crashpc
Veteran Member
This is lol discussion. TT chill out, some just refuse to see, and YOU are not that internet god, who will solve thing when others doesn´t want these to be solved...
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
For anyone who still thinks DxOMark isn't biased.
- Thread starter TTMartin
- Start date
brownie314
Senior Member
- Messages
- 2,139
- Reaction score
- 907
Well, lots of discussion on the this topic here. I do not think DXO is biased, but cannot prove it (can't prove a negative right? (well maybe, but that is another topic)).
Image from DxOMark.com , used for educational purposes.
Evidently DxOMark's scoring system gives two points simply because it's mounted on a Nikon camera.
--
All statements in my posts represent my interpretation of data, research opinion or viewpoints.
The opinions expressed are not representations of fact, and are subject to change without notice.
All images are used for educational purposes.
What I like about DXO is that they are the best site to compare many, many different cameras. And it is a direct comparison because they use the same testing methods for everything. Unlike other review sites that just give you a qualitative "it looks a little better than X to me" - this gives real numbers to go by. It is probably not perfect, but it is the best we have right now.
What you're seeing is not bias, except possibly your own 
Tests and ratings are based on lens + camera combos. Not the bare lens by itself.
Some combinations produce stronger results than others.
For example, here the exact same lens (the famous Otus), on the exact same mount (Nikon) gets two wildly different scores:
http://www.dxomark.com/Lenses/Compa...-ZF2-Nikon-on-Nikon-D7000___1242_792_1242_680
This is because one is on a D800, and the other is tested on a D7000.
If DPreview seems to give the Nikon combinations a consistently higher score, that's because right now Nikon makes better sensors than Canon, and better sensors = you get better results out of any given lens. That may change in a few months or a few years, but for now it's just reality. I'm a Canon user but I have no problem admitting that.
Tests and ratings are based on lens + camera combos. Not the bare lens by itself.
Some combinations produce stronger results than others.
For example, here the exact same lens (the famous Otus), on the exact same mount (Nikon) gets two wildly different scores:
http://www.dxomark.com/Lenses/Compa...-ZF2-Nikon-on-Nikon-D7000___1242_792_1242_680
This is because one is on a D800, and the other is tested on a D7000.
If DPreview seems to give the Nikon combinations a consistently higher score, that's because right now Nikon makes better sensors than Canon, and better sensors = you get better results out of any given lens. That may change in a few months or a few years, but for now it's just reality. I'm a Canon user but I have no problem admitting that.
Last edited:
crashpc
Veteran Member
I think this is faulty, because Canon combo is rated better in individual aspects (resolution) but gets lower overall score. What you all say is simply invalid. Don´t you see that?What you're seeing is not bias, except possibly your own
Tests and ratings are based on lens + camera combos. Not the bare lens by itself.
Some combinations produce stronger results than others.
For example, here the exact same lens (the famous Otus), on the exact same mount (Nikon) gets two wildly different scores:
http://www.dxomark.com/Lenses/Compa...-ZF2-Nikon-on-Nikon-D7000___1242_792_1242_680
This is because one is on a D800, and the other is tested on a D7000.
If DPreview seems to give the Nikon combinations a consistently higher score, that's because right now Nikon makes better sensors than Canon, and better sensors = you get better results out of any given lens. That may change in a few months or a few years, but for now it's just reality. I'm a Canon user but I have no problem admitting that.
brownie314
Senior Member
- Messages
- 2,139
- Reaction score
- 907
crashpc
Veteran Member
Yes, because that lens on the D800 resolves 29 P-Mpix and on the D7000 it resolves 14 P-Mpix so the score of the lens on the D800 should be much higher.What you're seeing is not bias, except possibly your own
Tests and ratings are based on lens + camera combos. Not the bare lens by itself.
Some combinations produce stronger results than others.
For example, here the exact same lens (the famous Otus), on the exact same mount (Nikon) gets two wildly different scores:
http://www.dxomark.com/Lenses/Compa...-ZF2-Nikon-on-Nikon-D7000___1242_792_1242_680
This is because one is on a D800, and the other is tested on a D7000.
In the case I posted the D7100 resolved 6 P-Mpix and the 70D resolved 7 P-Mpix, with the rest of the scores being equal. The overall lens score should have higher for the lens on the 70D, not lower by 2 points.
--
All statements in my posts represent my interpretation of data, research opinion or viewpoints.
The opinions expressed are not representations of fact, and are subject to change without notice.
All images are used for educational purposes.
There is no "resolution" category so I'm not sure what you're talking about.
Even if the lens attached to a canon, outscores the same lens on a Nikon
in one category, DxO's overall score is based on a weighted average of multiple categories.
Sharpness and Transmission are more important to the lens score than distortion or vignetting.
In fact, I think they may value transmission of light more than anything.
But I don't have their formula.
