Previous news story    Next news story

DxOMark introduces 'Perceptual MPix' score for lens sharpness

By dpreview staff on Dec 17, 2012 at 16:00 GMT

DxOMark has introduced a new metric of lens sharpness that it's calling the 'Perceptual MegaPixel'. This is designed to give a simple, easily-understandable indication of how a lens impacts on the resolution of the camera with which it's paired. The measurement is based on standard MTF (Modulation Transfer Function) testing, but also takes into account the sensitivity of the human visual system to detail. The data can be explored over at the DxOMark website, www.dxomark.com.

Our lens reviews are prepared in partnership with DxOMark, using test data shot in its labs. Do you think the 'Perceptual Megapixel' would be a useful number to include in our reviews? Let us know in the comments below.

Press release:

DxOMark introduces the Perceptual Megapixel as a new unit for measuring the perceived sharpness of lenses

A simple way to quantify the impact of lens sharpness on camera resolution

December 17, 2012 - DxOMark, the website of reference for measuring the image quality of digital cameras and lenses, announces the launch of the Perceptual Megapixel, its new measure of lens sharpness. Expressed in units already well-known to the general public, the Perceptual Megapixel provides photographers with a means of instantly understanding the performance of a lens by quantifying its impact on the perceived resolution of the camera with which it is coupled.

A measurement correlated to human vision

Designed by DxOMark experts, Perceptual Megapixel quantifies the sharpness "perceived" by the photographer while taking into account all the characteristics and defects of the lens and the camera.

"This new unit of measure replaces the MTF scores that we have been publishing up until now. Perceptual Mpix weights MTF measurements with the sensitivity to detail of the human visual system, thus providing a score for the sharpness perceived by our eyes. We think this new measurement will be more understandable and practical for photographers who use our website," explained Frédéric Guichard, Chief Scientific Officer for DxO Labs.

An average of 45% of all megapixels lost

The Perceptual Megapixel measurement allows for making numerous comparisons among cameras and for the first time quantifies the impact of lens sharpness on camera resolution. Thus the perceived resolution for a camera can be significantly different from the native resolution depending on the lenses with which it is coupled.

For example, when coupled with a Canon EOS 5D Mark II camera with a 21.1 Mpix sensor, a Sigma 35mm F1.4 DG HSM lens has a Perceptual Mpix score of 17.2, whereas a lens such as the Carl Zeiss Distagon T* 35mm f/1.4 ZE has a Perceptual Mpix score of 15.2.

The example above is based on data from DxOMark's database of test reults for more than 2,700 camera and lens combinations. These tests reveal that, on average, about 45% of the resolution is lost due to lens defects.

"The Perceptual Megapixel measurement allows photographers to step away from manufacturers' announcements about resolution and to answer a key question when the time comes to change equipment: would it be better to buy a new camera or a new lens?" observed Nicolas Touchard, VP Marketing, Image Quality Evaluation, for DxO Labs.

This new metric introduced by DxOMark aligns with the standardization work currently being undertaken by the IEEE, the International Imaging Industry Association (I3A), and the ISO - and in which DxO Labs actively participates. Perceptual Mpix is also based on the scientific findings of CNES (the French space agency) pertaining to the optimization of digital acquisition, notably in the field of satellite imagery.

More than 2,700 camera/lens combinations available

Freely available to all, DxOMark has been designed to serve the entire photographic community, from serious amateur photographers to photo experts and journalism specialists. It has a threefold mission: to provide objective measurements of digital cameras and lenses; to analyze camera and lens performance; and to provide a means for making meaningful comparisons.

DxOMark now has a database of more than 2,700 available camera/lens combinations. Dozens of DxO Labs engineers and technicians in six DxO laboratories perform these measurements, taking a photo every 10 seconds. Testing protocols are entirely reproducible and conform to the international standards established by the Camera & Imaging Products Association (CIPA), the International Imaging Industry Association (I3A), and the Camera Phone Image Quality (CPIQ) group.

