whether those results are expected (or better or worse than expected) for that type of lens.

23

EOS Bodies - For Stills / Re: Which would you pick for a wife camera sony nex 6 or eos sl1?

« on: March 23, 2013, 05:50:51 PM »

Never been a fan of the 4/3's cameras.

The latest offerings are worth another look even if you haven't "been a fan". Olympus switched to using Sony sensors, so sensor performance has gotten a lot better -- just a tiny bit behind Sony APS-C (due to the smaller sensor). You get the best of both worlds -- Sony sensors and the m43 lens selection. Take a look at the DxO mark scores for the OM-D (and the other new olympus cameras like the EPL5/EPM5 etc) compared to the previous generation (EP3/EPL3/EPM)

24

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 23, 2013, 01:53:33 PM »

I don't think it's really necessary on Rebels since the focusing on them tends to be inconsistent anyway (at least that was the case with my old T2i). AFMA wouldn't fix that.

Not quite. Yes, the phase detection AF on most Canons is inconsistent but you can increase your good image outcome by a lot by doing AFMA. Just to give you an example. Let's say that without any adjustment your average focus is located at +6 with the shot-to-shot error of +/-3 units (that means that the lower and upper bounds are +3 and +9). If you keep it at that you will never get a sharp image. Now, if you do microadjustment of -6 units, you are zeroing your average focus position and your range becomes from -3 to +3 - much better. You will get much better images in this case.

Yep. Even if you are within the error range, you will get improvements. Suppose the keep rate when properly calibrated is 95% (+-2 standard deviations). Let's say the standard deviation is 3 AFMA units. Even if you're off by half a standard deviation (1.5 units so 95% of your shots fall between -4.5 and +7.5) your miss rate increases by about 50%, and goes up to 16% (3 times as much) if you are off by 3 AFMA units.

If you are towards the edge of the 95% band, it becomes a crapshoot (at that point it's about 50/50 -- you basically need to hope that the random component of your AF errors are in the right direction to make up for the miscalibration)

25

EOS Bodies / Re: Congratulations Canon to another great Camera release!!

« on: March 23, 2013, 01:19:31 PM »

Should we expect new full frame bodies any time soon?

I heard the 5D Mark IV is just around the corner. The question is, whether to buy that or wait for the Mark V which will be even better.

26

Third Party Manufacturers / Re: DxOMark trashes the Leica M9 sensor

« on: March 23, 2013, 01:13:12 PM »

Yes "Lenses are a different matter" but lenses are an important part of a DSLR, no lens, no photo! -- no one is disputing DxO being a leading tester of sensors, however real world photography requires lenses, but when a "leading tester" like DxO easily screws up on testing lenses,  I too wouldn't necessarily trust "the leading tester's opinion, especially when what I need is a Camera and a lens to take real world photos

On this we agree -- I don't give much weight to their opinions, but thankfully they do a good job at publishing their measurements which I can read and understand.

Your point that the sensor is merely part of the camera which is merely part of the camera system is understood and well taken. That is why we don't just have sensor benchmarks -- lens reviews (measurements and qualitative/subjective reviews) as well as reviews of the bodies themselves (both the dpreview approach and more subjective testing) all factor into appraising gear.

So I'm not trying to say that the sensor is everything, but I am pointing out that DxO do a pretty good job at benchmarking  sensor performance. I also find hard cold numbers to be a refreshing alternative to exaggerated claims from manufacturers, fans, or new users trying to justify their purchase after the fact (instead of doing due diligence before the fact).

However, I do engage in arguments with people who keep insisting that everyone must accept the DxO test results as the absolute fact

Well, their measurements (unlike exaggerated claims from fans) are at least objective. You can argue about what the measurements mean (but even that requires some understanding, um,  of what the measurements mean)

As I said before, a good image requires a camera and lens

this we agree on

when companies like DxO or any other company gives out highly questionable test results, I do not necessarily refer to them as "objective".

I don't see how this is related to the first point. Regarding questionable test results, it's not at all clear that DxO's all-over-the-map lens numbers point to any bias (lots of variance but not much bias! but it does shed some light on why their lens tests don't have the same stature as their sensor benchmarks). It's like seeing a leading quarterback play a bad golf game and then wonder if they really are a good quarterback.  DxO's sensor benchmarks and methods have been studied and analyzed ad-nasueum and have stood up to scrutiny pretty well.

27

Lenses / Re: small primes to go with SL1?

« on: March 23, 2013, 12:55:16 PM »

I'm actually pretty excited about the announcement of the SL1. price-wise and function-wise it seems very competitive against m4/3 cameras which I've been considering for a while for a back-up/casual camera.

But that's why it's not really competitive against m4/3 cameras -- even with the reduced form factor, it still has the extra thickness because it's using a mount with a longer flange distance.

