[JimmyDranox]

Many people expressed doubts about DxO Mark test results. And by good means. Now we had an answer for this. Still, I don't know how a camera like Canon 1Dx is behind Nikon 3300D.

You Tube

You Tube

Many reason to doubt those measurements, but especially No 9, where I have to make a break to read all of this parameters that are ignored in DxO tests.


PS. ''Trust nobody'' seems to be the only thing you can trust this days.

[Giklab]

No. 9 is a piece of cr@p. There's a reason it's called DxOMark SENSOR ratings, not CAMERA ratings.

Same goes for No. 7.

Nos. 6 - 4, see No. 9 above. Also, it doesn't matter if Canon sensors almost reach the "excellent" x or y for some measurements. The other sensors reach higher numbers, and are therefore more excellent, as reflected in their higher scores.

No. 3 is shown without any explanation, then No. 2 is shown. Pretty ironic IMHO.

[Alcazar]

You post this video as if it was revealing something shocking the public needs to know.

When in fact, this is a video for people like you who didn't bother to read on the dxomark website what those scores actually represent, and need a video laying it out in front of them.

I think it's fine that the scores don't reflect the cameras' ergonomics or autofocus accuracy, since that's never what dxomark was set out to do.
They give you only the sensor information, it's your decision what to do with that information and how to value it against the other aspects that make a camera great.

So... it's not that the numbers aren't correct, it's that until now you were mistaken what those numbers meant.

[Rondec]

They are what they are. The fact that Canon's sensors don't have as good dynamic range at base iso seems to be true. If you are someone who seldom shoots at base iso, then that is probably unimportant to you. DXO Mark graphs are probably more useful than their scores, as they show multiple points and what they measure at various iso settings. You also can decide with the screen or print buttons whether to see a normalized graph (for 8 megapixels) or a per pixel graph (less useful in my opinion).

Normalization is probably the best way to truly compare sensors. The assumption is that your print/viewing size will remain constant, whatever camera you are shooting with.

People probably do over value the DXO Mark scores. When the K5 scores two points better than the K3, it was a big deal, but the reality is that they are basically equivalent, but the K3 has more resolution (not really factored into the DXO Mark score) and the K3 does have a lot more features.

These scores are one (very small) factor that folks should take into account when purchasing a camera. There are other reviews that talk about the other things like auto focus, ergonomics, frame rate, and other features. Certainly if I owned Canon lenses, I wouldn't take these scores as a reason not to shoot Canon, although there is no doubt that up till recently Canon cameras have had less shadow detail and have been more prone to banding.

[JimmyDranox]

Alcazar, You can disagree with my opinion, but you cannot assume what I have read or not, what I understand or not. Keep this for yourself.

For your peace of mind, I can tell you that I read those DxO ratings, and camera reviews, for more than 3 years. And my doubts about those ratings came from this readings.

More than this, I always download test pictures from at least two sites, DPR and Imaging Resources, to make comparison between cameras.

[ZeljkoS]

If you like Canon more than Nikon or Pentax, feel free to buy it and use it.

I like Pentax more than Nikon or Canon, mainly because in-body share reduction, great legacy lenses and great quality of RAW images.

Don't really care what Canon or Nikon users rave about, I'm happy with my K-5IIs, best camera in the world for what I paid for it.

DXO Mark is testing sensors, not cameras. If you don't know how to use that data, don't visit their website. You'll sleep better.

[audiobomber]

I learned nothing new in the video. I understand what DXOMark scores mean, how they are arrived at, and what they do not measure. This video is a bunch of FUD, intended to deny the substandard performance of current Canon sensors.

[JimmyDranox]

Thank you for your advice, It seems that nobody remembers that was a time when Pentax users had a lot of complains about DxO marks of Pentax cameras. OK.

A good day to you, gentlemen.

[audiobomber]

Thank you for your advice, It seems that nobody remembers that was a time when Pentax users had a lot of complains about DxO marks of Pentax cameras. OK.

A good day to you, gentlemen.

I don't remember complaints about camera scores. Pentax sensor scores have always matched the competition. I certainly remember lots of complaints about DXOMark lens scores. Maybe that is what you are referring to?

[ChristianRock]

Audiobomber is correct, I think... The issue with DXO doesn't seem to have much to do with what's being explained in the video, Jimmy. And I think their sensor testing is also not what's usually being questioned. I think what people usually question is that the DXO lens tests are performed in a basement under artificial light and a pre-set working distance in a flat field. While these parameters work well to test macro lenses (flat field), they tell nothing about lenses created for 3D effects, or lenses created for the purpose of landscapes. Also, these tests tell nothing about microcontrast and color rendering.

[BrianR]

10 questions about this video YOU don't know the answer to:

1. Why is it even a video?
2. Seriously, why not just a static list of text?
3. If I overlay this post with that thrashing tune, will it make me more credible?
4. Can people really not understand that 'sensor score' is a rating independent of hyper fast auto-focus or the ability to shoot at 27 fps, or integrated coffee maker?
5. Or that distilling a bunch of tests down to 1 single number at the end throws away information you personally may have found interesting?
6. Can I get my 3 minutes back?
7. Has a closing aperture replaced star wipes as the transition of choice?
8. Is it possible to get attention on the internet without titles aimed at insulting and enraging your viewers?
9. What do you do when you agree with the general thesis of a document ('don't read more into dxo scores than there really is') but it's presented in an annoying manner making you want to disagree?
10. What's with the blonde chick by the fire?

[Rondec]

I don't remember complaints about camera scores. Pentax sensor scores have always matched the competition. I certainly remember lots of complaints about DXOMark lens scores. Maybe that is what you are referring to?

I think people made a big deal about the difference between the K20 and K7 scores on DXO Mark, when they were basically equivalent (not great).

Obviously a single number doesn't tell you much. The break down of the numbers tells you more -- SNR, dynamic range, color and resolution are basically what you get with a sensor. Sometimes people read way too much into the scores. x camera has a sports iso of 977 and y camera has a sports iso of 1065. So what? You'll never see that difference in real world shooting.

The above video though feels like a Canon user expressing sour grapes. So Sony sensors score better than Canon sensors, that's not DXO Mark's fault. Basically Canon sensors haven't changed a whole lot over the years. The 7D and the 7D Mk II have the same base iso dynamic range (11.7 versus 11.8) and really similar high iso scores (854 vs 1082). At the same time, Toshiba, Sony and Samsung have really had significant improvements in sensor tech.

Obviously you can take great photos with a Canon camera. I hope no one says otherwise, but I hope that over time Canon decides to up their game, as it would put pressure on the other sensor makers to continue to do so as well.