Comments

Total comments: 182
12
Timmbits
By Timmbits (Dec 17, 2012)

DXOmark stated that the Fuji X-S1 has a smaller sensor than the Pana LX5, Oly XZ1, etc, and that the Fuji XS1 has no optical image stabilization. This is one example, of where they don't even know what they are talking about. Their "reviews" are littered with errors, that they don't respond to or correct.
It seems like serious research and data come second to marketing hype and plugging their name everywhere they can. It's all just a tactic to get our attention so they can sell their software product... but unless they can get the most basic, easiest facts straight, they will have no credibility with me.

Comment edited 4 minutes after posting
3 upvotes
Nick-
By Nick- (Dec 17, 2012)

Measurement are averaged over the lens aperture range so lenses with large ranges will be penalised using this method.

For example the Nikon 105 macro with a minimum aperture of f32 has now slipped down in the rankings.

I think the min. aperture included in the average should be fixed to something like f16 for every lens.

0 upvotes
Donald Duck
By Donald Duck (Dec 17, 2012)

Nonsense. What they need to do is to report what they measure, not to encrypt it.

Let us hope that this is just a marketing trick and they will still report the MTF data.

3 upvotes
Mssimo
By Mssimo (Dec 17, 2012)

They only tested one lens with the D800 and D800E.

Tamron AF 28-300mm F/3.5-6.3 XR Di VC LD Aspherical [IF] Macro Nikon

Results for sharpness:

D800 8
D800E 11
D600 9
D3x 8
D4 7
D700 6

D3200 5
D90 4
D40x 4

Im not sure what to make out of this data. Its a superzoom lens..but that is all the data we have for these high MP cameras.

Note: added a few more cameras

2 upvotes
Erik Magnuson
By Erik Magnuson (Dec 17, 2012)

D800 8, D800E 11, D600 9, D3x 8 seems anomalous. Interesting that the D600 and D800 versions also have different vignetting (1.2 vs.. 1,0 stops) while the D800/D800E are the same. Lens copy variation perhaps?

1 upvote
rrccad
By rrccad (Dec 18, 2012)

they are certainly lacking on the one camera body (or two variants) which probably need this measurement more than any others on the nikon side - which lenses perform the best on the highest resolution 35mm sensor.

1 upvote
digitac
By digitac (Dec 17, 2012)

Yes, interesting.

0 upvotes
Erik Magnuson
By Erik Magnuson (Dec 17, 2012)

I think this score will correlate pretty well with lens sharpness in general, i.e. a lens with a P-Mpix 15 (of 18) will be noticeably better than a lens with a score of 8. However, it may not correlate as well for small differences, e.g. a score of 14 and a score of 15 may mean lenses that merely have different strengths (i.e. center vs. edge, best vs. wide open performance) where the weighting used does not match the user preference.

If a user merely wants to know "what's an excellent vs. merely good lens", then this number will be useful. Buying the lens with the highest number is rarely going to be "wrong" but it may not be optimal. If you want to know "what the best lens for 35mm f/2 corner to corner" then this number will not be as useful. Digging down into the detail graphs may help, but if you have specific questions you should not expect average numbers to be the answer. Also remember YMMV so small differences in measurements may not be significant in field use.

4 upvotes
random78
By random78 (Dec 17, 2012)

"Digging down into the detail graphs may help, but if you have specific questions you should not expect average numbers to be the answer"

That is true for those users who already understand about the detailed data available. However most new users will just look at the score and use that as the yardstick of which lens is "better". People like ratings because life is much simpler if someone could tell you that thing X is superior than thing Y. Unfortunately once such ratings are in place, most people just use them without realizing how they are just an arbitrary mix of a larger number of detailed parameters. In reality comparing lenses requires looking at many factors and in my opinion it is better if the users are exposed directly to all of them so that they know that it is not a simple question of which one is better. By feeding a single number you give the false impression that somehow this one number is encapsulating the overall goodness of a lens versus the other.

Comment edited 1 minute after posting
1 upvote
Erik Magnuson
By Erik Magnuson (Dec 18, 2012)

> However most new users will just look at the score and use that as the yardstick of which lens is "better".

