5dIII review at DxOMark

[Dr Croubie]

I'm also reading the graph of ISO, SNR 18%.
On Print, comparing photo to photo, the D800 and 5D3 are almost exactly equal, both above the 5D2.


On Screen, comparing pixel to pixel, 5D3 beats 5D2 beats D800.


How does the D800 score 2800 and 5D3 score 2200? Maybe only 1/3 - 1/2 stop, but I call shenanigans on DxO.

[Dr Croubie]

In other news, canon *still* have not fixed the iso 100-800 plateau on DR (although above iso1600 5D3 beats 5D2 and D800).
Tonal range Print just gives the edge to D800, on Screen 5D3 wins easily.
Bit depth D800 wins on Print, 3-way tie on Screen though.
Why oh why can canon not make a camera with more DR at iso100 than it does at iso800?

Anyway, confirms what we already know:
- 5D3 is a wedding/event camera
- D800 is a Landscape camera
- If you don't do either, both are good enough all-rounders
- 1DX and D4 are sports/photojournalist cameras
- And I still don't care about video

Looks like i'm sitting out this round of camera upgrades. If I keep the 7D, buy a Hartblei 1006 and a few 6x45 film backs for $500, I wonder how many rolls of Velvia I can get for the other $3000...

[Daniel Browning]

I don't get the D800's high score for Sports (Low Light ISO) performance.

Makes sense to me. The D800 has more read noise, but it detects 14% more light. So depending on exactly what light level (and tonal level) you measure, results will vary. At the 30dB chosen by DxO (which reflects the "common" photographer pretty well IMHO), the 14% light is more important than the read noise, so it's 1/3 stop ahead. At higher ISO (say, 25,600), the lower read noise of the 5D2 will compensate for having less light, and so it would turn out a better result.

I didn't see where they offered graphs or anything to back this up.

It's in the same place as all their other cameras, which is to say their slow, enormous, and labyrinthine web site. Personally, I prefer http://sensorgen.info/, which is simply a conversion of DxO data into industry standard sensor metrics such as read noise, QE, and FW. (They do it with a "view source" on the right URL, which gives you DxO's data.)

[dsiegel5151]

It's pretty simple. You measure three areas of performance (DxO's portrait, landscape, and sports scores) and then average them together. The D800 sensor did better in all three, so it had a higher combined score.



No. Say Johnny gets 65% in Chemistry, 65% in English, and 100% in Math. While Sally gets 70% in all three. Sally beat Johnny in two out of three subjects, but Johnny's GPA is still higher. Are you to infer that Math is the only subject that matters when calculating someone's GPA?

Any time you try to reduce something down to a single number, it will *not* reflect the full reality and complexity of the situation.

Summarizing something as complicated and nuanced as sensor performance into a single number is impossible. But a lot of people *want* a single number, because they don't have the time or inclination to learn all that stuff. So DxO provides the solution for everyone.

At one end of the spectrum (technically-minded folks), it provides the full charts and data so you can go in and see how it really does for yourself. Make up your own metrics that reflect what you do with the camera, and use their data for it.

Towards the other end of the spectrum, DxO has chosen three particular measurements, that while they may not match your shooting exactly, they do reflect the taste and perception of many photographers. You merely choose how important each of these three are to you and weight them yourself (for one, low light may be twice as important as dynamic range and color depth; for another photographer, the reverse).

Then at the very end of the spectrum you have folks for whom even three numbers is too much. They only have time or inclination to compare based on a single number. DxO doesn't know how important the three factors are for every person, so they just weight them equally. I think that's a good choice, even though it doesn't match the weighting that I would use for myself, personally.