If you think there's bias, link to a comparison where a certain lens attached to a canon outscores the same lens on a nikon in every category, yet the nikon is rated higher anyway. You won't find it.
Even if the lens attached to a canon, outscores the same lens on a Nikon
in one category, DxO's overall score is based on a weighted average of multiple categories.
Sharpness and Transmission are more important to the lens score than distortion or vignetting.
In fact, I think they may value transmission of light more than anything.
But I don't have their formula.
If you think there's bias, link to a comparison where a certain lens attached to a canon outscores the same lens on a nikon in every category, yet the nikon is rated higher anyway. You won't find it.
Last edited:
According to DxOMark's Perceptual Megapixel Score with that lens the 70D resolves more than the D7100.
crashpc
Veteran Member
Yes! We don´t have the formula, and it doesn´t look we can count it as results are wild at times. That´s all we need to know.
I don´t have to fulfill this condition to think it is biased. Because of logic. There can be SINGLE important parameter to decide a thing. It´s like "unless you paint five dots on the paper, I won´t take it as there is even a single dot, because you didn´t draw them all". Nonsense. Wrong formula. And IF you draw LESS dots for certain situation, i´ll take your result, while more dots will lead in poor result. Cool!
--
Why does he do it?
Last edited:
The sharpness category is based on DxOMark's Perceptual Megapixel Score, which is resolution score.
If the overall lens score is not based on the lens component scores of how that lens performed on that camera what is it based on?
In the example I posted all those scores where the same.
See above.
Reilly Diefenbach
Forum Pro
Stop reccing yourself.
sportyaccordy
Forum Pro
LOL!!!!
Also I'm pretty sure some of the IQ tests normalize for resolution.
Also I'm pretty sure some of the IQ tests normalize for resolution.
brownie314
Senior Member
- Messages
- 2,139
- Reaction score
- 907
Why would it be hard to figure out this weighted average? There are what 4 catagories. Pick four independent tests, then it is a simple matter (remember linear algebra) to figure out the weights. Then, once weights are known - THEN you can go and see if those weights are changing to "help" Nikon cameras. I am too lazy to do this (or I guess I don't care enough) - but this is an elementary exercise.
Sharpness
D7100 = 6P-Mpix
70D = 8P-Mpix
Transmission
D7100 = 6.6T-Stop
70D = 6.6T-Stop
Distortion
D7100 = 0.1%
70D = 0.1%
Vignetting
D7100 = -0.3EV
70D = -0.3EV
Chr. aberation
D7100 = 10um
70D = 10um
DxOMark Lens Score
D7100 = 12
70D = 10
It's actually five categories, sharpness, transmission, distortion, vignetting, and chromatic aberration. But unless you add a sixth Nikon bonus category you don't get any further than here and the simple math breaks down.
brownie314
Senior Member
- Messages
- 2,139
- Reaction score
- 907
crashpc
Veteran Member
What is this?
Some parade of drop outs with learning disorder unable to count that bigger number is actually BIGGER?
Lol TT don´t feed hungry trolls....
Some parade of drop outs with learning disorder unable to count that bigger number is actually BIGGER?
Lol TT don´t feed hungry trolls....
Before assuming the worst, I'd assume it is simply a flaw in how either she sharpness or overall number is obtained. A rounding or math error. There's an old quote - "Never attribute to malice that which is adequately explained by stupidity."
Not that I'd call the scoring issue stupid, just an honest mistake that should be brought to their attention. I think the sharpness figure is where the problem lies:
They test the lens at 150, 200, 250, 300, 400, and 500mm.
For each of those lengths they shoot at 4 or 5 apertures.
The overall score number might be based on some average of those 30+ shots.
The single Sharpness number may be based on just a limited set of those shots - the lens under optimal conditions. They have a note in small text right next to those scores that says "best at f=150mm & f/5". So maybe the canon copy of the lens wins at that specific setting.
In fact if you look at the sharpness graphs and measurements, you can see that's almost certainly what's happening.
- At 150mm, the Canon wins at any aperture.
150, f/5... canon (right) shows more green.
However at 200, 250, 300, 400, and 500... using f/11 or higher, the nikon wins. So due to some quirk of lens or camera design, the nikon is diffracting less.
f/11 and f/29 are all better on Nikon, for every focal length except 150.
One more possibility - the P/MP is resolution dependent, and they fail to explain that.
Meaning 6P/MP on a 24 megapixel camera, beats 7P/MP on an 18.
I really doubt the site is just run by Nikon fanboys getting secret backroom payoffs from nikon to just arbitrarily add points to nikon's scores. I think they just need to revamp or add a disclaimer to their sharpness number so it better reflects what the graphs already show:
Visual graph - Nikon (right) shows more green and less red in center sharpness.
tkbslc
Forum Pro
Speaking of clueless, can you explain why they gave lower perceptual sharpness resolution on the camera with the higher score?
Keyboard shortcuts
- f
- Forum