And what happens when they do this? They get a good lens. Maybe not the absolute best lens but it's likely to be darn close. It may not be the best value either but that's what happens if you ignore price. Can you suggest anything better that lazy users should use as a guide?

0 upvotes
Entropius
By Entropius (Dec 18, 2012)

Not being lazy and actually reading a real review, assuming someone will write one :)

2 upvotes
random78
By random78 (Dec 18, 2012)

> "And what happens when they do this? They get a good lens"

They can also be paying a lot more than needed just to get a lens which "scores" better, even though for their uses it might not even be the better lens.

>Can you suggest anything better that lazy users should use as a guide?

What Entropius said - don't be lazy. I don't think promoting laziness by making it easy to make misinformed decisions is a good idea. We have already seen this with DxoMark sensor scores where people on forums talk about one sensor being rubbish because it is 10 points lower than the other one on DxoMark, even though looking at the detailed graphs tells a very different story. The issue with "scores" is that you always want the one with the highest "score" and feel dissatisfied with anything with lower score. Whereas by looking at detailed reviews you actually understand which product better meets your needs.

0 upvotes
Erik Magnuson
By Erik Magnuson (Dec 18, 2012)

Most sites cater to both the lazy and non-lazy reader: for the lazy there is the single number score (e.g. 73% or 3.5 of 5 stars or 19) and for the non-lazy the details behind those numbers (but rarely the formula that unites the two.)

0 upvotes
random78
By random78 (Dec 18, 2012)

And I disagree with every site that does that :) including dpreview's percentage scores for cameras which I think are even more useless than DxoMark's scores. If I am lazy when reading say dpreview or photozone, I just read their conclusion paragraphs, but not the scores which can be safely ignored. Also with dpreviw and photozone the score comes at the end after the whole review, as sort of a summary. DxoMark on the other hand emphasizes its scores and de-emphasizes the detailed results which this score is based on. Also by the way not every website does that - for example imaging resource, dcresource or steves-digicams do not have any scores for cameras, even though these websites do a lot more point and shoot reviews as well. Similarly lenstips or slrgear do not have any lens scores. They expect their readers to read the review not reduce the whole review into one meaningless number.

Comment edited 3 minutes after posting
0 upvotes
Erik Magnuson
By Erik Magnuson (Dec 18, 2012)

> They can also be paying a lot more than needed just to get a lens which "scores" better,

You can't protect the buyer against themselves. If they want to buy something that's the "best" regardless of price, then they will do it.

My main concern is not lazy buyers, but fanbois who will think that a lens that scores 15 "blows away" a lens that scores merely 14 not realizing that because of weighting & rounding the lenses may not differ much at all.

0 upvotes
Donald Duck
By Donald Duck (Dec 18, 2012)

They may not get a good lens. It is easy for the manufacturers to compromise the design to score higher at some aperture, in the center. Fast lenses have to deal with different challenges and that often hurts a bit sharpness at f/4-f/5.6 across the frame.

If you do this, you are giving an incentive to the manufacturers to compromise the design. You are the most powerful photo review site on the web, owned by the biggest photo equipment retailer. If the average Joe makes his purchase decisions based on that stupid number, the manufacturers would have to make him happy.

0 upvotes
Erik Magnuson
By Erik Magnuson (Dec 18, 2012)

> compromise the design to score higher at some aperture, in the center

It's a weighted average over the entire field so your concern here is misplaced.

> Fast lenses have to deal with different challenges and that often hurts a bit sharpness at f/4-f/5.6 across the frame.

And if you are looking for the "sharpest" lens, then enses that compromise peak sharpness for speed are not the best choice (e.g. the Canon 50mm f/1.0L) and this metric will reflect that.

Comment edited 46 seconds after posting
0 upvotes
PowerG9atBlackForest
By PowerG9atBlackForest (Dec 17, 2012)

Pomposity!

1 upvote
VA-ArtG
By VA-ArtG (Dec 17, 2012)

Interesting. I will have to review their data before reaching any conclusion on the value.

