Kent Rockwell on 1DX and 5D3

[duydaniel]

Hi,

I know what you probably don't like Kent but before bashing him, let check his writing here
about 1Dx and 5D3:

My 5D Mark III is about a half-stop faster at any given ISO setting than my 1D X.

In other words, at any given ISO, my 5D Mark III gives the same image with a half-stop less exposure. In other other words, my 5D Mark III actually runs at a half-stop higher ISO than marked, or my 1D X is actually running a half-stop slower than marked, or somewhere in between.

It's common for camera makers to cheat a little here, since Canon needs to make the 1D X look like it has better high ISO performance than the 5D Mark III. In this case, it's stacked the deck a little, since when anyone compares at the same indicated ISOs, the 1D X appear to have a half-stop advantage since it's really only operating at a half-stop less ISO than the 5D Mark III at any given ISO. When comparing images, you won't notice unless you shoot at the same manual exposure, in which case my 1D X is a little darker than my 5D Mark III, or if you get both to match, you'll notice that your 5D Mark III only needed about a half-stop less exposure.

src: http://www.kenrockwell.com/canon/comparisons/5d-mkiii-vs-1dx.htm

Can someone check if this is true?

 

[Jim Saunders]

Even if it is true it is academic, and in any case the 1Dx has exposure compensation (like AFMA). He's probably right about the fact - but so what?

Jim

 

[duydaniel]

Even if it is true it is academic, and in any case the 1Dx has exposure compensation (like AFMA). He's probably right about the fact - but so what?

Jim

If that is true then it would influence how I shoot from now on.
I would dial down 1/2 stop on ISO since the 5D3 is calibrated "incorrectly".
People report the 5D3 has more noise than D800 in low iso probably because the above reason.

 

[Jim O]

I know what you probably don't like Kent but before bashing him, let check his writing here
about 1Dx and 5D3

Well I know how I feel about Ken Rockwell..

He sometimes makes me laugh. He's generally unbiased, sometimes outrageous. He takes no money or free equipment from Nikon or Canon so he speaks his mind - for better or worse. He holds some patents on chips and optics. His site is more like a series of random thoughts loosely organized. Anything he says is his opinion and should be taken with a large grain of salt.

I don't have a 1D X to compare. I believe that the results that he presented are accurate for his copies of the two bodies in question, which are almost surely retail copies. Whether they hold true in general would be interesting to know.

 

[xvnm]

I'm not sure how much I trust DxOMark, let's take this as a second piece of anecdotal evidence: http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/795|0/(brand)/Canon/(appareil2)/753|0/(brand2)/Canon

 

[duydaniel]

I'm not sure how much I trust DxOMark, let's take this as a second piece of anecdotal evidence: http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/795|0/(brand)/Canon/(appareil2)/753|0/(brand2)/Canon

If you go to measurement / dynamic range
after iso 160, the 1dx has lower measured iso at any given manufactured iso than the 5d3

BUT THEN THIS WOULD THROW THE ISO 160-THEORY OUT OF THE WINDOW

 

[privatebydesign]

If you are shooting RAW then the histogram should be dictating your exposure compensation, not the iso or some over hyped self admitted page hit seeking internet troll.

If you are shooting jpeg, as he also recommends, then it is even more important the histogram should be dictating your exposure compensation, not the iso or some over hyped self admitted page hit seeking internet troll. The exposure adjustments just need to be different for the different formats.

The ISO standard for the iso rating specifically allows for rounding up and down, the iso values on different camera models are not identical, and they don't need to be. This has caused some angst between very vociferous DR adherents and myself with regards shadow recovery, a "same" exposure across models, let alone brands, is not, necessarily, the same.

Because of this allowable "rounding" DxO (another flammable topic around here) run their tests after working out the actual sensor sensitivity, they subsequently give each camera a true measured iso speed. The 1DX has a true iso speed of http://www.dxomark.com/index.php/Cameras/Camera-Sensor-Database/Canon/EOS-1Dx#measuretabs-1 and the 5D MkIII has a true iso speed of http://www.dxomark.com/index.php/Cameras/Camera-Sensor-Database/Canon/EOS-5D-Mark-III#measuretabs-1 as you can see from these two sets of measurements, the two cameras are very very similar and do not display the symptoms Mr Rockwell described.

 

[risc32]

i'm going to have to disagree with you there privatebydesign. i don't know what you mean by the ISO dictating the exposure compensation. sure it does,it all dictates the exposure.
ISO values should match across formats and brands. Not saying they do, but they should. all canons used to run a bit higher than indicated ISO values, but i think they brought them within spec some time ago. I've never looked into the 5dmk3-vs-1dx exposure value thing, but it shouldn't be to hard to find some comparable test shots on the net. 1/2 a stop slower than indicated isn't earth shattering news, but that's not all together insignificant either. people pay big bucks for a 1 stop faster lens.

 

[Jim O]

some over hyped self admitted page hit seeking internet troll.

Tell us how you really feel...

