[Emacs]

They also don't understand lens testing
Should I believe my 70mm Limited is no better than a Canikon kit zoom?

DXO? These guys surely understand how to test glass. Unlike Klaus, BTW.

[monochrome]

DXO? These guys surely understand how to test glass. Unlike Klaus, BTW.

It isn't the testing that's the problem. They surely do know how to test.

Its the scoring.

[westmill]

I will stick and trust to photozones reveiws and opinions where lenses are concerned

[ElJamoquio]

DXOMark doesn't even attempt to include Bokeh. DXOMark doesn't seem to consider the application of the lens when it gives the (most highlighted) score.

Neither DXO nor photozone really do what I would consider 'adequate' contrast testing. That said, Photozone has never steered me wrong.

---

Last edited by ElJamoquio; 04-02-2012 at 04:22 PM.

[glanglois]

Nor does it weigh the application of the camera. Not intended to. The technical part is one piece of a very complex puzzle that will never be expressed as a number. Even a complex one.

[Laurentiu Cristofor]

Many times I've put the 15mm on, looked through the viewfinder, and just said 'wow'.

I'm still waiting for that moment.

Neither DXO nor photozone really do what I would consider 'adequate' contrast testing. That said, Photozone has never steered me wrong.

They do rank the Tamron 70-300 as sharper than the DA 55-300.

I like photozone too, but no one is perfect.

[pingflood]

I don't know where I picked it up, but from what I was told by someone very knowledgeable, that digital camera sensors are made to operate at a certain level of amplification without much noise. It usually is the default lowest and that raising or lowering the ISO from there, either forced or through turning on extended ISOs, creates more noise. So if a camera has an ISO that starts at 100, but can be brought to 80 or 50 with extensions, you would suffer a noise penalty. Any salt to that?

Well, the "extended" ISOs are accomplished through over- or under-exposing, which means you will possibly clip data at one end of the spectrum. If your lowest sensor ISO is 200 and the camera as an extended ISO 100 setting, it gets there by overexposing ISO 200 by a stop and then taking it down a stop when processing the signal. It should not yield more noise, but less dynamic range.

As for the rest, assuming the same exposure (i.e. amount of light hitting the sensor) you are actually better off cranking the ISO (up to around 1600 or so depending on the body) as long as you can avoid blowing out the highlights.

[Kunzite]

DXO? These guys surely understand how to test glass. Unlike Klaus, BTW.

Because... you like their results (i.e. Pentax getting bad scores) better?
I know what I have in my bag. Btw, Falk iirc proved the little 70mm is more than very sharp.

[Smeggypants]

As expected, the performance is pretty close to the K-5.

On the other hand the aesthetics are light years away from the K-5

[Smeggypants]

Lens score should be considered with the testing platform (camera). The top lenses scoring 30-34 are on full frame cameras. All their lenses appear to get a +10 or +12 factor when shown on a full frame. If you take the lenses tested on the D3x and drop to the D7000, you'll get a -12 to the score, putting it closer to what the Pentax lenses are scored at.

Now, that being said - the top scoring Pentax lens is the DA 35 F2.4 on the k-5. It outscored (at 15, so full frame would be 27-28):

DA 70mm Limited
DFA 100mm WR
DFA 50mm F2.8
DA 40mm Limited

Just a quick comparo of the 35 F2.4 and the 100 WR, both reached 46lpmm, the 100 WR has 0% distortion vs 3% on the 35mm, the 100 WR has 0.5ev less vignetting than the 35 F2.4, and it's chromatic abberations are 3 micrometers smaller.

The only thing the 35 F2.4 has on the 100WR is light transmission, which seems to be at .6Tstops better. Now, the general mark of the Limited lenses are the lower light transmission, and it seems to be the big factor in the DXO score. In fact, the 40mm F2.8 is also 46lpmm, 2% distortion, same vignetting, and less chromatic aberrations, but lower light transmission and it scores 4 points lower than the 35 F2.4.

It's weird.

Who cares? I can't imagine a serious photogpapher even worrying in the slightest about all that . They just get on and take pictures and don't let numbers distract their artistic minds

[JinDesu]

Who cares? I can't imagine a serious photogpapher even worrying in the slightest about all that . They just get on and take pictures and don't let numbers distract their artistic minds

If a "respected" testing company that is used by many photographers to determine which lenses are better and which sensors are better, and thereby allowing them to select the best of the best, is skewing the results wrongly for one company (in this case ours) - I would care at the minimum.

[ElJamoquio]

Some of the numbers they give Pentax lenses are a joke. I've seriously considered writing my own software to analyze lenses, it's not that hard. Sharpness and CA are rather easy to quantify, and no one seems to test for 'contrast' although tests for this would be straightforward as well. No one does a standard autofocus speed test. These are all straightforward.

At the end of the day the human is much more discerning on things that really seem to separate lenses - bokeh, ease of use, etc.

[Smeggypants]

If a "respected" testing company that is used by many photographers to determine which lenses are better and which sensors are better, and thereby allowing them to select the best of the best, is skewing the results wrongly for one company (in this case ours) - I would care at the minimum.

Indeed, but that's exactly my point. Using numbers to select which lens is better!! Such a practise goes against the ethos of photography IMO.

For example 'the numbers' were against the Sigma 30/1.4 and many 'photographers' who didn't understand the concept of shooting with a shallow DOF and nice Bokeh were parrots on forums to these numbers without even trying the lens. fantastic lens and good that Sigma had the sense to optimise it for centre sharpness wide open and not worry about sharpeness at the edges which in 99% of images are bokeh anyway.

testing companies and especially 'reviewers' are never impartial and a great many of them talk nonsense too. Anyone who relies upon these people for their purchases is doing themselves a disservice IMO. Also hose who parrot the propaganda as fact need to stop it.

[JinDesu]

But therein lies the issue - nowadays, everyone will start off by looking at numbers first. They look at online reviews. They rarely come on forums and have a full discussion. A lot of these people do not get anywhere special, but quite a few learn quickly and become great photographers. However, since they started by looking at these skewed numbers first, our brand was not given the fair playing ground to be put in the hands of these good photographers. These photographers go out with their Nikons, their Canons, gain fame and work - but what about new Pentax users?

I don't imagine the older days were that different - but information was much harder to come by. So people would go to photography class that would say "Canon only" or "Nikon only" and so come out with a skewed mind as well. They are not in the wrong - but the providers of information are in the wrong.

I think I should care that DXOmarks is skewing the Pentax results. I would write a complaint - "Why do your measurements not seem correct with Pentax lenses". Maybe I may make a difference, maybe I won't. But it wouldn't be so detrimental to think: Hey, what if Pentax had a fair playing field on DXOmarks when people look at the review pages?

It's easy to sit back and say "the great photographers do not measurebate" - but one doesn't become a great photographer without going through trials and errors of learning. And if the new generation grow up on Nikon and Canon based on pure review numbers - Pentax will not survive.

[Na Horuk]

Hm, so to test the lenses better they should use the same sensor for all of them, right? Or maybe the same film. And then just cut it up when using APSC lenses. Now that would be an interesting comparison. Film also wouldn't be affected by different camera brands and in-camera processing either.