5D Mark III at F/8 and smaller Tests

Status
Not open for further replies.
Mar 25, 2011
16,847
1,835
I started testing my 5D MK III for autofocus beyond f/5.6 yesterday. I decided to start by using my 1.4X and 2X Mark II TC's stacked together rather than taping pins. The day was overcast, and dim, but not exceptionally dark either.

I had my 135mm f/2 L handy, so I put both TC's on it (F/8) and it had no problem autofocusing about 40 feet away, but it was quite slow. Then when I tried it on some flowers 6 feet away, it chattered rapidly, focusing back and forth as though in a loop. Back to a 20 ft distance it focused slowly again. I tried it numerous times with the same results.

Then I tried my 70-200mm f/4 IS (F/11 equiv), it did not even try to focus (my 1D MK III would snap in with this combo when focus was close).

Next, my old non is 70-200mm f/2.8, (f/8 equivalent). It autofocused reasonably fast and reliably every time with the 2 TC's stacked.

Finally, my 100-400mm L at 400mm (f/16 equiv). No go not even a attempt to focus as expected.

Next, I'll tape the pins and try the 70-200mm f/4 + 2X TC (f/8) and the 100-400mm L + 1.4X (f/8). I'm thinking it might work.

i'll post the results.
 
Mar 25, 2011
16,847
1,835
Update: 5D Mark III at F/8 and smaller Tests

Its stopped snowing, and the 1 inch of snow has melted, so I went out to the studio and pulled out my 70-200 f/4L IS and 100-400mm L and taped the contacts on the lens rather than the TC, simply because it was temporary and easier.

First, I tried the 100-400mmL and 1.4X TC (f/8 equivalent). At 100mm which is a little faster than f/8, the lens did focus fairly reliably but not fast. At 400mm, it hunted, but did manage to focus. I suppose its usable, but pretty slow. I panned the lens at 400mm from distant to near, and the focus did follow fairly well.

Then, I put the 2X TC on my 100-400 (f/11) and it tried and erratically found focus, but I'd call it a failure, because it sometimes focused incorrectly (fuzzy image).

Next, I taped the 70-200mm f/4 L IS and attached the 2X TC (F/8). I tried it at 70 and 200mm as well.

At 70mm, the lens was unstable, but eventually found focus for near objects. For far objects, it focused slowly. At 200mm, the lens would not focus on near objects, but wanted to chatter like my 135mm L. At intermediate and far distances, it failed to focus except sporadically, so I call that a failure.

To summarize,

My 70-200mm f/2.8L (Non IS) focused fine, quickly at 70mm and 200mm every time with a 2X and 1.4X TC stacked (f/8 equalivent).

My 100-400mmL + 1.4X TC worked, but not well, and would focus slowly with considerable hunting.

My 70-200mm f/4 L IS was basically unusable with a 1.4X TC.

My 135mm L worked fine at longer distances with stacked TC's, but was slow to focus.

I'd say that a firmware fix would not work for reliable f/8 AF, a new sensor design would be needed, but some combinations of taped TC's and lenses might just work.

I'd be interested to here if the new 70-200mm f/2.8L IS MK II works like my old one. Mine worked suprisingly well.
 
Upvote 0
Re: Update: 5D Mark III at F/8 and smaller Tests

Mt Spokane Photography said:
I'd say that a firmware fix would not work for reliable f/8 AF, a new sensor design would be needed, but some combinations of taped TC's and lenses might just work.
That mirror's the opinion of the CPS rep at Focus a few weeks ago. He didn't know, but based on how the system had been explained to him, his feeling was that a firmware update wouldn't fix it, as he thought it was due to the larger sensors in the new system, preventing any room for f/8 sensor points.
 
Upvote 0
dr croubie said:
Sounds like what happens with my 7D, 70-300L, and Kenko 1.4x T/C (which doesn't translate aperture).
Sometimes it focusses, a lot of the time it just hunts, even when I MF it to the correct spot it will go to one extreme or the other.

Do you have a 5D2 or 7D lying around to compare results?

I tried 7d and 5d2 with 70-200 f4 LIS and TC2.0. It didn't work - hunted for AF like MTSpokane described. Pins covering doesn't help.

EDIT: And with TC 1.4 both work fine. AF of course doesn't work when stacked.
 
Upvote 0
Mar 25, 2011
16,847
1,835
dr croubie said:
Sounds like what happens with my 7D, 70-300L, and Kenko 1.4x T/C (which doesn't translate aperture).
Sometimes it focusses, a lot of the time it just hunts, even when I MF it to the correct spot it will go to one extreme or the other.

Do you have a 5D2 or 7D lying around to compare results?

My 5D M II worked with the 70-200mm f/4L and my 1.4X TC reasonably well. I don't recall if my 7D worked. It already has a 1.6 advantage, but is noisy for high ISO which is needed for high shutter speeds and small apertures.

My 1D MK III would usually work at f/11.
 