0 upvotes
Lan
By Lan (Dec 17, 2012)

Questions to DxO:

1) How do you source your lenses for testing?

2) On average, how many copies of a lens do you test?

2 upvotes
Mssimo
By Mssimo (Dec 17, 2012)

This is why I love lens rentals blogs. They also post result with large batch of lenses.

0 upvotes
Entropius
By Entropius (Dec 18, 2012)

Roger Cicala also understands the optics and mathematics well enough to write very competently about the technical things.

0 upvotes
PoohBill
By PoohBill (Dec 17, 2012)

From the perspective of an admitted "lazy" and "total amateur" photographer the pairing values make sense. Saves me from ploughing through the minutiae (yup, "lazy") offered by the "pros" (yup again, "total amateur") just looking for good matches with my next buy.

Please do. Thank you for asking.

-Bill

2 upvotes
lester11
By lester11 (Dec 17, 2012)

DxO oversell (or possibly undersell, I don't know) Mpix as a "single" number. It isn't, but hooks many of those commenting here into a knee-jerk rubbishing of a "single" figure of merit. Mpix is in fact a range of numbers for a given lens, one for each body it mounts on. Now that is seriously useful, and tells me something I would never have got from its MTF scores.

Comment edited 36 seconds after posting
0 upvotes
Imagerymayhem
By Imagerymayhem (Dec 17, 2012)

Where can these scores be accessed?

I looked all over their website and couldn't find it.

Comment edited 18 seconds after posting
0 upvotes
Erik Magnuson
By Erik Magnuson (Dec 17, 2012)

Via the lens comparison tool (or the lens selection tool), e.g.
http://www.dxomark.com/index.php/Lenses/Camera-Lens-Database/Canon/Canon-EF-40mm-F28-STM/(camera)/483
It's the sharpness score units. You can more details in the "Measurements" tab under "Sharpness"

0 upvotes
Imagerymayhem
By Imagerymayhem (Dec 17, 2012)

Thanks!

0 upvotes
Tape5
By Tape5 (Dec 17, 2012)

Not my definition
or unit
or measurement
or measurement devices ( if any )
or lab
or question
or interest.

I am jealous. I have always wanted to define a unit and have the only device in town to measure it.

3 upvotes
Timmbits
By Timmbits (Dec 17, 2012)

hahaha well said!

0 upvotes
craig66
By craig66 (Dec 18, 2012)

Yes, that thought did cross my mind too. The article mentions that DXO has been active on IEEE, ISO and other standards bodies committees. Before jumping to too many conclusions there are some questions to be answered:

Are any such standards published and freely available?

Is the DXO implementation rigorously conformant to any such standards?

Are any such standards encumbered by patents?

1 upvote
onlooker
By onlooker (Dec 17, 2012)

One thing that I would still like to see for each lens is bokeh evaluation. I understand people smarter than I can use MTF measurements for that, but there is probably an even more sophisticated mathematical algorithm to evaluate smoothness of out-of-focus blur.

Hit "Like" if you would like to see it.

12 upvotes
Timmbits
By Timmbits (Dec 17, 2012)

I think that a more significant value, is the DEVIATION from center sharpness, rather than some average number. Even something like chromatic aberration would amplify that, so it could be used as a better metric.
That is where a good lens stands out.

Comment edited 2 minutes after posting
0 upvotes
Richard Murdey
By Richard Murdey (Dec 18, 2012)

Lawrence Olivier famously asked of Dustin Hoffman "have you tried acting?"

With regards to your attempts to evaluate bokeh, I would ask you, "have you tried photography?"

0 upvotes
onlooker
By onlooker (Dec 18, 2012)

Oh, Richard, please tell me, is it okay to consider bokeh when selecting gear, or is it the antithesis of being a true photographer to be informed before purchasing? I don't know, I apparently missed that memo. I mean, I have only been photographing for the past 40 years, so what do I know.

Does that answer your question?