I'm just never going to get it. I noticed that the Nex7 scored the same score as the 5dIII. The Nex7 dominated the 5dIII in dynamic range at 2 iso levels (a little over one stop Ev). In pretty much every other category at every iso, the 5dIII dominated the Nex7 (http://www.dxomark.com/index.php/Cam.../(brand2)/Sony). But b/c the Nex7 had a little over one stop better dynamic range at two isos, it tied the 5dIII? It makes sense to me to combine and average percentages in terms of school grades. We all know what these percentages indicate. Is it really appropriate to combine and average three different sensor measurements into one overall score? I'm sure that DxO has some sort of conversion to make all the measurements relative (or are they just relative to some optimum?), but why is a little over 1 stop dynamic range at two iso settings weighted so heavily, especially at low iso levels? At five higher iso levels, the 5dIII dominates the Nex7 by approximately one stop of dynamic range. How is it possible that these sensors score the same?!?! It makes my head want to explode; but fortunately, I'm never buying either of these cameras, so it really doesn't matter.

[Daniel Browning]

It makes sense to me to combine and average percentages in terms of school grades. We all know what these percentages indicate.

Personally, I think combining and averaging scores tells you just as much about a student as it does about a camera (which is to say, not very much). A 4.0 GPA may seem impressive on the surface, but what if all the classes are basket-weaving and the like? Or it's from University of Nigeria (Mailorder Dept). A 2.5 may not seem as good, but what if it's for classes that are 6 years ahead of the student's level? Or what if the GPA is poor but the student is the best pianist in the world, or an Olympic gold medalist? The number is only useful if the way that it was calculated also corresponds with a way that will be useful to you. Often times they don't match up, so it's necessary to have more information than one number. If you want to know which student is the best at math, the GPA may not help.

Is it really appropriate to combine and average three different sensor measurements into one overall score?

Yes, just as much (or more) as any single number score is appropriate in any type of summarization of performance or value -- like rottentomatoes, amazon review stars, etc. Oblig. xkcd: http://xkcd.com/937/

Personally, I would never use the single DxOMark score, because a single number really cannot tell anyone enough to make a good decision. But if someone asked me to rank sensors by a single performance number, without knowing what the sensor would be used for, I would probably use DxO's. Incidentally, their three measurements (Landscape, Portrait, and Sports) correspond very closely to the three primary performance metrics that image sensor designers optimize for (read noise, full well, and quantum efficiency).

I'm sure that DxO has some sort of conversion to make all the measurements relative (or are they just relative to some optimum?),

The scores are open-ended, not relative to some optimum. If I recall correctly, they are normalized to "stops" (1 bit of color depth is 2/3 stops), for photographers.

but why is a little over 1 stop dynamic range at two iso settings weighted so heavily, especially at low iso levels?

Why isn't it weighted *more* heavily? To me personally, low iso dynamic range is even more important than the weighting that DxO gives it. In fact I switched from Canon to Nikon primarily for low ISO dynamic range.

Doesn't that illustrate the point? Everyone has different preferences and desires. You don't care so much for low ISO dynamic range, so you'd prefer it to be weighted less. I'd prefer it to be weighted more. DxO had to pick *something*, probably what they thought the average photographer would think.

I know plenty of photographers who have *never* changed the ISO setting from 100 (in fact I was at his house this evening), nor ever used more than 6 stops of dynamic range. For them, the portrait score is pretty much all they need.

At five higher iso levels, the 5dIII dominates the Nex7 by approximately one stop of dynamic range. How is it possible that these sensors score the same?!?!

Because the dynamic range at high ISO does not factor into the *landscape* score. Most landscapes are not shot at high ISO.

It makes my head want to explode; but fortunately, I'm never buying either of these cameras, so it really doesn't matter.

[Daniel Browning]

I'm also reading the graph of ISO, SNR 18%. On Print, comparing photo to photo, the D800 and 5D3 are almost exactly equal, both above the 5D2. How does the D800 score 2800 and 5D3 score 2200? Maybe only 1/3 - 1/2 stop, but I call shenanigans on DxO.

Good question. It's probably because the SNR 18% is calculated without consideration for chroma noise, while the Sports score is. For example, you can improve SNR 18% by making the green pixels accept more red light, and the red pixels accept more green light. But then when you apply a color matrix to get accurate colors back out, chroma noise goes up. That is precisely what Canon has done with some of their cameras.

