6D vs 5Diii vs 5Dii - Speedlite AF Focus Beam Assist Tests

[Chris Burch]

It's the amount of time it takes to lock focus when using the AF beam in low light. Pretty much everyone agrees the focus is spot on when it locks, but it takes considerably longer than it did with the 5DII.

 

[AudioGlenn]

so according to those numbers, the mk2 was taking about 1.5 sec/shot (45 sec/30 shots), and the mk3 was taking 2.33 sec/shot.

Noticeable difference but I'm not sure what the fuss is about. I'm coming from a 60D though so maybe you guys noticed that .83 second difference much more coming from a mk2.

 

[Louis]

"not sure what the fuss is about" it really is something to worry about when your working with the camera, You have a time limit to deliver photographs and you cant take the shot you need because of this, also there are not allot of golden opportunities to grab a great shot all the time, so you need fast AF when the moment arrives,

this is a problem and Im so hoping the 5D3 gets a load of Love this 2013

 

[AudioGlenn]

I am talking about "working" with the camera... I shoot weddings, kids, candid moments, shows... for money most of the time.

I understand that speed is an issue and you want the camera to be as fast as possible. It just seems there's a bit of an uproar over the issue but the difference in time is less than 1 second (at least for this test). I'm not trying to say there isn't an issue or that we shouldn't have a camera that's at least as fast as the mk2 for the money paid for the mk3. It just sounds like some users need to check their settings and play with the "focus priority/release priority" settings.

@Mike (OP), by chance, did you or can you try running the tests with different settings under the AF2 Menu for both 1st and 2nd shots, as well as the AF3 Menu under one-shot AF release priority?

 

[MichaelTheMaven]

Just an update on the tests. I've done tons more with different speed lites, targets, etc, most of this has been 6D vs D600 type stuff. Ive even recorded video of how I do the tests so you guys can see and I will post the video here when it is done.

I am not seeing any difference between the 600 or 580 Speedlites. I did look into the auto focus settings on the 5Diii, but as another forum reader pointed out, most of it has to do with AI Servo type shots (which doesnt require a focus lock) to take the shot.

I should probably add a note, as I did on the Canon Forum about why I am going back and forth between the 2 differnent targets at different distances; Target 1 serves mainly to defocus. If im just aiming at target 2 and focusing over and over, its not a good measurement of the cameras ability to get that initial focus.

Another point is that the average focus times for each shot work out to:

5Dii - 1.5s
6D - 1.7s
5Diii- 2.33s

But this is not what is happening. The focus lock on Target 1 (in good light) is always very quick, almost instantaneous.

Focusing on Target 2 is taking longer than these average times. In fact, most of the time is spent on Target 2.

It is hard to be precise, but I would say the times are closer to:

5Dii - 2.2s
6D- 2.5s
5Diii - 3 + s

I have a Sekonic 558 Light Meter that seems to be unable to measure the light Im shooting in which would be ideal to know, and am waiting on my new meter - that should help to know exactly at what point it starts to struggle.

M

 

[ScottyP]

Just an update on the tests. I've done tons more with different speed lites, targets, etc, most of this has been 6D vs D600 type stuff. Ive even recorded video of how I do the tests so you guys can see and I will post the video here when it is done.

I am not seeing any difference between the 600 or 580 Speedlites. I did look into the auto focus settings on the 5Diii, but as another forum reader pointed out, most of it has to do with AI Servo type shots (which doesnt require a focus lock) to take the shot.

I should probably add a note, as I did on the Canon Forum about why I am going back and forth between the 2 differnent targets at different distances; Target 1 serves mainly to defocus. If im just aiming at target 2 and focusing over and over, its not a good measurement of the cameras ability to get that initial focus.

Another point is that the average focus times for each shot work out to:

5Dii - 1.5s
6D - 1.7s
5Diii- 2.33s

But this is not what is happening. The focus lock on Target 1 (in good light) is always very quick, almost instantaneous.

Focusing on Target 2 is taking longer than these average times. In fact, most of the time is spent on Target 2.