3 upvotes
memau
By memau (Dec 19, 2012)

you may find this page useful
http://toothwalker.org/optics/bokeh.html

good bokeh is the combination of several effective aberration corrections, like spherical aberration will affect foreground and background bokeh, chromatic aberration will ruin the transition zone of hilghlight bokeh, and most of us all know, sometimes bad aspherical lens will leave some ugly signature.

So I will say it's not quite practical to offer such a ”bokeh  test” because there is no formal definition for "good bokeh", although major companies already noticed it's importance

0 upvotes
onlooker
By onlooker (Dec 20, 2012)

Good read, thanks, memau.

0 upvotes
fad
By fad (Dec 17, 2012)

After reading the article now on LL concerning DXO mark scores, I recommend they hire a good technical writer to make all their testing clear to non-technical public. The LL piece does not simplify, but it makes things much more clear.

Writing and communication is just as technical as optics, and requires just as much skill. Engineers rarely have these skills. Clearer communications would increase the value of DXO scores perhaps 20 fold, and give the company much more good will.

The great novelist Thomas Mann once said that a professional is someone for whom writing is much more difficult than it is for other people. That's because a good writer sees many things that other people ignore and that they do not normally realize defeat them.

It's hard for people to realize that a university education and being a native speaker does not give one professional communication skills. But it's true.

4 upvotes
Cy Cheze
By Cy Cheze (Dec 17, 2012)

Did Mann ever write anything under 500 pages that a typical camera nut would understand? If DxOMark can turn a magic mountain of babble into a score, does that not perform a great service?

1 upvote
fad
By fad (Dec 18, 2012)

I quoted a novelist, but said they should hire a technical writer.

Please don't confuse the engineering types. :)

0 upvotes
john Clinch
By john Clinch (Dec 17, 2012)

Is it just one number for a lens?

If it is then I'd say that it will only lead to confusion.

There are just so many factors. Lets imagine two different 70 200 zooms. One constant F4 the other f2.8. How do we compare them.

If we compare them at f4 then a consumer buying the f2.8 might be shocked to find poor performance wide open

If you build in a factor to allow for being soft wide open then almost all fast lenses will score worse

So my conclusion is that simplifying lens ratings will be a waste of time

I have to say I thought the DPreview graphic was good. I particulalry like seeing the crop of the chart. I also like photozone.de and slrgear.com as well

However i never seemed to get to grips with the DXO mark lens reviews

In summary I believe that this will a simplification to far

PS why mega pixels, thats proportional to resolution squared. Why use a hard to understand number

DXO mark seems to be great science over simplified into single score by an ignorant manager

2 upvotes
Erik Magnuson
By Erik Magnuson (Dec 17, 2012)

The "one number" for a lens is it's DxOMark score. P-Mpix are the units for the "Sharpness" measurements on DxOMark - either a single number (weighted score for best aperture) or with more details (e.g. plotted by by aperture & focal length or by distance from center for each aperture similar to the DPR graphs.)

0 upvotes
Tilted Plane
By Tilted Plane (Dec 17, 2012)

Why complain? More is more. Don't read it if it isn't worth your time. For me it's all interesting. Thanks all around. Now just give us a link to this particular new metric (rather than just the general site) and I'll be really happy.

2 upvotes
RadPhoto
By RadPhoto (Dec 17, 2012)

Agree with you

0 upvotes
Matthew Miller
By Matthew Miller (Dec 18, 2012)

Because if the metric steers people the wrong way, it will eventually steer the market the wrong way, which will in turn make more new lenses designed around an arbitrary and meaningless score rather than attractive real-world performance.

0 upvotes
dszc
By dszc (Dec 17, 2012)

YES!!!
As long as DXO uses TOTAL lens performance, all the way to the corners.

I am a full-time pro of 35+ years and can be a super-geek with Airy disks, circles of confusion, Nyquist frequency and the rest. I find the DXO sensor scores a pain and generally misleading. But a simple single realistic MP score that includes total lens performance, from corner to center to corner, would be very helpful at a glance, save me a lot of time, and encourage manufacturers to start making lenses that are up to the demands of the current sensors.