Upvote 0
Re: Update: 5D Mark III at F/8 and smaller Tests

Kernuak said:
Mt Spokane Photography said:
I'd say that a firmware fix would not work for reliable f/8 AF, a new sensor design would be needed, but some combinations of taped TC's and lenses might just work.
That mirror's the opinion of the CPS rep at Focus a few weeks ago. He didn't know, but based on how the system had been explained to him, his feeling was that a firmware update wouldn't fix it, as he thought it was due to the larger sensors in the new system, preventing any room for f/8 sensor points.

off course it is not fixable. can they claim it? sure but the reliability would be soo poor with some lenses they may as well save themselves the PR fiasco it would cause. such a drastic change would have had to happen years ago before production started. it is simply too late.
 
Upvote 0
Mt Spokane Photography said:
I had my 135mm f/2 L handy, so I put both TC's on it (F/8) and it had no problem autofocusing about 40 feet away, but it was quite slow.
If this is with 1.4x + 2.0x, wouldn't this result in f/5.6? Can you also try with two 2.0x TC's?

Mt Spokane Photography said:
Next, I taped the 70-200mm f/4 L IS and attached the 2X TC (F/8). I tried it at 70 and 200mm as well.
...
My 70-200mm f/4 L IS was basically unusable with a 1.4X TC.
So you mean unusable with 2.0x TC (and probably also with two 1.4x TC's).
 
Upvote 0
Jul 21, 2010
31,229
13,092
Re: Update: 5D Mark III at F/8 and smaller Tests

Kernuak said:
That mirror's the opinion of the CPS rep at Focus a few weeks ago. He didn't know, but based on how the system had been explained to him, his feeling was that a firmware update wouldn't fix it, as he thought it was due to the larger sensors in the new system, preventing any room for f/8 sensor points.

Looking at the AF sensor for the 1D X/5DIII, I remain unconvinced that there is a technical reason ('not enough room') for eliminating f/8 sensitivity. I suspect this was a conscious decision on the part of Canon, to eliminate f/8 and induce people to buy longer lenses.

I suspect it's not something that can be corrected with a firmware change, although that remains a slim possibility. Despite the results of TC stacking and pin taping, keep in mind that there's a difference between those 'tricks' and Canon re-writing the firmware to support f/8 - presumably, optimized algorithms would make a difference.

It's also possible that a slight re-working of the AF optics (the lenses which split the incoming light) can add f/8 capability to the existing sensor (I presume that's how the previous 1-series bodies achieve f/8 AF, since the sensor lines for the center point have the same physical spacing on the chip as the other f/5.6 lines).

I think Canon was taken by surprise with Nikon's multiple f/8 cross-type points. I hold out a small amount of hope that they'll address the issue now, for the 1D X, rather than make users wait 4+ years for a new pro camera to bring the return of f/8 AF.
 
Upvote 0
W

WarStreet

Guest
The previous cameras used to focus @ -1 EV, and the new ones @ -2 EV. I would like to ask if it is correct to say that the 1DX and 5DIII can focus with f/5.6 @ -2 EV or F/8 @ -1 EV if Canon decide to do a firmware update with a new algorithm ?

A different algorithm done for the purpose of f/8 focusing should make a significant difference than just trying to trick the current firmware code with the tape. I think, there might be a chance that this get fixed. Maybe ready before the 1DX release.

Did you got any different results when changing the focus limiter switch ?
 
Upvote 0
Jul 21, 2010
31,229
13,092
WarStreet said:
The previous cameras used to focus @ -1 EV, and the new ones @ -2 EV. I would like to ask if it is correct to say that the 1DX and 5DIII can focus with f/5.6 @ -2 EV or F/8 @ -1 EV if Canon decide to do a firmware update with a new algorithm ?

A different algorithm done for the purpose of f/8 focusing should make a significant difference than just trying to trick the current firmware code with the tape. I think, there might be a chance that this get fixed. Maybe ready before the 1DX release.

You're talking about unrelated issues. The EV sensitivity of the sensor is independent of the f-number threshold. The f-number specification is related to the 'baseline' of the sensor - f/2.8 lines are further apart (physical spacing on the AF sensor), and require a wider aperture lens to cover that larger area.

Also, the -2 EV spec applies, "Using a single central AF point with an f/2.8 lens..." So, you'd need four times as much light (EV 0) to focus with an f/5.6 lens in extremely dim conditions.
 
Upvote 0
W

WarStreet

Guest
neuroanatomist said:
You're talking about unrelated issues. The EV sensitivity of the sensor is independent of the f-number threshold. The f-number specification is related to the 'baseline' of the sensor - f/2.8 lines are further apart (physical spacing on the AF sensor), and require a wider aperture lens to cover that larger area.

Also, the -2 EV spec applies, "Using a single central AF point with an f/2.8 lens..." So, you'd need four times as much light (EV 0) to focus with an f/5.6 lens in extremely dim conditions.

Thanks for the clarification, I was always pondering about it and this thread was a good place to ask.
 
Upvote 0
Status
Not open for further replies.