5D3 vs 6D AF in low light

[dlleno]

Zlatko can you express what you mean by "low light" in terms that can be correlated to the results of others', as well as the published specifications for AF sensitivity?  short version:  What   does "low light" mean to you in terms of EV(100)?

[Zen]

The definition of "low light" is critical to these conversations, and because it is so subjective, the debate is almost meaningless. In my case, I'm sitting here in my den, appx. 12' by 24', with one 100 watt tungsten bulb in the corner lamp. The lamp is behind me and when I focus on the opposite wall, containing well filled book cases of aged cherry, my 5D3 clicks to focus quite easily and without any delay. I call this low light, and have no problem at all. I cant imagine a faster or sharper focus! And when I take the time to set the custom WB after use of a gray card, the images come out perfectly.

Perhaps my standards are lower than others, but I couldn't be more satisfied with my 5D3. 

Zen

[Marsu42]

The definition of "low light" is critical to these conversations

Furthermore "achieving focus" needs definition, because in problematic lighting conditions the camera will take more time to focus - I hope the detailed dpreview.com 6d review will tell us at last if the -3lv rated 6d is getting a focus lock faster @-2lv than the 5d3 which is maxed out in this level.

[J.R.]

The definition of "low light" is critical to these conversations

Furthermore "achieving focus" needs definition, because in problematic lighting conditions the camera will take more time to focus - I hope the detailed dpreview.com 6d review will tell us at last if the -3lv rated 6d is getting a focus lock faster @-2lv than the 5d3 which is maxed out in this level.

It should based on the theory. As I normally shoot in the evenings / night, I am really looking at the 6D as a replacement for the 7D if this is confirmed. Too bad in my country there is no option of renting the 6D or I would have checked it out by now.

For me the AF issues have arisen only in near darkness situations. The AF seems to continually hunt from end to end. Otherwise the AF system is near perfect.

[dlleno]

 why is it so hard to state the light level in terms of EV(100)?  if you have taken a picture, that is exposed properly,  you have the necessary information

[Zen]

I wonder . . . do we expect the camera to "see" better than we ourselves can? For example, I can barely make out shapes in "near dark" situations, yet some expect the camera to do better than that. Is that logical?

Zen

[dlleno]

I wonder . . . do we expect the camera to "see" better than we ourselves can? For example, I can barely make out shapes in "near dark" situations, yet some expect the camera to do better than that. Is that logical?

Zen

we expect the camera to af in the light levels specified by canon, not those specified by our eyes.  why is it so difficult to express the light levels of interest in the terms specified by canon? thats the comparison of value.

[Zlatko]

Zlatko can you express what you mean by "low light" in terms that can be correlated to the results of others', as well as the published specifications for AF sensitivity?  short version:  What   does "low light" mean to you in terms of EV(100)?

Sure, looking at recent photos in Photo Mechanic, I would say that typical "low light" photos were in the range of EV 1.2 to EV 4.0.  These would be interior scenes at night, photographed without flash and without an AF assist light.  Those EV numbers correlate to these exposures:
EV 1.2 = ISO 3200, f/1.2, 1/50th
EV 4.0 = ISO 800, f/1.6, 1/50th

[RLPhoto]

Lol. I just used my 5D3 strobist with no AF assist beam with just a distant street lamp for a source of light to AF. If the 6D is any better, wow but the 5D3 is still the best AF camera I've ever used.

[dlleno]

Zlatko can you express what you mean by "low light" in terms that can be correlated to the results of others', as well as the published specifications for AF sensitivity?  short version:  What   does "low light" mean to you in terms of EV(100)?

Sure, looking at recent photos in Photo Mechanic, I would say that typical "low light" photos were in the range of EV 1.2 to EV 4.0.  These would be interior scenes at night, photographed without flash and without an AF assist light.  Those EV numbers correlate to these exposures:
EV 1.2 = ISO 3200, f/1.2, 1/50th
EV 4.0 = ISO 800, f/1.6, 1/50th

Ahh thank you.  So 3-6 stops brighter than published minimum AF sensitivity

[Wildfire]

Animals often also aren't very enthusiastic about getting the af assist in their eye (and that's why I usually focus) - unfortunately the 60d (af rated up to +0.5lv) uses af assist very often, so I'm looking forward to the 6d.

The 60D doesn't have an AF assist beam, it just uses the built-on flash for AF assist.

Try mounting a 430EX or higher model speedlight to your 60D and you'll get an unobtrusive red AF-assist beam.

[Marsu42]

Animals often also aren't very enthusiastic about getting the af assist in their eye (and that's why I usually focus) - unfortunately the 60d (af rated up to +0.5lv) uses af assist very often, so I'm looking forward to the 6d.

The 60D doesn't have an AF assist beam, it just uses the built-on flash for AF assist.

Thanks for enlightening me :->

Try mounting a 430EX or higher model speedlight to your 60D and you'll get an unobtrusive red AF-assist beam.

If you look at my gear list (to the left) you'll see that I've got a 430ex2 & 600rt ... it's just my experience the red af assist is not as "unobtrusive" as you seem to think. And I never do pre-flash, it's just so annoying.

I just stated "60d" because a af system rated to up to +0.5lv is bound to use af assist more often than say the 6d that is supposed to focus up to -3lv.

[jonathan7007]

All,
The wild card in these discussions is what appears to be some variability in the 5Dmk3 copies'AF abilities. I feel I got a flawed body and it seems slower than it "should". Unscientific, absolutely, and I should work up a test here that Canon Service can replicate. These discussions are helpful for establishing what the limits should be. My comparison is my 5Dmk2. For any of the tests described the next level is bringing a lot of 5Dmk3 bodies into the same setup and looking for variations in "lock" time. I bought my body in May last year.

I didn't figure this out in time to exchange it. My bad! Canon has already looked at this body once.

I need to find an acquaintance with a 5Dmk3 and time to compare. That would be interesting.

Aaaaarg. <anger>

jonathan7007