All I really care about is getting good sharp pictures as expediently as possible. If the lens has an adequate MP score, then I'll look at it. Check the 3-D plot that slrgear has; the checkerboard targets at the corners that dpreview has; and the bokeh and flare examples that dpreview and photozone have.
Then I'll order one and test it for myself.

0 upvotes
Mescalamba
By Mescalamba (Dec 17, 2012)

In theory its interesting, unfortunately it has many flaws.

In practice you show unprocessed photo only if you are total amateur or review writer. So obviously, even bit of sharpening will increase percieved sharpness. And other kinds of processing too.

Then there is that Zeiss. Zeiss has unique feature called micro-contrast which increases percieved sharpness of photo taken using this lens. Simply put even that sample graph is bullsh*t. Sigma is excellent lens (most likely one of best Sigma ever did), but its except being technically perfect, its just "another lens". No special rendering, no look, only bit of Sigma colors. Which thanks to being really sharp lens is fixable in post-processing. Zeiss has look, rendering, micro-contrast. Which is obviously possible to improve even further in post-processing.

To sum it up. We all postprocess, if we value our work or ourselves. Which makes this "perceptual MPix" pretty much useless. I think MTF is good way to measure lens. Lets keep it.

1 upvote
kelpdiver
By kelpdiver (Dec 17, 2012)

Goodie - what we all needed was another proprietary benchmark from DXO that will oversimplify a larger collection of proprietary tests they do and lead to really stupid discussions. bravo!

5 upvotes
PowerG9atBlackForest
By PowerG9atBlackForest (Dec 17, 2012)

Advertisement:
I will swap my two DXO-rated 10 lenses for one 20.

1 upvote
random78
By random78 (Dec 17, 2012)

Sigh. DxoMark's insistence on hiding lots of useful data behind much less useful "scores" is frustrating. I understand that they want to make it easy for average non-tech consumer who may not be willing to look at all the graphs and resolution maps etc. However in an attempt to do so they give this false impression that a single number somehow can give an accurate impression of the relative performance of two lenses (or sensors). It just cannot and they risk misleading and confusing the users.

1 upvote
Erik Magnuson
By Erik Magnuson (Dec 17, 2012)

Detail is there - you have to drill down a bit through the tabs to find it. It's as if DPR presented it's summary chart first and you had to click on parts of that to see the detailed interactive graphics.

0 upvotes
Otto Fabricius2
By Otto Fabricius2 (Dec 17, 2012)

Sharpness is important but far from the only parameter. How are those measurements done? Will there be M-Pix scores for both lens center and corners? And for different apertures? If it is only the center, or just the "best" aperture the data is not enough as a help in choice.

It also seems necessary to get some more information about the weighting procedures mentioned - and just how "the human vision" is being incorporated in the model.

2 upvotes
AP7
By AP7 (Dec 17, 2012)

More interpretations of the test data is better than the less. I believe that this M-Pix score is not going to replace other test score, such as, standard MTF, etc.

So, I don't see anything wrong. DPReview can certainly use this.

It'll be certainly useful.

0 upvotes
micksh6
By micksh6 (Dec 17, 2012)

Did you read the article?
"This new unit of measure replaces the MTF scores that we have been publishing up until now"

So, if press release is formulated correctly, there will be no MTF anymore.

0 upvotes
AP7
By AP7 (Dec 17, 2012)

Then, it is an extremely bad idea.

0 upvotes
AP7
By AP7 (Dec 17, 2012)

@micksh6: You are not completely right.

I just checked dxomark website and checked few tested lenses. They only replaced Lens Metric Scores for Sharpness. Previously, they had average resolution lp/mm, now they have P-Mpix. Both are single valued (So, no problem of accepting P-Mpix instead of lp/mm).

However, all the detailed measurements for Sharpness (Global Map, Field Map, Profiles), Transmission, Distortion, Vignetting and Chromatic aberration are still there. Nothing is changed in that section.

So, I am not worried at all for the change. Its kind of renaming. I welcome the change.