That said, Canon's wide non-L primes are pretty tiny (among Canon's smallest lenses) and optically quite decent (the 35mm f/2 and the 28mm f/2.

28

Reviews / Re: Most Objective and Less Objective REVIEWER?

« on: March 23, 2013, 08:40:24 AM »

So that pushed me to make this Poll, to see the users opinion and maybe the credibility of the reviewers.

Seems to me that many don't understand what "objective" means. "Objective" is not the same thing as "credible" or "useful". The luminous landscape reviews for example are largely subjective -- they generally do not contain any measurements or test chart shots, they are largely descriptions of the author's experience with and impressions of the lens.

Lens Rentals is a very useful site, and contains a large amount of both objective tests as well as purely subjective "Roger's take" comments. 

TDP have the test charts, but I don't see how that website has any business leading this poll. Objective doesn't mean "subjective and agrees with me". Ironically, the two highest scoring sites in the poll are two of the more subjective reviewers (e.g. lenstip and photozone are more objective than LR or TDP)

29

Third Party Manufacturers / Re: DxOMark trashes the Leica M9 sensor

« on: March 23, 2013, 08:15:44 AM »

+1  ... actually several people at nikonrumors echo similar sentiment, recently DxO ranked some Sigma and Tamron zoom lenses ahead of some of the better Nikon zoom lenses ...  also the Nikkor 70-200 f/4 scores higher than Nikkor 70-200 f/2.8  ... lot of the Nikon users are not very thrilled by it. I'm no expert at DxO scores but ranking Sigma 120-400 lens ahead of some of the Nikon lenses clearly shows that real world results and lab tested results are very different.

Lenses are a different matter -- DxO are the leading tester when it comes to sensors, but there are a number of sources who test lenses and appear to do a better job than DxO. Having said that, I wouldn't necessarily trust the popular opinion at a camera "rumors" site (which tend to be heavily stacked with "fans") over an objective reviewer.

30

Third Party Manufacturers / Re: DxOMark trashes the Leica M9 sensor

« on: March 23, 2013, 08:12:07 AM »

And I will reiterate again, that their ISO score is weighted toward the base ISO, and not at more moderate levels

Completely wrong, please review their method for ISO score. The dynamic range and color depth scores use base ISO. The ISO score is the highest ISO that meets a number of image quality criteria (the highest for which a certain noise level and dynamic range are maintained)

And then there's the downsampling to 8 megapixels...certainly that helps the D800...a ton!  It's also a meaningless method of scoring noise.

Also not true. Normalizing to a common resolution is completely sensible unless you're always viewing 100% crops. The typical use case is to ultimately print images at the same size or otherwise rescale the image to the same size. To do otherwise is a "pixel peeping" approach. They have "screen DR" so that you can also see the per-pixel results, but these are not as meaningful.

 

If it scores less than 10 points behind the D800E, then in reality, it's actually 10 points ahead..

I don't think so -- perhaps you can show us your industry leading benchmark that demonstrates this.

.and why give the 800E one more point just for not having an anti aliasing filter?  The sensors are THE SAME... 

The anti-aliasing filter obviously has some impact on noise properties of the image. You wouldn't expect an enormous difference but then the test doesn't show an enormous distance so I don't really understand this complaint.

If average people like me can predict the outcome of a test, 6 months or more before the product is even announced in the public domain, while it is still a rumor...then you have to admit, it does smack of favoritism and intentionally skewed test results.

Maybe it just means that average people can predict how well the next sensor from each major manufacturer will perform. Given the relatively slow trajectory in sensor improvements, today's performance is a pretty good predictor of tomorrow's performance.

I mean, the little Nikon D5200 scores 2 points better than the 1Dx...GET F***ING REAL...

Canon's low ISO noise hurts them because the aggregate score is heavily weighted towards low ISO performance. The 1DX has substantially better high ISO performance but that is not the only factor the overall score takes into account. The 1DX also has a number of features that are important to serious photographers, but these don't enter into the test either.
 

That's like saying a Prius scores higher than a Ferrari 458 because it gets better gas mileage,

Well, what do you mean by "scores higher" ? Trying to assign a single numerical "score" to something complex is always a difficult task. Here's one -- should a Mazda Miata score "better" or "worse" than a Honda accord ?

The reason they do report a single aggregate score is that many users do not want to dig deeper. But for those who are, they publish all the underlying measurements.

or because the driver tends to obey the speed limit more closely.  The test is just wrong, skewed to portray a "truth" that is too narrow in its scope to have the meaning that many think it does.

The test is just fine. You are confusing the test with the score. You may disagree with the way the factors are aggregated, but they do provide their measurements for you so that you can get the most out of their tests even if you don't agree with the weighting scheme (or if you want to understand why the results came out the way they did).

The notion that the test is intentionally biased in favor of Nikon is just silly. DxO devised their test score and methods before Canons struggles with low ISO performance became the elephant in the room.