DXO Lenses for D800

tcole1983

So I have looked through the test and obviously noticed some of the big guns missing from the reviews as have many. Do you find the tests and results useful though? I still just feel like the scores are skewed and don't tell you the whole story. I think say the Nikon lens vs Samyang that the Nikon should get points for being fully functional while the Samyang doesn't autofocus and the camera can't correct for distortion or CA. Not to sound like a total Nikon snob...as I know some of the other lens manufactures make good lenses, but even with my Tokina lens that was very sharp it didn't perform at the same level as my Nikon lenses in regards to flare, ghosting and CA (I know they "test" for CA, but it doesn't seem to change the score much).

Any other thoughts?

PB_PM

One thought, don't rely on DXO marks to make decisions.

haroldp

Reducing something as complex as the many facets of lens performance into a single number is beyond silly.
... H

JJ_SO

Is it really necessary to judge this as "silly"? What else does everybody of us do when it comes to choose a lens out of a whole bunch of available ones? We also have our individual priorities. I would be silly to rely on DxO's lens points for my own purposes, but it's not silly to favor certain characteristics as long as it's comparable and understandable. Plus, it's a kind of overview.

I think it isn't more silly than giving marks on a scale to 5 points - in each case it's good to read the whole test, but if a lens gets "only" 2 1/2 points and there are lenses with 4 points, it's clear on which test results I spend more time.

Also, we should not forget: mostly they're testing a single copy. Now, that is really much more something to be called "silly" although I understand, testing a whole bunch of copies multiplies the efforts and is mostly not well paid and difficult to finance.

So, to me that number is an orientation - but not a verdict.

Paperman

Reducing something as complex as the many facets of lens performance into a single number is beyond silly.
... H

Agree with that partially but you don't have to settle with just the general score. One can look at aspects which matter most to him. If it is sharpness, nothing will give a better idea than line tests/figures. It sure will be more reliable / more objective than one's eye.

PB_PM

Also, we should not forget: mostly they're testing a single copy.

I don't know any review sites that test more than one copy. Photozone for example only tests a second sample if they detect major centering issues.

JJ_SO

Then give lensrentals.com a go. Roger Cigala states he's not doing testing but only looking to resolution. But that he does with sometimes dozen of copies on various bodies. And the results are worth to be considered, 'sides his great writing style

haroldp

Then give lensrentals.com a go. Roger Cigala states he's not doing testing but only looking to resolution. But that he does with sometimes dozen of copies on various bodies. And the results are worth to be considered, 'sides his great writing style

Absolutely, lens rental is a great source for this.

Many of the differences between closely ranked lenses are within sample variation ranges.

... H

Symphotic

I find the DxO ratings to be an enjoyable use of internet time. However, part of my job is to come up with optical performance tests and specifications for video camera systems. (Your "Hi-Res" and my "Hi-Res" should be the same.) MTF curves or DxO ratings are possible ways of evaluating camera and lens systems, but our goal is to come up with simple, meaningful tests that production can run and certify before they ship a camera, and that the customer can replicate when they receive their camera system. Since Nikon and Sigma can ship a camera and lens combination that DxO gave a number like "39" but we can't put our copies on a tripod and see we can also get a 39, we can't use it. I just read the DxO ratings for fun, not for profit.

PB_PM

Then give lensrentals.com a go. Roger Cigala states he's not doing testing but only looking to resolution. But that he does with sometimes dozen of copies on various bodies. And the results are worth to be considered, 'sides his great writing style

Guess I never think of lenrentals, never looked at any of their reviews (or the site for that matter).

TaoTeJared

DXO's combined scores are one step above a "Joke." I have never been able to find the correlation of their scores with what they "test." If it is not repeatable by others, then they are hiding something. If you are confident in your processes, you have no reason to hide anything. I do find it odd that they 1- have waited so long to test the D800, 2- if they want to be the "go to" site for lens tests, why they don't keep a better stock of lenses or test more. There are a basic set of lenses that most people have, that they seem not to.

The only tests I have found they do well is their Vignetting and Distortion results (which makes since from their editing software) but as a "this is better than that" not so much. I just use it to begin to base some lens profiles up in LR and change it as I feel necessary
.
It seems they weigh most heavily on their own made up "Perceptual MPix" which is some mysterious number that says what the human eye can see measured in Mega Pixels. Now anyone who actually prints images, owns the lenses they test and are able to compare them for your self, knows it is complete BS. You don't have to be a rocket scientist to figure that out - you just have to be willing not to assume "just because it is on the internet it is true" and do the tests yourself.