Comment edited 2 minutes after posting
0 upvotes
Cacophonic Visions
By Cacophonic Visions (Dec 17, 2012)

If it's possible to view the tested Perceptual MPix score for each lens on each body that it can mount to, then it may be useful. As long as the other test values (distortion, vignetting, etc.) remain in place, adding another rating might be ok.

1 upvote
Raist3d
By Raist3d (Dec 17, 2012)

I looked at some lens scores. I think from a relative scale point of view this seems to get an idea where lenses fall in the lens continuum.. but as far as the score itself goes, it seems to suggest a lens with say a 5 perceptual megapixel score would not get more resolution on say a camera from 8 megapixels to 16 megapixels....

But I am seeing lenses there that clearly do get a bit more resolution as the camera goes up past a 6 perceptual megapixel score... so I am not sure how this helps in "just looking at the score" terms...

Of course, maybe I misunderstood what this metric means, but then that defeats the purpose of being simple/non confusing.

Comment edited 40 seconds after posting
0 upvotes
DrugaRunda
By DrugaRunda (Dec 17, 2012)

I like this score, it equals the resolution field a lot across the cameras and across the formats, except showing that FF can resolve more with a given lens, which is a fact of life. Also M4/3rds are about the same as APS-C for that metric.

So hopefully this bring the MP race into perspective a bit more.

Will be interesting what it shows for Foveon cameras, once the tests are made.

3 upvotes
brunobarolo
By brunobarolo (Dec 17, 2012)

One single number to characterize the resolution of a lens at all f-stops, center, border and corners, at close and far distance, and for zooms at all focal lengths?

How meaningful is that single number? Not a lot, IMO.

6 upvotes
Erik Magnuson
By Erik Magnuson (Dec 17, 2012)

It's "just one number" only on the summary page. In that sense it's like DPreview's single "Optical Quality" bar in their review summary chart. In the details pages, DxO plots this value vs. aperture & focal length and also center vs. edge if you want to see it.

4 upvotes
Dennis Watts
By Dennis Watts (Dec 17, 2012)

MTF seems to be the most consistent measurement. Although when you consider the number of people that are new to photography this measurement seems to have some value.

I guess it won't be of any direct value to me, but it will help me explain to my friends why buying good glass is better than constantly upgrading to the latest body.

4 upvotes
lensez
By lensez (Dec 17, 2012)

This will be useful only if photographers agree that it corresponds to "the sensitivity of the human visual system to detail." We have to see DXO measures and compare them with what we already know about certain lens/camera combinations. Let's see how DXO rates the DP2M.

1 upvote
Richard Schumer
By Richard Schumer (Dec 17, 2012)

Yes! Yes! Yess!

To me, this metric is *more* important than sensor resolution alone.

Unless the lens produces the image the photographer had in mind, the sensor size is meaningless.

I've noticed on my Pentax k10d I can pull "sharpness" up beyond the Bayer limit but only on certain lenses and never on a zoom. I know this is "false" sharpness, but to me it shows the lens is outperforming the sensor. I can blow these "falsely sharp" images much bigger than ones taken with an inferior lens, even if perfectly focused and without motion blur.

I would find a test that measures this very useful.

3 upvotes
Uwe Steinmueller
By Uwe Steinmueller (Dec 17, 2012)

Interesting concept but one number will hardly tell it all (corners, f-stops)

Here is a shocking graph from the DxOMark document:

http://www.dxomark.com/itext/articles/intro-pmpix/03.jpg

Would like to know why the excellent Sigma 35mm f/1.4 perform that low on the 7D?

0 upvotes
DxO Labs
By DxO Labs (Dec 17, 2012)

The graph has just been corrected a few minutes ago.

P-Mpix is based on a weighted average across the field and for each focal lens, we use the performance of the best f/stop.

For further details, please go to www.dxomark.com

The DxOMark team

0 upvotes
Uwe Steinmueller
By Uwe Steinmueller (Dec 17, 2012)

Now corrected. Now it makes sense to me.

Comment edited 7 minutes after posting
0 upvotes
DxO Labs
By DxO Labs (Dec 17, 2012)

You can refresh or reload your browser window.