 

[JPAZ]

I have found that my particular 5d3 produces "better" images with a 1/3+ exposure compensation dialed in. It is possible that this could vary from copy to copy (but I am not sure if that indicates a production issue or just the ways things look to my eyes). But, standing on the wet Tundra in the early morning light shooting the fall colors as the sun rises over the mountains, my iso 8000 exposures look pretty good. The guy next to me has a 1dx, and his photos look pretty good as well. His settings and mine are a little different but the output is what counts.

So, is a possible difference in the iso sensitivity something I should care about?

 

[privatebydesign]

i'm going to have to disagree with you there privatebydesign. i don't know what you mean by the ISO dictating the exposure compensation. sure it does,it all dictates the exposure.
ISO values should match across formats and brands. Not saying they do, but they should. all canons used to run a bit higher than indicated ISO values, but i think they brought them within spec some time ago. I've never looked into the 5dmk3-vs-1dx exposure value thing, but it shouldn't be to hard to find some comparable test shots on the net. 1/2 a stop slower than indicated isn't earth shattering news, but that's not all together insignificant either. people pay big bucks for a 1 stop faster lens.

And that is your prerogative.

And in an ideal world you are correct, everything should be equal, this is not an ideal world.

As you say iso makes up one third of the exposure triangle, when was the last time you had your shutter adjusted or even tested? Back in the day we had to send our cameras in for CLA's pretty much annually, that would be a clean, lubricate, and adjust. The adjust part would be shutter timing, so does your shutter expose for 1/1127 sec or 1/963 sec when it is set to 1/1000? Your apertures, are they exactly where they should be, probably not as they rarely are, not to mention vignetting and the huge exposure variations that can induce. This is part of the reason top level cameras allow you to adjust for these metering differences.

So built into our exposure triangle are manufacturing tolerances for aperture and shutter speed, and the ISO allowable "rounding" for iso values, add in the confusion between 18% grey, you know the value everybody assumes is metered middle grey, and the fact that Canon meters (along with many others) at 12-13% grey +/- meter manufacturing tolerances, plus the fact that focal lengths are normally only close to accurate at infinity and you can see why, after 30 years, I am fairly dismissive of doing things by the numbers.

The point to my first post was don't sweat the numbers, your histogram, along with a knowledgeable interpretation of it, is the most powerful tool you have for getting your exposure where you want it to be.

 

[duydaniel]

If the post is true, this would mean the following:

Supposed the 5D3 gave the following exposure:
iso 6400
1/200
f5.6

According to the theory, 5D3 over exposes ISO, the correct one would then be:
iso 5000
1/200
f5.6

which would yield a slightly darker image but that would be similar to one suggested by the 1DX

 

[Skulker]

I can't say that I'm very worried about this. I have both and will adjust exposure to either over or under expose to suit the image I'm trying to capture. I have never noticed anything affecting how I take shots.

As an engineer I wouldn't expect identical results even from aparently identical cameras. But I'm not sure I would expect 1/2 a spot variation.

Its almost a shame I'm not interested enough to do a test. ;D

 

[Northstar]

I've always felt that my 5d3 and 1dx both underexpose by about a 1/4 to 1/2 stop if I use auto iso in manual. (I've never tried ap or shutter priority so can't speak to that)

My 5d3 underexposes more than my 1dx.

Just my 2cents.

 

[comsense]

Yaaaaaaaaawwwwwwwn

 

[DavidKM]

Dunno... I own both bodies. Real world, put an image from one next to the other and you'd be hard pressed to tell where they came from. It's the only thing that really matters to me. Ken(t), meh.

I do use my 5D3's the majority of the time. Love the silent shutter and a smaller rig!

 

[risc32]

And that is your prerogative.

And in an ideal world you are correct, everything should be equal, this is not an ideal world.

As you say iso makes up one third of the exposure triangle, when was the last time you had your shutter adjusted or even tested? Back in the day we had to send our cameras in for CLA's pretty much annually, that would be a clean, lubricate, and adjust. The adjust part would be shutter timing, so does your shutter expose for 1/1127 sec or 1/963 sec when it is set to 1/1000? Your apertures, are they exactly where they should be, probably not as they rarely are, not to mention vignetting and the huge exposure variations that can induce. This is part of the reason top level cameras allow you to adjust for these metering differences.

So built into our exposure triangle are manufacturing tolerances for aperture and shutter speed, and the ISO allowable "rounding" for iso values, add in the confusion between 18% grey, you know the value everybody assumes is metered middle grey, and the fact that Canon meters (along with many others) at 12-13% grey +/- meter manufacturing tolerances, plus the fact that focal lengths are normally only close to accurate at infinity and you can see why, after 30 years, I am fairly dismissive of doing things by the numbers.

The point to my first post was don't sweat the numbers, your histogram, along with a knowledgeable interpretation of it, is the most powerful tool you have for getting your exposure where you want it to be.

i agree completely. cheers.

 

[Mt Spokane Photography]

I find that different lighting can fool the in camera exposure, so the photographer needs to be aware of this and adjust accordingly. I tend to get more accurate exposures outdoors then indoors, for example. My D800 was the same.