The big "hoopla" over what is best on the D800 is nothing more than what they say scores the best with their sub-par roll-up scoring. It has nothing to do with the D800 or any other camera. The same lenses that tested well on the D300 are the same lenses that tested well on the D800. There is no "holly Cow" in there stuff.

Imatest is a company/software that Lens Rentals & Photozone use which actually does test lenses well and it is repeatable. I feel that that software is much better and gives much more detail and they are transparent in how they test.

The only thing DXO does better is to keep their site updated, and publish articles. Just because they do that, doesn't make them the authority of lenses, sensors or anything else, just a better marketing company than others.

Golf007sd

@TaoTeJared +1 Well said.

tcole1983

I personally don't use it to make decisions on what I have purchased. I have looked at it on occasion but never as an absolute to decide on a lens. I was mostly curious what others thought about these tests and conclusions. I would probably prefer to see side by side photos to compare for myself.

proudgeek

I never look at that stuff. Just makes me dizzy. I'm guessing most of us couldn't tell with the naked eye the difference between two lenses whose scores varied by 1-2%. Maybe I skew too far AWAY from the tech side, but I try not to get caught up in raw numbers. One man's pov I guess.

donaldejose

If DxOMark lens testing is invalid as some people in this thread believe why did DPReview adopt it for their lens testing calling it "the trusted industry standard for independent image quality measurements and ratings?"

PB_PM

Most likely because Dprivew (owned by Amazon, if you didn't know) likely gets a kickback.

donaldejose

Oh my, conspiracy theories again. Often such allegations are all too convenient ways to explain something which contradicts our opinions. Perhaps it is just cheaper for DP review to use DxOMark data than to test lenses themselves? I wouldn't call that a kickback. Yet, if those tests were so obviously "junk" DPReview wouldn't want to let that "junk" downgrade the reputation of their business.

PB_PM

Dpreview isn't what it once was. Yes it is big, but that's about it. DXO products have been given more and more notice there over the last year or so. If you don't think DXO is paying to get more attention (advertising), you must be dreaming.

donaldejose

How about the D7100 scoring the same as the D7000 when it clearly should be better from the images I have seen? Does that prove DxOMark is wrong or does it prove my eyes are mistaken?

JJ_SO

Actually, I don't get it: DxO still doesn't support D7100 RAW files. But is able to give D7100 marks or points?

When I asked them today why I can't see my RAW files although a lot of functions are already supported and only some are planned for April 2013, I was told, yes it is supported if I shoot JPGs. I was impressed by the tone of the supporter and by the fact I got answers in German. From a French company. That alone scores 10 points...

donaldejose

If a French company is outsourcing to Germany they are moving up! )

JJ_SO

donald, the scores were 83/84/80 with D7100/D5200/D7000 and one score difference is really not much. And sensorwise, the D7000 was not better than D5100 - I don't see a IQ difference between the two.

haroldp

These scores are likely within sample variation tolerances.

At the performance level of any modern sensor, factors like lenses, and focus accuracy and speed have more impact on real IQ than these sensor variables.

... H

PB_PM

Since we have no idea what DXO's test bench is like, we have no way of knowing if they even use a consistent lighting method for their tests. Considering how close those cameras are, might as well save some cash and get a D7000.

JJ_SO

That was exactly what I was told today: "I don't know if you have any idea how we are working, but blabla..."

I replied, you're right, I have no idea and share this faith together with most people outside DxO which are only a couple of billions.

Never had such a weird conversation with supporters.

Save some cash and get a D7000? What would I miss now? The multiple usage of OK-button, the high ISO thing, the mode dial lock, the AF module I will miss once it's working equally on both sides, but then I hesitate to make a bigger list of all the progressy things. So, why not?

donaldejose

Well, I think haroldp is correct. I do suspect we will see a significant sharpness difference between the 24 mp no AA filer D7100 and the 16 mp D7000 AA filter when the same subject is shot with the same lens. These sensor only tests of color depth, dynamic range and ISO just aren't measuring the added resolution of the D7100 sensor when coupled with very sharp lenses.