0 upvotes
Carbon111
By Carbon111 (Dec 17, 2012)

I just *knew* that Zeiss Distagon was junk! ;)

3 upvotes
Mescalamba
By Mescalamba (Dec 17, 2012)

1) Sigma 35mm f1.4 is actually amazing lens
2) you know nothing

3 upvotes
Carbon111
By Carbon111 (Dec 17, 2012)

1) Winky = Joke.
2) Sense of humor = Get one.
Thanks for the ad hominem attack though, you must be a blast at parties. :D

7 upvotes
Mescalamba
By Mescalamba (Dec 17, 2012)

I dont go to parties.

1 upvote
Carbon111
By Carbon111 (Dec 19, 2012)

An acerbic personality will put you at the bottom of the invite list in no time.
FWIW, I *like* the Sigma 35. :)

1 upvote
nekrosoft13
By nekrosoft13 (Dec 17, 2012)

Dxo up to it again, annoucing/making up another standard that doesn't help anything.

"answer a key question when the time comes to change equipment: would it be better to buy a new camera or a new lens" and sorry, but that is just big BS. resolution is not the only thing that matters.

1 upvote
Tonio Loewald
By Tonio Loewald (Dec 17, 2012)

"a key" not "the key".

It's an interesting idea although linear resolution is probably more useful (MP after all goes up as the square of linear resolution, which is something careful reviewers have to explain over and over again).

0 upvotes
Kinematic Digit
By Kinematic Digit (Dec 17, 2012)

Do we really need another standard that frankly is just as confusing? Lines Per Height has always worked well for me. It covers both resolution of the body, the lens. Perceived resolution is just as confusing if not more so.

Looking at those numbers seem relatively meaningless to me considering that much of the Perceived numbers are changed by lens profile correction of in camera and also software applications that correct for some of those issues without any compromise to image quality.

I don't think this makes it any easier to decide which lens (or body) is better to pick from at all.

3 upvotes
Dennis
By Dennis (Dec 17, 2012)

I agree - this is just another attempt to simplify complex issues for people too lazy to learn about them. Sort of like the dpreview scores of 78% ... how that score helps anyone pick a camera is beyond me, unless they're just too lazy/ignorant to think for themselves. What good is a lens test that shows a score based on the "best f-stop" when I'm buying it to be good wide open ? How do I know if it's sharp corner to corner if it's got stellar center performance and the average number ends up high ? Anyone willing to make decisions based on this number deserves what they get.

1 upvote
Ronj2
By Ronj2 (Dec 17, 2012)

From most of the lens scores it looks like we don't need more than 10-16 MP. Zoom lens scores are mostly under 10mp the new 28-300 is about 5mp. Even some newer primes score low. Interesting how the best score is close to wide open rather then around F 5.6 -8.

Lets see some 100% crops to prove it.

Comment edited 2 times, last edit 2 minutes after posting
1 upvote
Danielvr
By Danielvr (Dec 18, 2012)

I think this 'Perceptual Mpix' metric has practical value for a lot of people and as such I welcome it. However, if it should become very popular, I'd be concerned that manufacturers will optimize new lenses to get a good Mpix score at the expense of other qualities (and, frankly, I think they've been optimizing them for 'sharpness' way too much already -- I really miss the beautiful, 'analog' rendering of detail that the properly resolving but lower acutance lenses of the 80s still had).

0 upvotes
louisjaffe
By louisjaffe (Dec 18, 2012)

I've noticed the painful optical shortfalls of many lenses when used with higher-end camera bodies. Sony NEX-7 for instance: the Zeiss 24mm f1.8 is the only lens to do justice to this 24.3 megapixel sensor. Sony's lenses, zooms and primes, are mediocre by comparison. I own this equipment and have seen the evidence on my monitor and in prints.

Same story with my 5d Mk II. The Canon 35mm f1.4 L uses every pixel this camera has to offer. Most of my other Canon lenses can't. Granted, there's more to lens performance than resolution. For starters, contrast is very important. But the Perceptual Mpix score makes a lot of sense to me, presuming it's well implemented.

2 upvotes
Total comments: 182
12