Canon 6D Dominates Low Light Focusing Test vs 5Dii, 5Diii & Nikon D600

[dlleno]

back on topic --

Isn't the 6D rated at -3ev while the 5d3 is rated at -2ev?

the single AF point in the center is rated at -3EV sensitivity, yes.  However, per Canon product support the other 10 AF points are -1EV sensitive.  so -- the stationary "focus, then frame" approach with only the center AF point should yield a not-surprising one-stop advantage in very low light compared to the 5D3.  Note for interests sake that -3EV is four stops less light than requried for the exposure meter itself to perform up to its rated specifications!  so you can focus, but you  might want to get to know how the metering system behaves under those same conditions. 

Someone correct if wrong, but Canon also confirmed to me that all 61 AF points of the 5D3, on the other hand, are sensitive to -2EV.

Don't forget that EV sensitivity is not the only measure of success for the AF system.  for example, Roger over at lensrentals put these puppies to the test and discovered a wider variability in AF accuracy (std dev = 28 I believe) for the 6D compared to the more consistent 5D3 at std dev = 17.   thats a considerable difference. 

[Helevitia]

Thanks for the results.  I'm considering upgrading from a 7D.  How does the 7D compare in these tests?  Does anybody know?  I'm sure there have been similar tests done between 7D and 5D2/3?

[rafaelsynths]

I'm thinking of renting the canon 6D myself . Can anyone help me on my other post it has 3 replies and all of them are from me
I need some pro tips on whether to rent a 6D or 5D

[Marsu42]

Thanks for the results.  I'm considering upgrading from a 7D.  How does the 7D compare in these tests?  Does anybody know?

The af is supposed to work from -0,5lv ... my 60d from 0lv, so not much of a difference, and the af slows down considerably and I have to use af assist frequently. So the -3lv from the 6d (or -2lv from the 5d3 for that matter) should make a very, very large difference in speed and low light capability.

Note for interests sake that -3EV is four stops less light than requried for the exposure meter itself to perform up to its rated specifications!  so you can focus, but you  might want to get to know how the metering system behaves under those same conditions. 

Interesting point, I wondered about that, too and would like to see some reviews on that, but afaik it only matters for available light w/o en e-ttl flash.

for example, Roger over at lensrentals put these puppies to the test and discovered a wider variability in AF accuracy (std dev = 28 I believe) for the 6D compared to the more consistent 5D3 at std dev = 17.   thats a considerable difference. 

... but only if you use a large aperture lens and shoot with a thin dof, for other shots a little variation doesn't matter - 60d and esp. 7d are much less precise than the 6d.

[dlleno]

Thanks for the results.  I'm considering upgrading from a 7D.  How does the 7D compare in these tests?  Does anybody know?  I'm sure there have been similar tests done between 7D and 5D2/3?

yes, the 7D fared among  the worst of all the cameras he tested.  FYI pardon my asssumption of background here, but for those unfamiliar, high standard devition is bad.  7D was 41, 5D3 was 17. 

http://www.imaging-resource.com/news/2012/08/01/autofocus-reality-part-3b-canon-cameras


and Marsu42 yes good reality check. 

[Chosenbydestiny]

Thanks for the results.  I'm considering upgrading from a 7D.  How does the 7D compare in these tests?  Does anybody know?  I'm sure there have been similar tests done between 7D and 5D2/3?

yes, the 7D fared among  the worst of all the cameras he tested.  FYI pardon my asssumption of background here, but for those unfamiliar, high standard devition is bad.  7D was 41, 5D3 was 17. 

http://www.imaging-resource.com/news/2012/08/01/autofocus-reality-part-3b-canon-cameras


and Marsu42 yes good reality check.

Could it be they had a bad copy of the 7D for testing? I've always remembered my old 7D focusing a bit better than my 60D.

[Dylan777]

I'm always surprised when people put forth results and a conclusion, and then are surprised when others critically evaluate the methods used to generate those results, and weigh the quality of the methods when evaluating the reliability of the results and the strength of the conclusions. 

Something else to think about.

I welcome and appreciate criticism, but how does the discussion of semantics in this test somehow mean that the 6D doesn't focus well in low light?

Some people appreciated the amount of time & energy it took to do this. Some don't. But I do know this:

That center focus square on the 6D is very, very good in low light. Ill grab it before the others for low light situations.

If you cannot walk away from this post/discussion with that very basic, heart of the matter idea, than nothing I can say will matter to you.

No hard feelings either.

Thanks!

MM

Michael,
I'm not an expert in testing.

I did have a chance to play with 6D outside parking lot, behind business building, around 7PM westcoast. It was quite dark and the light sources were street lights. I compared to my 5D III and I didn't see the advantage of 6D has to offer in term of -3EV AF center point. Maybe it wasn't dark enough?  I shot with my 24-70 f2.8 II lens at f2.8. The outer points were bit slower though.

[Marsu42]

Could it be they had a bad copy of the 7D for testing? I've always remembered my old 7D focusing a bit better than my 60D.

I guess you didn't do any (semi-)scientific tests? I don't have a 7d, but I've read many times that the af on the 7d is the least precise, and the lensrentals test confirms it - though by a margin you may not notice in day-to-day shooting because unfortunately the 60d is rather unprecise, too, and of course the lens also matters.

I compared to my 5D III and I didn't see the advantage of 6D has to offer in term of -3EV AF center point. Maybe it wasn't dark enough?  I shot with my 24-70 f2.8 II lens at f2.8. The outer points were bit slower though.