Go to the individual data (not a comparison) for a camera and click on the "Measurements" tab and then "Color Response". You'll see that the 5D3 red pixels have almost the same amount of green as they do red. While D800 red pixels have a bigger difference between red and green. Same with blue. The 5D3 is almost a YGB (Yellow Green Blue) sensor instead of an RGB.

[neuroanatomist]

I don't get the D800's high score for Sports (Low Light ISO) performance. I didn't see where they offered graphs or anything to back this up. I would have thought the 5D III would have been the victor here.

I'm going to repost something I put up on CR (modified slightly):

Honestly, everyone should take a step back and consider what the DxOMark scores mean - and don't mean. They are quite transparent about their scoring metrics and methods, and there's no reason to suspect there's collusion or favoritism occurring. At the same time, it's important to remember that the sensor score is just that - a score of the sensor itself, not a 'camera score'.

Furthermore, keep in mind that all of their overall scores are based on reducing the image to 8 MP - thus, the greater the starting resolution the more downsampling, which means lower apparent noise. That lower apparent noise means apparently better ISO performance and apparently more dynamic range (lower apparent noise lowers the 'floor' for the DR estimate). If you like, you can substitute 'artifically inflated' for 'apparently'.

They have a page describing the mathematics of the normalization to 8 MP, and on that page, there's the following statement:

What should be remembered is that doubling the resolution adds:
3dB to the normalized SNR
0.5 bit to the normalized DR
0.5 bit to the normalized TR
1.5 bit to the normalized CS.

So, compared to 8 MP the D800 is 4.5-fold higher (just over two doublings), whereas the 5DIII is 2.75-fold higher (just over one doubling). What that means, mathematically, is that the D800 has approximately one extra doubling of resolution relative to DxOMark's 8 MP normalized value - that accounts for all of the differences in the scores for both ISO and Color Depth, and part of the difference in dynamic range.

Note that DxOMark does provide the non-normalized data, they just don't use those data to calculate the overall scores, the rationale being that normalizing to 8 MP allows appropriate comparisons. In one sense, it does - if you're going to print 8x10" images all the time, then their scores actually apply pretty well.

Let me give a specific example for color sensitivity, which is the basis for DxOMark's Portrait Score. That's one area where the D800 with 25.3 bits 'beats' the 5DIII with 24-bits. In the comparison without normalizing to 8 MP, the D800's advantage pretty much entirely disappears.

Rick, that piece about reducing to 8 MP is what makes the D800 'better' on their Sports (Low-light ISO) score - the higher MP of the D800 means relatively greater effective noise reduction when reducing from 36 MP to 8 MP, compared to going from 22 MP to 8 MP.

This is why I quit reading DxO: http://www.dxomark.com/index.php/Cam...(brand2)/Nikon

I don't see how these measurements give one sensor a score of a 95 and the other a score of an 81.

Here's another example: http://www.dxomark.com/index.php/Cam...brand3)/Pentax

Notice that the Pentax K5 has scored an 82 overall; however, it is inferior to the 5d Mark III in ever category except Dynamic Range at low Iso settings. Thus, am I to infer that Dynamic Range at low Iso settings is the only measurement that matters when rating a sensor?

Again, it's about understanding what the measurements mean and how they're derived. It looks like the Landsacpe (Dynamic Range) and Portrait (Color Depth) have a much greater impact on the overall score than Sports (Low-Light ISO), and that actually is the case - for a reason. The overall score is an 'average' of the three use-case scores, but two of those three, Portrait (units are bits) and Landscape (units are Ev) are on log scales. ISO is a linear scale, so for example, comparing the ISO (Sports) values for the K5 and the 5DIII, the 5DIII 'score' is almost double that of the K5 (2293 vs. 1162), but when you log transform that difference, the difference is less than one stop (Ev).