It is hard to be precise, but I would say the times are closer to:

5Dii - 2.2s
6D- 2.5s
5Diii - 3 + s

I have a Sekonic 558 Light Meter that seems to be unable to measure the light Im shooting in which would be ideal to know, and am waiting on my new meter - that should help to know exactly at what point it starts to struggle.

M

Very interesting and very topical stuff Michael. A lot of people are very interested to see what you are doing and really appreciate your efforts.

I think the "two targets" idea is very solid thinking. My only "helpful suggestion" would be to not make one of your 2 focus points be in good light. I think it basically gives the cameras a big boost to their average focus time in bad light. I am sure the 5D3 is super fast in good light, but I think the question is how slow is it in bad light? I would be interested to see the same tests done with two targets, but with both targets being in bad light (none in good light). I think you would see a much more dramatic difference between the cameras if indeed 5d3 struggles in low light AF while 5d2 (or 6D) would not have. THAT would make a real statement.

 

[Bennymiata]

I use my 5D3 for weddings, receptions etc, and mostly with a 580EXII and find the auto focussing almost instantaneous - and very accurate, with only about 4 out of 600 shots being soft (and those are probably because of me and not the camera), even ln very low-light situations.
I've never measured the time it takes to aquire focus in thousanths of a second, but it's plenty quick enough for me.

 

[zim]

I would be interested to see the same tests done with two targets, but with both targets being in bad light (none in good light).

+1 I'd really like to see those numbers

 

[risc32]

So it's looking like my thought that now that the 5d has gotten 1D level AF, it gained 1D level AF problems. All of these numbers look like what I would have guessed from my hands on use with this stuff.

 

[Chris Burch]

Has anyone sent their 5DIII into CPS for repair based on the slow focus lock? Clearly some people don't have the problem (or they just haven't noticed), so it's a case of build quality. If enough people start doing it, perhaps Canon will be forced to at least acknowledge the problem.

 

[AudioGlenn]

This is a copy of my post from another thread on the same topic:

So I've been doing some further testing. I can create a situation on my mk3 and 600 ex-rt where I can notice a much less snappy AF lock. I was not able to re-create the lag when trying to focus in a very dark (almost black) setting. Maybe, that's where I messed up in my prior testing. Maybe a completely black environment creates enough contrast between the environment and the red focus assist beam to allow focus whereas a dimly lit situation is lit just enough to give the camera trouble focusing because it can't really see the red AF assist beam as well.

I'm able to notice a significant lag when the ambient light is dim (proper exposure in camera at f/2.8, 1/50, ISO 5000).

Seems to me like I would just turn the beam OFF when I notice the ambient light is about at this level. Am I still over simplifying the issue? I get that pros need their focus to be snappy to catch those moments. Wouldn't that mean a few test shots in the environment first to gauge the settings required....including wether or not to turn the AF assist beam on/off?

Maybe I don't understand the depth of the issue because I'm not coming from a 5D mk2 where the focus is supposed to be faster. That said, I'm an engineer by trade. We are trouble shooters by nature. It seems like there's a pretty easy fix for this (just based off of what I've been able to test). If this really is an issue for a shooter, just turn the beam off when you feel like it's slowing you down. Does ambient light in a reception hall or your venues change constantly for you guys? Isn't this just a matter of your experience telling you how your camera should be set to capture moments? I'm not trying to be condescending. promise. I still don't understand what the big deal is.

 

[steliosk]

its very simple

5D2 focuses faster with af assist beam while 5D3 doesn't
same thing in 6D in 1Dx and all the new aps-c cameras

In low light situations where there is not enough ambient light for the focus point to find target AF beam is a saviour and canon screwed that up with their new cameras..
That costs moments and costs money and if you're shooting an event, people are filled with worries like "did he get the picture or is he still focusing?"

i shoot weddings too and light is a pain in the ass sometimes

 

[AudioGlenn]

thats probably the most sense any one has made about it. Thank you.... so there is a grey area where you really need the AF assist beam to focus and can't do it with just ambient light. got it

 

[dlleno]

Canon has made an important annoncement on this topic:

http://forums.usa.canon.com/t5/EOS/5D3-AF-assist-beam-slower-focus/m-p/13547/highlight/false#M1486