The good news for 6d buyers is that obviously you didn't see any disadvantage, too - and yes, I guess it wasn't dark enough because -3lv is rather dark, though not different enough from -2lv to tell a difference w/o a lightmeter (or you can look inside your picture files, the measured light value is in there). From kr http://www.kenrockwell.com/tech/ev.htm

LV 3 Brightly lit night street scenes
LV 2 Typical night street scenes
LV 1 Dark scenes outdoors at night
LV 0 LV Zero is defined as the light level that requires a 1 second exposure at f/1 with ISO/ASA100 speed film.
LV-1
LV-2
LV-3
LV-4
LV-5 Scene lit by the full moon
LV-15 Scene lit only by starlight.

[dlleno]

...I did have a chance to play with 6D outside parking lot, behind business building, around 7PM westcoast. It was quite dark and the light sources were street lights. I compared to my 5D III and I didn't see the advantage of 6D has to offer in term of -3EV AF center point. Maybe it wasn't dark enough?  I shot with my 24-70 f2.8 II lens at f2.8. The outer points were bit slower though.

folks this is not complicated.    The way to test the advantage that the 6D has to offer in terms of a -3EV center point, is to test in -3EV light! 
 
were you shooting in that level of light? I would say never mind trying to correlate with KR or Wikipedia or any other attempt to describe qualitatively what -3EV is.  Maybe its a firefly in a coal mine during power outage, I don't know --  but none of that matters: it turns out  we know what -3EV is!     Does proper exposure require 4 seconds at f/2.8 and  ISO 1600? 

[MichaelTheMaven]

Hi Dylan-

  I appreciate you taking the time to do some tests of your own. This is what I wanted to read, other low light focusing tests.

  It's going to be very difficult to notice the difference between the 5Diii and 6D for single focusing type tests, both are going to focus in low light in just over a second. Even when going for 30-50 rounds of focusing, the 6D was only marginally better in the central point. Set up a timer and focus back and forth between 2 high contrast targets 50 times and then you will start to see the differences in speed easier.

As a side note, most of the 5Diii's central, top to bottom, middle row, did very well in low light, where the 6D's other squares failed.

As a whole the 5Diii's focusing systems are the best of the bunch (not even close), but when it comes to low light, yes, it does seem the 6D has the edge in that one central square. Its a true pity Canon didn't have more of an upgrade, but the 6D absolute dominated the 5Dii and D600 on that central point in low light.

M

[ScottyP]

Good-natured helpful criticism is great.  But what compells someone to just plain savage someone else's efforts?  What button got pushed?  There is room for more than just one or two experts on this forum, right?  Why waste such considerable talents for precise measurement conducting vain (and metaphorical) comparisons of the size of one guy's "monopod" vs. the other guy's "monopod"?

If someone knows the only proper and relevant way to look at this question, and yet he sits back casting only stones instead of light, is he not denying humanity by not investing the time/effort to do a proper comparison themselves? 

Or, put another way, "is this how we welcome new members to this forum"?

 

[zim]

I actually really do appreciate any testing especially real world as I’m very much in the 6D – 5D3 quandary right now but have time on my side. I’m not questioning the effort put in or the results, I just don’t understand the results, apologies for not articulating that very well.

I thought the 6D autofocus in -3EV conditions was a one stop improvement over the 5d Mark III and a two stop improvement over the 5D Mark II? I don't know about the Nikon.
I don't really get testing beyond the rated limit, maybe I missunderstand further explaination of the -7EV thing would be much appreciated. I also seem to recall a video from DigitalRev (in the boxing ring) where the 5D3 excelled against the D600 no where near -3EV?

I watched that vid ---- it was testing the servo tracking of the 3 though, this is one shot focus mode

and that’s the sort of thing that confuses me how can a camera that by all account blows the socks off the D600 in AF Servo tracking not do the same in single point. Does the sensitivity of the points change somehow or is it down to software algorithms in which case isn’t single point the simplest scenario?

[MichaelTheMaven]

I think the thing that may have caused the confusion is that EV is often used interchangeably with "stops". There are countless cameras out there that adjust exposure compensation in EV values relative to the current setting, not the absolute standard EV value. I should have described it as relative stops instead, but though the results were still meaningful. I approached it from the perspective of an experienced wedding photographer who has been frustrated with the single point in low light.

As to your quandary about the 5Diii vs 6D, let me ask you this. What kind of shooting are you primarily going to do?

MM

[dlleno]

What we're learning here is consistent with the published specs.  I admit I'm one of those who "stopped reading" when the test wasn't correleated to the standard I was expecting and I couldn't (immediately) tell if the test actually tested the -3EV condition or not.    But that doesn't mean  I quit watching for additional info

What isn't so clear to me  is the aparent (empirical evidence only, maybe even just anictotal) slow behavior of the 5D3 to focus in low light situations compared to the 6D (this is not about the AF assist beam with flash).  I'm hearing that the 6D AF (even if slightly less accurate) is faster than the 5D3 under certain situations (which themselves have not yet been acurately described).   Should this turn out to be true,  I wonder if it has anything to do with the fact that the 5D3 AF system is remarkably consistent (low standard deviation per the lensrentals test):  It may have a more elaborate retry/credence algorithm,  and as such require a greater number of handshakes with the lens to acheive that level of accuracy.  It could be that, assuming the evidence is consistent, that the more elaborate algorithm is not as easily satisfied in low light.  To the extent that such a suggestion is true, this strikes me as something Canon might fix or improve silently with a firmware update without officially acknowledging the issue. 

[neuroanatomist]

That center focus square on the 6D is very, very good in low light. Ill grab it before the others for low light situations.

If you cannot walk away from this post/discussion with that very basic, heart of the matter idea, than nothing I can say will matter to you.

Already done, several posts back...   

Having said that, a generic conclusion that the 6D center point AF performs very well in low light is quite reasonable.