________________________

To summarize, one key point about DxOMark's scores is that they are all normalized to an 8 MP image, and this strongly impacts the overall and use-case scores, giving an advantage to sensors with higher MP counts. Likewise, since DR and Color Depth are measured on a log scale, they have a relatively greater weight in the overall score than low-light ISO.

Try the following: click on the first link above (5DIII vs. D800), then in the comparison click the Measurements tab. I'd recommend skipping the ISO Sensitivity plot (it has nothing directly to do with the Sports/Low Light ISO score, despite the name of the test - it's really looking at ISO accuracy relative to the actual International Organization for Standardization criteria). But...look at SNR (signal-to-noise ratio), Dynamic Range (basis for the Landscape score), Tonal Range, and Color Sensitivity (basis for the Portrait score), and for all of them, look at the Screen plots - those are the data that are not normalized to 8 MP (vs. the Print plots, which are normalized and used to generate the overall and use case scores). When you do that, you'll see the following:

• SNR - 5DIII wins (when normalized they tie)
• Dynamic Range - 5DIII loses up to ISO 1000 but wins at higher ISOs (when normalized, the 5DIII loses up to ISO 1600, then they near-tie)
• Tonal Range - 5DIII wins (when normalized they tie)
• Color Sensitivity - they tie (when normalized, D800 wins)

So, for all of the above measures, the higher MP count of the D800 gives it an advantage when downsampling the images to 8 MP.

Bottom line is that I think DxOMark's measurements are more useful than their scores, but even their scores are useful - as long as you understand how those scores are generated, and the inherent limitations and caveats about them. The Overall Score is something the have to have (how can you have a ranking site and not actually actually assign ranks). It's like looking the Gross Domestic Product by country, and concluding that the USA is the best country in the world simply because it has the highest GDP. Or, to use a photographically relevant analogy borrowed from Bob Atkins, it's represeting the Mona Lisa by it's average color value.

[Jon Ruyle]

Only 2293 on the low light ISO score. The D800 has one third stop better performance (2853), despite a 76% higher pixel count. Both are still excellent cameras, of course.

I was comparing the 5DIII to the D700, not the D800

Both are excellent cameras, but there is no question in my mind which I would choose if I was starting fresh. The D800 is better at low iso and high iso, and has higher resolution, and it costs less to boot. I honestly can't see why anyone would go with a 5DIII over a D800 if they weren't already heavily invested in canon equipment (as many 5DIII buyers of course are ). Of course, many who choose the 5DIII know a lot of stuff that I don't.

I got the 5DIII because I can't bear to part with a few lenses (85 f/1.2, 70-200 f/2.8 IS II, 65mm and 100mm IS macros, 135mm f/2 in particular), not because I think it a better camera. Even so, had canon not hit a home run with the 5DIII autofocus (which, as far as I am concerned they did) I don't think I would have stayed with canon.

I agree with those who say that the text part of the DxO review is pretty useless (I couldn't tell from the review that they had ever actually touched the camera). But their data (assuming it is accurate) seems to me extremely useful and interesting.

[Kayaker72]

Still laughing at the "Single Number" representation of the Mona Lisa.....

BTW, if looking at the DXO single number representation of a sensor or even the detailed charts is the only way to selected a camera I wasted a lot of money buying my 7D over the 60D or a Rebel. Not surprisingly, since they essentially have the same sensor, they score very closely. But, we all know, and I hope anyone looking DXO also knows, that while the sensor is important, there is more to a camera than the sensor.

What interests me and why perhaps this has created a bit of an uproar (other than brand pride) is the fact that most other comparisons between the 5DIII and D800 have them pretty even. DXO's total score is the first that makes one body look like a "winner" over the other.

[Daniel Browning]

most other comparisons between the 5DIII and D800 have them pretty even. DXO's total score is the first that makes one body look like a "winner" over the other.

I can't imagine that there are any other comparisons out there that say the 5D3 is "pretty even" with the D800 in dynamic range and color depth. If there are, they are most likely incorrect.